/llbacmillan's Science /Iftonoorapbs

TRANSPIRATION AND THE ASCENT
OF SAP IN PLANTS

MACMILLAN AND CO., Limited

LONDON . BOMBAY . CALCUTTA
MELBOURNE

THE MACMILLAN COMPANY

NEW YORK . BOSTON . CHICAGO
DALLAS . SAN FRANCISCO

THE MACMILLAN CO. OF CANADA, Ltd

TORONTO

D7.

TRANSPIRATION AND

THE ASCENT OF SAP

IN PLANTS

BY

HENRY H. DIXON, Sc.D., F.R.S.

University Professor of Botany in Trinity College, Dublin ; Director of
Trinity College Botanic Gardens

MACMILLAN AND CO., LIMITED
ST. MARTIN'S STREET, LONDON

W 1914

COPYRIGHT

b~ % U

PREFACE

The Ascent of Sap is a problem of old standing. As
was the case with several other biological problems, a
peculiar alternation may be observed in the hypotheses
formed to explain the phenomena. At first the process
of the ascent of water in trees was, almost without serious
thought, assigned to the vital activities of the plant,
and, in general with other vital processes, put outside
the domain of physical investigation. Later, advances
in physics and chemistry, introducing rationality into
observation, emboldened hardy spirits to assign the
whole process to various physical forces or to combinations
of them. Doubtless these philosophers somewhat trans-
gressed legitimate deduction, and their daring met its
punishment in the overthrow of their successive physical
theories. It was then again the turn of the Vitalists,
and during the latter part of the last century they erqoyed
their heyday of dogma. However, the leaven of ration-
ality still worked on in even their theories, and several
physiologists must have felt what Strasburger stated,
that the physical forces developed in, and the physical
configuration exhibited by, the water tracts would supply
a complete explanation when properly understood.

In the present monograph, an account is given of a

VI

PREFACE

physical explanation of the rise of water in trees. This
theory rests on the knowledge of a property of liquids,
which, although discovered in the middle of the last
century, was little recognised and seldom referred to in
physical literature. It now appears that a full appreci-
ation of this property is essential for the realisation of
the manner in which water is raised in plants and of
the meaning of the structure of trees as a mechanism for
lifting water.

In the formation of the theory and in much of the
earlier work carried out to test its validity and to illustrate
its applicability, Dr. J. Joly and the author worked in
collaboration. The work on the cryoscopy of sap was
carried out with the help of Mr. W. R. G. Atkins, to whom
the author is further indebted for his help in reading the
proofs and for his useful criticisms. Prof. R. A. Gregory
the editor of this series of monographs has also kindly
supplied several valuable emendations and suggestions.

Henry H. Dixon.

School of Botany,

Trinity College, Dublin.
1914.

CONTENTS

CHAPTER I

PAGE

THE NATURE OF TRANSPIRATION 1

CHAPTER II

ASCENT OF SAP IN STEMS. CRITICISM OF PHYSICAL THEORIES . . 27

CHAPTER III

ASCENT OF SAP IN STEMS. CRITICISM OF VITAL THEORIES ... 47

CHAPTER IV

COHESION THEORY OF THE ASCENT OF SAP IN STEMS .... 81

CHAPTER V

TENSILE STRENGTH OF THE SAP OF TREES 101

CHAPTER VI

ESTIMATE OF THE TENSION REQUIRED TO RAISE THE SAP . . . .116

CHAPTER VII

OSMOTIC PRESSURES OF LEAF-CELLS 139

CHAPTER VIII

THE THERMO-ELECTRIC METHOD OF CRYOSCOPY . ... 155

vii

viii CONTENTS

CHAPTER IX

PAGE
METHODS OF EXTRACTING SAP FOR CRYOSCOPIC OBSERVATIONS . . 175

CHAPTER X

OSMOTIC PRESSURES IN PLANTS 190

CHAPTER XI

ENERGY AVAILABLE FOR RAISING THE SAP 201

INDEX 213

TRANSPIRATION AND THE ASCENT
OF SAP IN PLANTS.

CHAPTER I

THE NATURE OF TRANSPIRATION

Transpiration and evaporation. Comparison of
the transpiration from the upper side and of that from
the under side of hypostomatous leaves has shown that
transpiration is principally effected through the stomata.
The total area of the stomata, however, bears but a very
small proportion to that of the whole leaf. For example,
in a leaf of Catalpa bignonioides it is but 0*9 per cent, of
the total leaf surface. In a leaf of Helianthus annuus
the area of the stomata appears to be about 3 per cent,
of the whole leaf surface.

Many years ago, Unger showed that under similar con-
ditions the amount of water evaporated from a free surface
was 2 - 8 to 13*8 times that transpired from an equal leaf
surface. If we take the lower limit, this means that
35 sq. cm. of water surface will give off as much water
vapour as 100 sq. cm. of leaf surface under similar con-
ditions. Of this 100 sq. cm. as much as 3 sq. cm. may
be represented as stomata. So we arrive at the surprising
result that the combined area of the stomata is at least
ten times as efficient in giving ofT water vapour as a free
water surface of equal area.

B

2 TRANSPIRATION AND ASCENT OF SAP ch.

Diffusion through stomata. The beautiful re-
searches of Brown and Escombe on diffusion through
perforated septa have satisfactorily cleared up this diffi-
culty. They showed experimentally that the amount of
vapour which diffuses through a perforation in a septum
does not diminish proportionally as the cross section
diminishes, but only in proportion as its diameter
decreases. This result, which is at first sight by no
means obvious, has been mathematically explained by
both Stephan and Larmor. Brown and Escombe quote
both writers ; but as their methods may not be easily
followed by all botanists the following simple explanation
may be of interest.

When diffusion, undisturbed by convection currents, is
taking place from a perforation which is kept filled with
water vapour, or from a surface of water, into a space
less rich in water vapour, it is evident that after a time
a certain gradient of concentration will establish itself
over the aperture. The concentration will be greatest near
the perforation and less more remote from it, whilst shells
of uniform concentration will be found over-arching the
perforation. At a little distance from the aperture, say
when the diameter of the shells is about four times that of
the perforation, these shells will be approximately hemi-
spheres.

To possess any given concentration p a shell must have
an area bearing a definite proportion to the area of the
perforation ; because the supply maintaining its concen-
tration is derived from the water molecules leaving the
aperture, and the number of these will be proportional to
its area.

Let R and A denote the radius and area respectively of
the shell having a concentration p, and r and a the radius
and area of the perforation. Let p l indicate the concen-
tration of water vapour in the perforation.

A = La or & = lr-.

i THE NATURE Oh TRANSPIRATION 3

The flow into the shell = the gradient x area = x p x A

R r K 'k

Therefore the flow from the aperture is proportional to
its radius and not to its area.

Under the " static " conditions to which these calcula-
tions apply, the shape of the stomata also contributes to
their efficiency. It will easily be understood that separated
as they are from one another by distances relatively great
compared with their diameters, the diffusion of water
vapour from adjacent stomata will not interfere. Conse-
quently the rate of diffusion at the margins will be greater
than over the middle of the apertures. Therefore, an
opening having the longest margin relatively to its area
will be the most efficient ; and the slit-like form of the
stomata is seen to be particularly advantageous.

The consideration that the margins are the most effec-
tive parts of the stomata in diffusion suggests another way
of looking at the " diameter law." It is evident that for
very small holes the marginal region bears a very large
relation to the whole opening. For circular apertures the
area decreases as the square of the radius while the margin
is reduced only as the radius. In the case of slit-like
apertures the whole opening may be regarded as marginal.
But it is this marginal region which is most effective ;
and therefore we should expect the amount of water vapour
diffusing through such an aperture to be approximately
proportional to the margin (and therefore to the diameter)
rather than to the area of the aperture.

The diffusion of water vapour from the intercellular
spaces of the leaves through the stomata has been thus
explained on simple physical principles. It remains to be
seen how the supply of water to these spaces is to be
accounted for.

Functions of evaporation, osmosis, and im-

R 2

4 TRANSPIRATION AND ASCENT OF SAP ch.

bibition in transpiration. The leaf trachea? contain-
ing the rising sap are separated from the intercellular
spaces by a layer of one or more thin-walled cells. The
thin wall, which is formed of cellulose, is permeable to
water and dissolved substances ; at the same time it
is tough and of considerable tensile strength. It is
lined by a layer of viscid protoplasm which in turn
surrounds, and is completely filled by, a solution the
vacuole. The latter contains various carbohydrates and
electrolytes in solution, and round it the protoplasmic
layer forms a fairly perfect semi-permeable membrane.
It is evident that, when sufficient water is available,
the solutes in the vacuole will exert an osmotic
pressure on the protoplasm which will be forced outwards
against the wall and distend the latter. Ultimately,
the tensile stress in the wall will balance the osmotic
pressure of the solutes of the vacuole and equilibrium will
be established and the cells will be tense and rigid. The
walls of the trachea? are also quite permeable, but they
are more rigid than those of the cells and they are further
prevented from collapsing by the presence of the internal
supports in the form of rings and spirals. The trachea?
do not enclose a semi-permeable membrane of protoplasm.
We may now consider how this mechanism will act in
the transference of water. The imbibitional or capillary
forces of the cell-walls in contact with the intercellular
spaces of the leaf will draw off water from their vacuoles
through the protoplasm until the concentration of solutes
in the vacuole is such that the vapour pressure of water
in it is equal to that obtaining in the cell-wall. Supposing
now that the vapour pressure of the water menisci in the
reseau of the cell-wall is greater than that obtaining in the
intercellular space, water will leave the wall and the menisci
will retreat into it. This will cause their curvature to
increase and will raise their capillary forces so that they
will endeavour to extract water from the solution in the

i THE NATURE OF TRANSPIRATION 5

vacuole. A concentration of the vacuole results, and con-
sequently the osmotic pull on the water in the trachea
is increased. Hence it follows that a transference of water
from the tracheae will take place so long as the vapour
pressure of water in the tracheae is greater than that in
the intercellular spaces of the leaf.

With such a mechanism before our minds we can easily
realise that the whole process of transpiration may be
purely physical and one into which vital actions (viz.,
actions connected with living substance the intermediate
steps of which are little understood) need not enter.

These considerations also explain several observations
otherwise difficult of interpretation. When colouring
materials {e.g., eosin) or other substances, poisonous or
non-poisonous {e.g., copper sulphate, picric acid, tartaric
acid, sodium bicarbonate, ferricyanide of potassium, etc.)
are supplied in watery solution to the cut surface of a
transpiring branch, it is found that the rate of transpira-
tion continues without much diminution till the solution
can be detected in the leaves, but then suddenly falls off.
This may be seen either by direct measurement of the water
drawn up, or by the fact that the leaves of these branches
usually remain fresh until the arrival of the solution in
them, but then rapidly lose their turgor, flag, and
finally become dry and crisp. These observations are
easily explained if we may assume that at first the solu-
tion is drawn up in the conduits under the tension set up
through the osmotic forces of the cells of the leaves, but
afterwards, when the upper portion of comparatively pure
water has been eliminated from the plant by transpiration,
the solution comes into contact with the living cells and
modifies or destroys their osmotic properties. The latter
soon lose their turgor, and the leaf droops.

That the loss of turgescence will necessarily be attended
by a diminution of the amount of water evaporated from
the leaf will appear evident when it is considered that :

6 TRANSPIRATION AND ASCENT OF SAP ch.

First the transpiring surface of the evaporating cells will
be diminished owing to the contraction of the cells when
they cease to be turgescent. Secondly, the diffusion of
the water-vapour from the inner tissues of the leaf will be
hindered by the collapse of the intercellular passages caused
by the drooping of the leaf ; and, finally, the evaporating
films which on the surface of the wall of the turgid cell
may be supposed to be but slightly concave, will under the
new conditions tend, by retreating inwards into the sub-
stance of the cell-walls, to form a series of more concave
menisci from which, as is known, evaporation will proceed
more slowly. However, as the walls of these cells are im-
bibed with water, and a continuous column of water
extends back from the outer evaporating walls down
through the conduits into the plant hanging from the
surface film formed on the outside of the evaporating cell-
wall, a slow movement upwards will take place of the
solution in the conduits even after the loss of turgor by
the osmotic cells. The supply is in most cases inadequate
to prevent the drying of the leaf.

To the evaporation at the surface film formed on the
cell-walls must also be referred the ascent of water in
pieces of dead wood which have been soaked and injected
with water, as described by Strasburger ; for in this case,
of course, the intervention of osmotic forces is excluded.
It is to be noticed that in this experiment also the wood
dries downwards from above.

From the foregoing considerations it seems quite feasible
to explain the process of transpiration as a purely physical
one in which the intervention of living matter, as such, is
not necessary ; for we might substitute a dead osmotic
cell in the plant for the living one without having reason
to expect a difference in the action of the mechanism.
This being so, it appears highly probable that transpira-
tion is often a purely physical phenomenon.

However, we must not shut our eyes to the fact that

i THE NATURE OF TRANSPIRATION 7

several observations seem to support the view that under
normal conditions some sort of vital action intervenes.

Transpiration of living and dead leaves. It
was noted above that when the leaves of a branch
are killed, not only is the flow of water upwards greatly
reduced, from causes which have been already explained,
but ultimately the leaves dry up and finally fail entirely
to raise water in the branch. It is true that in this case,
not only are vital actions removed, but also one of the
most important features of the mechanism, viz., the semi-
permeable membrane, is destroyed by the coagulation of
the protoplasm. It is evident, however, that after this the
capillary forces of the cell- wall of the leaf- cells alone are
unable to continue to raise the water under the new con-
ditions, and this would suggest that unaided they may be
insufficient in the living leaf.

This line of reasoning would indicate that the protoplasm
may not only act as a semi-permeable membrane by allow-
ing water to pass through to a region of diminished pressure,
but it may also actively secrete water, or a solution, on
its outer surface. Such glandular action of protoplasm is
well known, and examples from the coenocytic fungi may
serve as illustration. In Phycomyces, Pilobolus, and
Mucor, water which is absorbed by the submerged part of
the coenocyte is expelled on its aerial surface. The same
process may be witnessed in the nectaries and all the
glands of higher plants.

In these cases there is some reason to believe that the
liquid exuded is not pure water ; and hence it seems pro-
bable that the protoplasm first secretes soluble substances
on its outer surface and that these act osmotically when
concentrated by evaporation, on the vacuoles within and
draw water from the cells. Where the process is very
rapid, as, for example, the expulsion of water from the
cells of the pulvinus of Mimosa in response to a stimulus
such a sequence of events can scarcely be imagined. The

8 TRANSPIRATION AND ASCENT OF SAP ch.

extremely rapid action seems only explicable by assuming
some sudden change in the permeability of the proto-
plasm.

Secretion of water. The active secretion of water
by cells is particularly well illustrated by the water glands
in the tips of the leaves of Colocasia antiquorum. As is
well known, when this aroid is under suitable conditions
of moist soil, saturated atmosphere, and favourable
temperature, a succession of drops (often as many as two
per second) may be seen issuing from its leaf tip. A
similar exudation of fluid may be often observed on the
leaves of grass-seedlings when surrounded with a satu-
rated atmosphere.

It is evident that this exudation of fluid would be easily
explicable on purely physical processes if the fluid con-
tained any considerable quantity of dissolved substances.
Then the external solutes might be assumed by virtue of
their osmotic pressure to extract water from within and so
keep up the supplies on the outside of the cell. The pres-
ence of dissolved substances may be sometimes demon-
strated in the exudations of the fungi by evaporating
drops of the fluid on polished glass. How far these dis-
solved substances are crystalloids, and so competent to act
osmotically, remains to be determined. At the same time,
it should be noted that, even admitting there are sufficient
crystalloidal solutes on the surface of the cell to account
for the exudation, it seems we must assume a secretory
action in the protoplasm to bring these crystalloids out
of the cells and expose them on the outer surface, and this
action must be a continued one as the exudation will
constantly be carrying off its solutes.

If, however, the fluid exuded is pure water, or contains
practically no dissolved crystalloids, it is evident that
osmosis cannot account for the process, for the osmotic
action of the solutes within the vacuoles cannot cause the
elimination of water on the outside of the cells. In this,

i THE NATURE OF TRANSPIRATION 9

energy is being expended, and the energy apparently must
be supplied by the metabolism of the protoplasm.

Fortunately the quantities exuded from the leaf tips of
Colocasia are quite large enough to allow us carefully to
test the purity of the water. With this intent some 10 c.c.
were collected from the leaves of this plant during one
night ; and, by means of the thermo-electric method of
cryoscopy, which will be described later, were tested with
regard to their freezing-point. It was found not to differ
sensibly from that of distilled water. The electrical con-
ductivity was also determined for the sample, and it was
found to be less than that of tap water. These tests show
that osmotically we may regard the exudation as pure
water, and, consequently, the process must be one of secre-
tion involving the intervention of living protoplasm and
the expenditure of stored energy.

In the case of Colocasia the dripping ceases when the
leaves are surrounded with an unsaturated atmosphere.
Evidently the secretion is no more than able to keep pace
with the demand of evaporation, but at the same time it
seems reasonable to assume that the water has been largely
supplied for evaporation by a secretory process. It seems
also reasonable to suppose that when evaporation is acceler-
ated beyond the capacities of protoplasm for secretion
the cell-walls will dry and the capillary forces of the
menisci in their substance will lend their aid in separating
the solvent from the solutions in the vacuoles.

How far can we transfer the glandular functions of the
leaf tip of Colocasia to the mesophyll cells of the trans-
piring leaves ? An answer to this question has been
sought by different ways.

Transpiration into various gases. If the cells
of the mesophyll secrete pure water on their outer side
we should be justified in expecting that this secretion,
in common with other vital actions, would be inhibited
when the protoplasm was anaesthetised, or would be

io TRANSPIRATION AND ASCENT OF SAP ch.

reduced when its activities were depressed. Starting
from this idea an endeavour was made to test the effect
on transpiration of an application of chloroform vapour,
ether vapour, and carbon dioxide gas. Although each of
these three agents, when they were brought round the
transpiring leaves, led to a marked falling off in the amount
transpired, consideration showed that their actions are so
manifold that the effect observed cannot with certainty
be attributed entirely to their direct action in checking the
vital actions of the protoplasm. At the same time, experi-
ments were made to test the effect of an increased supply
of oxygen on transpiration ; the marked acceleration pro-

FlG. 1.

duced by this gas possibly is due to its stimulating effect
on the secretion, but in any case is of considerable interest
in this discussion.

The method of experiment was as follows : The rate of
transpiration of a branch enclosed in a large receiver, and
supplied with a constant current of dried air, was observed.
This rate was then compared with the rate of transpiration,
when a similar current of some other dried gas, or dried
air, carrying with it some anaesthetising vapour, was
passed through the receiver.

The rate of transpiration was estimated, either by the
motion of an index moving in a capillary tube sealed

i THE NATURE OF TRANSPIRATION n

hermetically to the cut end of the branch, or by directly
weighing the amount of water transpired. In the latter
case, which was found to be the more satisfactory, the
branch, inserted through a caoutchouc cork into a test-
tube containing water, was hung from one arm of a
balance. The arrangement is shown in Fig. 1. In this
figure b is a tower ' containing calcium chloride, and
c is a sulphuric acid bulb for drying the gas supplied.
Before passing through the drying materials, the gas enter-
ing at e is led into an inverted flask d, which is provided, in
addition to the tubes of entry and exit, with a U-tube, /,
filled with oil. The supply of gas is adjusted until the
oil in the longer arm of the tube / is brought to a certain
level. By this means the pressure, and consequently the
flow, of gas through the apparatus, can be adjusted and
compared. When vapours are to be supplied, the liquid
from which the vapour is derived is placed in a sulphuric
acid bulb, like that in the figure, but inserted in the train
between the air supply and the flask d. In each case
the supply was passed into the apparatus from a cylinder
containing the compressed gas.

The chamber containing the branch was exposed only
to a very feeble light so that the stomata were closed
throughout the experiments.

The first experiments made were with the index method
of estimating the rate of transpiration. A modification
of the apparatus, as figured, which is readily understood,
was then used, and the branch, sealed hermetically to a
capillary tube containing the index, was inserted from
below into the receiver.

With these arrangements there soon appeared to be a
marked difference in the rate of transpiration in oxygen
and carbon dioxide. Thus, to quote the mean of a number
of observations with a branch of Cytisus laburnum in
carbon dioxide the index moved 1 cm. in 38 sec. ; with
the same branch in oxygen it moved 1 cm. in 28 sec.

12 TRANSPIRATION AND ASCENT OF SAP ch.

When the rate of transpiration in air was compared
with that in oxygen, it was found that when a branch
was surrounded with the latter gas, transpiration was
slightly more rapid. The index for the branch in air
moved 1 cm. in 36 sec. ; for the same branch in oxygen
it moved 1 cm. in 33 sec.

If, while a current of air was passing through the receiver,
a piece of cotton wool soaked in chloroform was intro-
duced, a much more marked difference in the rate of
transpiration became apparent in a short time. Thus the
index was traversing 10 cm. in 50 - 8 sec. before chloroform
was introduced. Thirty minutes later the index took
516 sec. to traverse the same distance. The chloroform
was then removed, and the air current maintained for
60 minutes. At the end of this time the index moved
10 cm. in 120 sec. This result is the mean of a number
of experiments made with a small branch of Acer macro-
phyllum.

With another branch in air the motion of the index
was 10 cm. in 127 sec. When surrounded with chloroform
vapour for 45 minutes the index took 642 sec. to traverse
the same distance.

A similar diminution in the rate of transpiration is
observed when the branch is surrounded by ether vapour.
Thus, with a branch of Acer macrophyllum in air, the
index moved 10 cm. in 205 sec. ; with the same in ether
vapour it moved 10 cm. in 265 sec.

These experiments indicate a large difference in the rate
of transpiration in the different gases. The figures given
here will serve only as examples of the results of such
experiments, for, although they were the means of a
number of observations, the latter are made so precarious
by various circumstances, that they can only be taken as
indicating a difference, and not as giving a measure of it.
The sticking of the index in the capillary tube, and the
opening of the receiver to introduce the anaesthetics, bring

THE NATURE OF TRANSPIRATION

13

in errors, which render the method unsuited to exact
observation.

In order to eliminate these sources of inexactness,
recourse was made to the arrangements shown in the
figure. The results obtained by this method are displayed
in the following table. The difficulties of keeping the flow l
of gas exactly constant through the apparatus, and other
experimental errors, lead to variations between the indi-
vidual observations often amounting to 10 per cent. ; but
by multiplying these observations, an approximation to
the actual alteration in the rate of transpiration has been
obtained. The numbers here given are the means of a
large number of observations. In each case, the branches
experimented on were from a bush of Syringa vulgaris,
except in the experiments where effect of ether vapour
was observed. In these, branches of Cytisus laburnum
were used.

If the amount transpired in air be taken as 100, the
amounts transpired in the other gases are as follows.
These figures may be said to denote the specific transpiration
for the gases :

Table 1.

Medium.

Specific
Transpiration.

Oxvsen . ....

1358

1000

87-3

82-3

66-4

Ether

The first source of error affecting these experiments,
and one which it seems hard to eliminate, arises from the
fact that the effects of the different gases may be more or
less rapid. Thus it is very certain that the light gases

1 It is to be observed the rate of flow of the different gases will be
different, even if the pressure be the same.

i 4 TRANSPIRATION AND ASCENT OF SAP ch.

will diffuse into the intercellular spaces more quickly than
the heavy gases, and so come into contact with evapora-
ting cells more quickly. Besides this, it seems probable
that the poisoning and anaesthetic effects of one may be
more rapid than those of another. The observations, on
which the numbers given above are based, were com-
menced in each case after the branch had been surrounded
by the gas for five minutes, and were discontinued before
any lethal effects could be observed in the leaves ; for such, if
arising, would cause the osmotic pressures obtaining in the
leaf-cells to become diminished, by rendering the proto-
plasmic membranes permeable. These effects were usually
visible within 45 minutes after starting the experiment.
It is possible that the denser vapour could not, within this
time, diffuse into all the intercellular spaces of the leaves.

An error arising from this possibility is most unsatis-
factory, as it seems extremely difficult to make proper
allowance for it. It seems impossible, at present, to
decide how soon the surrounding gas will come into con-
tact with the evaporating cells, and, also, when the
anaesthetising or stimulating action will cease, and the
lethal effects will begin, if, indeed, there is any sharp line
of distinction.

Next we come to an error which can, in some degree, be
eliminated. It is known that the rate of diffusion of a gas
will be influenced by the nature of the gas occupying the
space into which it is diffusing. Thus water-vapour will
diffuse more slowly into carbon dioxide gas than into
oxygen. This difference depends on the relative sizes of
the molecules of the gases into which the water-vapour
has to diffuse. For the same pressure and temperature,
there will be the same number of molecules of these gases
in the surrounding space ; but if their sizes are different,
it is plain that the water-molecules will less readily diffuse
into the space occupied by the gas composed of the larger
molecules.

THE NATURE OF TRANSPIRATION

r 5

In order to form some idea of this effect, I suspended a
shallow dish containing water in the receiver, previously
occupied by the transpiring branch; and in connection
with the train of apparatus previously described, successive
weighings gave approximately the loss of water by evapora-
tion from this dish. During the experiment a stream of
gas, dried as before described, was kept up through the
apparatus. The rate of evaporation, when this current
was composed of air, oxygen, carbon dioxide, and largely
of ether and chloroform, was observed.

Again denoting the loss of weight of a vessel of water
in air as 100, the loss in the other gases was found to be
as follows :

Table 2.

Medium.

Specific
Evaporation.

Air .

104
100

89
81
59

Carbon dioxide ....
Ether

From these observations it would appear that the rate
of transpiration is diminished when the leaves are sur-
rounded by carbon dioxide, ether vapour, or chloroform,
much in the same degree as the rate of evaporation would
be diminished by the presence of these gases ; and this
diminution is in the inverse order of their densities. In
the case of oxygen, however, the rate of transpiration is
increased much more than the rate of evaporation would
be from a liquid surface.

It must be understood that these numbers only apply
to the first effects of carbon dioxide, ether, and chloro-
form ; for when these gases begin to exercise a lethal
action on the cells, the rate of transpiration is very

1 6 TRANSPIRATION AND ASCENT OF SAP ch.

markedly diminished, presumably owing to the reduction
of the osmotic pressure in the cells.

The experiments both on transpiration and evaporation
are exposed to two common errors e.g., a certain amount of
gas will be dissolved in each case by the liquid present,
and this will reduce the loss of weight, and so diminish
the rate in both cases. Again, this solution of the gas in
the liquid will alter the surface tension, and so modify the
rate of loss.

With regard to oxygen the case is different. The in-
crease in the percentage of this gas, or even possibly some
impurity carried with it, increases the rate of transpira-
tion much more than that of evaporation ; and so furnishes
experimental evidence in favour of the view that water is
brought forward to the seat of evaporation by secretion.

With the other gases there is practically no difference
between the specific transpiration and specific evaporation.
The logical conclusion from this seems to be that these
gases were without perceptible effect on the vital actions
of the leaf-cells, so far as transpiration is concerned, during
the experiment.

Thus the problem as to how far secretory actions, taking
place in the leaf-cells at the expense of the stored energy
of organic compounds, accelerate transpiration is not
decided by these experiments ; but I think it will appear
that their evidence, although by no means unequivocal,
favours the view that such actions have some function in
the elimination of water from the transpiring cells.

Although these experiments cannot be regarded as quantitatively exact,
I think they are not without their bearing on plant physiology.

It is a matter of frequent observation that many plants which are natives
of arid regions secrete a relatively large amount of ethereal oils. It has
been urged that the vapours of these ethereal oils form a screen which
arrest the heat radiations, and thus the leaves of the plant are kept cooler
than they otherwise would be. It might, however, be said against this
theory that such an absorptive screen in contact with the leaves (and it
would evidently be most effective at the surface of the leaves) would rather
tend to raise their temperature. Be that as it may, it seems that the

i THE NATURE OF TRANSPIRATION 17

notion of vapours in checking evaporation, emphasised by this research,
affords a simpler explanation of the function of these oily secretions. When
the vapour of the ethereal oils is liberated from the leaf-tissues, it will
surround the leaves, and fill the intercellular spaces. In these positions we
might expect that it will exert a retarding action on transpiration
and evaporation, in accordance with the experiments quoted above. I
have only been able to make a few experiments on the matter, but
these indicate the surmise given here is correct. I found that the
vapour given off from chopped-up leaves of Artemisia absinthium
reduced the rate of transpiration very considerably. Thus, if we denote
tlie rate of transpiration of a branch of Syringa vulgaris, in a current of
dry air, by 100, this rate will be reduced to about 87 if we allow the air-
current to pass over chopped leaves of this Artemisia, and so carry
some of the vapour given off by these leaves round the transpiring branch.
The air is, of course, dried after passing over the leaves. In a similar
manner I found that the same vapour reduced the rate of transpiration
of a branch of Gystisus laburnum from 100 to 93. In these experiments
the temperature lay between 16 and 17 3 C. At higher temperatures, it is
possible that the effects would be more marked.

Information as to the nature of the forces effective in
bringing forward the water to the seat of evaporation was
also sought by examining the possibility of transpiration
into a saturated space.

Transpiration into saturated spaces. In normal
circumstances transpiration is effected under conditions
favourable to evaporation. The transpiring surfaces are
at such a temperature that the vapour pressure in the
surrounding space is less than at the surface of the tran-
spiring cells. To maintain this temperature, the leaves
of the plant are free to receive light and heat radiations,
and heat may be conducted into them, as evaporation
tends to lower their temperature below that of their
surroundings. This inflow of energy from the external
world must, in ordinary circumstances, be taking place
during transpiration. In addition to these sources of
energy, the cells of the leaves may do work at the expense
of the potential energy of the store materials they possess.
This stored energy, which is, of course, ultimately derived
from the radiant energy entering the plant, is the only
remaining source of energy available for the leaves.

c

1 8 TRANSPIRATION AND ASCENT OF SAP ch.

If, when the radiated energy is cut off, and the con-
ditions are such that water tends to condense on the leaves
from the surrounding space, the cells of the leaves still
continue to draw up water in the capillaries, then the
work done must be at the expense of the stored energy ;
and, if this work is no longer continued, when the leaves
are killed, we may fairly ascribe it to vital actions pumping
or drawing up water from the conduits of the plant.

It may be pointed out that this energy could only be
made available when the store materials can obtain the
requisite oxygen from the plant's surroundings, or from
its own substance, and so, in common with other vital
actions, it would cease when oxygen is not available.

Supposing, then, we find that the upward motion of
the transpiration current continues when radiated energy
is cut off, and when the leaves are surrounded by a space
saturated with water vapour, we are driven to conclude
that the traction exerted on the ascending water is exerted
by a vital action, and we can no longer assume that simple
physical processes, exactly corresponding to the actual
inflow of energy, at the moment, can account for the
elevation of water in such a case. On the other hand,
the converse will be true if no elevation of water occurs
in the plant when it is submitted to the conditions described.

To put this matter to an experimental test, the following
arrangements are made : A small branch about 30 cm.
long is cut and set in water in a cool, dark cupboard.
From this it is transferred, still standing in water, to
a glass receiver. The internal walls of the receiver are
kept wet. After remaining one hour under the receiver,
and still screened from light, it is assumed that any
reduced gas pressure existing in the water conduits has
become equalised to that of the atmosphere, and that,
consequently, the external pressure exerted at the base
of the branch has ceased to move the water upwards.
An open beaker, containing water at 100 C, is now intro-

THE NATURE OF TRANSPIRATION

19

<^

n

felUfe

duced under the receiver, and the branch is transferred
from the water to a watery solution of eosin. A wooden
screen is set to cut off the direct radiation of the beaker
from the branch. These arrangements are made in a
dull light, and, when complete, the whole is set in total
darkness.

As soon as the beaker containing the hot water is intro-
duced under the receiver, the space included will immedi-
ately be filled with cloud and water vapour. Water is
freely deposited on the walls of the receiver and on the
surfaces of the leaves of the plant. The space is com-
pletely saturated, and remains so, as it continues to fall
in temperature, owing to the
gradual cooling of the whole ;
and, as the water is always
at a higher temperature than
the leaves, a constant distil-
lation goes on from the
beaker to the leaves. The
arrangements are shown in
Fig. 2.

When these arrangements
have been made, the ap-
paratus is left for one hour.

At the end of this time, it will be found that the eosin
solution has been drawn up very markedly into the plant,
thus showing that the elevation of the water in the
conduits may be effected by vital action. For in this
experiment the immediate energy relations of the' plant to
its surroundings cannot account for the rise. I have
performed this experiment, obtaining the same result, with
Chrysanthemum sinense, C. lacustre, Myrtus communis,
Eucalyptus globulus, Escallonia macrantha.

As we should expect, it was found that, when dead leaves
and branches were set in this saturated chamber, no rise
of the eosin was observed, although simultaneously eosin

c 2

taWi

Fig. 2.

20 TRANSPIRATION AND ASCENT OF SAP ch.

was drawn up into living specimens placed side by side
with the dead ones. The dead branches which I used
had been killed by chloroform vapour, or by immersion
for some minutes in water at 90 C.

In these experiments when the coloured fluid was drawn
up only into the capillaries of the stem, the pumping action
raising it may have been exerted either by the cells
bordering the conduits in the stem, or by those in a similar
position in the leaves. But when the veins of the leaves
become injected, it is evident, since no cells interrupt the
continuity of the water-conducting capillaries, that some
of the cells exerting the traction in the fluid must be
situated in the leaves.

This fact may be demonstrated more directly by experi-
ments in which the ascent of watery eosin in a branch
stripped of its leaves is compared with that in a similar
branch provided with leaves, when both are placed in the
saturated chamber. It will be found I have performed
the experiment with Chrysanthemum sinense, Escallonia
macrantha, Cheiranthus cheiri that the leafv branch will
draw up the eosin rapidly, while in similar circum-
stances the colouring matter will rise but slightly a few
cms. per hour in the branch deprived of its leaves. The
rise observed may be easily explained by the supposition
that, in the green parts of the young branches and the
buds, the cells probably act like those of the leaves, and
draw up water ; or, again, the action of the cells border-
ing the capillaries of the stem wood-parenchyma and
medullary rays may be responsible for the elevation
observed. In any case the rise is but slight, 3-5 cm. in
the stripped branches, compared with 20-30 cm. in the
leafy branches during the same time.

That the elevating force is chiefly located in the leaves
may also be shown by the fact that large leaves detached
from the stem are capable of quickly injecting the finest
veins at their apices when set upright in watery eosin in

i THE NATURE OF TRANSPIRATION 21

the saturated chamber. For this purpose I used the leaves
of Eucalyptus globulus, and found that their apical veins
were injected often after standing only 30 min. in eosin,
when surrounded with a saturated atmosphere. The eosin,
to do this, had risen 20 cm. in the leaf above the level of
the solution in which the leaf stood. In this case it is
evident that the cells of the leaf must have been solely
responsible for the observed elevation.

But the directed pumping actions which cause the eleva-
tion of the coloured fluid in these cases, although mostly
confined to the leaf, do not appear to be restricted to any
special cells forming water-glands on the surface of the
leaf. It seems most probable that most or all of the cells
bordering on the vascular capillaries, both in leaf and
stem, are able to exert a tractional force on the water in
the conduits, and are able to expel water, when thus drawn
in, on their outer surfaces. It may be, however, that the
cells of the water-glands of plants are more highly special-
ised for this function, and hence the exudation of drops
on leaves of plants in a moist atmosphere takes place over
these glands or hydathodes, as Haberlandt prefers to call
them.

The following observation shows that the elevation of
the water is not solely due to the functioning of these
water-glands, even in plants possessed of these structures.
The leaves of Escallonia macrantha, Chrysanthemum sinense,
and Chrysanthemum lacustre have water stomata on the
margin of the leaf ; but if these glands are removed by
cutting away the whole margin with a scissors, it will be
found that water will be drawn up into these leaves through
the stem almost as quickly as into leaves which are left
intact.

Another observation which shows that the traction is
exerted by cells of the leaf, which are not visibly differ-
entiated, may be made on Cheiranthus cheiri. The leaves
of this plant, so far as I can make out, have no specialised

22 TRANSPIRATION AND ASCENT OF SAP ch.

water-glands. However, the extreme apex often withers
away in the older leaves, as if some substance had been
exuded there from the leaf. In case this tip be the seat
of a water-gland, it was removed from all the leaves of a
branch which was set in the saturated chamber. After a
suitable time it was found that the coloured fluid had
risen into all the veins of the leaves, and it was seen in
the ultimate blind terminations of the vascular bundles.
In Cheiranthus cheiri these terminations are surrounded
by cells undifferentiated from the other cells of the
mesophyll of the leaf. The coloured fluid must have been
drawn into the terminal portions of the veins by these
cells, and not by any specialised water-glands. We may
conclude that the similar cells along the conduits have the
same function.

It was usually found at the end of all the experiments
conducted in the saturated chamber that the surfaces of
the leaves had a copious deposit of water upon them, and
so it seemed probable that water was actually extruded
from the cells of the leaf even after water had begun to
condense on them from the surroundings.

The actual presence of free liquid on the surface of the
leaves apparently did not markedly diminish the rate of
rise of the coloured fluid in the branch, and so, if the branch
was immersed in water before commencing the experi-
ment, it was found that the eosin mounted notwithstanding
into the dripping leaves.

In these cases, the pumping cells, being surrounded by
water, must possess a directed action, which enables them
to draw the water in on one side from a liquid supply, and
to expel the water on the other into free liquid.

This directed action may be more strikingly demon-
strated by the following experiment : A branch is fixed
water-tight into the lower narrow opening of a glass
receiver, so that its upper part and leaves project into the
interior, while its base extends beyond the cork in the

-

3Mr

ZS~-

^vff ~

;

=B

Z^

Fk;. 3.

i THE NATURE OF TRANSPIRATION 23

neck, and is supplied with a solution of eosin {see Fig. 3).
If the receiver be filled with water, so that the leaves of
the branch are completely
submerged, it will be found
that, notwithstanding the
presence of the water in
contact with the leaves,
and the hydrostatic pres-
sure due to its depth, the
eosin will mount rapidly
into the branch.

In some of my experi-
ments the pressure of the
water was sufficient to
drive liquid back into the
intercellular spaces of the
leaves of the branch. So
that it appears that the pumping action can raise water
against a considerable external hydrostatic pressure.

In carrying out this experiment, of course, care must
be taken that the gas-pressure in the branch has become
equalised with that of the atmosphere. With this pre-
caution, however, the result seems conclusive, i.e., that
secretory actions, and not evaporation, cause the rise of
the eosin into the branch.

It will be found that, if the water in the receiver is warm
(25-30 C), and if the apparatus is placed in a strong
light, the ascent of the eosin will be rapid ; if, on the other
hand, the water is cold (below 12 C.) and the light is not
strong, the eosin will rise but slowly in the branch. If
the apparatus is placed in darkness, the eosin will rise but
little or not at all.

It seems probable that the increased rate is, in part, due
to the quickening of the vital processes caused by the
rise in temperature when the water surrounding the
leaves is warm.

24 TRANSPIRATION AND ASCENT OF SAP ch.

The stimulating action of the light is indirect, and pro-
bably is effective by the increased supply of oxygen set free
by assimilation. The upward movement of the eosin is most
rapid when bubbles of oxygen are being evolved in quan-
tity at the surface of the leaves. This observation, then,
constitutes another proof that the lifting action is due to
a vital process, and decreases when the supply of oxygen
is diminished. In this respect the action resembles other
vital phenomena, such as growth, irritability, etc. The fact
that a small rise does take place in the dark is explained
by the presence of oxygen in the water, and also of that
derived by intra-molecular respiration.

The combination of this oxygen will of course lead to a
minute rise in temperature which will favour a distilla-
tion of water from the leaves. This effect, however, would
probably be so small that it could not account for the
rapid rise of water in plants in a saturated space, as has
been just described.

Over short periods osmosis may, however, be adequate
to draw up water into leaves even though they are sur-
rounded by a saturated space ; but this seems only possible
if we start with the leaf-cells incompletely distended.
Until they are fully turgid they will absorb water on every
side, and naturally some of this will come from the tracheae,
while some may at the same time be derived from water
in contact with their outer surfaces.

Summary. From what has been here detailed, I
think we may with great confidence assert that the
elevation of the sap, when plants are situated in satu-
rated places, is effected by directed actions taking place
in the living cells of the leaves. Simple osmotic and
evaporative forces cannot be continually effective in
raising the water in the conduits in these circum-
stances. With regard to the elevation of water, when
the leaves are surrounded by an unsaturated atmosphere,
we cannot as yet be dogmatic. But the fact that,

i THE NATURE OF TRANSPIRATION 25

when the leaves of plants are killed, they dry up and
are unable to furnish themselves with sufficient water
from an unlimited supply at the base of their stem, argues
that surface tension and evaporation forces at their sur-
faces are in themselves inadequate. And, when we couple
with this, the observations on the directed vital actions
taking place in the leaf-cells when they are surrounded with
a saturated atmosphere, I think we may, with great pro-
bability, assume that these directed vital actions are
responsible to a great extent for the raising of water in
plants even in unsaturated spaces. In any case the present
evidence shows that directed vital actions are capable
of replacing and supplementing the more simple physical
actions, e.g., evaporation, capillarity, and osmosis.

When, however, the pressure of water vapour is further
reduced and evaporation is so rapid that secretion cannot
keep the supply equal to the demand, then it must be sup-
posed that capillarity (imbibition) and evaporation draw
the water to the surface of the cells. Whether this state
of affairs can be prolonged indefinitely without injury to
the leaves is not at present known.

Literature.

Brown, H. T., and Escombe, F., "Static Diffusion of Gases and Liquids
in Relation to the Assimilation of Carbon and Translocation in Plants,"
Phil. Trans. Roy. Soc. London, vol. 193, B., p. 223.

Dixon, H. H., "Note on the Role of Osmosis in Transpiration," Proc.
Roy. Irish Acad., vol. iv, ser. 3, p. 61, and Notes from the Botanical School
of Trinity College, Dublin, vol. i, p. 35.

Id. " On the Effects of Stimulative and Anesthetic Gases on
Transpiration," Proc. Roy. Irish Acad., vol. iv, ser. 3, p. 618, and Notes
from the Botanical School of Trinity College, Dublin, vol. i, p. 97.

Id., "Transpiration into a Saturated Atmosphere," Proc. Roy. Irish
Acad., vol. iv, ser. 3, p. 626, and Notes from the Botanical School of Trinity
College, Dublin, vol. i, p. 106.

Id., "On the Physics of the Transpiration Current," Notes from the
Botanical School of Trinity College, Dublin, vol. i, p. 57.

Id., "A Transpiration Model," Proc. Roy. Dublin Soc, 1903, vol. x
(N. S.), p. 114, and Notes from the Botanical School of Trinity College, Dublin,
vol. i, p. 217.

26 TRANSPIRATION AND ASCENT OF SAP ch. i

Id., "Note on the Supply of Water to the Leaves on a dead Branch,"
Proc. Roy. Dublin Sue, 1905, vol. xi (K S.), p. 7.

Id., "Transpiration and the Ascent of Sap," Progressus Rei Botanicae,
Bd. iii. s. 1.

Dixon, H. H., and Joly, J., "The Path of the Transpiration Current,"
Ann. ofBot, 1895, p. 403.

Henslow, G., " Origin of Plant Structures."

Joly, J., "Contribution to a Discussion on the Ascent of Water in
Trees," Brit. Assoc. Report, 1896 and Ann. of Bot, 1896, vol. x, p. 647.

Strasburger, E., " Ueber den Bau und Verrichtungen der Leitungs-
bahnen in den Pflanzen" (Jena), 1891.

linger, F., "Neue Untersuchungen tiber die Transpiration der
Vftnnzen," Sitzuivjsb. d. Wien. Akad., 1861, Bd. 44, p. 362, and But. ZUj.,
1872, p. 62.

CHAPTER II

ASCENT OF SAP IN STEMS. CRITICISM OF PHYSICAL

THEORIES

Early writers. Since the ringing experiment of
Hales (1727) 1 and the experiments of Magnol and De la
Baisse, who about the same time supplied cut branches of
plants with coloured fluids and thus mapped out the con-
ducting tracts, physiologists have been agreed that the
upward movement of water from the roots takes place in
the woody tissues of plants. In contrast to this unanimity
concerning the path of the upward current are the very
divergent views which are held as to the nature of the
process by which the water is raised.

Of the views of the earlier writers it is hard to obtain
a clear conception ; their point of view was so utterly
different from that of the present day. Much of their
work is vitiated by the fact that they constrained them-
selves to see in plants a circulation of fluids similar to
the circulation of the blood of animals. Little attention
was paid to the forces causing this circulation.

It is true that Christian Wolff (1723) believed that the
forces involved were the expansion of air and capillarity ;

1 Ic is remarkable that this classic experiment was not devised by Hales
to trace the upward path of the water current, but to prove that there is no
circulation of sap in trees comparable to that of the blood of animals.
Indeed, from Hales's own account, it appears that he thought the upward
movement of water was slightly interfered with by the ringing..

27

28 TRANSPIRATION AND ASCENT OF SAP ch.

and Hales, a little later, attributed the rise of water in
the vessels to capillarity. These forces have ever since
been called in again and again to play their part in the
various theories formed to account for the ' circulation
of juices," or as it is called in more recent times, the Ascent
of Sap in Plants.

Gas pressure theory. Direct descendants of these
theories were the air-pressure and the gas-pressure theories
of Bohm and Hartig, and these, like their predecessors,
were soon rendered untenable by quantitative examination.

Jamin's chain. Nor did the comparison of the dis-
tribution of the water and gas in the plant to a Jamin's
chain avail to save these doomed hypotheses, for it was
early recognised that any support rendered to the water
columns by this configuration must at the same time act
as a resistance to upward motion.

A recent attempt to rehabilitate the Jamin's chain hypo-
thesis, or rather a modification of it, was not more happy.
It was suggested that the peculiar intermixture of air and
water in the conducting tracts brought into play obscure
physical forces which lead to the elevation of the water.
It has been shown, however, that the experiments upon
which the view was based, were completely vitiated by
the neglect of a property of plaster of Paris which enables
it to continue to absorb water long after it has set. This
absorption gave rise to an upward movement of water in
the experiment, which was erroneously believed to demon-
strate that the ' suction ' of less than one atmosphere
applied to the top of a continuous column of water sus-
pended in a porous substance with air bubbles intermingled
could still operate as a suction more than 12 metres lower
down.

Boehm's views. It will be understood that all that
was needed to refute the foregoing theories was a clear
statement of the theories themselves and a quantitative
estimate of the forces they asserted were adequate to

ii ASCENT OF SAP IN STEMS 29

lift the water columns in trees. Such a clear statement
is, however, not possible in the case of Bohm's final
theory to account for the ascent of sap. Owing to
his contradictory expressions and obscurity in description
it will always remain impossible clearly to understand
what his hypothesis was. He assigned a part to the
capillary forces of the tracheae and to atmospheric
pressure. The latter he conceived as driving the water
from the tracheae into the leaf-cells, and also he saw no
difficulty in the height of the water columns owing to the
cohesion of water ; but without doubt he was quite astray
as to the conditions under which cohesion could act.

Function of the tracheal walls. With Sach's
imbibition hypothesis, a new factor was introduced into
the discussion on the ascent of sap. The assumed
mobility of water in the tracheal walls was directly
negatived by the experiments of Elfving and Vesque,
who showed that only negligible amounts of water
passed upwards when the lumina of the tracheae were
plugged with cacaobutter. It was objected to Elfving's
experiments that cacaobutter was of too greasy a nature,
and might enter the wall, so gelatine was substituted by
Errera and Strasburger. It might fairly be urged, how-
ever, that there is less danger of a greasy substance enter-
ing the water-saturated wall than of a substance miscible
in water such as gelatine, which, even if entering the wall
in minute quantities, might be very injurious to its trans-
mitting properties.

Penetration of gelatine into walls. In this
connection experiments were made to test the possibility
of gelatine permeating the cell walls. A few of these
may be quoted. They all agreed in showing the passage
of warm dilute gelatine through the cell-walls possibly
in some cases through the closing membranes of the
pits only and consequently suggest the probability that
it would alter the capability of the wall for transmitting

30 TRANSPIRATION AND ASCENT OF SAP ch.

water. A length of 10 cm., straight and free from side
branches, was cut from a branch of Taxus baccata, the mean
diameter being 25 cm. This was deprived of its bark, and
affixed by an india-rubber ring at one end to a glass tube
communicating with an air-pump ; a little water in the

tube covered the upper end of the
wood. (Fig. 4 shows the arrange-
ments.) On exhausting the tube,
bubbles rose from the surface of
the wood. These could be stopped
by simply immersing the lower
end in mercury. Hence it was
concluded that continuous air-
passages existed in this piece of
wood, which should be stopped
before any tests could be made
as to its permeability by gelatine.
Accordingly the lower end was
dipped in melted paraffin at about
70 C, the melting point of the paraffin being 56, and the
whole length of the stick jacketed by water which was
maintained at 70 for 45 minutes, a vacuum being pre-
served in the tube attached to its upper end during this
time. Finally the stick was cooled slowly from above
downwards by lowering the water-bath to allow of the con-
traction of the paraffin being made good by supply from
below. When all was cold, the end was pared to expose
lumina free from paraffin. The stick now drew up water
freely, 3 or 4 c.c. in 15 minutes, but allowed neither air
nor mercury to pass up. The water pumped through was
next tested by a solution of tannin, but remained perfectly
clear. We conclude therefore that no direct air-passages
remain open, and that nothing is yielded by the wood of
the yew giving an obscuring reaction with tannin.

Some gelatine which had been cut up into fine threads
and soaked in repeatedly changed water for two days

Fig. 4.

ii ASCENT OF SAP IN STEMS 31

was now melted and diluted till it set weakly at 13. At
a temperature of 30 to 40 this was supplied to the lower
end of the yew, the latter as before being kept warm
throughout its entire length by a water-jacket which was
never raised above 40. At the expiration of four hours
the liquid within the vacuous tube had risen by about
5 c.c. The experiment was then stopped, and the contents
of the tube tested with tannin. There was an opalescent
precipitate. Comparison with the solution below showed
that much of the gelatine had been held back by the wood.

Starting the experiment a second time with the same
piece of yew, it transmitted 3*5 c.c. in four hours ; the
liquid drawn up affording this time a much denser preci-
pitate. A final test showed the wood to be still impervious
to air when a vacuum was maintained in the tube.

A similar experiment with the wood of Pinus austriaca
gave a like result. It was observable that if the dilute
gelatine was not raised some few degrees above its melting-
point i.e., till the solution almost ceases to be opalescent
its passage was much less marked ; indeed in some
experiments only traces were transmitted through the wood.
This appears to be due to the fact that in solutions pre-
senting an opalescent or milky appearance, the gelatine is
probably still in the solid or gelatinous state ; the hetero-
geneous distribution and difference of refractive index
giving rise to the milky colour. In all cases a considerable
quantity of the gelatine is held back. One quantitative ex-
periment on Taxus gave the percentage of gelatine in the
transmitted liquid as only half that in the original solution.

In one experiment the gelatine was stained with Klein-
enberg's hematoxylin. The solution was made of such
strength as to set at about 20, and was supplied at 40 to
the wood of Taxus baccata. It passed out colourless into
the glass tube, about 1 c.c. in two hours, the length of the
wood traversed being 2 5 cm., and its cross-section 22
sq. cm. This wood had not been treated with paraffin,

32 TRANSPIRATION AND ASCENT OF SAP ch.

as it revealed no direct air-passages upon trial. As the
hematoxylin does not stain wood, this experiment points
to a mechanical separation from the gelatine owing to the
passage of the latter through membranes or walls. It is
possible, however, that some of the stain was taken up by
the cellulose walls of the medullary rays and the tori of
the pit membranes.

Microscopical examination of branches choked with gela-
tine mixed with Indian ink, after the manner of Errera
and Strasburger, showed that the closing membranes of
the pit had exerted a straining action, accumulating Indian
ink upon the one side, so that the pits were picked out
very sharply as black objects. This filtering action is
suggestive of the passage of the medium carrying the
precipitate ; and although, so far as this observation
is concerned, there might have been nitration of the
gelatine from the water in which it was dissolved, still,
taken in conjunction with the other observations, we
think it supports the view led to by those observations, i.e.,
that dilute melted gelatine can pass through the substance
of the closing membranes, and, if so, is very probably
capable of penetrating into the cell-wall, or otherwise we
must suppose perforations to exist in the pit-membrane
or its torus.

Effects of paraffin and gelatine compared.
The effect of using paraffin wax of low melting-point as
the material for choking the lumina was also tried and
compared with that of gelatine. Four similar branches of
lime, Tilia europaea, were cut (May 9), and put standing
for twenty minutes in water at 50 C, immersed to a
depth of about 20 cm. These were called A, B, C, D.

A was preserved in water at 50.

B was transferred to melted paraffin at 50 (melting-
point 48).

C was transferred to gelatine coloured with Indian ink
at 50.

ii ASCENT OF SAP IN STEMS 33

D was transferred to gelatine coloured with haematoxylin
at 50.
Each was immersed to a depth of 20 cm. and placed in
bright light, the air temperature being 16. At the expira-
tion of forty minutes all were transferred to water at 13.
Then the end of each was thinly pared, and at 5.30 p.m.
all were left finally standing in water at 13. At 6.30 all
were fresh. At 11 a.m. on the 10th, i.e., after 15j hours,

A was still quite fresh,

B ,, very much flagged,

C ,, less flagged than B,

D ,, ,, ,, B, but more nagged than 0.

All were now transferred to a strong solution of saffranine,
and put in full sunshine for lj hours, when they were
washed and sections made for microscopical examination.
So far as C and D were concerned, it is only necessary to
observe that they revealed that only some of the lumina
were actually stopped with gelatine. The walls of many of
the gelatine -filled vessels were found stained with saffranine,
which attained to 26 cm. in C, and to 5 cm. in D. The gela-
tine in the lumina had become coloured with the stain.

Transverse sections of B close to the base showed all
lumina choked with paraffin, while the walls between were
deeply stained with the saffranine.

In polarised light with crossed Nicols the appearance was
very striking, the crystalline paraffin showing out strongly.
Transverse sections, 2 cm. from the end, showed the large
vessels still filled with paraffin. In some places neigh-
bouring vessels apparently quite filled with paraffin had
the intervening walls deeply stained ; at this level, however,
where the vessels were filled with paraffin the staining was
not quite so dark as elsewhere, though the colour was still
strong. The paraffin finally attained a height of 12 cm. in
one or two vessels. In no case was there any visible appear-
ance of shrinkage of the paraffin from the wall, although
in some sections, as might be expected, the action of the

D

34 TRANSPIRATION AND ASCENT OF SAP ch.

razor was to compress it from the cell-wall upon the one
side over the section.

Similar experiments were made on elm and lime, with
the added precaution of removing the paraffin or gelatine
at the ends without cutting or removing any of the wood.
This was effected by careful use of the razor, the object
being to avoid as far as possible laying open the lumina
of conduits the terminal walls of which might lie upon the
surface of the section. In the case of elm and lime again,
sections taken about half a millimetre from the end showed
areas over which the filling with paraffin was complete, and
yet deep staining of the intervening walls. Longitudinal
sections near the end confirmed this appearance ; the
lumina seemed quite filled. In these cases the removal
of the branches from the hot paraffin was effected gradually,
to secure, so far as possible, that solidification and shrink-
age should proceed slowly from above downwards, and
thus guard against shrinkage leading to the withdrawal
of the paraffin out of contact with the wall. Again, the
branches of lime treated for comparison with gelatine
revealed areas in the cross-sections completely injected
with gelatine and having the walls deeply stained. Thus
we see that both in those experiments in which the lumina
were choked with paraffin and in those in which the
lumina were choked with gelatine there was at least
a feeble upward motion of the solution of saffranine
in the walls. Lime branches treated with paraffin, in
some places close to the cut surface, showed the pene-
tration of this into the protoplasm-filled cells, permeating
their contents. High up, only the larger vessels were
filled with paraffin.

The result of these experiments may be summed up as
follows :

The stoppage of the lumina and the freedom of the cell-
wall is preserved by the use of paraffin, and, possibly, by
th at of gelatine.

II

ASCENT OF SAP IN STEMS

3S

The flagging of the leaves appears to be the more rapid
the more completely the closing of the lumina has been
effected.

When the lumen is closed an upward passage of liquid is
still maintained in the wall, but this is probably much too
feeble to meet the wants of the leaves.

Paraffin casts of tracheae. Owing to its extreme
mobility the penetration of the paraffin is very complete in
these experiments. Thus it was found easy by its means
to demonstrate the continuity of the tracheal elements
forming the vessels in lime, sycamore, and elm. For this

id

/y - : //' ..V

%m > U\Vt

V\W >

&j!

Paraffin Casts of Vessels.

Fig. 5. Tin a

MICROPHYLLA.

Fig. 6. Tima
microphylla.

7.

Fig

Ulmus campestris.

lengths of 35 cm. were used and the wood of the branches
injected as described above was removed with sulphuric
acid. It is necessary to anchor the branch by a leaden
weight in a deep vessel of the acid. A single night suffices
in many cases to remove the wood and leave the paraffin

r> 2

26 TRANSPIRATION AND ASCENT OF SAP ch.

casts of the vessels streaming upwards from below like a
sheaf of fine white threads. The examination of these
threads under the microscope reveals many features of
interest. Figs. 5, 6 and 7 represent portions of some
of these casts.

Some further experiments were made bearing upon the
ascent of water in the wall. All confirm the fact that an
appreciable quantity of water ascends in branches most
carefully choked with paraffin.

Thus, while flagging will inevitably overtake a paraffined
branch left furnished with the same number of leaves as
it bore upon the tree, yet if the greater number of these are
removed, the remaining leaves will generally hold out fairly
well. This experiment was tried with a control paraffined
branch upon which all the leaves were left standing.

If after injection we remove part of the branch at a fork
and, keeping the one part which is attached to the
paraffined extremity in water, insert the extremity of the
other through a cork into a dry vessel, the latter will flag
much the more rapidly. Still more direct is the following : a
paraffin-injected branch of Tilia micropkylla, with nine
leaves, was put standing, from 4.15 p.m., May 11, till noon
on the 12th, in a vessel of water which had been carefully
weighed and so closely corked round the stem as to preclude
possibility of loss by evaporation at its surface. In this
period of nearly twenty hours the branch drew up 1*005
grammes of water. This same branch, after it had flagged,
and had been put out into breeze and intermittent
sunshine from noon till 3.30 p.m., drew up 0*161 gramme.

Again, of two paraffin-injected lime branches, one
scraped to free the surface and placed in water, the other
left closed with its cap of solid paraffin ; the latter flagged
much more quickly, although it bore a smaller number of
leaves. In two days the second was, indeed, dry and
shrivelled, while the former had preserved much of the
freshness of its leaves.

ii ASCENT OF SAP IN STEMS 37

Generation of gas in the lumen. Bearing on
this same point the partial passage of water through
the walls the following experiments were carried out,
in which it was sought to replace the paraffin or
gelatine by a gas developed in the plant. Thus a cut
branch first supplied from a solution of tartaric acid and
subsequently from a solution of sodium bicarbonate will
have carbon dioxide evolved in the lumina of its conducting
tissues in consequence of the interaction of these substances.

A preliminary experiment upon a lime branch (Tilia
microphylla) which had stood for two hours in a solution of
tartaric acid, and then one hour in sodium bicarbonate,
before finally being transferred to pure water, showed rapid
flagging of its leaves and soft shoots as the result. But
as this was possibly a direct consequence of the action of
the reagents, and not of the evolved gas, a more careful
experiment was carried out upon five branches of elm cut
from the same tree, with similar precautions, and, as far
as possible, of like dimensions.

A and B were placed in sodium bicarbonate solution.

C and D were placed in tartaric acid solution.

E was placed in a solution of a mixture of tartaric acid
and sodium bicarbonate which had ceased effervescing.

After 1 J hours A and C were interchanged in the solutions;
thus, in these two only was carbon dioxide developed.
B served as a control regarding the effects of sodium
bicarbonate alone, D, as a control for tartaric acid, E, for
the effect of the mixed solution without development of gas.
In five hours A and C were very much, and about equally,
flagged, while the rest remained fresh. Next morning,
however, all had drooped, showing that prolonged treat-
ment with either or both of these substances is injurious in
any case. It was evident, also, that the stoppage of the
lumina by the gas had greatly accelerated the flagging.

It was not probable that the check upon aeration of
the tissues involved in all of the foregoing experiments

38 TRANSPIRATION AND ASCENT OF SAP ch.

wherein the lumen is choked, could account for so rapid
a flagging of the soft parts. However, to set this doubt
at rest, we inserted branches in water which had been
boiled and cooled in vacuo, and coated with oil after the
insertion of the branch. These, however, remained per-
fectly fresh ; indeed, they seemed in no way affected.

Lumen blocked with ice. In order to investigate
this subject more fully, an additional series of experiments
was devised and carried out upon the passage of water
through the wood of Taxus at low and at high temperatures.
For it was very certain that in the one case the formation of
ice, and, in the other, the formation of steam, would occur
in the lumen before occurring in the wall, rendering the
former non-conducting without the introduction of any
foreign substance.

It was necessary to determine first of all the freezing-
point of water in the lumen by direct microscopic observa-
tion. To effect this a special form of cold stage was used,
the construction of which will be easily understood from
Fig. 8. In this stage the object under examination is

Fig. 8.

completely surrounded by the cooling liquid, which also
flows round the bulb of the thermometer, t. The tempera-
ture is, therefore, accurately known. The bottom of
the cell is of glass ; a ring screwing out upon the top
serves to permit the lifting of a cover-glass acting as a
water-tight window, this being luted on the edge with a
little white lead. The object is luted between two cover-
glasses, and carried upon an open support within. It is
necessary to protect the upper window from moisture pre-

ii ASCENT OF SAP IN STEMS 39

cipitated from the atmosphere ; this is done by the loose
metal ring surrounding the object-glass, and packed round
with a little cotton wool. The thermometer enters by a
tubulure in front ; its bulb appears in cross-section at t in
the figure. The regulation of the temperature is very
simply effected by retarding or accelerating the current of
cold liquid (brine) by means of the pinch-cock.

The section of the wood to be examined is cut, and
with the addition of as little water as possible, is luted up
between the cover-glasses, so that it is surrounded by air,
but contains water within its substance. The close
proximity of the section to the upper window, some
T5 millimetres, allows of considerable magnification.

The cold cell, after the introduction into it of the section
sealed up between the cover-glasses, is placed on the stage
of a microscope, and then, by the arrangement already
described, the temperature is caused to fall gradually,
while the water within the section is carefully observed.

The phenomena attending freezing were perfectly definite,
the clear liquid in the lumina assuming the aspect of solid
paraffin. In two experiments in which the reduction of
temperature was effected very gradually, the freezing-
point was found to lie between - 10 and - 11 \ Freezing
spread with great rapidity all over the field in both wide
and narrow lumina. Air-bubbles present exhibited im-
mediate reduction of volume, and often distortions of
shape, and it was important to observe that an exudation
of sap occurred upon bare cell-walls, which, appearing in
drops, instantly turned to rough-shaped ice-crystals.

1 Of course, this is not the true freezing-point of the liquid in the wood,
but the temperature which it attains by supercooling before solidification takes
place. It thaws consequently at a higher temperature, which again is not the
true freezing point, but one which approximates to the eutectic point. It is
evident that much variation may be expected in the temperature at which
crystallisation occurs in the supercooled liquid.

The true freezing-point of the sap in the tracheae of other plants is much
higher than either of these temperatures, and there is no reason to believe
that the sap of Taxus is peculiar in this respect (see page 45).

4 o TRANSPIRATION AND ASCENT OF SAP ch,

Thawing occurred at a higher temperature than freezing,
no signs of melting being exhibited till 4 or 5 was
reached. This specimen of wood was removed from a
branch which had been standing some days in water. A
freshly cut branch of Taxus afforded - 14'5 as the freezing-
point.

Owing to the pressure-effect of the ice upon the wall,
visibly shown by the forcible expression of drops, there
appeared some doubt whether this method would afford
any result of value. However, the experiments were per-
sisted in, and a length of 22 cm. by 6 mm. in diameter of
a yew-twig, carefully washed, was attached to the appa-
ratus shown in Fig. 9, in which the passage of liquid

Fig. 9.

through the vertically placed stick (due to a diminished
air-pressure in the vessel above) is shown by the move-
ment of mercury in the horizontal capillary tube. The
rate of transmission of water was observed while the tem-
perature of the jacket was varied. The general results
were as follows :

In cooling, the current had almost ceased at 7 and,
completely at 11; in ivarming, it recommences feebly
at 5. It was impossible to fix upon any temperature
as the actual freezing temperature in the lumina from the
observations, but as all current had ceased at 11, at
which temperature the water in the walls was almost
certainly not frozen, we must conclude that these observa-
tions reveal no current in the walls, even of the feeblest

II

ASCENT OF SAP IN STEMS

4i

intensity, for the method of observation is very delicate.
However, the method is beset by the doubt involved in
the evident ice-pressure upon the walls. The large
increase in the viscosity of water at the lower temperatures
would also greatly reduce or stop the flow.

Lumen blocked with vapour. Experiments in
which the wood of Taxus was exposed to high tem-
peratures above 100- appear to show that water, carry-
ing a dye in solution, can be drawn through the wood
when this is at a temperature so high as 125, and
very certainly filled with water-vapour everywhere in
its lumina.

Fig. 10 shows and explains the arrangement of the
experiment. The vessel into
which the branch dips con-
tains mercury heated from
beneath. A glass tube sur-
rounds the branch, the space
between branch and glass
being filled with mercury.
To resist the pressure of the
vapour evolved from the
surface of the wood at this
temperature, it was necessary
to bind the stick into the tube with air-tight rubber
rings overlaid with wire. The following experiment was
made :

A small branch of Taxus baccata, 24 cm. long, having
a woody cylinder of 5-6 mm. in diameter, and com-
posed of nine annual rings, was jacketed with mer-
cury at 125-130 for eight minutes, while its basal
end was attached to an air-pump so that the atmospheric
pressure forced through water supplied to the distal end
of the branch. The water was then replaced by a strong
solution of eosin, and the whole, still kept at 125-
130, was left for two hours. Then the experiment was

Fi.;. 10.

42 TRANSPIRATION AND ASCENT OF SAP ch.

stopped. The eosin being first removed, the surface to
which it had been applied was pared and dried. The
branch was then detached from the air-pump and allowed
to cool. On microscopic examination it was found that
the eosin-solution had passed 22 cm. up the wood, and at
this height was seen in cross-section as two irregular patches
occupying quadrants in the seventh and eighth rings. The
walls of these were uniformly coloured. At the level of
the mercury jacket, and throughout the 7 cm., where the
branch was immersed in mercury, the colouring was most
intense in the limiting membranes. At the end where the
eosin was applied, the walls were scarcely coloured, except
those adjoining the medullary rays and immediately round
the bordered pits.

Small transmission in the walls. The simplest
interpretation of these results is that the coloured water
moved in the wall, while the lumen was occupied with
vapour ; the intenser coloration of the limiting membrane
is strongly in support of this view, for it is very probable
that for some distance from its surface the wall was so
far choked with vapour as to impede the motion of a
liquid.

These experiments then, so far as they go, are in perfect
agreement with the previous set in which the lumina are
choked by the introduction of foreign substances (cacao-
butter, gelatine, air, in the experiments of other authors,
or by paraffin and carbon dioxide in our own) ; and they
show that the freedom of the lumina is necessary for the
rapid transmission of water, but that a slow current may
pass through the walls even when the lumina are completely
blocked.

Negligible amounts transmitted as vapour.
There appeared the possibility that the nagging of the
branches having closed lumina might be due to the stoppage
of them as vapour-conduits, and not as water-conduits ;
that is, the experiments were not yet conclusive as to the

ii ASCENT OF SAP IN STEMS 43

actual function of the lumina, although showing clearly
that their freedom is essential to preserve the turgescence
of the leaves. The well-known phenomenon of the
equilibrium vapour-pressure varying with the curvature
of the meniscus suggested the possibility that a trans-
port of vapour of considerable importance might occur
in the conduits, the menisci high up in the trees possess-
ing a lower equilibrium vapour-pressure than the menisci
lower down. By successive condensations beneath and
evaporations above the pit-membranes, this current might
be maintained throughout the conduits unoccupied by
liquid water.

This idea led to experiments in which cut branches were
fed entirely upon water- vapour in the following manner :
The branch had its cut extremity fixed in a short glass
vessel containing water at the bottom ; and the cut surface of
the wood (which was cut at an acute angle in order to expose
a larger surface) was raised some 5 or 7 cm. above the
surface of the water. A side tubulure to the vessel enabled
a vacuum to be maintained within by means of a Sprengel
pump. The vacuum was so complete that ebullition
occurred upon placing the hand round the lower part of
the vessel. Such experiments were made upon elm and
lime, using control branches, some of which were simply
left with their cut surfaces exposed to the air, others with
their ends sealed into tubes containing air, but no liquid
water. In no case was any result. obtained going to show
that the vapour-fed branch possessed any advantage over
the others.

It appears then that the movement upwards of water in
the form of vapour through the lumina is insignificant, and
the imbibition theory cannot receive effective help from
this direction.

It is needless now to go into the many arguments which
overthrew the imbibition hypothesis. It is enough to say
that the ingenuity of the theory, and the reputation of

44 TRANSPIRATION AND ASCENT OF SAP ch.

its elaborator and defender, made it survive an incredibly
long time, despite the accumulation of crushing evidence
from structure, and of clear inference from experiment.

Osmotic hypothesis. Various writers have endea-
voured to explain the rise of the transpiration stream by
appealing to the action of osmotic phenomena. In
almost all the vital theories the hypothetical pumping
actions of the cells are supposed to be effected by osmosis.
Consequently, the osmotic hypothesis, so far as the osmotic
actions in question are supposed to be manifested in con-
nection with the cells of the wood, has been discussed
with those theories. It is, in fact, hard to see how osmotic
pressures can be generated in the conducting tracts apart
from these cells. In the cells only are to be found semi-
permeable or approximately semipermeable membranes.
Larmor's suggestion that an upward movement is deter-
mined by a gradient of concentration in the solutions
contained in the tracheas from below upwards, could only
apply to an ideal state of things, and is negatived by
the facts : (1) The membranes of the tracheae are freely
permeable to dissolved substances ; (2) no such differences
in concentration are found ; (3) the resistance to flow
upwards and downwards is the same.

With regard to the second objection enumerated above,
some recent experiments on the concentration of the sap
in the roots and in the stem at various levels are of par-
ticular interest.

It has been found possible to extract the sap from the
wood of roots and stems of transpiring trees by centri-
fuging short lengths cut from these organs. Considerable
quantities of the sap, quite unaltered, may be obtained in
this manner, and the molecular concentration of the solutes
in them accurately measured by means of cryoscopic and
electrical conductivity determinations. In each case it has
been found that the concentration of the sap is sensibly less

II

ASCENT OF SAP IN STEMS

45

above than in the lower parts of the tree. The record
of a few measurements will illustrate this. Under A in
Table 3 is given the depression of the freezing-point, which
is proportional to the concentration of the total dissolved
contents of the sap ; while under C is recorded the con-
ductivities expressed as reciprocals of the resistances
measured in ohms. These give a measure of the content of
electrolytes.

Table 3.
Showing the concentrations of Wood-sap in Stems and Roots.

A.

C x 10. 6

,, Stem at 30 ft )
level .... J

Populus alba : Root

,, ,, Stem at 40 ft. level . .

,, ,, Stem at 3 ft. level

0-070 r C.

049

0-072
047

0-099 3
0-082

693

410

518
339

Literature.

Boehm, J., "Ueber die Ursache des Saftsteigens in den Pflanzen,"
Sitzungsb. d. Alcad. d. Wiss. in Wien, 1863, Bd. 47.

Id., "Wirddas Saftsteigen in den Pflanzen durch Diffusion, Capillaritat
oder durch Luftdruck bewirkt"? Sitzungsb. d. A lead. d. Wiss in Wein, 1864,
Bd. 50.

Id., "Les causes de l'ascension de la seve," Ann. des Sciences Nat. Bot.,
1878, vi, 6.

Id., "Ueber die Ursache der Wasserbewegung und der geringen Luft-
tension in transpirirenden Pflanzen," Bot. Ztg., 1881, s. 801.

Id , " De la cause de mouvement de l'eau et de la faible pression de l'air
dans les plantes," Ann. des Sciences IS r at. Bot., 1881, vi, 12.

Copeland, E. B., "The Rise of the Transpiration Stream," Bot. Gazette,
1902, pp. 161 and 260.

Dixon, IT. H., "The Cohesion Theory of the Ascent of Sap," Proc. Roy.
Dublin Soc, 1903, vol. 10 (N. S.) p. 48, and Notes from the Botanical
School of Trinity College, Dublin, vol. i, p. 203.

Id., "Transpiration and the Ascent of Sap," Progressus Bei. Bot., 1909,
Bd. Ill, s. 1.

Dixon, H. H. and Joly, J., "The Path of the Transpiration Current,"
Ann. of Bot., 1895, ix, p. 403.

Elfving, F., " Ueber die Wasserleitung im Holz," Bot. Ztg., 1882, s. 707.

46 TRANSPIRATION AND ASCENT OF SAP ch. ii

Errera, L., " Une experience sur l'ascension de la seve chez les plantes,"
Compt. rend, de la Soc. Rot/, de Bot. de Belgique, 1886, Bull, xxv, 2, p. 28.

Hales, S., " Vegetable Staticks," London, 1769.

Hartig, R., " Ueber die Vertheilung der organischen Substanzes, des
Wassers und des Luftraumes in den Biiumen, und iiber die Ursache der
Wasserbewegung in transpirirenden Pflanzen," Untersuchungen d. Forst. />'<</.
Ind. Miinchen, II. Berlin, 1882.

Id., "Die Gasdrucktheorie," (Berlin) 1883.

Id., ' ; Die Wasserbewegung in den Pflanzen," But. Zt<j., 1883, Bd. 41,
s. 250.

Id., " Holzuntersuchungen," Berlin, 1901.

Larmor, J., "Note on the Mechanics of the Ascent of Sap in Trees,"
Pioc. Roij. Soc, London, 1905, vol. 76 B, p. 460.

von Sachs, J., "Lectux-es on the Physiology of Plants," translated by H.
Marshall Ward, Oxford, 1887.

Strasburger, E., "Ueber den Bau und Verrichtungen der Leitungsbahnen
in den Pflanzen," Jena, 1891.

Vesque, V., " Recherches sur le mouvement de la seve ascendante," Ann.
des Sciences Nat. Bot., 1814, ser. vi, 19, p. 188.

CHAPTER III

ASCENT OF SAP IN STEMS. CRITICISM OF VITAL THEORIES

With the overthrow of the imbibition hypothesis, and
the establishment of the fact that the major part of the
transpiration current moves in the lumen of the vessels
and tracheids, the pendulum of opinion among botanists
swung over to the vital theory of the ascent of sap.

Godlewski. Thus in 1884 Godlewski formulated
a view which won many supporters, and, with modifica-
tions, still seems . to appeal to some. He assumed
a periodic change in the permeability of the osmotic
membranes of the parenchymatous cells contained
within the wood in order to bring about a pumping
action which would account for the raising of water
in the tracheae of the stem. Thus, supposing a cell
of a medullary ray in contact with eight tracheae-
four on each side to draw water into itself and to
increase its turgor so that its protoplasmic membrane
is considerably stretched, and assuming the osmotic pres-
sure of the cell and the resistance to filtration of the mem-
brane opposite to one tracheae to be periodically and
suddenly diminished owing to a chemical change, then it
is evident that the contractility of the protoplasm will
cause water to escape through the most permeable spot
of the membrane, viz., into the trachea opposite to which
filtration is most easy. Once in the tracheae, Godlewski

4 8 TRANSPIRATION AND ASCENT OF SAP ch.

assumed it to move upwards until it was drawn into a
medullary ray cell lying at a higher level in the stem.
The reason given for the motion upwards in the tracheae
rather than downwards in obedience to the gravitational
force, is because the air pressure in the tracheae above is
less than in those at lower levels. For this difference of
pressure in the air bubbles in the stem Godlewski relies
on Hartig's results. Hartig had, indeed, previously shown
that the percentage of air in the higher parts of stems is
less than that in the lower, but this does not necessarily in-
volve a lower pressure. Godlewski claimed for his hypothesis
that it explained the relations of the tracheae to the paren-
chymatous tissues, the radial position of the bordered pits,
which facilitates a staircase motion of the water upwards
in the stem, and the radial intercellular spaces along the
medullary rays, which afford the aeration necessary for
the respiratory liberation of energy in these cells.

Janse supported Godlewski in a general way, and was
one of the first to point out that, if the lower part of a
branch be killed, the leaves above fade, as a rule, within
a few days. This result is obtained even when the branch
remains attached to its supporting stem. Janse, at first,
assumed that this wilting is due to the interference with
the vital sap-raising functions of the wood parenchyma,
and the consequent failure in the water supply. He also
conceived that the pumping action of the medullary rays
is polarised, water being regularly absorbed on one side
and expelled on the other.

Westermaier, who stated his views about the same time
as Godlewski, also maintained the water-raising function
of the cells in the wood. But, while Godlewski believed
the major part of the motion to take place in the tracheal
tubes, Westermaier considered the upward passage to be
effected in the wood parenchyma, while the vessels and
tracheids acted as water reservoirs rather than as con-
ducting pipes. Godlewski contrasts his view and Wester-

in ASCENT OF SAP IN STEMS 49

maier's by stating that lie holds the wood parenchyma
cells to correspond to the piston, and the tracheae to the
tube of a pump, while the latter holds that the cells act
as tube and piston, and the tracheae are the reservoirs.
Furthermore, Westermaier assumes that the water is held
by capillarity in the tracheae in the form of short columns
of water alternating with columns of air.

It is evident that Godlewski and Westermaier saw the
difficulty of assuming that the water, once set free from
the cells, would move upwards rather than downwards.
Godlewski supposed that the difference in air pressure
above and below decided its movement. Westermaier
believed it is held by capillary forces in the tracheae at
the level to which it is raised by the cells.

According to the first assumption, however, we are
reduced to the theory that the whole upward motion is
due to a difference of air pressure above and below. Inas-
much as this difference is never so great as one atmosphere,
a rise of more than about 10 metres cannot be accounted for.

Neither could Westermaier's hypothesis be followed, in-
asmuch as it was known that the major part of the motion
of water upwards takes place in the tracheae and not in
the living cells, which, as a matter of fact, offer a great
resistance to the passage of water. It was also rendered
invalid owing to the fact that water is free to stream past
the bubbles, and that they do not offer support for the
water in the wood, but only increased resistance to its
downward motion. This was made evident by the " drop
experiment " ascribed to Sachs and Hartig, and also by
Schwendener's observation, that the bleeding sap does
not carry the bubbles with it from the tracheae. Wester-
maier's hypothesis also, as has been often pointed out, is
quite inapplicable to Conifers, for in their wood there is
no vertical connection of cells placing the medullary rays
of various levels in communication with one another.

Boucherie's experiment. One of the most

E

50 TRANSPIRATION AND ASCENT OF SAP ch.

remarkable features in the discussion of the problem
of the ascent of sap is that long before these vital
hypotheses were conceived, they had received decisive
experimental refutation. Boucherie, in 1840, while
experimenting on methods of injecting timber for
various technical ends, found that if a tree were cut across
at the base and supplied with a poisonous fluid, it not
only drew this fluid up to its highest leaves, but afterwards
would draw up a second solution when the latter was
supplied.

It is strange how the bearing of Boucherie 's experiments
appears to have escaped the notice of botanists completely,
and it may be noted that Biot, when commenting on
Boucherie's work, seems to find no special interest in it
in connection with the problem of the elevation of the
water of the transpiration current, but occupies himself
with other questions. Boucherie's experiments did not
wait long for confirmation. A few years afterwards
J. Schultz injected trees in the same manner, but his
results were also overlooked by botanists.

It appears, however, to have remained for fStrasburger
to point out the full significance of these experiments, and
to confirm them with many more of his own, carried out
with all possible precautions.

Strasburger further demonstrated the needlessness of the
vital hypotheses by experiments in which stems more
than 10'5 metres long, in a vertical position, continued to
draw up water after they had been completely killed by
exposure to a temperature of 90 C.

From these results it was abundantly proved that water
can rise, and has in these experiments risen, without the
assistance of the living cells of the stem; and, if forces
exerted by these cells do intervene in raising the water
in living plants, they are accessory to, and can only assist
the purely physical forces in play which are able to perform
the task unassisted.

in ASCENT OF SAP IN STEMS 51

Evidence from structure. The very structure of
the conducting wood of trees, far from supporting
Godlewski's contention that the cells assist in
elevating the transpiration stream, offers the strongest
evidence against it. These cells, placed as they are
beside the transmitting tubes, can, by their pumping
actions, in no way exert a lifting force on the stream.
The water, as experiment shows, is free to move down-
wards as well as upwards. Nor would a polarised or uni-
directional action of these cells, as hypothecated by Janse,
help, owing to their relation to the tubes. In that case,
these cells would regularly draw in water on one side and
expel it on the other into tubes, where it is exposed both
to downward forces and to resistance to upward motion.
In fact, to utilise the pumping action of living cells in
raising the transpiration current would require that the
continuity of the conducting tracts should be here and
there completely interrupted by the pumping cells ; but
if flow be possible in permeable tissues round these groups
of cells, pumping actions on the part of the cells will be
futile in assisting to raise water. As is well known, no
such interpolation of cells cutting the continuity of the
woody tissues is revealed by the most careful study of
the conducting tracts. The structure of the latter
tissues is, therefore, fatal alike to the earlier views of
Godlewski and Westermaier, and to the less precise form
under which they have been recently resuscitated by
Ursprung.

Exception has been taken to the statement that the
structure of the wood is against the vital hypothesis, and
it has been maintained that this statement is based on
ignorance of the resistance offered by the wood to the
nitration of water.

Assuming the resistance to be considerable, Janse be-
lieves that a pumping action in the medullary-ray-cells is
able to account for the rise of the water. This action,

e 2

52 TRANSPIRATION AND ASCENT OF SAP ch.

according to Janse, consists in an absorption of water
on one side of the cell and a giving off of it on the other,
a result which is effected, he suggests, by the harmonious
action of enzymes and protoplasmic streaming. An enzyme
fixes water in the cavities of the protoplasm on one side
of the cell, and streaming brings it to the other side, where
a reversing enzyme sets it free.

In order to get a clearer idea of the demands of this
theory, it is interesting to apply to it the figures already
obtained for the velocity of the transpiration current, and
those for the rate of protoplasmic streaming, and at the
same time to take into account the structure of stems.

Ewart has shown that the rate of upward movement
during transpiration in stems of Conifers is about 7 cm.
to 10 cm. per hour. Higher figures are sometimes
given. According to Janse, this movement must be
due to the difference between the actual amount raised
by the medullary-ray-cells and the leakage back by filtra-
tion. The pressure under which this filtration takes place
is the pressure of a head of water equal to the height of
the tree. As will appear later, this head causes a flow
approximately equal to the rate of the transpiration
current. Therefore the pumping action of the cells must
continually pass an amount of water which would, if
there was no leakage, be twice that of the transpiration
current. It is obvious, however, that the motion
through the cells must be much more rapid than this,
as they occupy but a fraction of the cross-section of the
wood.

A cross section of the wood of Pinus silvestris shows
that

The lumina of the medullary ray cells occupy 6 9% i r . ,1

,, ,, tracheids ,, 61*2% V ,

The walls of all 31'9%J cross sectlon -

I am indebted to Mr. W. R. G. Atkins for these
measurements. They were made by cutting out and

in ASCENT OF SAP IN STEMS 53

weighing accurate camera lucida tracings on uniform
paper of sections of the wood.

It follows that to produce the 7 cm. per hour velocity
observed, an actual rate of 14 cm. per hour must be main-
tained, while the velocity of the water through the cells
must be about nine times that, or 126 cm. per hour, i.e.,
about 2 cm. per minute. But this figure has still further
to be increased, inasmuch as at most half of the cells'
lumen can at any time be active in transferring the water
from the lower to the upper side. The other half will,
by hypothesis, be occupied by the return current of the
protoplasm. Thus we must assume a motion across the
cells of at least 4 cm. per minute. This will be still an
under-estimate of the velocity, for it is evident that only
a portion of the total protoplasm can be occupied by
water, and also only a portion of the lumen is occupied
by protoplasm. Therefore, the velocity of 4 cm. per
minute must be regarded as an under-estimate of the
velocity of the current in the protoplasm required by the
theory. It is needless to say that such a velocity of
protoplasmic streaming has never been observed. Janse
himself states the streaming in the endodermis cells of
Iris, which he believes to act in the same way, amounts
at a temperature of 19 C. to 30(V or 0*03 cm. per minute.
It appears, then, that this theory which attributes the
lifting force of the water in the stem to the protoplasmic
streaming of the medullary-ray-cells requires a rate of
protoplasmic streaming at least 100 times that hitherto
observed.

The adhesion of writers to the vital hypothesis since
Strasburger's results were published is so remarkable that
we must devote some space to examine fully the grounds
for their contention.

Transmission through dead stems. When a
considerable length of a branch, still attached to an
uninjured plant, is killed by surrounding it with steam or

54 TRANSPIRATION AND ASCENT OF SAP ch.

hot water, it is found that the leaves above the killed
portion sooner or later fade and wither. The vitalists'
interpretation of this observation is that, when the vital
actions of the wood parenchyma and medullary-ray-cells
are removed, the supply of water to the leaves above is
so reduced that they fade and dry from want of water.
It is significant that the discoverers of the phenomenon
saw in it no support of the vital hypothesis. Weber
recognises that the reduction of the water supply may be
attributed to the stoppage of the tracheal tubes. Janse,
although supporting Godlewski's vital theory, quickly saw
that this observation, being traceable to the blocking of
the supply tubes, could not be quoted in support of the
vital hypothesis. Finally, Vesque sees no support in it
for the vital theory. Ursprung, however, one of the
most recent champions of this hypothesis, relies on the
observation with great confidence to support his view, but
even his observations show that, out of some twentv
species in which the phenomenon was observed, in fourteen
stoppages after heating in the conducting tubes were actu-
ally observed usually above the heated region. Ursprung,
in spite of this, maintains that the reduction in the water
supply is not caused by stoppage because (1) the leaves
may begin to die before the stoppage of the tubes is ob-
served ; (2) even when stoppages do occur, fading may
be postponed ; (3) fading does not occur when a piece of
the wood is removed from an uninjured branch corre-
sponding in size to the plugged portion of the heated
branch.

Stoppages in the tracheae. It is to be noted
that the failure to observe stoppages microscopically
does not negative their existence. A transparent material
might form quite an effective plug in the tubes, or
render the transmitting pits almost impermeable, while
it might be almost or quite invisible. Such a trans-
parent substance Weber did actually detect in his experi-

in ASCENT OF SAP IN STEMS 55

merits on Picea excelsa. This investigator's observation
that the resistance of the region immediately above the
killed part is enormously increased, is a piece of positive
evidence in favour of stoppage which cannot be put aside.
He found that the lower part (1T8 cm. long) of the heated
stem, including both boundaries between the dead and
the living regions, refused to transmit water under a head
of 62 cm. of mercury. Microscopic observation shows that
clogging is greatest in these border regions. The perme-
ability of the upper parts of the same stem was much
greater, as was seen from the fact that the upper 16*5 cm.
supporting the faded leaves transmitted sufficient water
under a head of 59 cm. of water to render the leaves
turgescent once more. In several other experiments Weber
has shown that the resistance of the conducting wood is
enormously increased where the uninjured part borders on
the heated region. Janse confirmed and extended Weber's
observations of this increase of resistance even when the
temperature to which the branch was exposed was as low
as 60-64.

Tt is easy to show that the water traversing a piece of
stem killed with heat is contaminated in its passage.
Thus, if distilled water is forced through a piece of a
stem freshly cut from a tree, it is transmitted without
sensible coloration so long as morbid changes in the stem
do not take place. But if the same stem is surrounded
with steam so that its cells are killed, the water which
emerges is no longer colourless, but is tinged, more or less
deeply, with brown. When attached to the tree the
killed piece must contaminate the rising transpiration
stream in the same way. The record of this contamination
first appears on the walls of the tracheae of the leaves,
which become stained ; then in their lumina, which become
filled with the coloured fluid concentrated by evaporation.
This substance is finally caught in the walls of the con-
ducting tubes of the stem and forms plugs, filling their

56 TRANSPIRATION AND ASCENT OF SAP ch.

lumina immediately above the killed region. Even with-
out Weber and Janse's direct determinations it would be
hard to believe that the deposit of this coloured substance
in the walls and lumina of the tubes could be without
effect on their efficiency in transmitting water.

From what has been said it will appear that even in
those cases where no visible stoppage has been found, we
have ample reason to suspect its presence. Furthermore,
in many cases, both where the gum-like clogging material
has been observed, and where it has not been found, the
bordering living cells develop tyloses, and so more or less
completely stop the flow of water in the tracheae. The
occurrence of tyloses has been recorded in a large number
of cases.

Thus the fading and drying of the leaves above the killed
region gives no support to the idea, that the lack of water
from which the leaves suffer is due to the removal of vital
forces which are required to raise the water-supply, but it
rather indicates a great increase in the resistance of the
stem, due in part to the stoppage of the lumina and to the
clogging of the walls of the transmitting tracheae. Janse's
and Ursprung's observation, that the greater the length of
the killed portion the more rapid the fading, is quite explic-
able on this view, as it is natural that a greater amount of
the clogging material would be set free by the larger number
of killed cells.

Contamination of the sap. It also seems certain
that some of the materials liberated into the transpiration
stream by the killed cells act deleteriously on the cells in
the leaf, and bring about morbid changes in them so
that in many cases these cells actually lose their turgor
and die even before they suffer severely from the reduc-
tion in their water supply. It actually, then, appears,
as Vesque puts it, that the leaves dry because they die. 1

1 Here it may be pointed out that the leaves which fade after their sup-
porting branch has been killed by heat, fade in a different manner from

in ASCENT OF SAP IN STEMS 57

A priori sucli a result seems inevitable. When the heat
has killed a portion of the stem, the cells adjoining the
water-tracts become permeable ; and hence the dissolved
substances in their vacuoles are set free into the upward
current of sap in the tracheae. The vacuoles contain acids,
carbohydrates, and salts, so that, even in the absence of
corroborative observations, we should expect the sap to be
enriched with these substances Furthermore, very pro-
bably substances in the cells ordinarily not in solution
would be brought into solution, and introduced into the
sap by the higher temperature ; possibly, too, some bodies
might be precipitated from the rising sap by the higher
temperature. Yet another change is to be anticipated.
The heat will destroy any thermolabile substances in the
sap and in the adjoining cells. Coagulation changes may
also be expected.

It is not difficult to test these surmises experimentally ;
and, indeed, a colour-change in the sap issuing from heated
stems has before now been recorded and commented upon.

The sap extracted from various trees by means of a
centrifuge provided material suitable for this investiga-
tion. Short lengths of the branch to be investigated
(9-10 cm. long by 2-2'5 cm. diameter) are placed in gilt
buckets of a centrifuge, and the sap yielded after about

those which wilt owing to a lack of water. In the former case the margin
of the leaf first becomes darkened and this darkened region gradually invades
the leaf between the veins. It then dries and shrivels while the green
parts immediately round the veins remain comparatively fresh. As this
change is taking place these veins usually become pink and finally brown.
This coloration is particularly noticeable when the leaves are viewed with
transmitted light. Shrivelling and withering of the leaf, except at the
edges, does not occur until after these changes are complete.

On the other hand, when leaves fade simply from an insufficient water
supply, e.g., on a branch severed from a tree, shrivelling comes on while
they are still green. Blackening appears only after shrivelling and occurs in
irregular patches. The veins do not change colour and the walls of the
trachea? do not appear coloured in transverse section. The first colour
change is when the cell-contents of the mesophyll and parenchyma of the
veins colour brown after death.

58 TRANSPIRATION AND ASCENT OF SAP ch.

five minutes' rotation is collected. The quantities of sap
obtained in this way are surprising. Whether in spring, mid-
summer, autumn, or winter, I have found that four such
pieces of the various woods used yielded about 3-5 c.c.
In the same way, sap was collected from pieces of steamed
branches. These samples of sap could now be compared
physically and chemically. In every case, as was antici-
pated, profound differences were found to exist between
the characters of the saps drawn from the fresh and the
steamed branches. Some of the results may be seen in
Table 4 : under A the depression of the freezing-point, under
C the electric conductivity is given. Also the reaction
of the sap to litmus and the presence of oxydase are
noted :

Table 4.
Changes introduced into Wood-sap by Steaming.

Name

Fresh.

A.

C x 10 4

Acidity.

Colour.

Oxydase.

Fagus silvatica . .
Populus alba , . .
Hex aquifolium . .

0-083"
0-055"
072"

3

2-7
6-4

very faint
very faint
very faint

colourless
colourless
pale grey

+
+

Name.

Steamed.

A.

CxlO 4

Acidity.

Colour.

Oxydase.

Fagus silvatica . .
Populus alba . . .
Hex aquifolium . .

0-509"
0231
0-321

19
11

28

marked
marked
marked

brown
pale brown
pale brown

The change in A brought about by steaming is due to
the total increase of the dissolved substances, and indicates
that the concentration of the sap has increased four to six
times. The changes in conductivity (expressed as the

in ASCENT OF SAP IN STEMS 59

reciprocals of the resistances measured in ohms) indicate
the relative richness of the saps in electrolytes. 1 From
the table it appears that the concentration of electrolytes
has become four to six times greater by steaming. The
development of strong acidity during the heating (observed
in every case so far examined) shows that the increase of
electrolytes is partly due to the introduction of acids into
the sap. 2 In Fagus and Populus an oxydase was present 3
in the sap of the unsteamed stem, which coloured guaiacum
tincture faintly blue. The blue was intensified by the
addition of hydrogen peroxide. The oxydase was, of
course, destroyed with heating. The oxydase was not
looked for in the case of Ilex.

Qualitative tests on the sugars of the saps from the fresh
and steamed branches indicated changes in these bodies also.
Where non-reducing sugars are present, they are, of
course, hydrolysed by the steaming of the branch, and the
acid liberated ; and they appear after the heating as reducing
sugars. Examples of this were found in Ilex aquifolium,
Salix babylonica, and Cotoneaster frigida. In the last-
mentioned, however, larger quantities of reducing sugars
were found present in the sap of the steamed branch
than could have been formed by the inversion of the non-
reducing sugars present in the sap of the fresh branch, so
that we must assume that they were introduced into the
sap from the neighbouring cells.

These tests are sufficient to substantiate the surmise that
the physical and chemical nature of the sap is profoundly
altered by steaming the branch through which it passes.

It is evident that the substances thus introduced into
the sap must be swept along in the rising current till they
reach the leaves, except for what material is absorbed by

1 I am indebted to Mr. W. R. G. Atkins for the determination of these
conductivities.

- The very faint acidity of the sap from the fresh branches may probably
be ascribed to the sap set free from the injured cells at ends of the pieces.

J It is possible that this oxydase was also derived from the cut cells.

60 TRANSPIRATION AND ASCENT OF SAP ch.

the walls of the trachea?, and by the cells adjoining the
water tracts above the heated region. In the leaves, those
which are not in a form suitable for assimilation must
accumulate ; and, if sufficient of the branch has been
killed, the accumulation will ultimately without any
other poisonous action plasmolyse the cells of the leaf. 1

Reduction in the water-supply may be also brought
about by the coagulation of colloids in the sap, and the
consequent formation of plugs in the conducting tubes.
This condition, as has been noted above, has been observed
by several investigators.

Poisonous action of contamination. It seemed
of interest to essay to find out if the sap in steamed
branches contained any substance which acted as a
protoplasmic poison, and not merely as a plasmolysing
agent by simple accumulation. To test this point, saps
extracted from branches subjected to various treatments
were applied to severed leaves of Elodea canadensis, and
the effect on the cells of these leaves microscopically
controlled.

In the first place it was found that the cells remained
normal, and protoplasmic streaming continued undimin-
ished in the sap from fresh branches for at least five days,
and probably much longer. This point was verified in
the case of the sap of Ilex aquifolkim, Prunus cerasus,
Syringa vulgaris, Cotoneaster frigida, and Salix babylonica.
In contrast to the sap from the fresh branches, that from
the steamed branches of all these, with the exception of
Ilex aquifolium produced lethal changes in the leaves of
Elodea within two or three days. These changes consisted
in a cessation of protoplasmic streaming, in the discolora-
tion of the veins and margins, and in the contraction of
the protoplasts of the cells all over the leaf, and their

1 It may be noted that Ursprung looked for the production of plasmolysing
effects in the root-hairs of Impatiens svltani by a decoction of the same plant,
but did not tind any. Here, of course, concentration would not take place.

Ill

ASCENT OF SAP IN STEMS

61

ultimate blackening. The contraction which occurs is not
of the nature of plasmolysis ; for more than a day is often
required for its production, and it cannot be undone by
the transference of the leaf to water.

In each case a sample of sap centrifuged from the fresh
branch was tested ; other similar tests were made upon
that centrifuged from a branch immediately after steaming,
or centrifuged from a branch steamed a day or two pre-
viously, or with the liquid centrifuged from a branch which
had a day or two previously been steamed, and at once
depleted of its sap by centrifuging, and refilled with dis-
tilled water. These two last tests were made in order to
see if the poisonous materials are set free immediately
into the sap on steaming, or whether they are produced as
subsequent degradation-products of the cells.

Table 5.

Day of Experiment.

1st.

Ilex aquifvlium, fresh

,, ,, steamed direct ....

,, ,, ,, indirect : ...

,, + water 1
Primus cerasus, fresh

,, ,, steamed direct

,, ,, ,, indirect + water
Cot oneast er frigida, fresh

,, ,, steamed indirect . . .

,, ,, ,, 4- water
Salix babylonica, fresh ! . . .

,, ,, steamed indirect . . . .

,, ,, ,, ,, + water
Sijringa vulgaris, fresh

,, ,, steamed direct

,, ,, ,, indirect . . . .

+ water

?

1

1

2nd. 3rd.

4th.

+

1

+

+
+
+

+
+

+
+

2

+
+
+

+
+

+
+

+
+

+
+
+

5th.

+
+

+
+

+
+

+
+
+

In Table 5, the sap obtained by centrifuging immedi-
ately after steaming is termed ' steamed direct " ;

1 When the sap had stood in contact with the dead cells for two or more
days, discoloration of the veins and blackening of some of the cells without
marked contraction occurred on the fifth day.

62 TRANSPIRATION AND ASCENT OF SAP ch.

that which was centrifuged some days after steaming
is called "steamed indirect"; while the liquid obtained
from the steamed branches which had been emptied
of their sap, and subsequently filled with water, is
tabulated as ' steamed indirect + water." The ciphers
in the table indicate that no effect was observable on that
day of the experiment under which the figure is placed,
while a plus mark shows that an effect was observed.
Query-marks indicate that only some of the leaves tested
were affected, or that only slight protoplasmic contraction
was observable. In each case three or four leaves were
immersed in the liquid and examined.

It may be noted that, in the case of Cotoneaster frigida,
Syringa vulgaris, and Ilex aquifolium (see footnote, p. 61), the
liquid centrifuged from the steamed branch, after it was
emptied of sap and filled with water, is more rapidly
poisonous than the sap itself. In these cases probably a
poison is formed in the cells after death, which is not suffi-
ciently concentrated in the sap centrifuged immediately
after steaming. The same explanation probably applies
to the observation that the sap extracted from the Syringa
branch immediately after steaming is not so quickly lethal
as that drawn off a couple of days after death.

The slow generation of poisons indicated in these experi-
ments probably affords an explanation of the fact that,
even when steamed branches are washed out immediately
after the heating, as described later, some of the leaves
above perish from poisoning.

But it is possible to demonstrate the poisonous action
of the substances liberated in the killed portions of branches
on their own leaves without at the same time curtailing
their water supply in any way. This was found
feasible in the following manner : One branch of
a bifurcated shoot of Syringa vulgaris was killed by
immersion in water at about 90 C. for ten minutes.
After this the dead branch was stripped of its leaves, and

in ASCENT OF SAP IN STEMS 63

cold water supplied through it to the leaves supported 011
the uninjured branch. These leaves then drew supplies
of water from two sources, viz., from the roots, and through
the dead branch. To facilitate this latter supply a fresh
surface was occasionally cut on its distal extremity. The
length of the dead branch in my experiments varied from
30 to 40 crn. Notwithstanding the double supply of water,
the leaves on the living branch in each experiment showed
signs of wilting. When the supply through the dead branch
was cut off, either by the withdrawal of the water from its
end, or by its own clogging, the wilted leaves partially
recovered. But in several experiments the edges of the
leaves were too far injured to recover, and the injury
persisted as a brown margin on the leaf.

The experiment may also be carried on with a straight
branch from which the upper leaves and side branches
are removed. The stripped upper portion is immersed for
a short time in water at about 95, and after death
so caused, water is supplied through it to the lower
leaves and side branches, which have been left un-
disturbed. If the supply of water through the upper dead
part of the branch is kept up, fading and partial withering
of the leaves below will be noticeable in a few days.

Inasmuch as the effect on the leaves depends on the
amount of harmful matter carried from the dead cells,
it is evident that immediate or complete withering is not
to be expected ; for the supply from the dead part is largely
diluted with the supply from the roots through the living,
and, furthermore, it is difficult to make the supply through
the dead part considerable owing to the clogging at the
cut surface, and to the internal stoppages caused by the
exudations into the water-capillaries from the dead cells.

Another and simpler method of observing the withering
effects of the substances liberated from cells, when
killed by heat, may be carried out as follows : A decoction
is made by boiling small pieces of a stem in water for a

64 TRANSPIRATION AND ASCENT OF SAP ch.

short time. This decoction, after repeated filtering, is
supplied to cut branches of the same tree. It will be found
that the leaves of the branches supplied with the decoction
fade and wither much more rapidly than those of con-
trol-branches supplied with water. For example, three
branches of Syringa vulgaris, set in a decoction of the
stem of the same plant, lost their turgidity within two
days, while the leaves of three control-branches were still
fresh after five days.

Taken alone, this last observation would not be sufficient
to prove that in intact branches the withering of the leaves
is due to deleterious substances emerging from the killed
cells ; for it might be urged that colloid substances in the
decoction aggregating on the cut surface obstruct the free
transmission of water, and thus cause the fading by partially
cutting off the water-supply. The probability of this ex-
planation is lessened by the fact that the decoction causes
the fading even after it has been repeatedly filtered. In
any case the observation, taken along with the previous
experiments, may be regarded as confirmatory of the view
that the fading is largely due to the plasmolysing or poison-
ous effects of substances extracted from the dead cells.

The contaminated nature of the water supplied through
a dead branch may be demonstrated by collecting some
of the water transmitted through a branch killed by the
application of hot water, as in the previous experiments.
If Syringa vulgaris is used, the water is of a dark
brown colour, and quite different in appearance from
what is transmitted through a living branch in similar
circumstances. If this brown fluid is supplied to cut trans-
piring branches, the latter rapidly fade and wither. When
making this experiment, I killed the lower 30 cm. of a
straight branch 40 cm. long, by immersion in hot water.
The branch was then inverted and water forced through
it under a head of 20 cm. The water transmitted was
thus filtered through 10 cm. of living wood after its passage

Ill

ASCENT OF SAP IN STEMS

65

through the dead portion. Subsequently, it was twice
filtered, and supplied to cut branches. These latter faded
in two and a half days, while controls did not show signs of
fading for several days later.

Removal of Contamination. It has been found
possible to confirm this explanation of the fading
of the leaves in another very striking manner. It
was to be expected that if poisonous materials gene-
rated by the application of heat in the supporting stem
caused morbid changes in the leaves above, it should be
possible partially or completely to prevent the changes
by washing out these materials from
the stem as they are formed. We
should expect the leaves above such a
heated and washed-out stem to last
fresh much longer than those above
a similar piece of stem which had been
heated but was not washed out.

The experiment was performed on
a pot plant of Prunus cerasus. The
stem of this plant, at a level of about
40 cm. above the soil, bifurcated into
two equal branches, B and C (see
Fig. 11). B produced two lateral
branches, E and D, at 11 and 34 cm.
respectively above the bifurcation,
while C had two smaller leafy branches
about 15 cm. above the bifurcation,
and terminated with a tuft of leafy
branches above. The top of B above
the base of D was removed, and, with suitable precautions
to minimise clogging, a rubber tube was attached
to it. The whole of E, except a few centimetres of its
base, was cut away. The cut surface at the top of
B was now supplied with distilled water under a head of
33 'cm. When the cut surface of E, by becoming moist,

F

Fig. 11.

66 TRANSPIRATION AND ASCENT OF SAP ch.

showed the arrival of this stream below, the intervening
space of 23 cm. between the bases of D and E was
lapped in a cloth and sprayed with boiling water.
Meantime the leaves were protected from injury by being
covered with damp cloths. The hot spraying lasted 10 mins.
After it ceased the cloths were removed from all except
the leafy branch D. The transpiration of this branch was
thus kept at a minimum during 24 hours, while the supply
of distilled water was kept up to flush out the materials
exuded into the waterways of the heated region. During
this time about 35 c.c. were passed in at the top of B.
After this the plant stood in a cool greenhouse under
conditions favourable to transpiration.

Three days later it was observed that the lower leaves
of D and a few of the leaves on the lower branches of C
were slightly discoloured and curled at the edges. Evi-
dently the heating of the part of B had affected not only
the leaves above it, but also some of the leaves on a branch
springing from below its base.

After seven days from the beginning of the experiment,
the leaves on C and D, which before had shown the slight
changes just mentioned, had become quite withered and
curled.

Fourteen days later, i.e., twenty-one from the beginning
of the experiment, there were still four living leaves on D,
but their veins were coloured red, and the edges of two
were discoloured brown. The remaining leaves on D were
dry and crisp. On C all the lower leaves had fallen, or,
if still attached, were crisp and much discoloured. The
remaining upper leaves were apparently still quite healthy.

The lower killed leaves on C were now removed, and the
lower 23 cm. of the branch was killed with hot water in
the same manner as B had been, while the healthy upper
leaves were protected by enveloping them in a damp cloth.

Five days after this treatment, all the leaves on C were
stained and curled, their cells being evidently dead.

in

ASCENT OF SAP IN STEMS

67

In this experiment we see that all the leaves above a
length of stem of 23 cm. which has been killed by heat,
show strongly marked morbid changes after five days,
while in a similar case, if the heated region is washed out
with water, these changes are postponed in some for
twenty-one days, and even then are not complete. The
experiment also demonstrates that leaves on an uninjured
branch may be caused to wither by supplying them with
water which has passed through a heated stem.

In another experiment a plant of Cotoneaster frigida was
used. Distilled water was supplied at
the cut end of the main stem. Below
this four lateral branches, A, B, C, and
D, took origin, separated from one
another by distances on the stem of
18 cm., 5 cm., and 22 cm., respec-
tively, from below upwards (see
Fig. 12). From each of these leafy
secondary branches sprang. D sup-
ported only one secondary branch, and
it was cut short by an old injury. C
was lopped at the start of the experi-
ment, and left with only one secondary
branch. B and A were not interfered
with. During twenty-four hours dis-
tilled water was supplied at the cut
top of the main stem, and meanwhile
the region of the main stem, viz.,
22 cm., between C and D was killed
with hot water as before, while the leaves were suitably
protected from injury.

Three days later blotchy discolorations appeared on
the lower leaves of D, and less markedly on some of those
of A, B, and C.

On the fourteenth day the four lowermost leaves of D
were completely discoloured, and a small blotch had

F 2

68 TRANSPIRATION AND ASCENT OF SAP ch.

appeared on the fifth leaf from the base, while the seven
leaves above were quite healthy. On A there were five
leaves dead ; on B two small ones and two large leaves
blotched, on C there was one withered and crisp, and one
blotched.

At this stage, 14 cm. at the base of C was heated, the
leaves above being protected.

On the next day all the leaves of C were stained and
beginning to curl. No further changes were noticed in
A and B.

Six days after the heating of C all its leaves were dead,
while still the uppermost leaf of D was unaffected, although
its supply was drawn through 22 cm. of stem which had
been killed twenty days previously.

In this experiment, evidently the flushing of the killed
branch removed much of the deleterious substances from
the supply to D, for its changes were at first less notice-
able than those of the branches below, into which evidently
the major part of these substances were distributed. The
changes observed in C after its base was heated, indicate
the extent and the rapidity of injury we might have
expected in D had the killed region not been washed
out.

Thus we see there is a large mass of experimental evi-
dence showing that the fading of the leaves on a killed
stem is due to the introduction of poisonous or plasmo-
lysing substances into the transpiration sap. In these
branches the freedom of motion of the upward water
current may be reduced by plugs formed above the heated
region and in the veins of the leaves themselves ; but, if
sufficient length of the stem has been killed, poisoning
will apparently always supervene, whether plugging occurs
or not. With these facts established, this fading affords
no evidence in support of the view that the vital activities
of the cells of the stem are essential to the raising of the
transpiration stream.

in ASCENT OF SAP IN STEMS 69

Attempts to detect the action of vital forces.
In another form the vital hypothesis has been lately
supported by Ewart. Adopting Janse's method of deter-
mining the resistance opposed to the transpiration current,
Ewart obtained results which indicate that, in order to
move water in stems of plants at the velocity of the trans-
piration current, pressures equivalent to a head of six to
thirty-three times the height of the plant are required.
To overcome this resistance, which in the case of the highest
trees would amount to 50-100 atm., Ewart feels constrained
to fall back on the vital activities of the wood parenchyma.
He supposes them to lend a helping hand in some way,
and to overcome the resistance of the stem all along its
length.

These hypothetical forces, however they are imagined
to be exerted by the living cells, Ewart believes to be but
feeble, and has not, as he himself admits, in any case
obtained unequivocal evidence for their existence. And
yet, according to his own figures, these forces should be
easily demonstrable. According to him, the pressure re-
quired to raise water at the transpiration-velocity in an
elm tree 12 m. high would be equivalent to a head of
75*6 m., i.e., about 7 '5 atmospheres. Of this he admits
about 2 atmospheres might be supplied by the tension set
up by the transpiring leaf-cells, leaving about 5*5 atmo-
spheres to be made good by the lifting forces of the cells
in the 12 m. of stem. Therefore, the lifting force of the
cells of this stem must amount to 0*45 atmosphere per
metre of stem, or to a head of water equal to four and
a half times the length of stem.

A lifting force of this magnitude should be easily re-
vealed if the velocity of flow through a branch in the
normal direction for a given head were compared with the
flow in the reverse direction, or, again, if the amount
transmitted downwards in a living stem were compared
with that transmitted after death. As is well known,

7o TRANSPIRATION AND ASCENT OF SAP ch.

experiments have not been able to demonstrate a sensible
difference in either case. Furthermore, Ewart, working
very carefully by a different method, has failed to detect
the existence of these pumping actions in stems. Con-
sequently it is quite impossible to admit that any large
amount of work falls on the cells of the stem in the raising
of the sap.

While these considerations show that forces of any great

magnitude are not exerted by the cells
in the wood on the transpiration-cur-
rent, it seemed desirable by some more
careful method to test the matter, and
see if some much smaller force were
not assisting the upward flow of water.
In the ordinary methods of testing
this question, uncertainties arise from
the fact that conditions are not the
same before and after the reversal of
the current, or before and after the
death of the branches. These differ-
ences are principally due to changes
in temperature, which, as Ewart has
pointed out, entail large differences
in viscosity, and to clogging in the
experimental stem.

In order to eliminate these sources

T" of error, and so be in a position to

C //)X' ]'[" detect the effect of even a very small

force exerted by the stem-cells in
lifting water, I carried out some ex-
periments in the following manner :
Two straight branches (A and B, Fig. 13), about 80 cm.
long, without lateral shoots, and as similar to one another
as possible, 1 were passed through tubulures (a and b) in

Fig. 13.

1 I used Syringa vulgaris, as similar and straight branches of this shrub
are readily obtained.

in ASCENT OF SAP IN STEMS 71

the bottom of a metal cistern, about 65 cm. deep. The
upper ends of the branches projected above, and the lower
ends below, the cistern. The joints round the lower ends
were rendered water-tight by binding on a rubber tube
overlapping the tubulures and the projecting ends of the
branches. The cistern was then filled with water which,
being in motion, secured that both branches were at the
same temperature, and so differences in viscosity in the
water passing through the branches did not arise.

In order to avoid irregularities in transmission, much
care is needed in the preparation of the branches. After
selection of the branches, the upper leafy part was cut
away, and the cut surface of the lower part still attached
to the tree was moistened by a jet of water. This part
was then cut off under water, and while still submerged,
was removed to the laboratory. Fresh surfaces were next
prepared at each end, and smoothed off by a razor, under
a stream of distilled water. A wide glass tube about
20 cm. long was attached to the upper end of each branch.
This was kept full of distilled water, which acted as the
supply and head, driving the water downwards through
the branches. If it was desired to apply picric acid or
some other poison as a killing agent, the simple glass tube
was replaced in each case by one which was provided with
a side tubulure, bent in a J form, and with two stop-cocks
placed as shown in Fig. 13.

The rate of transmission from above downwards was
first observed for the two branches by weighing the amount
of water transmitted in a given time (say 10 min.). It is
evident if vital actions were at work tending to raise the
water in the branches, the rate of flow downwards would be
reduced by this activity ; and inasmuch as both branches
are under similar conditions, the reduction of flow would
be the same for both. It may here be noted that an altera-
tion in the head of 10 cm. makes a very sensible difference
in the amount transmitted, raising it, to take an example,

72 TRANSPIRATION AND ASCENT OF SAP ch.

from 0*450 gram to 0*500 gram per 10 min. If now one
branch be killed, the vital lifting-force, if present, will be
removed, and we should expect the amount transmitted
by the killed branch to increase correspondingly. This
increase, even though small, would be easily seen by com-
paring the flow through the dead and the living branches,
when both have been again brought to the same conditions.
For killing the branch I used either a jacket of steam or
an injection of poison.

When it was desired to kill the experimental branch
with steam, the water in the cistern was run off through
a small side-tubulure (c), and, when the cistern was empty,
steam was passed by the same tubulure into a wide tube
(d), now placed round the experimental branch, and fitting
tightly into a socket made for its reception in the bottom
of the cistern. The space round the top of this branch,
and between it and the tube, was packed with cotton-
wool. The supply of steam was kept up for 20 min. ; the
tube (d) was then removed ; and the cistern was filled
with water through the tubulure (c). After some time,
during which the water in the cistern was kept stirred,
when it was judged that the experimental and control
branches had come to sensibly the same temperature,
measurements of the amounts transmitted by each were
resumed. In this way it was easy not only to compare
dead and living branches under the same conditions, but
also rough manipulation and shaking were avoided. These
latter are known to cause irregularities in the amounts of
water transmitted by cut branches, probably owing to the
displacement of, or compacting of, some clogging material
on their upper surfaces.

Observations made in this way showed in each case that
the amounts transmitted by a branch before and after
killing by steam were sensibly the same ; or, if they
differed in the experimental branch, the same difference
was observed in the amounts transmitted by the control

Ill

ASCENT OF SAP IN STEMS

73

branch at the same time, the observable differences being
due to changes in conditions which affected the flow in
the living as well as in the killed branch.

In Tables 6 and 7 is recorded an example of one of these
experiments, which is graphically recorded in Fig. 14.

Particulars of Experiment.

A and B, two similar branches of Syringa vulgaris, each
with five year-rings.

A : length, 84'5 cm. long. Upper diam. of wood,

HI

1 -30O

2.5 i

200

B

_ -

&"'

.. I. J

B

2?

| 10

<3

* .. x_

A
Steam

A

11-6

119

12-7

5 Time in hours
12-0 Temp, of
jacket

Fig. 14.

- 8 cm. ; of pith, 0"15 cm. Lower diam. of wood, 0'9 cm. ;
of pith, 0*23 cm. Head, 9*0 cm. of water.

B : length, 83*5 cm. Upper diam. of wood, 0*73 cm. ;
of pith, 01 cm. Lower diam. of wood, 0*85 cm. ; of pith,
0*1 cm. Head, 9 3 cm. of water.

Table 6.

Time

Amount transmitted

Amount transmitted

from start

Temp, of Cistern.

per 10 minutes

per 10 minutes

in min.

by A.

byB.

5

11-6 C.

0-164 gr.

0-267 gr.

28

117

0-168

0-273

49

11-75

0172

0272

54

11-85

0-172

0-272

80

11-9

0-173

0-274

The cistern was emptied when the experiment had run
for 120 min., and B was surrounded with steam for

74 TRANSPIRATION AND ASCENT OF SAP ch.

20 min. When the cistern was refilled and stirred, the
measurements were resumed.

Table 7.

Time

Amount transmitted

Amount transmitted

from start

Temp, of Cistern.

per 10 minutes

per 10 minutes

in mm.

by A.

byB.

164

12-7 0.

203 gr.

0-215 gr.

187

12-8

0185

0-230

212

12 8

0184

0-268

226

12-85

0-181

0-240

240

129

0-184

244

258

12-9'

186

0253

Immediately after the experimental branch is surrounded
with steam, the water transmitted through it becomes
coloured, at first amber, changing to brown. 1 This change
probably indicates the introduction into the transmitted
water of some clogging material ; for, instead of
an increase in the amount transmitted downwards
after the removal of the supposed vital lifting forces,
we see from the table that the amount is diminished.
According as the downward stream washes out this sub-
stance, the original rate of transmission is approximated
to, but, during the experiment, is not attained. It is
evident that if there was any considerable length of branch
below the steamed part, this material would accumulate
there.

The small rise in the amount transmitted by the control
branch, noted immediately after the observations were
renewed, is probably to be attributed to a rise in tempera-
ture and consequent reduction of viscosity.

The effect of steaming on the experimental branch, and
the uniform behaviour of the control, are clearly brought
out in Fig. 14, in which the ordinates are grammes of

1 The production of the brown colour [may with probability be assigned
to the action of the oxydases of the stem on a colourless chromogen
before the enzymes^were destroyed by heat.

in ASCENT OF SAP IN STEMS 75

water transmitted per 10 min., and the abscissse are the
times in hours. The full line connects the successive
observations on the control branch ; the dotted line joins
the observations on the experimental branch. The gap
indicates the interval during which the cistern was emptied,
and the steam was applied to the experimental branch.

Change taking place in the experimental branch, demon-
strated by the exit of coloured liquid below after the
application of the high temperature, renders the experi-
ment somewhat unsatisfactory ; for, although the experi-
ment gives no indication that the removal of vital processes
from the stem does away with lifting forces opposing the
downward motion of water, yet it is just conceivable that,
if these forces previously existed and were removed, the
clogging of the branch during the steaming might just
compensate for their removal. The possibility of such a
coincidence, though very improbable, suggested the use of
picric acid as a killing agent in place of steam.

The experiment was at first arranged in exactly the
same manner as before ; and the initial rates of trans-
mission of the experimental and control branches were
determined. Then some dry picric acid was introduced
into the water-tube of the experimental branch, and the
observations were continued. No change, which could be
assigned to the removal of vital actions by the picric acid,
could be detected in the rates of transmission, even when
the acid appeared at the lower end of the experimental
branch.

The gradual killing of the branch and the slow penetration
of the picric acid in this method are open to objection, and
would tend to render a change due to death less noticeable.
To remove this objection, a modification was introduced
by means of which the picric acid is quickly forced through
the stem under pressure ; and, in order to place the control
under similar conditions, distilled water is simultaneously
forced through it. This is arranged by having the tubes

76 TRANSPIRATION AND ASCENT OF SAP ch.

containing the water supplies to each branch provided
with a side-tubulure connected with a J -shaped glass tube
containing a mercurial column. The J -tubes and the tops
of the water-supply tubes are provided with stop-cocks
(e, f, g, and h respectively, Fig. 13, p. 70). At first the
side-tubulures (e and /) are closed ; and the rate of trans-
mission of distilled water under a low pressure is measured
for each branch ; then picric acid is introduced into the
supply-tube of the experimental branch, and the stop-
cocks at the upper ends of the supply-tubes (g and h) of
both are closed, and the lateral tubulures (e and/) opened ;
so that the picric acid is forced through one, and distilled
water is forced through the other. When the picric acid
appears below, by the suitable manipulation of the stop-
cocks, the pressure in each is again reduced, and observa-
tions are recommenced.

Tables 8 and 9 give the figures of such an experiment,
and the results are plotted graphically in Fig. 15.

Particulars of Experiment.

A and B, two similar branches of Syringa vulgaris : A
with seven year-rings ; length, 80 cm. Upper diam. of
wood, 0'85 cm. ; of pith, 014 cm. Lower diam. of wood,
- 92 cm. ; of pith, 0*20 cm. Head, 24 cm. of water. B
with four year-rings ; length, 80'5 cm. Upper diam. of
wood, 0*85 cm. ; of pith, 0'16 cm. Lower diam. of wood,
0*95 cm. ; of pith, 0'20 cm. Head, 24 cm. of water.

Table 8.

Time
from start Temp, of Cistern.
in min.

Amount transmitted

per 10 minutes

by A.

Amount transmitted

per 10 minutes

byB.

10 11-7 C.

23 IIS
44 11-8 3

0-455 gr.
0-445
450

0-470 gr.

0-4(58

0-471

Ill

ASCENT OF SAP IN STEMS

About 30 min. after the last observation, picric acid was
supplied to B, and forced through under a head of 41 5 cm.
of mercury. At the same time, water was forced through
A under a head of 44 cm. of mercury. After twenty-
five minutes when the picric acid had appeared in
quantity at the lower end of B, the head of 24 cm. of water
was restored to both.

Table 9.

Time

Amount transmitted

Amount transmitted

from start

Temp, of Cistern.

per 10 minutes

per 10 minutes

in nun.

by A.

byB.

107

11-8 C.

0-498 gr.

500 gr.

123

11-9

0-497

0-501

136

120 3

0-500

0-503

160

120

0-501

0-506

174

12-0

0-503

0-507

189

12-1

0-501

0-520

204

12-1

0-504

0-508

223

122

0-506

0512

244

12 3

0-504

0-515

263

12 4

0-507

0511

It will be seen that the mean rate of transmission before
the high pressure was applied was 0'450 g. and 0*469 g. per

c-600

500

5s,

400

S

g-300

c

3-200

<| -100

a

B

. ?

-9 -

8-

9-

A&B

B

A
Pier

'c acia

11-7

11-8"

11-8"

4 Time in hours

12-4 Temp, of
jacket

Fig. 15.

10 min. for the control and experimental branches respec-
tively. During this time the average temperature was 11*8.

78 TRANSPIRATION AND ASCENT OF SAP ch.

After the pressure was applied, forcing distilled water through
the control, and picric acid through the experimental
branch, the rates rose to 0*502 g. and 0*508 g. respect-
ively for an average temperature of 120. That is, the
rate of transmission of the killed branch has increased by
8*3 per cent., while that of the living control has increased
by 11" 5 per cent. Probably the rise in both cases is due
chiefly to the flushing-out of the branch and the washing
away of mechanical obstructions by the stream under high
pressure, and partly to the small rise in temperature
form 11*8 to 120, which would perceptibly diminish the
viscosity of the water. Inasmuch as the observed rise is
as great in the case of the living branch as in that which is
killed during the observations, it follows that there were
no vital actions in either retarding the transmission.

As was pointed out above, Ewart had recourse to the
vital theory, believing that the physical forces available
are insufficient to overcome the resistance to the trans-
piration current offered by the conducting tracts. It will
be shown later that this view was based on an erroneously
high estimate of both the resistance experienced and the
velocity attained by upward moving water.

Summary. The structure of the conducting tissues to
which the Vitalists appeal as supporting their views not
only does not render this support, but the relation of the
living cells to the tubes renders it even impossible for them
to exert an elevating force upon the water contained in
the tubes.

Experiments which had been quoted in support of the
vital hypothesis as demonstrating that when the vital
actions of the cells are removed the transpiration stream
is so slowed down, that the leaves above fade from lack
of water, have been shown rather to indicate that the
death of these cells is followed by a stoppage of the
trachea) and the more or less complete poisoning of the
leaves.

in SCENT OF SAP IN STEMS 79

Crucial experiments have failed to show directly the
presence of vital sap-lifting forces.

Experiments have abundantly shown that even when
the vital activities of the cells of the wood have been
eliminated, water under the action of purely physical
forces rises in the stems of high trees.

Literature.

Biot, J. B., "Remarques scientifiques a l'occasion de la lettre de M.
Boucherie inseree au dernier numero du Compfce rendu," Compt. rend., 1841,
T. 12, p. 357.

Boucherie, " Rapport sur une memoire de M. le docteur Boucherie relatif
a la conservation des bois," Compt. rend., 1840, T. 11, p. 894.

Dixon, H. H., "Physics of the Transpiration Current," Notes from the
Botanical School of Trinity College, Dublin, vol. i, p. 57.

Id., "Transpiration and the Ascent of Sap," Progressus Rei Botanicae,
1909, Bd. iii, s. 1.

Id., "Note on the Supply of Water to Leaves on a Dead Branch," Proc.
Roy. Dublin Soc, 1905. vol. xi. (N. S.) p. 7.

Id., "Vitality and the Transmission of Water through the Stems of
Plants," Proc. Roy. Dublin Soc, 1909, vol. xii (N. S.), p. 21, and Notes
from the Botanical School of Trinity College, Dublin, vol. ii, p. 5.

Id., "Note on the Spread of Morbid Changes through Plants from
Branches killed by Heat," Proc. Roy. Dublin Soc, 1914, vol. xiv (N. S.),
p. 205.

Id., "Changes produced in the Sap by the Heating of Branches," Proc.
Roy. Dublin Soc, 1914, vol. xiv (N. S.) p. 224.

Ewart, A. J., "The Ascent of Water in Trees," Phil. Trans. Roy. Soc
London, 1905, vol. 108 B, and 1908, vol. 199 B, p. 362.
* Id., "Protoplasmic Streaming." Oxford, 1903.

Godlewski, E., "Zur Theorie der Wasserbewegung in den Prlanzen,"
Jahrb.f. Wiss. Bot, 1884, Bd. 15, s. 569.

Hartig, R., " Ueber die Vertheilung der organischen Suhstanz, des
Wassers und des Luftraumes in den Baumen und uber die Ursache der
Wasserbewegung in transpirirenden Prlanzen," Untersucliuivjen aus d.
Forst. Bot. Inst, m Miinchen, ii and iii, 1882-1883.

Janse, J. M., "Die Mitwirkung der Markstrahlen bei Wasserbewegung
im Hulz," Jahrb.f. Wiss. Bot., 1887, Bd. 18, s. 1.

Id., "Der Aufsteigende Strom in der Pflanze, I.," Jahrb. f. Wiss. Bot.,
1908, Bd. 45, s. 305.

Id., "Der Aufsteigende Strom in der Pflanze, II.," Jahrb. f. Wiss. Bot.,
1913, Bd. 52, s. 509.

Id., "Die Wirkung des Protoplasten in den Zellen welche bei der
Wasserbewegung beteiligt sind," Jahrb. f. Wiss. Bot., 1913, Bd. 52, s. 603.

80 TRANSPIRATION AND ASCENT OF SAP ch. m

Overton, J. B., 'Transpiration and Sap-flow," Bot. Gazette, 1911, pp. 28
and 102.

Schidtz, J. A., " Neues, wohlfeiles und bewahrtes Verfahren, das Holz
zu conservircn, tlasselbe auch gegen Faulniss, Schwamm und Wurmstich
zu schiiltzen, u.s.w. Weimar, 1844.

Schwendener, S., " Zur Kritik der neuesten Untersuchungen iiber das
Saftsteigen," Sitzungsb. d. Preuss. Alad. d. Ills*., 1892,44, s. 911.

Strasburger, E., " Ueber den Ban und Verrichtungen der Leitungsbahnen
in den Pflanzen," Jena, 1891.

Id., " Ueber das Saftsteigen," Jena, 1893.

Ursprung, A., " Untersuchungen liber die Beteiligung lebender Zellen am
Saftsteigen," Beihefte z., But. Gen'ralblatt, 1904, Bd. 17, s. 147.

Id., "Die Beteiligung lebender Zellen am Saftsteigen," Jahrb. f. IViss.
Bat., 1906, Bd. 42, s. 503.

Id., " Studien Ueber die Wasserversorgen der Pflanzen " Biol. Centralblatt,
1907, Bd. 27, s. 33.

Id., "Ueber die Ursache des Welkens," Beihejtez. But. Centralblatt, 1907,
Bd. 27, s. 1.

Id., "Abtotungs- und Ringelungs-Versuche an einige Holzpflanzen,"
Jahrb. f. Wiss. Bot., 1907, Bd/44, s. 287.

Id., " Zur Frage nach der Beteiligung lebender Zellen am Saftsteigen,"
Beihefte z. Bot. Gentralb., 1912, Bd. 28, s. 311.

Vesque, J., " Sur le pretendu role des tissus vivants du bois dans
l'ascension de la seve," Comp'. rend., 1885, 101, p. 757.

Weber, C. A., "Ueber den Einfluss hoherer Temperaturen auf die
Fahigkeit des Holzes den Transpirationsstrom zu leiten," Ber. d. Deutsch.
Bot. Gesell., 1885, Bd. 3, s. 345.

Westermaier, M., "Zur Kenntniss der osmotischen Leistungen des
lebenden Parenchyms," Ber. d. Deutsch. Bot. Gessell., 1883, Bd. i., s. 371.

Id., "Die Bederstung todter Rohren und lebender Zellen fur Wasser-
bevvegung," Sitzb. d. Pretiss. Abut. d. Wiss., 1884, Bd. 48, s. 1105.

CHAPTER IV

COHESION THEORY OF THE ASCENT OF SAP IN STEMS

Origin of the theory.- In 1894 Dr. J. Joly
and the author published the first account of their
cohesion theory of the ascent of sap. Our attention
had been directed to the problem in the year 1892, when
the late Professor Strasburger had been good enough to
show us some of his experiments on high trees. After
more than a year's experimental work and discussion, we
were able to give an outline of our theory to the Dublin
University Experimental Science Association in March,
1894, and all the essentials were communicated to the
Royal Society in October of the same year.

Other hypotheses examined. In the work lead-
ing up to our theory we of course submitted the
theories of previous investigators, so far as we were
acquainted with them, to full consideration and ex-
perimental examination. In addition to these we sub-
jected to investigation various other hypotheses formed
by ourselves. As these investigations naturally lead us up
to the cohesion theory it may be permissible to outline
them here briefly.

Gravitational theory. In the first place it
seemed possible that gravitation itself might furnish
the force for lifting the upward moving water. This
at first seems paradoxical. Suppose the dilute sap
in the leaves to be concentrated by evaporation and

8i G

82 TRANSPIRATION AND ASCENT OF SAP ch.

by the addition of carbo-hydrates. The denser fluid
thus produced and passed into the tracheids would settle
downwards. As it passed down it would displace upwards
the less concentrated solutions entering at the root. An
accumulation of the denser material in the lower part of
the tree may be supposed to be prevented by the abstrac-
tion of materials from the concentrated sap all the way
down. In this way it is secured that the ascending " raw '
sap is just overbalanced by the denser descending column,
and the very dilute solutions brought into the root might
in this way be raised to any height. A model illustrating
the hypothesis is easily set up. A tube say 1 mm. bore
and closed at the lower end is filled with a solution of a
dye, e.g., fuchsin, and set upright. A small funnel con-
taining a denser salt-solution is attached to its upper end.
The heavy solution immediately begins to gravitate down-
wards, and in doing so displaces an equal volume of the
lighter fluid upwards. The rise may be noted by the pas-
sage of the coloured fluid upward into the funnel.

There is no doubt that this mechanism could work in
uninjured plants the roots of which continued to pass
comparatively pure water into the conducting tracts, pro-
vided there were an arrangement to prevent the mixing of
the descending and ascending fluids. In the plant, we may
suppose, the column is not supported below as in the model,
but is held up by the capillary forces of the imbibed cell-
walls. This would explain the presence of reduced air
pressure in the cavities of some of the wood trachea?,
which would be impossible if the water surrounding them
were in compression. But however promising for a time,
the theory had to be given up. The mingling of the dilute
ascending solutions with the concentrated descending liquids
which inevitably takes place in narrow tubes, would cer-
tainly destroy this gravitational action in the trachea? of
plants, and there is no evidence whatever of isolated
upward and downward currents.

iv COHESION THEORY OF ASCENT OF SAP 81

Electrical theory. An electrical theory also presented
itself. It is well known that when a colloid dispersed
through water is exposed to an electric field the colloid
tends to move to one pole or the other, depending upon
its electric sign. If the colloid is held stationary, the
water will be translated in the opposite direction. Con-
sequently, if there is a difference of electric potential at
the upper and lower extremities of a stem, we would expect
a tendency to motion upwards or downwards according to
the sign of the colloidal walls of the conducting tracts.
It seemed possible that atmospheric electricity, which main-
tains a potential gradient usually amounting to 50-150
volts per metre elevation from the ground, might produce
the necessary field.

It was disappointing to find that no experimental evi-
dence could be obtained in support of this hypothesis.
Thus, when leads coming from the terminals of a Wims-
hurst electrical machine were introduced into small reser-
voirs fixed to the opposite ends of a piece of stem about
2 m. long, actuation of the machine brought about no
observable motion of water from one reservoir to the other,
whether the branch and reservoirs were filled with water
or with a dilute solution.

Tensile film theory. Quincke's theory (which sug-
gested itself independently to us), viz., that the water is
drawn up in a tensile state over the surfaces of the walls of
the conducting tracheae in the form of a thin film, had also
to be laid aside. Our reason for discarding it was not
that which led Sachs to oppose it ; for he objected to it
on the grounds that there are not continuous tubes in
plants. In reality this objection is quite invalid, since
the water films may be regarded as continuous through
the imbibed material of the transverse and oblique
walls. Nevertheless, the theory had to be abandoned,
since, as we shall see later, such a film of water un-
supported on one side if exposed to tension, infallibly

G 2

84 TRANSPIRATION AND ASCENT OF SAP ch.

draws out thinner and thinner until it leaves no water
on the surface.

Theory of a tensile sap imbibed in the cell
walls.- A modification of this theory, combining it with
the Unger-Sachs imbibition theory, then occurred to us.
In order to escape the inevitable thinning out of the
unsupported water films, we assumed with Sachs that the
moving water is located in the substance of the walls, and
that the surface-tension forces developed at the surface of
the fine-textured substance of the wall prevent the water
column from becoming indefinitely attenuated. Thus
the tension generated at the leaves is transmitted
downwards through the imbibed water in the walls.
This theory has undoubted advantages over the
imbibition hypothesis. It replaces the diffusion flow
by a movement under great tensions, and so the rate
of transmission may be increased proportionately to
the increased tension. But it is open to many of the
objections which overthrew the imbibition hypothesis, viz.,
the lumina are known to transmit the major part of the
current, and it seems improbable, even where we can invoke
forces only limited by the tensile strength of water, that
they could suffice to drag an adequate water supply through
the fine-grained cell-walls.

When we found ourselves compelled to give up these
hypotheses, the one as assuming conditions inimical to
the transmission of tension in the water, and the other
because it did not agree with the ascertained fact that the
water moved in the lumina, it was an easy transition to
arrive at the conclusion that the water passed up in the
lumina in a state of tension. How, in the lumina of the
conducting wood, the necessary conditions for the pro-
duction of tension are fulfilled, we shall now proceed to
inquire.

Cohesion of water. Perhaps the easiest method of
realising the cohesion of water and the conditions

IV

COHESION THEORY OF ASCENT OF SAP 85

necessary to bring this property of liquids into play is
by performing the following experiment.

The vessel in which the liquid is to be enclosed is a J-
shaped glass tube about 1 cm. in diameter (see Fig. 16).
The long limb of the J is about 90 cm., while the shorter
one is about 20 cm. long. On the shorter limb there is
a bulb with a capacity of about 60 c.c. The shorter limb
is continued beyond the bulb as a narrow tube drawn out
to a point. The whole tube is carefully washed out in
the manner to be described in the following chapter, and
about 100 c.c. of repeatedly boiled water is introduced
into it. In order to be certain that the glass is thoroughly
wetted, and also to make sure that the water is in perfect
contact with any dust particles con-
tained in it, the liquid is again
repeatedly boiled after introduction
into the tube. Before sealing off the
fine tube the whole of the space
unoccupied by the liquid is filled
with steam by bringing the water to
ebullition, and, when the steam has
expelled the air, and is issuing
through the narrow tube the latter
is sealed off. When the whole has
cooled, it will be found that the

J -tube acts as a water-hammer, i.e., if, by inclining the
tube the water is made to travel from end to end, its
concussion makes a metallic ring. This is owing to the
fact that very little air has been included when the tube
was sealed, and water-vapour at normal temperatures is
unable to act as an elastic pad in the same way as air at
normal atmospheric pressure would. The clicking metallic
ring, then, may be taken as an indication that the gas-
pressure within the tube is very slight. Care, indeed,
must be taken not to let the concussion become too violent,
as in that way the tube may be easily shattered.

86 TRANSPIRATION AND ASCENT OF SAP ch.

If now, by gradually inclining the tube, the long limb
is completely filled with water (Fig. 16 B) and all the bubbles
are chased out of that limb by holding the bent end upper-
most, so that no breaks, even the most minute, remain,
we shall find, on inverting the tube, that the water
remains in the long limb and does not, under the force
of gravity, take up the lowest possible level in both limbs
(Fig. 16 C). From the level in the two limbs it is evident
that the hydrostatic pressure of the shorter column cannot
possibly balance the pressure of the column in the longer
limb ; the one is about 85 cm. higher than the other. The
water in this case hangs in the tube. The liquid in the
long limb is in contact with the glass all over, and, since
it wets it perfectly, it adheres to it. To the film of water
adhering to the glass the rest of the water coheres, and this
cohesion is well able to sustain the weight of the column
of water which is counterbalanced by no other upholding
force. In this way the lower part of the water in the
longer limb of the tube transmits a stress through the
upper part to the glass equivalent to its gravitational
pull.

The reality of this pull becomes all the more evident
when, by destroying the cohesion at one spot, a rupture is
started. This rupture, which may at first be invisibly
small, rapidly spreads across the whole column. The
rupture may usually be started by a sharp knock admini-
stered to the side of the longer limb ; but, when the
cohesion is very perfect, to produce a rupture may require
a shock so violent as to be liable to shatter the tube.
When the rupture is started, the lower part tears suddenly
away from the upper part of the column and falls into the
bend of the tube. The upper part follows it more slowly,
trickling down the inside of the tube, and all the water
comes to occupy a position in the lower part of the tube
(Fig. 16 A).

It is instructive to note how the cohesion of the water

iv COHESION THEORY OF ASCENT OF SAP 87

in these experiments is overcome. The rupture starts as
an extremely small discontinuity in the water. Surface
tension forces develop immediately at the surface of
this bubble. At its inception, being extremely small,
these forces are very great, but if the bubble enlarges, the
surface tension forces tending to close it rapidly diminish.
In our experiments the forces tending to open it are (1)
the momentum of the water conferred on it by the shock,
and (2) the gravitational pull giving rise to the tension in
the liquid. We may neglect the vapour pressure of the
bubble, since it is balanced by the vapour in the other
limb. If the break opened by the shock is so small that its
surface tension can withstand the tension in the liquid,
it will close again. But if once the bubble formed is so
large that its surface tension is overcome by the inertia
and weight of the liquid, an unstable condition is entered
on, and the bubble is continually enlarged till the tension
in the liquid is nil. It is, however, evident that if at any
moment we could confine the bubble and prevent it from
enlarging, the liquid would again pass into a state of
tension due to the weight of the lower parts.

Cohesion theory. The theory of the ascent
of sap, which Dr. Joly and the author advocate,
assumes that the water in the conducting tracts
of high trees hangs there by virtue of its cohesion,
just in the same way as the water hangs in the
experiment with the J -tube described above. The ad-
hesion of water to the walls of the tracheae we have shown
to be very great. For, as will be seen, if a fresh piece
of wood from the conducting tracts is enclosed in a
vessel filled with water in a state of tension, in every
case rupture will tend to occur at the surface of the
glass rather than at the walls of the tracheae, showing
that the adhesion of water to the walls of the conducting
tubes is thus always greater than the adhesion of water
to glass. This is quite to be expected, when we take into

88 TRANSPIRATION AND ASCENT OF SAP ch.

account the manner in which water permeates the sub-
stance of the walls of the trachea? when brought into
contact with them.

The teaching of these experiments obviously is that
water under suitable conditions can transmit a tension just
like a rigid solid. In the liquid, however, the stress is
hydrostatic, and, like hydrostatic pressure, is transmitted
equally in all directions. It is not sustained consequently
by a single point but affects the whole internal wetted
surface of the containing vessel. In another particular
the stressed liquid differs greatly from the stressed solid ;
it is much more unstable. A small flaw (i.e., a bubble)
in the tensile liquid rapidly spreads and almost instan-
taneously severs the whole column ; it matters not how
large the cross section of the unbroken part may be, a
comparatively feeble tension will tear it across. In the
solid a metal wire, for example on the other hand, if
the cross section of the unbroken part is sufficient, a
small discontinuity in its substance is immaterial, and the
stress may be successfully resisted by the intact part.
This difference in the behaviour of the two forms of matter
when submitted to a stretching force is to be referred to
the fact that the particles of a liquid are perfectly mobile
and are free to move round each other without being
opposed by any sensible internal forces, whereas in solids
there is a great opposition to the relative motion of the
parts. To this property solids owe their rigidity. In fact,
in tension experiments the liquid becomes capable of sus-
taining and transmitting tensile stresses only when it is
adhering completely to a rigid envelope which confers on
the liquid a pseudo -rigidity. The state of tension then
persists because the stretching forces act solely against
the cohesive properties of the liquid (i.e., in an endeavour
to separate the water molecules from one another a
separation which a liquid is able to withstand as well as
a solid). If, however, the liquid is free to change its

iv COHESION THEORY OF ASCENT OF SAP 89

shape, not adhering to any rigid envelope, the smallest
forces, whether of compression or of tension, spend them-
selves in leading to a readjustment of form, to which the
liquid owing to its mobility readily submits, and no stress
is produced. On the other hand, if a pull is exerted on
a liquid which thoroughly wets and adheres to the internal
surface of a rigid vessel, and, if there are no bubbles or
discontinuities in the liquid, a state of tension inevitably
supervenes.

We have seen that the evaporation taking place from
the outer surfaces of the mesophyll cells is continually
abstracting water from the tracheae of the leaf. It is a
matter of common observation that these tracheae are
constantly filled with water, and they enclose no bubbles.
Experiments on pieces of the conducting tracts of plants to
be described later, show that the adhesion between their
walls and water is at least as great as, and probably much
greater than, the adhesion between glass and water. Hence,
if water is given off from the cells more rapidly than lifting
forces raise it in the trachea), the water in the latter
must inevitably fall into a state of tension.

Apart from root-pressure, investigation has shown that
the only force from below which could be effective in raising
water in plants is the pressure exerted by the atmosphere.
The amounts of water forced up by root-pressure are in-
significant compared with the losses due to transpiration.
Atmospheric pressure can supply the evaporating cells at
most only up to a level of about 10*3 metres. When allow-
ance is made for the resistance opposed by the conducting
tracts to the motion of water in them, we must conclude, that
the supply of water raised by these two forces to a height
of 10 metres above the roots, must be exceedingly small.
It follows that the water in the tracheae above this level is
at all times in tension, and, in times of vigorous trans-
piration, whenever the loss cannot be made good by the
lifting pressure of the atmosphere, the water in the tracheae

90 TRANSPIRATION AND ASCENT OF SAP ch.

of leaves, at lower levels also, is in a tensile state. This
tensile state is no less inevitable at the top of a column
of water unsupported at the base, such as is found in a
high tree, than is the state of compression at the bottom
of a deep vessel filled with water. The former is caused
by the weight acting against the cohesive forces of the
water, while the latter is necessitated by the weight acting
against the resistance of the water to crushing.

Owing to the permeable nature of the walls, the water
in one trachea is continuous with that in its neighbours,
and, consequently, the tension in one is transmitted to
the water in adjacent trachea?. Thus the tension applied
at the mesophyll cell-surfaces is transmitted downwards,
through the water in the tracheae of the leaf and of the
petiole, to the water in those of the stem.

Effect of bubbles. While air bubbles are found
extremely rarely in the tracheae of the vascular
bundles of the leaf, investigators seem agreed that
they are of common occurrence in the conducting
tissues of the stem. It is evident that in the tensile
sap of plants these bubbles will behave in exactly
the same way as we have seen bubbles behave in the
experiments on tensile liquids. If they are sufficiently
minute they will have a very small radius of curvature,
and the surface tension forces preventing them from en-
larging will be correspondingly great. When these forces
balance, or are greater than, the tension in the water,
the tension will be transmitted past the bubbles, and, if
the bubbles adhere to the walls of the tracheae, the tensile
stream will be drawn past them. Kammerling has shown
that a bubble having a radius of 0*01 mm. is in equili-
brium with a pull equal to the hydrostatic head of T65 m.,
while one having a radius of O'OOl mm., = 1/a, could resist
the tension exerted by a column of 16*5 m. of water.
Bubbles having a radius of 1/* would just be visible with
the highest dry objectives commonly in use, their diameter

iv COHESION THEORY OF ASCENT OF SAP 91

being about one-fifth of the diameter of the lumen of the
finest tracheids of the pine. Bubbles of this minute size
are scarcely ever observed in the tracheae of plants. In
fact, the methods of preparation, involving as they do the
relief of the existing tension, or even the exposure to
atmospheric pressure, would cause bubbles of this magni-
tude to disappear. A tension anything greater than that
exerted by a column of water T65 m. will overcome the
surface tension of bubbles having a diameter of 0*02 mm.
and they will tend to expand indefinitely under its action.
Tensions as great as this must frequently occur in plants.
On first thoughts it might appear then, that one bubble
having a diameter of 0*02 mm. or more would destroy the
possibility of tension in the water of the conducting tracts.
A moment's consideration, however, will show that the
structure of these tracts sets a limit to the enlargement
of the bubble. In the conducting tracts, after the forma-
tion of a bubble, the sequence of events will be as follows :
The water round the bubble is drawn away by the tension,
and the surface of the bubble comes to rest against the
wall of the trachea in which it has developed. The retreat-
ing surface is held by the wall, and, as more water is
drawn away, the bubble can enlarge only longitudinally.
At this moment the surface tension of the spherical bubble
is replaced by the capillary forces of the tubular trachea,
and, the capillary forces developed in these tubes being
insufficient to withstand the tension, the bubble gradu-
ally pulls out till it completely fills the trachea. When
this stage is reached the bubble can enlarge no more ; its
surface is restrained on all sides by the walls of the trachea
which, as is well known, though very permeable to water,
are so fine-grained that their capillary or imbibitional
forces are enormous and hold the surface of the water,
limiting the bubble close to their inner surface. Sur-
rounded thus by the imbibed and rigid wall of the trachea
the bubble becomes just like a wetted solid or rigid body

92 TRANSPIRATION AND ASCENT OF SAP ch.

in the tensile current. No doubt it diminishes the effec-
tive cross section of the flow, but, owing to the fact that
the conducting tracts are subdivided into such numbers
of minute compartments, the development of even a large
number of bubbles is unable to wreck the stability of the
tensile column of water in the wood.

The state of affairs in the conducting tissues is illus-
trated in Fig. 17. For the
sake of simplicity a longitu-
dinal section of a conifer's
wood is represented. The
shaded tracheids are sup-
posed to be rilled with water,
whilst the light spaces indi-
cate those containing air-
bubbles, which have been
expanded by the tension of
the transpiration stream till
they completely fill the
tracheids in which the bub-
bles occur. It is evident that
even when a large number
of tracheids are blocked with
air, the water column in the
wood need not be broken,
but may be drawn up round
the bubbles enclosed in, and
rendered harmless by, the
walls of the tracheids. In
the figure, for example, 50 per
cent, of the tracheids contain
bubbles, and yet a consider-
water might be drawn up in the
The imbibitional properties of the

Fig. 17.

able volume of

remaining tubes

walls of contiguous water-filled tracheids render the water

throughout the stem continuous. Consequently, the

iv COHESION THEORY OF ASCENT OF SAP 93

tension developed above is transmitted round the air
bubbles and draws. the stream past them, to use Schwen-
dener's figure, like islands in a river. Hence it is evident
that it would be impossible to sever the continuity of the
water in the conducting tracts, i.e., to prevent evapora-
tion above from transmitting a pull to the water in the
roots, unless tracheae containing bubbles were to form
in some place an unbroken diaphragm across the conduct-
ing tissues of the stem.

Number of air-containing tracheae. From
these considerations it appears that, unless an exceed-
ingly large number of the conducting tubes contain
air and are arranged in a special manner, there is no
likelihood of the tensile column being broken. On the
other hand, the amount of water transmitted in the stream
will be affected by the number of tracheae which contain
bubbles and are consequently put out of action in
the transmission of water drawn upwards under tension.
Hence it is of interest to inquire into our state of know-
ledge as to the air-contents of the conducting tracts.

Results like Hartig's, where the amount of air present
is estimated as a percentage of the volume, cannot be
utilised here. These results do not tell us how the air is
distributed, and it is evident that 10 per cent, of air occur-
ring in each trachea might effectively destroy the cohesion
of the transpiration current, while 50 per cent, placed in
half the tracheae would only diminish the maximum trans-
missibility for a given tension to one half.

Attempts to estimate the number of the vessels and
tracheids which contain air have been made on various
occasions ; but, unfortunately, all the methods hitherto
devised are probably open to error. Of these, Stras-
burger's results seem to be the most reliable.

His general conclusion is that, while a limited amount of
air does not make the conducting tracts impassable to water,
yet in the peripheral parts, which are principally used in

94 TRANSPIRATION AND ASCENT OF SAP ch.

the transport of water, the number of air-bubbles is a
minimum.

On the whole, Ewart's estimate of the number of
tracheae completely filled with water is lower than that of
Strasburger. On the other hand, Strasburger's results
are much more numerous, and possibly Ewart's were
made when the water content of the branches was extremely
low. The methods employed by both investigators seem
open to criticism.

When a branch is cut, even under water, it is possible
that bubbles are formed in the tracheae by the act of
cutting. Bubbles may be formed anywhere close to the
knife, but naturally mostly in the tracheae in contact with
the knife on either side, as the knife introduces a discon-
tinuity, and the water adheres feebly to it. Probably
some of the bubbles observed were thus formed at the
moment of making the preparation for examination, and
were non-existent when the plant was transpiring.

In Ewart's experiments the internal and external pres-
sures were not given time to come into equilibrium ; conse-
quently, supposing 20 per cent, or 30 per cent, of the
vessels contained continuous water, while the remaining
70-80 per cent, contained gas at reduced pressure,
as soon as the branch was cut across, atmospheric
pressure would drive the water from the full vessels
opening on to the cut surface into those which contained
gas at a low pressure, and vessels which had been full
and transmitting a tensile stress during transpiration, would
appear almost empty after cutting. Ewart himself con-
siders that 10 per cent, of the vessels of last year's wood
would transmit enough to cover the losses of the most
vigorous transpiration. Even if a much larger percent-
age than this were completely filled with water at the
time of cutting, they would elude observation in this
method of investigation.

Both Strasburger and Ewart have shown that coloured

iv COHESION THEORY OF ASCENT OF SAP 95

liquids rise most rapidly in the trachea? containing un-
broken columns.

It will be seen that although our knowledge as to the
actual proportion of trachea) containing bubbles during
transpiration is very unsatisfactory, yet observation
supports the view that always during transpiration there
are continuous tracts of tracheae free from air of
considerable cross section. It is also to be remembered
that the periodic flooding of the tracheae with water forced
upwards by root-pressure will bring the bubbles into
solution and will re-establish the conditions for tension
throughout the water-tracts.

Evidence from structure. Here it will be interesting
to consider the structure of the conducting tracts, and
to see how far their details bear out the theory of the
tensile sap.

The salient feature of this structure is the subdivision of
the water-ways by an immense number of longitudinal
and transverse partitions into minute compartments the
vessels and tracheids. For a system the function of
which is to conduct fluids, this is evidently a most unex-
pected configuration. It is true that the partitions are
permeable to water ; but when a considerable distance is
to be traversed, the sum of the resistances opposed by the
walls to the flow becomes important. This becomes clear
from the experiments of Bohm, Elfving, and Stras-
burger, comparing the conductivity of wood in tangential
and longitudinal directions. From their experiments it is
seen that the pressure required to force water in a tangential
direction is immensely greater than that needed to urge it
longitudinally in the wood, although in both cases the water
is free to move through the pits. In the tangential direction,
however, in the same distance the number of walls traversed
may be hundreds of times greater than in the longitudinal
path. It is evident that the persistence of the walls in the
development of the water-conduits of plants introducing,

96 TRANSPIRATION AND ASCENT OF SAP ch.

as they are shown to do, an immense resistance to flow-
is inexplicable on any view which regards the water as
being forced through the stem. Viewed, however, in the
light of the tension hypothesis, this structure becomes a
most beautiful adaptation to confer stability on the
tensilely stressed transpiration stream, and one which
transforms the water, despite its mobility, into a sub-
stance which is stable while sustaining very great stresses,
just as if it were a rigid body. True, the tensile stream
experiences the resistance opposed by the numerous walls,
but the presence of the partitions, conferring, in the manner
just pointed out, a new property on the water, renders
available such an enormous source of energy at the evapo-
rating surfaces in the leaves for the lifting of the sap, that
the amount of energy which is spent in overcoming the
resistance opposed by the walls is relatively insignifi-
cant.

The elongated form of the conducting elements secures
that the resistance shall be small consistently with the
stability of the water ; for, of course, if the tension is great,
a bubble in a long tube renders a larger portion of the
conducting tissues useless than does one confined in a
short vessel ; but on the other hand, when the long tube
is completely filled, it transmits more readily than if it
were subdivided into a number of tracheids. Hence we
may regard the tissue formed of long vessels as the path
of the most rapid part of the transpiration current when
the plant has an abundant supply of water, while the
tracheids transmit the slowly moving water and continue
in function even when supplies are very limited. It is
also evident that the small cross section of the tubes,
though introducing resistance, is most essential. In this
way each bubble which is formed occupies only an infini-
tesimal part of the cross-section of the whole water current.

The structure of the walls themselves is also in com-
plete harmony with the tension hypothesis, and finds its

iv COHESION THEORY OF ASCENT OF SAP 97

most natural explanation viewed in the light of that
hypothesis.

It has long been recognised that the thickenings found
on the walls of the trachea?, viz., the internal supports in
the form of annul i, spirals, and networks, are of such a
nature that they are pre-eminently suited to resist crush-
ing forces. Such strengthenings are quite meaningless
from the point of view of the imbibition and the various
vital hypotheses ; and even according to those views
which regarded the sap pressed upwards by gas or atmo-
spheric pressure they are needlessly strong. For it has
been shown that it is impossible to crush the tubes of a
leaf by an external pressure amounting to 30 atmospheres,
when, according to the theories just alluded to, they would
be exposed at most to one atmosphere. The presence of
these thickenings in the tracheae of the leaves forbids us
accepting Elfving's view that they protect the tubes from
the pressure of the growing tissues. If needlessly bulky
they are disadvantageous because they produce friction
and introduce turbulent motion into the upward stream.
Ewart finds that, owing to the presence of these thick-
enings and to the transverse walls, the flow of water
through the capillary tubes of plants (viz., tracheae) is
only about half what we would expect to find, calculating
the flow by Poiseuille's formula. Consequently, for ordi-
nary methods of transference assumed in earlier theories,
the tracheae of the plant cannot be regarded as efficient.
For the transmission and stability of a tensile stream,
however, these thickenings are essential. And their
strength, so far from being superfluous, is probably often
tested severely in times when the transpiration removes
large quantities of water, and so develops high tensions
in the sap. The whole wall is not thickened uniformly,
because the permeability of the thinner parts is essential.
The thickenings confer on the thin walls the rigidity
necessary to support the tensile stresses in the sap.

H

98 TRANSPIRATION AND ASCENT OF SAP ch.

It is interesting to find that we often have indications
that the unsupported wall would not in itself have sufficient
rigidity to bear the crushing forces it is exposed to. These
indications are particularly frequent in the protoxylem.
Here, commonly, when elongation has widely separated
the rings and spirals, the thin part of the walls of the vessels
is drawn in as a constriction between the spiral or annular
supports, and often the whole vessel is collapsed if the
supports have become too oblique. That this is not due
to the pressure exerted by the growth of the surrounding
tissues follows from the fact that these instances are most
frequently found in leaves.

The most perfect adaptation, to secure the advantages
of ease of flow without seriously reducing the rigidity of
the tracheae, is to be found in the most general of all the
wall-structures, viz., the bordered pit. The membrane and
torus of each bordered pit in the conducting tracheae is
able to take up three positions a median position,
symmetrically dividing each domed chamber of the pit
from the other, and two aspirated or lateral positions.
The median position is naturally assumed by the more
or less tightly stretched membrane when it is not acted
upon by lateral forces. In the aspirated positions the
membrane is deflected against one dome or the other, and
the torus lies over and fills the opening into the dome. The
membranes of pits in the common wall separating two
adjacent tracheae filled with water, naturally take up the
median position. Pappenheim found that an immense
rush of water through the pit was needed to deflect the
membrane to one side. A moderate flow does not disturb
it from its median position. The reason for this is to be
found in the fact that the membrane round the torus is
very permeable to water and, consequently, water moving
at a moderate speed passes through it easily without
displacing it.

The normal transpiration current never possesses the

iv COHESION THEORY OF ASCENT OF SAP

99

velocities which Pappenheim found were necessary to
deflect the membrane, and, of course, hydrostatic tension
in the liquid on each side of the membrane will not tend
to displace it. Hence it is that the tensile transpiration
current, passing from one trachea to another through the
bordered pits, experiences only the very small resistance of
the porous and thin membrane. But the very delicacy
and porosity of the membrane render it unsuitable for
sustaining any severe stress. Hence we find, when a bubble
develops in a trachea and is gradually distended by the
tension in the liquid, or by a difference of gas pressure,
till it fills the trachea, the membranes
of the pits in the walls of the trachea
become aspirated away from the
bubble, and the membrane is sup-
ported by the dome, while the torus
lies over the perforation in the latter
like a valve on its seat (.see Fig. 18).
In this position of the membrane the
tension of the water and the gas pres-
sure are withstood, not by the thin
and delicate membrane, but by the
surface of the water supported by the
denser and more rigid material of the
wall and of the torus, while the delicate
membrane is shielded from all stress.

Thus, from the point of view of the
tension hypothesis, we regard the bor-
dered pits as mechanisms to render the
walls as permeable as possible to con-
tinuous water streams, while, when conditions require,
they provide, by an automatic change, a rigid support
to the tensile sap and oppose an impermeable barrier to
undissolved gas.

Fig. 18.

H 2

ioo TRANSPIRATION AND ASCENT OF SAP ch. iv

Literature.

Bohni, J., " De la cause du mouvement de l'eau efc de la faible pression de
I'air dans les plantes," Ann. des Sciences Nat. Bot., 1881, 12, p. 233.

Id. " Ueber die Ursache der Wasserbewegung und der geringeren
Lufttension in transpirirenden Pflanzen," Bot. Ztg., 1881, 49, s. 801 and
817.

Dixon H. H., "On the Physics of the Transpiration Current," Notes from
the Bot. School Trin. Coll., Dublin, vol. i, p. 57.

Dixon H. H., and Joly J., "On the Ascent of Sap " (abstract), Proc. Roy.
Soc. London, 1894, vol. 57 B, p. 3.

Id. "On the Ascent of Sap," Phil. Tram. Roy. Soc. London, 1895. vol.
186 B, p. 563.

Ewart, A. J., "Ascent of Water in Trees," Phil. Trans. Roy. Soc. London,
1905, vol. 198, p. 41.

Id. " Ascent of Water in Trees," Phil. Trans. Roy. Soc. London, 1908, vol.
199 B, p. 362.

Hartig, R., "Ueber die Vertheilung der organischen Substanz, des
Wassers und des Luftraumes in den Bitumen, und iiber die Ursache der
Wasserbewegung in Transpirirenden Pflanzen," Berlin 1882.

Joly, J., " Report of a Discussion on the Ascent of Water in Trees," Brit.
Assoc. Report, 1898.

Kammerling, Z., " Oberflachenspannung und Cohasion," Bot. Centrulb.,
1898, 73, s. 369, 439, and 465.

Pappenheim K., " Zur Frage der Verschlussfahigkeit der Hoftiipfel in
Splintholze der Coniferen," Ber. d. Deutsch. Bot. Gesell., 1889, 7, s. 2.

von Sachs, J., "Lectures on the Physiology of Plants," translated by H.
Marshall Ward (Oxford 1887.)

Strasburger, E., "Ueber den Bau und Verrichtungen der Leitungsbahnen
in den Pflanzen " (Jena 1891).

CHAPTER V

TENSILE STRENGTH OF THE SAP OF TREES

Even in text-books of Physics the cohesion of liquids
is seldom discussed, and the conditions necessary to pro-
duce a state in which liquids may transmit a tensile stress
are not adequately treated.

Cohesion of liquids. Donny in 1846 showed that
it was possible for a column of sulphuric acid
1255 m. high to hang in a vertical tube closed at
its upper end, when atmospheric pressure was not
allowed to press the liquid upwards from below. He
compares the phenomenon to the well-known experience
that the mercury of a barometer may be retained above the
actual barometric height, if the tube, filled by inclining it,
is raised gradually to a vertical position. He further states
that this phenomenon has been explained by Laplace as
being due to the cohesion of the mercury and to its adhesion
to the glass. Donny also looked for the cohesion or tensile
strength of water. He appears, however, to have failed to
demonstrate it in the same way in which he had successfully
showed it in the case of sulphuric acid. He observed, how-
ever, the tensile strength of water in the following less direct
manner : If a vertical glass tube one metre long, partially
filled with water and sealed at both ends, is struck vigor-
ously on the lower end with the palm of the hand, bubbles
open in the liquid and instantly close again with a metallic

102 TRANSPIRATION AND ASCENT OF SAP ch.

click. A blow on a tube, which has been similarly set up,
but from which the air has been removed by careful ex-
haustion, produces no bubbles ; nor is a click heard.
Donny explained that in the first case the blow causes
minute bubbles to be opened against the forces of surface
tension and of atmospheric pressure, therefore, in the
second case where bubbles are not formed the cohesion
must be greater than these two forces together. Donny
believed that even a little air in solution suffices to reduce
the cohesion of a liquid to an insensible figure. This error,
soon to be corrected, has been frequently copied by writers
on this subject. Donny also pointed out that the boiling
of liquids is retarded when air is removed, owing to their
increased cohesion, and it is the sudden rupture of the
liquid which causes explosive boiling.

As will presently be made clear, the absence of dissolved
air from the water is a condition by no means necessary
for its cohesion, and in these experiments it appeared neces-
sary only because by the removal of the dissolved air perfect
contact with the glass and complete wetting of the dust
particles suspended in the liquid was secured.

In his memoir Donny l points out that when one with-
draws a plane disc from contact with a surface of water
the tensile strength of the latter does not come into play.
As the disc is raised, water adheres to its lower surface,
but the column of water connecting the disc with the
liquid below grows gradually thinner, until at a moment
when the disc is removed a certain distance above the
general level of the lower liquid, the column spontaneously
draws in from the edges, and, when its diameter becomes
extremely small, breaks in two. He shows also that, in
a tensile liquid column, a bubble, sufficiently small to have

1 The presence of undissolved air, or unwetted surfaces, or both, probably
prevented Janse from obtaining considerable tensions in the experiment
quoted by him : " Der attfsteigende Strom in der Pflanze," Jahrb. f. M iss.
Bat., 1908, 45, 3, p. 314.

v TENSILE STRENGTH OF SAP OF TREES 103

surface-tension forces capable of supporting the hydro-
static head of the liquid below, will not destroy the tensile
state.

Berthelot's estimate. Berthelot a few years
afterwards succeeded in showing directly that water
has a very considerable cohesive strength and, under
proper conditions, can sustain a very great tensile
stress. His procedure was as follows : He rilled a
strong capillary tube, which was sealed at one end and
drawn to a fine point at the other, with water at a
temperature of 28 or 30. He allowed it to cool to 18,
and, as it cooled, to draw in air. Then the fine-drawn
end was sealed. The tube was now heated to 28, or
over, and the air forced into solution in the water which
now occupied the whole of the internal space of the tube.
On cooling to 18 or lower, it was found that the liquid
continued to occupy the entire space enclosed by the tube.
From this he argued that the water preserved the same
density from 28 to 18. The dilatation needed to effect
this is very large, viz., l-420th of its volume at 18.
To produce a similar effect in the opposite sense would
require a pressure of about 50 atm. ; and it was concluded
that the experiment showed that neither the adhesion to
the glass nor the cohesion of the water is less than 50 atm.

Berthelot's experiment has been variously misquoted,
(1) with regard to the dilatation observed, and (2) as to
the effect of dissolved air on the tensile strength of water.
The dilatation has been quoted as amounting to 1-1 20th
of the volume instead of 1 -420th. This, of course, gives
a much too high result for the tension obtained in the
experiment. Hence Ewart's quotation of Berthelot in
support of the statement that air-free water can sustain
a tension of 200 atm. was illegitimate. The minor limit
obtained in Berthelot's experiment was 50 and not 200 atm.

Again, it is quite usual, when treating of the cohesion
of liquids, to state that Berthelot's experiments were

io 4 TRANSPIRATION AND ASCENT OF SAP ch.

carried out with air-free liquids. As a matter of fact,
his method of experiment shows that the water contained
air, and he expressly states that the water was, in his first
experiments, supersaturated, and that it was only at
Regnault's suggestion that he carried out experiments on
air-free water. His final conclusion is ' Le phenomene
[dilatation of water under tension] se produit done dans
le vide aussi bien que dans l'air, et est independant de la
sursaturation."

Cohesion of water containing dissolved air.-
Misled by the misquotations just alluded to, although
a priori there seemed no reason to suspect that the presence
of dissolved air would weaken the tensile strength of water,
Dr. Joly and the author considered it necessary to investi-
gate the point specially.

We used a cylindrical glass vessel with rounded ends
and provided at one end with a narrow tubulure. This
vessel was very carefully cleansed by washing it internally
successively with caustic potash solution, dilute acid, and
distilled water. Half filled with water it was boiled for
some time to make sure that the walls were thoroughly
wetted ; then it was almost completely filled with water
which had been previously boiled to get rid of undissolved
air and to wet thoroughly all dust particles which might
have been contained in the liquid. By subsequent
exposure, this water was allowed to become saturated
with dissolved air. During this care was taken to
shield the water from dust, which might not have been
completely wetted, or which might have introduced small
bubbles. To fill the vessel a small quantity of water in
it was raised to ebullition, and, while steam was issuing
from the attenuated tubulure, the latter was submerged
in the dust-free water. As the steam within condensed,
and the vessel cooled, the latter became completely filled
with water. A small bubble was then introduced, and the
vessel was closed by sealing off the tubulure.

v TENSILE STRENGTH OF SAP OF TREES 105

If the vessel was then cautiously heated, the water ex-
panded more than its glass envelope, and the air bubble
was compressed. The bubble became smaller and smaller
as the temperature rose and the contained gas was forced
into solution. When the bubble had reached very small
dimensions and was about to disappear, great care had to
be exercised in the further application of heat ; for, if the
water expanded too much and strained the glass beyond
its elastic limit, the whole experiment was rendered abortive
by the breaking of the glass. But, if the heating process
had been carried out successfully, all the air had been
dissolved so that the water had been made to fill the
vessel completely without breaking it.

At this moment the water in the vessel was either in
compression, being constrained by a tension in the glass-
walls, or it was quite unconstrained, just exactly filling the
envelope, and neither suffering compression nor causing
tension in the walls. As soon as cooling began, the water
and the glass commenced to contract. The coefficient of
thermal expansion of water being greater than that of
glass, the water tended to contract more. This contrac-
tion, however, was resisted by its adhesion to the glass
and its own cohesion, and consequently a tension, which
kept it sufficiently dilated to fill the glass, was set up.
As cooling proceeded the tension grew greater and
greater, till at last either the adhesion or cohesion
was overcome and a break appeared between the water
and the glass or in the substance of the water itself.
This rupture was signalised by a sharp click, and a
bubble sprang into existence, which rapidly augmented in
size as the water, now relieved from the stretching forces,
assumed a volume corresponding to its temperature at the
moment. Bubbles appeared round the original bubble and
passed into it.

By estimating the amount of deformation of the glass
envelope when strained by the contracting water, and by

106 TRANSPIRATION AND ASCENT OF SAP ch.

determining experimentally the pressure needed to produce
the same deformation, the amount of the tensile stress,
which was sustained by the water before rupture, was
determined. In an experiment, carried out in the
manner just described, water was subjected to a tension
equivalent to 7 '5 atmospheres before its cohesion was
overcome.

Cohesion of a soap film. It may here be pointed
out that every stretched water film not only gives us a
demonstration of the tensile strength of water, but also
enables us to set to it a minor limit, which does not fall
far short of that determined by Berthelot.

A film of soapy water stretched upon a rigid frame is

stable even when it is only 12^, or 12 x 10 " 7 cm. thick. 1

The thickness of the film is measured by the interference

phenomena of light from the opposite surfaces of the film.

When thus stretched, the film supports the stress of twice

the surface tension (T) of the soap solution. This force

has been determined as about 25 dynes per centimetre.

Hence it follows that the film which has a cross section of

50 x 10
12 x 10 " 7 sq. cm. supports a tension of 50 dynes or

dynes per sq. cm. An atmosphere pressure is equivalent
to 1*0132 x 10 6 dynes. So we find, according to this method,
that the cohesive strength or tenacity of water must at
least be equal to 41 3 atmospheres. 2

New determinations on water. Berthelot's ex-
periment with water enclosed in thick capillary
tubes, just quoted, is quite easy of repetition. As
some of my experiments with this method have given a
much higher minor limit for the tensile strength of water
than his, I have thought it of interest to record them here.

The lengths of the tubes used in my experiments varied
between 14*5 cm. and 220 cm.

1 In the second black film, Johonnot states the thickness may bo only (J ju/u..

2 I am indebted to Mr. J. lv. Cotter for this calculation.

v TENSILE STRENGTH OF SAP OF TREES 107

In each case the tubes were first cleaned with a solution
of caustic potash, which was afterwards removed by
repeated rinsing with boiled, distilled water. A piece
of the wood of the yew (Taxus baccata) was then
introduced, and, after being filled with boiling water,
the tube was kept submerged in boiling water for
an hour or so. Before sealing the drawn-out end,
the water was allowed to cool ; and a millimetre or
more of the bore was cleared of water by warming the
point in a flame. When all was cool, the fine end was
sealed.

In the table given on p. 108 is a record of my experiments ;
for each experiment the temperature t 2 at which the tube
is completely filled, and the temperature t v at which
rupture took place, are given in the fifth and sixth columns
respectively.

From these observations the tension may be roughly
determined according to the formula

a = JL Zizli

where /3 = the coefficient of compressibility of water, V^ =
volume under the lower pressure P l5 V, = volume under the
increased pressure P 2 . A tension = P, will bring about the
same change of volume in the opposite sense. Evidently the
apparent change of volume of the water due to this tension
must be corrected for the contraction of the glass cooling
through the range of the experiment. The correction is
allowed for by the formula

t- ( a -f/)(^'-i)

~/3[l-a(* 2 -*!>]'
where

T = tension ;

a = coefficient of expansion of water over the range ;

;/ = coefficient of cubic expansion of glass = 2 "4 x 10

t l = temperature of rupture ;

t 2 temperature when tube is full.

io8 TRANSPIRATION AND ASCENT OF SAP ch.

Introducing the correction for the elastic yield of the
glass, this becomes

T =

(a-gO(* 2 -*i)

0[i-a(* 2 -*i)] +

i?-'

k n

where

K = external radius of tube,

r = internal radius of tube,

k = compression modulus of glass (volume elasticity) = 4 x 10'" atm.,

n = torsion modulus of glass (torsional rigidity) = 3 x 10 5 atm.

The value of a was obtained from the table of the volume
of water at different temperatures in Landolt-Bornstein,
Physikalisch-Chemische Tabellen, by E. Bornstein and
W. Meyerhofer, Berlin, 1905, pp. 38 and 39. The com-
pressibility coefficients /3 for different pressures and tem-
peratures are given on p. 60 of the same tables.

Table 10.

1

2

3

4

5

6

7

8

9

No. of
Experi-

No. of
Tube.

R

r

h

h

/3xl0 r

a x 10 5

T

in atmo-

ment.

spheres.

1

I.

25

05

46-0

28-4-

437

36

125 8

2

I.

2-5

05

46

27

5

438

36

132

3

I.

2-5

0-5

47

9

25

8 a

439

36

157

5

4

II.

2-5

0-5

85

75

8"

444-8

65

120

5

5

11.

2-5

0-5

84

9

72

9

441-9

65

158

4

II.

2-5

0-5

84

4

72

9

4419

65

151

7

7

II.

2-5

0-5

83

6

74

5

443-5

65

119

5

8

III.

35

05

46

36

6

453

39

70

8

9

III.

3-5

0-5

45

9

36

2

453-0

39

73

1

10

III.

3 5

0-5

46

0'

36

6 3

453

39

70

8

11

IV.

3-5

0-5

34

5

25

6

464-0

30

49

4

12

IV.

3-5

05

34

5

27

V

464-0

30

37

7

13

V.

3-5

0-5

50

45

;

449

44

43

1

14

VI.

35

05

56

45

or-

427-0

46

104

4

15

VII.

35

0-5

69

3 3

01

8

456

57

83

8

16

VII.

3 5

05

70

2

02

8

457-0

57

82

5

17

VIII.

3-5

0-5

70

64

r

458-0

58

00

18

VIII.

35

0-5

70-2 3

59-3

427-0

56

127 4

1 I am indebted to Mr. J. R. Cotter for adapting these formulae and
rendering them suitable for application to these experiments.

v TENSILE STRENGTH OF SAP OF TREES 109

In Table 10 are detailed the experiments on eight
different tubes, and in it are recorded the radii R
and r, the observed temperature when the tubes were
full L, and the temperature at which the rupture took
place t 1} together with a, the coefficient of expansion
for water over the range (t^ t^, and finally the tension
calculated according to the above formula.

In making the observations on the temperatures, the
tubes were set in a large beaker of water. The tempera-
ture at which the tube became filled was then roughly
determined by warming the beaker up very slowly till the
bubble in the tube disappeared. The tube was then allowed
to cool, and the bubble to reappear. The beaker was again
raised to a temperature one degree below that at which
it was expected the bubble would disappear. The beaker
was kept at this temperature for five or ten minutes, and
the water within it was kept in motion to secure a fairly
uniform temperature. Fifteen minutes were occupied in
raising the water of the beaker through the next degree,
so that the water in the capillary tube must have very
closely approximated to the temperature indicated by the
thermometer in the beaker. By proceeding in this way
every effort was made to avoid exaggerating the tempera-
ture at which the tube filled. The large amount of
water in the beaker secured that the cooling should be
extremely slow before the reappearance of the bubble,
so that it is improbable that the thermometer gave
readings sensibly different from the temperature of the
tube.

In spite of these precautions, differences are observed
in the successive experiments with the same tube. These
can scarcely be due to errors of observation. It generally
happened that when a tube was heated on successive days,
lower readings were obtained for the " full " temperatures
in the later observations. See Experiments 4, 5, and 6
on tube II. Occasionally, however, the " full " tempera-

no TRANSPIRATION AND ASCENT OF SAP ch.

ture rises after a time, viz., Experiments 1, 2, and 3 on
tube I.

As a general rule, the temperature of rupture lowers
with time, suggesting that adhesion is improving. It may
be noted that some very high values of adhesion of water
to copper were accidentally obtained in tube II. , in which
by chance some minute shavings of copper were included.

These experiments amply confirm Berthelot's observa-
tions of the tensile strength of water containing air, and they
raise the minor limit obtained by him of its cohesion and
its adhesion to glass from 50 a tin. to more than 150 atm.
They further show that the adhesion to the walls of the
conducting tracts is also over this figure.

Tensile strength of sap. In many instances
sap adhering to the vegetable tissues introduced into
the tension tubes showed that its cohesion was no less
than the rest of the water, and, indeed, there was no
reason to suspect that it would have less tensile strength.
However, recently the doubt has been raised that sap,
especially if containing dissolved air, has not the same
cohesive properties as water. Consequently it seemed
proper to test the matter by direct experiment.

With this end in view, the sap centrifuged from pieces of
branches of Fagus silvatica, cut from about 70 feet above
the level of the ground, was enclosed in a tension-tube.
This sap, after collection, was boiled on three successive days
for about one hour in order to secure the complete wetting
of dust-particles fortuitously contained in it. After its last
boiling it was exposed for twenty-four hours as a thin layer
about 4 mm. deep to the air, but shielded from dust.
In this way it must have become practically saturated
with dissolved air. The capillary tube, into which it was
now drawn by alternate heating and cooling, had been very
carefully cleaned by successive washings out with chromic
acid, caustic potash, and boiled water. After this cleaning
the tube was boiled for about an hour on three successive

v TENSILE STRENGTH OF SAP OF TREES in

days in water, heating and cooling being effected in the
same water. The tube was emptied before each boiling,
and allowed to fill with the freshly boiled water. The
.object of this was to wet the tube thoroughly, and any
dust-particles it contained, by bringing all undissolved air
on their surfaces into solution. The tube, after filling with
the sap to within a few millimetres of its end, was sealed
off. The heating of the tube was effected, as in my pre-
vious work, in a large volume of water, and was very slow.

In the first tube submitted to experiment the air-
bubble disappeared at 63*5, which may be described
as the " closing ' temperature, and reappeared with the
characteristic click at 59 1. Three other observations
were made with this tube. All four agree in indicating that
the sap withstood a tension of more than 45 atmospheres
before rupture (cp. Experiments 1, 2, 3, and 4 in
Table 11, p. 113).

A second tube was charged with some of the same
sample of sap ; it was found to become completely filled
at 66*2, and ruptured at 59*5. Calculating the tension
developed in this case the result is more than 70 atmo-
spheres (see Experiment 5 in the Table). In another
experiment with this tube a tension of about 50 atmo-
spheres was produced (see No. 6).

It was thought that possibly, by keeping one of these
tubes after closing at a temperature close to that at which
the bubble disappeared, greater tensions might be attained.
This surmise was not realised. The tube used in the first
experiments described above was kept for two days at a
temperature of about 61. However, when ultimately
allowed to cool slowly, the rupture occurred at 59*2, a
temperature not quite so low as had sometimes before
been successfully passed. This experiment is recorded as
No. 3 in Table 11.

It mav be noted that there is no reason to believe that
the tensions produced in these experiments are indications

ii2 TRANSPIRATION AND ASCENT OF SAP ch.

of the maximum cohesion of boiled sap. The results
quoted happen to be the first obtained. Other experiments
were not made, as these are sufficient to demonstrate that
the boiled sap possesses cohesive properties of the same
order as those of water.

Having found that sap, free from unwetted nuclei, but
saturated with air, is able to sustain considerable tensions,
it seemed worth while trying if unboiled sap could be put
into the tensile condition. The consideration that heating
the enclosed sap in the glass envelope until the last visible
bubble disappeared would probably remove completely all
invisible bubbles encouraged me in this attempt. Accord-
ingly a quantity of sap was collected from a branch of
Ilex aquifolium by means of centrifuging ; and this after
exposure to air and without any special treatment was
introduced into several capillary tubes, which had been
prepared in a manner similar to those used in the other
experiments.

The first tube closed at a temperature of 78'2, and
ruptured on cooling to a temperature of 72*0 (see
Experiment No. 7 in Table 11). This rupture occurred
simultaneously with a slight shock accidentally dealt it by
the stirrer of the vessel of water in which it was immersed.
Had it not been for this, probably a lower temperature
would have been attained without rupture. Taking these
figures and the dimensions of the tube into account, the
tension developed must have been about 75 atmospheres.

Another tube containing some of the same sample of
sap completely filled at a temperature of 91 10 C. On
one occasion rupture took place only when a temper-
ature of 76*2 was reached, on another a rupture
developed at some temperature below 81*5. In the
latter case, when the tube had fallen to 8l'5, it was
withdrawn from the water for examination, and rupture
occurred some seconds after it was lifted from the water.
In the first instance the tension must have approxi-

v TENSILE STRENGTH OF SAP OF TREES 113

mated to 207 atmospheres ; while in the second a tension
of about 132 atmospheres was attained before rupture
occurred.

The former of these is, I believe, the highest tension yet
experimentally produced in any liquid. Possibly this very
good cohesion possessed by unboiled sap is due to the pres-
ence of colloids in it. It seems probable that when the
tension is just adequate to start a rupture, if the latter
remains sufficiently small, its surface tension will be able to
withstand the stretching action due to the contraction and
cohesion of the liquid. Thus, if the rupture, at its first
inception, can be delayed in spreading, it may be obliterated
and cohesion re-established. The presence of the colloid
may bring about the necessary delay. The appearance

Table 11.

No. of
Experi-
ment.

No of
Tube.

External

Radius

li.

Internal
Radius

Closing
Tempera-
ture t.

Tempera-
ture of
Rupture

Coefficient

of Com-
pressibility
pxW.

Coefficient
of Expan-
sion
a X 10*.

Tension
in Atmo-
spheres.

1

s,

3-57

0-50

63-5

59 1

455-9

54

1

47

2

Si

357

0-50

63-5

592 3

4559

54

1

45

3

s,

3-57

50

63 5 C

59 -2

455 9

54

1

45

4

s x

3-57

0-50

63 -4-

59-0

455 9

54

46

5

So

3 57

50

66-2 D

59 5

457

55

1

73

6

s.

3-57

0-50

66-2

61 -2

4576

55

7

54

7

s 3

3 50

100

78-2

72-0

464-4

63

1

75

8

s 1

3-50

1-00

91 1

77-2

453-0

68

6

192

9

o 4

3-50

100

91 -r

81-5 J

454-5

70

132

10

s 4

3-50

100

9i r

76-2 ?

440 5

68-3

207

exhibited occasionally in these sap-containing tubes may
be interpreted as favouring this view. The click of
rupture is not, in these cases, attended by the development
of a single bubble becoming surrounded by a group of
small visible bubbles, but, at the moment of rupture, a
milky, semi-opaque region develops in the tube. This
slowly rises, and clears away, as it turns into a mass of
excessively minute bubbles. Here, apparently at the de-

ii 4 TRANSPIRATION AND ASCENT OF SAP ch.

struction of cohesion, countless numbers of minute ruptures
have been simultaneously produced.

Experiment No. 8 gives the details of a third observation
with this tube.

The tubes S a and S 2 were filled with boiled sap of Fagus
silvatica, which was, however, subsequent to boiling, ex-
posed in a thin layer to the air ; while the tubes S 3 and
S 4 , on which Experiments 7, 8, 9, and 10 were performed,
contained unboiled sap of Ilex aquifolium.

In Experiment 3, after the bubble had been " closed '
at a temperature of 63, the tube was maintained at
about 61 for two days. During this time no rupture
appeared.

The foregoing shows that the sap of trees has consider-
able tensile strength, and in this respect does not differ
from water. In the few experiments made, the ease with
which tension was generated and its magnitude before
rupture occurred, possibly indicate that sap is somewhat
more stable under tension than pure water.

Literature.

Berthelot, M. , " Sur quelques phenomenes de dilatation forcee des liquides,"
Ann. de Phys. et de Ghim. 1850, 30, p. 232.

Dixon, H. H., "Note on the Tensile Strength of Water," Proc. Boy.
Dublin Sue. 1909, vol. xii (N. S.), p. 60, and Notes from The Botanical School,
Trinity College, Dublin, vol. ii, p. 38.

Id. " Vitality and the Transmission of Water through the Stems of Plants,"
Proc. Roy. Dublin Soc. 1909, vol. xii (N. S.), p. 21, and Notes from the
Botanical School, Trinity College, Dublin, vol. ii, p. 58.

Id. " On the Tensile Strength of Sap," Proc. Boy. Dublin Soc. 1914, vol.
xiv (N. S.), p. 229.

Dixon, H. H., and Joly, J., "On the Ascent of Sap," Phil. Trans. Boy.
Soc. London, 1895, vol. 186 B, p. 568.

Id. "The Path of the Transpiration Current," Ann. of Bot. 1895, 9,
p. 404.

Donny, J., "Sur la cohesion des liquides et sur leur adhesion mix corps
solides," Ann. de Phys. <! ('him. 1846, 16 Ser. iii, p. 167.

Evvart, A. J., "Resistance to Flow in Wood Vessels," Ann. of Botany,
1905, vol. 19, p. 442.

Johonnot, E. S., " Thickness of the Black Spot in Liquid Films," Phil.
Mag. 1899, No. 289, p. 50].

v TENSILE STRENGTH OF SAP OF TREES 115

Laplace, " Traite de Mecanique Celeste," Supplement au X libre, p. 3.

Poynting, J. H., and Thomson, J. J., "Text Book of Physies, Properties
of Matter," 2nded., p. 123

Ursprung, A., " Zur Demonstration der Flussigkeits-Kohasion," Ber. d.
Deutsch Bot. Gesell. 1913, Bd. 31, s. 888.

Id. " Ueber die Bedeutung der Kohasion fiir das Saftsteigen," Ber. d.
Deutsch But. Gesell. 1913, Bd. 31, s. 401.

Worthington, A. M., " On the Mechanical Stretching of Liquids," Phil.
Trans. Ron. Soe. London, 18 ( .2, vol. 183 A, p. 355.

1 2

CHAPTER VI

ESTIMATE OF THE TENSION REQUIRED TO RAISE THE SAP

In the previous chapter it has been convenient to quote
the various estimates hitherto obtained as to the tensile
strength of water. We have seen that, contrary to what
our everyday experience seems to teach us, water has a
very considerable tenacity, amounting at least to 150 atmo-
spheres, about 15 kilos, per sq. cm. Sap extracted from
the conducting tracts is, in this respect, not inferior to
pure water. Moreover, the presence of dissolved air in it
does not diminish its cohesion, which, in this condition,
has been demonstrated up to 200 atm., or 20 kilos, per
sq. cm.

It now remains to determine how far this tenacity will
be taxed, and what forces are required for moving the
tensile sap through the water tracts of the plant.

Ewart's high estimate of the resistance.
In a paper published in 1905, Ewart investigates the
question as to what force is required to move water
through the waterways of plants at the same velocity as
the transpiration current. His general conclusion is that
the resistance is so great that neither are sufficient forces
generated in the leaves to raise the water at the required
velocity, nor is the tensile strength of water adequate to
transmit these forces downwards, if such existed. In a
subsequent paper he seems to have modified this latter

110

vi TENSION REQUIRED TO RAISE THE SAP 117

view, and misquotes Berthelot as stating that air-free water
may support a tension of more than 200 atmospheres.

It is needless to criticise Ewart's calculations of the
resistance based on Poiseuille's formula, which he himself
admits is quite inapplicable to the case, owing to the
presence of transverse partitions and irregularities in the
cross-sections of the tracheal tubes. It may be noticed,
however, that when care was taken that discontinuities
were not present in the water columns of the wood experi-
mented upon, the flow observed was of the same order as
that calculated by the formula. In an experiment on a
piece of yew wood the approximation was very remark-
able. The actual amount transmitted through a length
of 15 cm. was 4'2 c.c. per hour, while the calculated amount
was 9*8 c.c. The average distance from one another of
the cross-partitions in the fine tubes composing the wood
is - 25 cm. Therefore, about 60 partitions must be
traversed in passing through a length of 15 cm. This
indicates that the resistance offered by the walls, or
rather by the pits in the walls, to the passage of water is
very slight.

First method. Ewart also endeavours to find experi-
mentally the resistance offered to the transpiration current.
His first method, that of Janse and Strasburger, was as
follows : " Leafy branches 4 to 8 feet in length were cut
under water and kept in darkness for half an hour.
Clean ends were then cut under water, placed in freshly
filtered eosin solution, and at once exposed in the open
on bright, cloudless, breezy June days between 10 a.m.
and 1 p.m., with a shade temperature averaging 18 to
22 C. The conditions for transpiration were, therefore,
optimal.

" After a timed period the stem was removed, and
rapidly sectionised from apex downwards until the eosin
solution was visible in the wood, the length of the remain-
ing portion of the stem giving the rate of flow during the

n8 TRANSPIRATION AND ASCENT OF SAP ch.

period of observation. Portions of the same stems, and
also similar ones from the same plants were then subjected
to varying water pressures until closely corresponding rates
of flow were reached."

By this method, and by assuming that the velocities
shown in these branches are maintained throughout the
stem, Ewart obtains results which indicate that, in order
to move water in the stems of plants at the velocity
of the transpiration current, pressures equivalent to a
head of water from 6 to 33 times the height of the plant
are required.

Objections. There are several reasons why this unex-
pected result of Ewart's must be regarded as incorrect.
(1) The velocity of flow given in E wart's experiment is
probably far in excess of even the maximum velocity of the
transpiration current in the intact plant. (2) The velocity
cannot be assumed to be uniform throughout high trees ; but
may fall off from below upwards. (3) E wart's results for
the resistance to flow in stems are not in agreement with a
large body of experiment to be quoted presently, but
appear to be excessive.

First, with regard to the velocities in Ewart's experi-
ment and in intact trees. In the latter the lifting forces
generated in the leaves must do work against the resist-
ance to flow all along the path of the current, and, if the
supply is inadequate from the roots, against other opposing
forces in addition, and against the whole hydrostatic head.
In Ewart's experiment, not only are the resistance of the
lower part of the conducting system, the other opposing
forces, and the hydrostatic head removed, but they are
replaced by the atmospheric pressure acting as a vis a
tergo, urging the water upwards. Naturally, then, a much
greater velocity is attained in the latter case than when
the branch is still attached to the tree. 1

1 It may be mentioned that Janse's method of observing the amount of
transpiration by successive weighings of a severed branch is not falsified by

vi TENSION REQUIRED TO RAISE THE SAP 1 19

The subsequent wilting of the leaves of cut branches
shows that clogging afterwards greatly reduces the flow ;
but, of course, in the experiment quoted it is only the
initial stage which is recorded.

The assumption that the velocity in the terminal
branches is as great or greater, than in the trunk, is also
not justified. The lower leafy branches have to do work
against a smaller hydrostatic head, and against a smaller
resistance than the outer and upper branches, and conse-
quently the flow will be faster through the lower parts
than above. Furthermore, in many cases the effective
cross-section supplying unit transpiring area is greater
above than below. A good instance of this was brought
under my notice by Dr. J. Joly, in a young specimen of
Abies excelsa, which had just been felled. The tree was
550 cm. high. The section at its base showed 17 year rings.
The lateral branches were almost uniformly clothed with
leaves, so that the length of the branches was approxi-
mately proportional to the leaf-area they supported. At
52 cm. from the apex the area of the cross-section of the
wood was 63 sq. cm., and the sum of the lengths of the
leafy branches above this was 170 cm. At 134 cm. from
the apex the cross-section was T29 sq. cm., and the sum
of the lengths of the leaf-bearing branches above was
1500 cm. At the higher level the cross-section of the
supply conduits was 0*37 sq. cm. per 100 cm. of leafy
branch, at the lower point the cross-section of the supply
was 0'286 sq. cm. per 100 cm. of branch. Therefore, in
the case of this tree, if all the leaves were transpiring
uniformly, the velocity at the lower level must be greater
than above. Hence, in Ewart's experiments it is quite
possible that the velocity below was considerably in excess
of that in the upper parts of his severed branch, and

this error, but probably gives too small an amount after transpiration has
proceeded for some time. Strasburger considers that this is corrected more
or less accurately by his over-estimate of the resistance.

no TRANSPIRATION AND ASCENT OF SAP ch.

certainly without special measurements it is not legitimate
to assume that the velocity throughout the length of the
branch was uniform. This is equally true with regard to
intact large trees.

With regard to the resistance, it will be seen later on that
Ewart's results are considerably too high.

Second method. The second method of deter-
mining the maximum velocity of the transpiration
current employed by Ewart is also open to objection.
It is described by him as follows : A small branch
bearing a small number of leaves, while still attached
to the tree, is led through a split rubber cork
into an air-tight glass chamber containing a weighed
quantity of calcium chloride. The gain in weight of the
latter gives the amount of water transpired by the leaves
in a given time. Assuming, then, that all the leaves of
the tree may act like those in the closed glass chamber,
the number of leaves on the whole tree will give the weight
of water transpired by the tree, and consequently the
amount of water which passes up the tree in a given time.
Then, by measuring the effective cross-section of the
trunk, the velocity of the transpiration current may be
estimated.

Criticism. The objections to this method are obvious.

1. When the branch is first introduced into the desic-
cated chamber it will lose water more rapidly than when
it was transpiring into the less dry external air. It will
continue this abnormally rapid rate of transpiration until
the concentration of the vacuoles of the evaporating cells
reaches a steady state, depending on the freedom of
supply from the water conduits and the vapour pressure
in the chamber, but, until this steady state is attained,
the amount of water entering the calcium chloride may
be largely in excess of that passing up the stem.

2. But a more serious source of error is the assumption
that all the leaves of the tree can transpire at the same

vi TENSION REOUIRED TO RAISE THE SAP 121

rate as those on the single desiccated branch, whereas the
supply, which, under the conditions of the experiment,
is available for the very actively transpiring branch,
would be largely encroached upon if all the branches were
under equally favourable conditions for transpiration. In
fact, the single branch in the desiccated chamber and still
attached to the rest of the tree, which is under normal
conditions of moisture, is under conditions of supply
approximating to those of a cut branch set in water and,
for the same reasons, cannot be assumed to give a correct
estimate of the velocity of the transpiration current
throughout the whole tree.

The validity of this objection may be demonstrated
experimentally by weighing the amount of water given
off by a given number of leaves in a desiccated chamber,
and comparing this amount with the quantity of water
transpired by the same number of leaves on the same tree
exposed to normal conditions of maximum transpiration.

It will be of interest to quote one of these experiments : A
small yew-tree was removed from the flower-pot in which
it had been grown, and the roots, and their surrounding
soil, enclosed in a rubber bag ; to prevent loss of water,
except from the leaves, the opening of the bag was tied
tightly round the stem. Periodic weighings gave the
amount of water transpired. At the same time, a branch
still attached to the tree was introduced into a hermeti-
cally closed flask containing calcium chloride. The flask
could be removed and weighed periodically. A rubber
bag, similar to that enclosing the roots, filled with moist
earth and closed, was exposed to the same conditions
and weighed simultaneously, thus giving a small correc-
tion for loss through the bag. In one of these experiments
the unenclosed branches supported approximately 9500
leaves, the enclosed branch 520, i.e., the proportion of
leaves on the single branch to those on the whole tree
was 1:18. When the tree was exposed in a hot sun and

122 TRANSPIRATION AND ASCENT OF SAP ch.

brisk breeze, air temperature being about 24 in the
month of July, the amount transpired was 5 '171
grammes per hour, or 0*544 gramme per 1000 leaves.
The amount transpired by the desiccated branch, on the
other hand, was 0*781 gramme per hour, or 1*502 grammes
per 1000 leaves per hour. Hence, the leaves in the desic-
cated chamber transpired nearly three times as much as
the others under normal conditions of maximum trans-
piration. In this experiment the loss of water from the
enclosed branch does not represent any temporary desic-
cation of the surface tissues, for before the weighings
were made the branch had been enclosed in the flask for
a day. If this initial desiccation had been included, as
it was in Ewart's experiments, the difference between the
amounts given off by the enclosed and unenclosed leaves
would have appeared greater.

In diffuse light the difference is not so marked. The
unenclosed leaves of the tree used in the last experiment
in diffuse light with a temperature of 21 transpired
1*250 grammes per hour, or 0*130 gramme per 1000 leaves ;
the enclosed leaves simultaneously transpired 0*099
gramme, or 0*192 per 1000 leaves per hour. The desic-
cated leaves are nearly one and a half times as active in
transpiration as those under the normal conditions.

Transpiration controlled by supply. These
experiments show that it is not justifiable to assume
that the rate at which water is given off by an
isolated branch under conditions of abnormal desiccation
is attained by all the branches when all alike are exposed
to conditions most favourable to transpiration. The
excess evaporation from the desiccated leaves will be
greater when the bulk of the isolated branch is but a small
fraction of the bulk of the whole tree ; for the greater
the preponderance of the latter the larger will be the
supply available for the branch, and, consequently, the
less the resistance to transpiration. In Ewart's experiment,

vi TENSION REQUIRED TO RAISE THE SAP 123

then, where the branch had only 500 leaves, while the
whole tree had 9,000,000 (1 : 18,000), it is probable that
the effectiveness of the former was much greater than
that of the remaining leaves of the tree. Such an over-
estimate in the amount transpired involves, according to
the method, an exaggeration in the velocity of the current
in the trunk.

The control of transpiration exercised by the freedom
of supply may be easily observed by means of the weigh-
ing method. The amount transpired will be found to
fall off as the plant exhausts the water in the soil round
its roots, and to rise when the soil is again rendered moist.
The following numbers (Table 12) illustrate this fact
in the case of a small yew-tree which was exposed to
conditions favourable to transpiration on seven successive
days. The conditions were fairly uniform, as throughout
the experiment the sky was lightly over-cast and a light
east wind blew.

Table 12.

Transpiration per hour.

Date.

Temperature.

2-65 gr.

July 24

18 -5 C.

2-14

July 25

20-5

325

July 26

220 D

0-87

July 27

19-5

0-97

July 28

21-5

Watered

2 23'

July 28

24-0'

Watered

1-59

July 30

19-0

3 05

July 30

22-5 1

The dependence of transpiration on the supply is prettily
illustrated by Darwin's experiment with the horn hygro-
scope. Darwin records that when the hygroscope is
applied to the leaves of a branch severed from a plant,
it indicates a gradually diminishing rate of transpiration
as the store of water in the branch is gradually exhausted.

i2 4 TRANSPIRATION AND ASCENT OF SAP ch.

A diminished rate of transpiration is also indicated by the
hygroscope when the supply to a branch is reduced by
the application of a clamp constricting the water conduits
leading to its leaves.

The diminution in the rate of transpiration in these
experiments is, no doubt, due to a rise in the tension of
the water columns supplying the evaporating cells. The
tension rises in the first case as the limited water store
is drawn upon, and the water surface is dragged into the
cut surface of the branch. The surface-tension forces
developed there then oppose the transpiration current.
In the second experiment the tension is increased by the
rise of the resistance brought about by the reduction of
the calibre of the conduits. Further reduction of trans-
piration is occasioned by the closing of the stomata
and by other phenomena, which are themselves direct
or indirect consequences of the rise of tension in the
water.

According to this view an interesting observation of
Darwin's receives a ready explanation. There is a
momentary increase of transpiration in these experiments
immediately after the separation of the branch, and im-
mediately after the application of the clamp. In the
first case, we may assume that the tension in the water
supply is reduced by fracture, in the second case, by the
compression applied when screwing up the clamp. Trans-
piration obviously soon re-establishes tension, which be-
comes greater in both cases than it was originally.

In many ways, then, we see it is established that supply
largely controls transpiration, and in neglecting this factor
Ewart considerably exaggerated the maximum velocities of
the transpiration through the stems of trees.

Determination of the resistance. We come now
to consider the amount of resistance experienced by the
transpiration current in its passage through the water
conduits.

vi TENSION REOUIRED TO RAISE THE SAP 12c

Advantage of using small pressures. To obtain
the resistance, Ewart forced water through lengths
of branches under various pressures. The pressure
which gave the same velocity in the branch as that
estimated by his methods for the transpiration current
he took to be equivalent to the resistance experienced by
the transpiration current. Ewart does not mention how
he prepared the pieces, or how he cleansed the water for
the experiment. The introduction of air-bubbles or of
any clogging substance at the cut surface would materially
exaggerate the pressure needed to obtain the observed
velocity. Indeed, if the branch were transpiring actively
when the experimental pieces were removed, it would not
be sufficient to cut it under water ; for the liquid in the
branch being in tension, bubbles would be formed at the
surface of the cutting knife. These bubbles would require
some time to dissolve and disappear. The slimy materials
exuding from the injured cells also clog the branch and
raise the apparent resistance of its conduits. The com-
paratively high pressures with which Ewart worked would
render the clogging from this source and from any
impurity in the water more marked. The curves
reproduced in Fig. 19 illustrate this point. The
ordinates of the curves there shown indicate weights
of filtered tap-water transmitted through 3 cm. of
wood of Abies 'pectinata per second. The abscissae in-
dicate time in minutes. The curves show the diminution
in the rate of transmission for each pressure, the fall-off
in the amount being much more rapid for the higher pres-
sures. From these it is seen that it is desirable (especially
when using colour solutions, which, from the nature of
the case, cannot be distilled) to employ low pressures in
order to determine the resistance of the conduits apart
from the surface resistance. For this reason, then, when the
resistance of the tracheidal tubes was being determined it
seemed preferable to experiment with very low pressures,

126 TRANSPIRATION AND ASCENT OF SAP ch.

assuming that the flow for higher pressures would be pro-
portional to the head. This point was examined later.

The wood used in these experiments was that of Taxus
baccata. The uniformity and comparative homogeneity of
structure (resulting from the fact that it is composed
solely of tracheids, and is not penetrated by resin ducts)
recommend it as by far the most suitable wood for experi-
ment, when the qualities of the water-conducting tissues

6P 70 80

Minutes

Fig. 19.

are being examined. The tracheids composing the wood
of Taxus baccata are elongate spindle-shaped chambers
1 to 5 mm. long, and approximately square in cross-
section. The cavity of these chambers is comparatively
small. In cross-section the lumen forms about a quarter
the entire area of the tracheid. The only other consti-
tuents of the wood are the cells of the medullary rays,
which are radial tiers of cells 0'4 to 0'02 mm. in height.
They do not differ materially in percentage from one level

vi TENSION REOUIRED TO RAISE THE SAP 127

in the stem to another, and consequently do not introduce
differences of serious magnitude. The fact, however, that
they exude small amounts of slimy materials sometimes
causes difficulties. But inasmuch as there are approxi-
mately the same amount of medullary-ray-cells per unit
area in every cross-section, the clogging introduced in this
manner is uniform and, it is to be noticed, tends to increase
the apparent resistance. The presence in other woods of
large medullary rays, of wood parenchyma, of vessels of
variable size, or even of resin ducts, renders them unsuit-
able for experiment.

Preparation of material. Certain precautions are
necessary in preparing the wood for these transmission
experiments. It is best to cut away the transpiring
leaves from the selected branch, and then to cut off
a short length of it (say 15 to 25 cm.) under water.
Five to ten centimetres are now cut from each end,
and the remainder, after lying in water for at least 30
minutes, may be used directly or reduced to smaller lengths
for experiments These precautions are necessary, so that
bubbles shall not be generated in the conduits. It seems
possible that Ewart's high estimate of the resistance was,
in part, due to bubbles being generated in the conduits
while preparing the wood for his resistance experiments.
He certainly does not mention having taken these or
similar precautions in this connection, and his results are
three to four times as high as those given when these
precautions are taken.

Measure of the resistance. When determining
the velocity of flow with colour solutions, I usually
worked with a head of water equal to, or less than,
the length of the piece of wood experimented with.
When using unit head {i.e., the head equal to the
length of wood used), a fresh surface was cut with a
razor on one end of a piece of a branch prepared as just
described. The piece was supported vertically with the

128 TRANSPIRATION AND ASCENT OF SAP ch.

fresh surface uppermost, and a little vaseline smeared
round the bark of that end. A drop of filtered concen-
trated eosin solution was placed on the upper surface, and
as the drops formed at the lower end they were drawn
away by lightly touching them with bibulous paper. As
the eosin sank through the wood it was kept constantly
replenished drop by drop above. After a definite time
the experiment was stopped, and longitudinal cleavage of
the piece of wood showed the distance travelled by the eosin
solution during the duration of the observation at unit head.
The following are the results of some of these observations :

Table 13.

Length of

Duration of

Distance

Velocity

Piece.

Observation.

Travelled.

per Hour.

Centimetres.

Minutes.

Centimetres.

Centimetres.

7-2

150

1-7

6-8

6

100

1-2

7-2

4-0

15

1-7

6-8

3-5

150

1-8

7-2

10-0

20

2-2

66

3

25-5

30

6-9

In the last experiment the fluid transmitted was a solu-
tion of ferrocyanide of potassium. Its presence below was
detected by ferric chloride. The mean of these experiments
gives a velocity of 6*9 cm. per hour under a pressure equiva-
lent to a head of water equal in length to the experimental
piece of wood.

Experiments were also made at lower pressures. For
these the water-pressure was applied at the lower end of
a vertical piece of wood, through a rubber tube bent into
a U -shape. This tube was filled with eosin solution, and
the surface of the solution was raised to the desired height
above the upper surface of the wood.

The mean of these observations is 8 \5 cm. per hour,
calculated to unit head. If we exclude the two extreme

vi TENSION REQUIRED TO RAISE THE SAP 129

observations we get 7*6 cm. per hour as the velocity under
a head equal in length to the stem. The last observation
of the series was made on a piece of a narrow branch
about 05 cm. in diameter, the others on pieces about 1 cm.
in diameter. The thin distal portions of the wood in
almost every case offer a greater resistance to flow than
the thicker parts. The high estimates of resistance are

Table 14.

Length of

Head.

Distance

Duration of

Velocity per Hour

Piece.

Traversed.

Experiment.

per I 1 nit Head.

Centimetres.

Centimetres.

Centimetre.

Minutes.

Centimetres.

8

2

0-7

20

8-4

6

2

0-8

20

7-2

G

3

1-8

15

144

6

3

09

15

7-2

6

3

0-7

15

5-6

almost always obtained with the former. This difference
appeared almost constantly in my experiments. This fact
is probably of importance in determining the total resist-
ance in the intact plant. I have included the third obser-
vation in the table, although it diverges so markedly from
the mean, because I could see no error in the experiment,
and it is quite possible that a maximum result like this
is the nearest to the actual velocity in the uninjured tree.

The higher mean in the second series of experiments for
the velocity of transmission is probably due to the fact
that clogging substances are less likely to accumulate
owing to the actually slower flow and to the position of
the surface of application.

With care, good results may be obtained with higher
pressures, if the supply is from below. In the following
experiments (Table 15) the cylinder of wood was fixed
in the short arm of a vertical J -tube filled with a repeatedly
filtered solution of ferrocyanide of potassium. The moment
of penetration through the wood, which was 3 cm. long in

K

130 TRANSPIRATION AND ASCENT OF SAP en.

each case, was determined by its reaction with ferric
chloride applied in a piece of bibulous paper to the upper
surface of the cylinder.

The mean of the entire series gives 6*9 cm. per hour as
the velocity at unit head. As all the known errors, such
as the introduction of bubbles, clogging and injury of the
tracheidal tubes, tend to reduce the result, it is probable
that the velocity in the intact tree would be at least 7 to
8 cm. per hour under the same pressure. The occasional
high results obtained indicate a still higher figure as the
probable velocity.

Table 15.

Head.

Time

Velocity per Hour

Traversing 3 cm.

at Unit Head.

Centimetres.

Minutes.

Centimetres.

30

2-0

90

30

3

60

27

2-0

10O

27

2-5

8-0

24

3 75

60

24

3-5

J-4

_'4

3-5

64

21

4-5

5-7

IS

5 25

5-7

18

4-5

6-6

18

4-7-~>

6-3

15

4-5

8-0

15

4-5

8-0

12

5-5

8-2

12

7-0

6-3

12

5-5

8-2

1)

10-25

5-8

i \

h? .*7C

1- .!-

'.)

t 1 >

1 i

!)

9 5

6-3

<;

12-75

70

<;

13

6-9

3

25 5

7

In Fig. 20 I have plotted these results. The ordinates
represent the lengths traversed in one hour, while the
abscissae indicate the pressures, considering a head equal
to the length of the branch as unity.

vi TENSION REOUIRED TO RAISE THE SAP i?i

It may be noticed that the most divergent observations
are those made at the higher pressures.

Comparison of results. When we compare these
results with Ewart's, a very wide discrepancy is appar-
ent. The results of three of his experiments allow
themselves readily to be compared with my figures.

(1) He found that water travelled in a piece of yew
stem, 35 cm. long, at the rate of 11*7 cm. per hour under
a head of 3 metres. The head here is nearly 8 - 6 times the
length of the transmitting wood. Assuming the velocity

90

80

70

60

5 50

Si

2 40

6

30

20

10

Ind

icates t

wo coin

ciclent

bservat

ions

c

)

> 1

(

*

/

*
*

*
*

*
*

D

f.

s
*

*

c

)

i

c

)
i

<

*

*

1'

)

s
*

)

c

*
*

>'

5 6 7

Units Head

Fic 20.

10

proportional to the pressure, at unit head the water would
travel at 1*36 cm. per hour. This rate is stated to be
above the average.

(2) On pp. 51 and 52 of Ewart's paper it is stated that
the rate of flow in a piece of yew wood, 25 cm. long, under
a head of 3 metres is 26 cm. per hour. In this experiment
the head is equal to a column 12 times the length of the

K 2

132 TRANSPIRATION AND ASCENT OF SAP ch.

transmitting branch ; when reduced to velocity under unit
head the result is 2*17 cm. per hour.

(3) Again, on p. 55 an experiment is recorded which is
suitable for comparison. A velocity of 19 cm. per hour
was observed in a branch 25 cm. long under a head of
4 metres. This becomes 1*19 cm. per hour under unit head.

The mean of these three observations gives T57 cm.
per hour as the maximum velocity of flow in the yew wood
under unit head.

The results of my own numerous observations, on the
other hand, made under very various conditions of pressure
and by different methods, point to a velocity exceeding
7 cm. per hour with the same head.

The only explanation of this discrepancy which appears
possible is that in Ewart's experiments sufficient care was
not taken to prevent bubbles forming in the opened con-
duits, and to obviate clogging at the surface. This last
effect would be exaggerated in his experiments, as he
worked apparently in every case at such high pressures.
Reference to Fig. 20 illustrates this point. There it appears
that the erratic observations are those made at high
pressures, although at both high and low pressures similar
precautions were taken. It is in only the first of Ewart's
experiments quoted above that it is mentioned that the
experimental branch was cut under water. Other pre-
cautions are not mentioned.

Whatever is the cause of the discrepancy, it is certain
that if Ewart had obtained my results, the difficulty of
resistance, which he finds to be fatal to the cohesion-
theory of the ascent of sap, would not have presented
itself to him, for the velocity of 7 cm. per hour, which he
demands in the stem of the yew, would not require a
pressure equivalent to a head of 65 metres of water, as
he supposes, but only to 11 metres. The recognition of
the fact that unit head produces a velocity of about 7 cm.
per hour invalidates his whole calculation, intended to

vi TENSION REQUIRED TO RAISE THE SAP 133

show that to raise the sap in trees 150 metres high would
require a pressure approaching 100 atmospheres. The
facts of the case would be more correctly stated by saying
that if the amounts transpired by isolated branches,
under exceptionally favourable conditions for transpira-
tion, were transpired by the remaining branches of the
yew, and if we further assume that the velocity in the
trunk is maintained out into the finest branches of high
trees, then the resistance to flow in the conducting tracts
would be about equal to a head of water the same height
as the tree.

My results for the resistance, which were obtained as
described already by directly measuring the velocity of
flow under a given head, were fully confirmed by other
experiments in which the amount transmitted under a
given head was observed. An estimate of the cross-
section effective in transmitting the current then gives
the velocity.

A piece of a branch of Taxus baccata, 4 cm. long and
having a woody cylinder T35 cm. in diameter, was placed
with its long axis vertical. Water was supplied at
its upper surface just as quickly as it percolated through
the wood, so that the upper surface of the wood was con-
tinually wet, but the water was never appreciably piled
upon it. The amount of water transmitted in this manner
under unit head was 1 '356 grammes per hour. In order to
find the effective cross-section, after this observation was
made, a solution of eosin was supplied under similar con-
ditions. By this means the transmitting portions were
coloured, and the area of their cross-section easily estimated.
A mean of three such estimations gave the effective cross-
section as 0"7 sq. cm. From this it appears that under
unit head 1*93 grammes is transmitted per square centi-
metre per hour. Assuming with Ewart that the lumina
occupy about 0"25 of the cross-section, the velocity to
secure this rate must have been 75 cm. per hour. This

i 3 4 TRANSPIRATION AND ASCENT OF SAP ch.

agrees well with the results obtained with the other
method.

Method of eliminating surface clogging. In
order to determine the amount of flow without danger
from the error of clogging at the surface of supply,
it was measured in a lateral branch springing from
a stout stem. Water under pressure was supplied at the
two cut ends of the stem. With this arrangement, owing
to the relatively large surfaces of supply, the clogging
taking place at these does not encroach upon the amount
necessary for the supply of the small lateral branch for
several hours, and consequently the rate of transmission
in the lateral branch remains constant for this period.
To quote one of these experiments : A piece of yew stem
7*3 cm. long, having a straight lateral branch about its
middle, was selected. The diameter of the stem was
0*85 cm. and 105 cm. at each end respectively. Its wood
was also laid bare by an oval scar where another lateral
branch was removed. This scar was 1*3 cm. by IT cm.
The small lateral branch was cut to a length of 10*5 cm.,
and its wood had a diameter of 0*47 cm. Not more than
a quarter of its cross-section was in a state suitable for
transmitting water the greater part being occupied by
blackened duramen. This lateral branch was fixed water-
tight in a rubber bung in an orifice in the bottom of a
tank in such a manner that the cut end of the lateral
branch projected from the tank, while the supporting stem
was immersed in the water in the tank. With the head
of 30 cm. 0-300 gramme was transmitted per hour.
The transmitting cross-section was about 0'043 sq. cm.
One-quarter of this area would be lumen : therefore
the velocity of flow must have been 27*2 cm. per hour.
If we reduce this to unit head the velocity becomes
about equal to 9*4 cm. per hour, which again falls
within the limits of the results obtained by the previous
method.

vi TENSION REQUIRED TO RAISE THE SAP 135

1-30

1^20

1-10

1-00

90

80

70

60

Velocity proportional to the pressure. Through-
out this discussion it has been assumed that the
velocity is proportional to the pressure. This is usually
done. Fig. 21 shows that this is certainly approximately
true for pressures up to 16 units head. In this diagram are
plotted the results of many experiments, the object of which
was to determine if the
amount transmitted in
a given time is propor-
tional to the pressure, as
the former assumption
would involve. The
utmost care was taken
to obtain clean water to
supply to the wood and
to free the latter from
clogging material as
much as possible. The
water used was distilled,
the vessels and tubes of
supply were repeatedly
washed with distilled
water before using, and
the surface of the wood
freshly cut and rinsed
with distilled water im-
mediately before the ex-
periment began. The woody cylinder was 3 cm. long. It
was supported horizontally, and a tongue of bibulous paper
was applied to the outer surface of the wood and hung
down vertically into a little phial which received the drops
of transmitted water. The time required for five drops
to fall from the tongue was noted by means of a stop-
watch, and the weight of these five drops determined by
weighing the phial. The mean of three such observations
is given in Table 16 for each pressure.

50

40

30

20

10

1
/

/

/
/

*

/
t

1

/

t
/
l

/
/
/
/

c

/

/

r

c

t
t

V

/
/
/

5

/
/

f

/
/

/(

/
/
/

)

/

6 8 10

Units Head

Fig. 21.

12

14 16

136 TRANSPIRATION AND ASCENT OF SAP ch.

These results confirm the received view that the amount
transmitted varies directly with the pressure. The slight
bending over of the line joining the observations at the
high pressures is amply explained by the gradual clogging
which takes place despite all precautions ; for the experi-
ments giving these results were made with the same piece
of wood after those at lower pressures had been carried
out.

Table 16.

Head.

Milligrammes

Transmitted.

Duration of Ex-
periment.

Milligrammes per
Second.

Centimetres.

Seconds.

6

44

246

0182

12

50

137

0-365

18

58 -0

.)'.)

0-535

24

50

7::

0-685

:;n

52-0

63

0-825

36

58-0

60

966

48

59-5

47

1-266

The proportionality of head to flow may be indirectly
investigated in another manner. A straight branch fixed
in a horizontal position is connected to a large vessel of
distilled water. Precautions are taken to clean the con-
nections and the freshly cut inner end of the branch. A
tongue of bibulous paper is applied to the outer end to
draw off the transmitted water. The head once adjusted
is kept constant ; but after each determination the branch
is shortened. Fig. 22 records such a series of experiments.
The head throughout these was 100 cm. The initial length
was 25 cm. At that length the flow was 1*18 mg. per
second. Five centimetres were then cut off the outer end
and the flow rose to T66 mg. per second. A shortening to
15 cm. increased the flow to 2 33 mg. When the branch was
10 cm. and 5 cm. long, the flow was 3*79 mg. and 6'70 mg.
respectively. The curve plotted in Fig. 22 is a rectangular
hyperbola in which M = K./1 ; M being the number of milli-

vi TENSION REQUIRED TO RAISE THE SAP 137

grammes transmitted per second, 1 = length of wood, and
K = average value of the product Ml observed. The
observations, it may be seen, approximate fairly closely to

70
60

1
1

(

1

1

\

\

*5

= 50

"3

"

\

%
\

e.

CO

|>3-0
20

\

-

\^ C

>

**<

>^

** "** * ..

"" - ~

10

10 15

Cms. long

Fig. 22.

20

25

this curve. When the weights of transmitted water are
plotted against units head the curve shown in Fig. 23 is
obtained. Here the proportionality of flow to head, or,

70

GO

50

a> 4-0

30

20

10

'

f

T

4 fa 8 10 12 14 16 18 20

Units Head

Fig. 2: j .

rather, the inverse proportionality of flow to length, is
immediately apparent. Up to 10 units the curve is
almost a straight line. The bending over which occurs

138 TRANSPIRATION AND ASCENT OF SAP ch. vi

after that point is to be attributed to the clogging, which
is practically unavoidable when the flow is rapid.

Summary. It appears that water may be moved
through a stem in a horizontal position with the velocity of
the transpiration current if urged by a head equal to the
length of the stem. To raise water in a vertical stem at the
same velocity, evidently twice the head will be required.
Consequently when the force is applied as tension at the
upper end, the greatest stress the water need be subjected
to is double the weight of the moving column. Even in
the highest trees this is vanish ingly small compared to the
tensile strength of water.

Literature.

Darwin, F., " Observations on Stoinata/' Phil. Trans. Roy. Soe. London,
1898, vol. 190 B, p. 539.

Dixon, H. H., "On the Transpiration Current in Plants,'' Proc. Roy. Sue.
London, 1907, vol. 79 B, p. 41.

Ewart, A. J., "Ascent of Water in Trees," Phil. Trans. Roy. Soc. London,
1905, vol. 198 B, p. 41.

Id. " Resistance to Flow in Wood Vessels," A tin. of Botany, 1905, vol. 19,
p. 442.

Strasburger, E., " Ueber den Bau und Verrichtungen der Leitungsbahnen
in den Pflanzen " (Jena, 1891.)

CHAPTER VII

OSMOTIC PRESSURES OF LEAF-CELLS

We have seen that the force required to move the sap
at the rate of the transpiration current must at least be
equal to the pressure produced at the base of a column of
water which is twice the height of the transpiring tree.
Is it possible to obtain any measure of the force actually
available to produce such a stress ? It will be of interest,
if this measurement can be made, to know whether the
force is taxed to its limits to produce the upward motion
of -the sap or whether there is plenty of reserve.

A gauge for measuring this available force is provided
in. the leaf cells of the transpiring tree.

Forces available for raising the sap. During
transpiration the cells of the leaves are normally in a
turgid condition. This distension is caused by the osmotic
pressure of the dissolved substances acting upon the
protoplasmic membranes of the cells and pressing them
against the cell walls. We have seen that secretion or
evaporation abstracts water from these cells, and so tends
to concentrate the solutions within them. This loss of
water can only be made good by drawing in water from
the adjacent tracheae, and this pull acting on the upper
ends of the cohering columns of sap is propagated down-
wards through the tree. We may then regard secretion
or evaporation as the force which actually exerts the

139

i 4 o TRANSPIRATION AND ASCENT OF SAP ch.

tension on the sap, and this tension is transmitted through
the leaf cells to the sap in the conducting tracts.

Pressure and tension in leaf-cells. The simul-
taneous presence of pressure and tension within these
cells, at first sight, appears paradoxical ; but a moment's
consideration will show that it is quite possible for
the solvent, water, to be in a state of tension, i.e., at
a negative pressure, while the dissolved substances may
be at a positive pressure and be active as a distending
force in the cell.

Although, by thus distinguishing the pressure con-
ditions of the solvent and of the dissolved substances,
it is easy to conceive how the water in a turgid cell may
be in a state of tension, it appeared of interest to show
experimentally in the following way that this peculiar
state of affairs is possible.

It is well known that when a small piece is cut from
the young stem of an herbaceous plant, and immersed in
water, its curvature will show if its cells are distended by
osmotic pressure or not ; for the outer surface, being less
extensible, will become concave, if the cells of its tissues are
distended by osmotic pressure, and it will remain straight, or
become convex, in the absence of these pressures. If, then,
such a piece of tissue assumes and retains this concavity
when immersed in a tensile water column, we may be
assured that an osmotic pressure is exercised by the solute,
while at the same time the solvent is in a state of tension.

The experiment may be carried out as follows : A long
piece of glass-tubing bent into a J -form is carefully cleaned
by washing with caustic potash solution, followed by
methylated spirit. Its upper end is then sealed, and it
is nearly filled with water which has been boiled for
some time. A piece of tissue cut from the stem of some
suitable plant (I used the peduncle of Doronicum
austriacum), after soaking for several hours in well-
boiled water, is introduced into the J -tube, and passed

vii OSMOTIC PRESSURES OF LEAF-CELLS 141

up to the upper end, where there is a small bend made to
receive it. The J -tube is now set in a vertical position,
and its short limb is connected with an air-pump. By
the action of the pump the atmospheric pressure is removed
from the lower end of the column of water in the tube,
and the weight of the lower parts of this column, hanging
from the upper parts, puts them in tension. As the piece
of tissue occupies the top of the tube, the water in it and
around it is in a tensile state. It will be noticed that,
although exposed to this tension for a considerable time,
the tissue will retain its curvature, indicating, as we have
seen, an osmotic pressure in its cells. I have exposed a
piece of the peduncle of Doronicum austriacum to a tension
of 50 cm. of water for two hours, without being able to
detect any diminution of curvature.

In order to expose the water surrounding the piece of
tissue to a greater tension, the lower part of the water
column may be replaced by mercury. Working in this
way I have submitted the osmotic cells of the peduncle
of Doronicum to a tension of 75 cm. of mercury for one
hour. During this time the turgor of the cells remained
unaltered.

These experiments show the possibility of realising
experimentally the conditions we have assumed of pressure
and tension in the transpiring cells of the leaves.

Osmotic pressure in leaf-cells a gauge of
tension in tracheae. From the foregoing considera-
tions it is evident that so long as the force applied to the
upper ends of the sap in the tracheae of the leaves is less
than the osmotic pressure of the vacuoles of the leaf-
cells, these cells will remain distended and the leaf
will appear fresh and stiff ; whilst if the force drawing
off water from the cells is greater than that which they
can exert on the water in the tracheae, they will collapse
and the leaf will fade. Under normal conditions of trans-
piration this collapse does not take place. Hence the

i 4 2 TRANSPIRATION AND ASCENT OF SAP ch

force applied to the sap in the trachea? during normal
transpiration does not exceed the osmotic pressure in the
leaf-cells ; and consequently, if we can determine the
osmotic pressure in the leaf-cells we shall have a measure
of the maximum stress which is applied to the sap during
normal transpiration.

Until recently the most usual way of determining the
osmotic pressure in cells was the well-known plasmolytic
method.

There are several reasons why the application of this
method is not suitable to leaf-cells. In the first place,
it is necessary to cut sections of the leaf in order to apply
the solutions and to allow of microscopic observation.
The injury involved in sectioning acts as a violent stimulus
to the tissues, which may in itself evoke a change in the
concentration of the vacuoles or a contraction of the
protoplasm. Secondly, accurate determination of the
plasmolysing concentration is very difficult, as the con-
traction of the protoplasmic membrane must be consider-
able before it can be observed microscopically.

It was owing to these objections that the plasmolytic
method was abandoned -and other means for estimating
the osmotic pressures in the cells of leaves were sought.
The first method devised was the following :

Osmotic pressure balanced against gas-pres-
sure. A branch bearing a number of leaves is enclosed in a
strong glass cylinder, capable of resisting high gas-pressure
{e.g., 50 to 100 atmospheres), and the pressure is raised in
this vessel by means of an air compression-pump, or by
attaching it directly to a cylinder containing liquid carbon
dioxide. The lower portion of the branch projects from
the cylinder and dips into a glass vessel containing a weighed
quantity of water. These arrangements are shown in
Fig. 24.

It is evident that when the gas-pressure in the glass
vessel surrounding the branch is raised and maintained

v.i OSMOTIC PRESSURES OF LEAF-CELLS 143

above the osmotic pressure of the cells of the leaf,
water will be forced from these cells back into the conduits
of the branch and into the vessel beneath. This will
become apparent in two ways : first, by the flagging
of the leaf, inasmuch as the rigidity of the leaf is due to
the internal pressure of these cells, so that when this
pressure is overcome by the external gas-pressure the leaf
will flag ; secondly, by the increase of weight in the vessel
beneath containing the

the

water into which
branch dips. For every
branch, then, we may ex-
pect to find a pressure
above which water will be
forced back from the leaves
into the stem by reason of
the squeezing out of the
osmotic cells, and below
which water will rise
through the conduits to
the leaves, on account of
the osmotic attraction of
the cell-sap and evapo-
ration from the outside of
the cells.

To carry out these observations, the form of apparatus
I used consisted of a strong glass cylinder of specially
well-annealed glass, 50 cm. long, 10 cm. in diameter, and
with walls 1 cm. thick. Such a glass cylinder should,
according to calculation, be capable of resisting an internal
pressure of at least 100 atmospheres. The ends of this
glass cylinder were closed by means of two heavy gun-
metal castings, which projected over the side of the
cylinder so as to take three long bolts with nuts, which
drew the castings together on the cylinder. Leather
washers, soaked in bees' wax and turpentine, were inserted

Fk;. 24.

i 4 4 TRANSPIRATION AND ASCENT OF SAP ch.

between the ends, which were ground flat, and the cylinder,
to make the joints air-tight. The lower end was per-
forated centrally, and in the perforation was sealed
hermetically a narrow brass tube, about 0*5 cm. in diameter,
projecting into the cylinder. This tube included the stem
of the plant to be experimented with, the lower end of the
stem projecting out of the cylinder while the leaves were
enclosed. To make an air-tight connection between the
tube and the stem, a stout rubber tube was first bound
on to the upper end of the brass tube. The branch was
then inserted into the rubber tube, and, before it had
been completely pushed down, a portion of it just above
the rubber was coated with thick glue, so that when it
was shoved down into its final position with reference to
the tube, it carried this glue down into the rubber tube.
When it was in position, a copper wire was bound tightly
round the rubber, and drew it into close contact with the
glue. To complete the joint, a little glue was smeared over it.

The upper end of the cylinder was also perforated cen-
trally to admit the gas coming from the pump or bottle. This
was a simple screw-joint, made tight by a leather- washer.
To the upper end, and on the inside, were also attached
three hooks, from which were suspended a wire basket,
carrying drying materials, and a manometer. The latter
consisted of a simple, straight glass-tube, closed at one
end ; the other end dipped into a small vessel containing
mercury. This tube was marked off with J, J, J, i, etc.,
of its length from its closed end, and the position of the
mercury index gave the pressures directly in atmospheres.
When the upper end of the glass cylinder was in position,
the drying materials and manometer hung in the cylinder.
The connection between the glass cylinder and pump or
bottle of carbon dioxide was made by means of a flexible
lead tube with screw couplings.

Observations with compressed carbon dioxide.
When making these observations I was unable to procure, by

vii OSMOTIC PRESSURES OF LEAF CELLS 145

the pump at my disposal, air-pressures above 8 to 10 atmos-
pheres. Higher pressures were obtained by means of liquid
carbon dioxide. At the time there seemed a priori no reason
to believe that the presence of carbon dioxide would falsify
the results of experiments, which were not continued for
a long duration. However, subsequent experimental work
showed that the presence of this gas profoundly modified
the behaviour of the leaves when exposed to high pressures,
and consequently rendered the experiments made with
carbon dioxide of little value in estimating the actual
osmotic pressures obtaining in the leaves under normal
conditions.

In the first experiment, a short branch of Acer macro-
phyllum was sealed into the apparatus, and the pressure
raised by means of an air-pump, and maintained for
fifteen minutes between 8 and 10 atmospheres. During
this time gas was continually bubbling out from the lower
end of the branch, showing that the pressure had been
transmitted to the inner tissues. No loss of turgescence,
however, of the leaves could be observed.

In a second experiment, a similar branch was exposed
to a pressure of 8, or nearly 8, atmospheres, for fifteen
minutes, and during this time showed no loss of turgescence.

From these two preliminary experiments, it appears
that the pressure within the cells of the leaves of Acer
macrophyllum, which confers rigidity on the leaves, was
greater than 8 atmospheres. The osmotic attraction
which would give rise to this pressure would be capable of
drawing up a column of water 240 feet high.

In a similar experiment, a branch of Cratcegus oxy-
acantha was exposed to a pressure of about 8 atmospheres
for fifteen minutes without showing signs of loss of
turgidity.

As the pump I had at my disposal was unable to compress
air above a pressure of about 10 atmospheres, I discarded
it in favour of a bottle containing liquid carbon dioxide.

L

146 TRANSPIRATION AND ASCENT OF SAP ch.

This was connected with the high-pressure apparatus by
suitable couplings ; and, by gradually opening the valve
at the mouth of the bottle, the pressure could be adjusted
at will to any pressure up to 60 atmospheres. This has
the additional advantage that careful observations are
possible while raising the pressure, which cannot be done
while using the pump unless an assistant is employed.

By means of this arrangement, the pressure was raised
round the same branch as was used in the last experiment,
to 16 atmospheres, and was maintained at this for fifteen
minutes. But even at this pressure the leaves showed
no loss of turgescence. When the pressure reached 10
atmospheres, the bubbling of gas through the stem became
very marked.

As it appeared possible that a certain amount of collapse
of the osmotic cells of the leaves might take place without
making itself noticeable by the flagging of the leaves, a
number of experiments were made in which the branch
dipped into a vessel beneath, the latter being weighed
before and after the experiment. Any increase in weight
of this vessel would be due to the forcing backwards by
the external pressure of the cell-sap contained in the cells
of the leaves, which would in turn displace a certain
amount of water from the conduits of the branch into the
vessel. A decrease, on the other hand, of the weight of
the vessel would show that the external pressure had not
crushed the osmotic cells, and that they had, in spite of
its action, drawn up water from the vessel.

The first experiment of this kind was made on a branch
of Acer macrophyllum, which bore 14 well-grown leaves.
This branch was sealed into the high-pressure apparatus,
and kept at a pressure of 8 atmospheres ; during one
hour of intermittent sunlight this branch drew up 0"1 gr.
of water from the vessel below.

A similar branch, similarly arranged, and exposed to
a pressure between 8 and 9 atmospheres, drew up, in one

vii OSMOTIC PRESSURES OF LEAF CELLS 147

and a half hour's sunshine, 0*342 gr. of water from the
weighed vessel.

From these experiments it follows that the osmotic
cells of the leaves of Acer macrophyllum were able to remain
turgescent and draw up water against a pressure of 8 atmo-
spheres. Consequently, the osmotic solution in the cells
must be capable of generating a tension equivalent to
8 atmospheres pressure, by attracting water from the
conduits.

All the trees with which I have experimented do not, how-
ever, show that their leaves possess such high osmotic pres-
sures when surrounded with carbon dioxide. Thus the
specimens of Cytisus laburnum, investigated by means of
the high-pressure apparatus, showed that their cells began
to collapse under an external pressure of 6 atmospheres.
Above this pressure the leaves faded, and water
was forced back from them into the stem. It is, how-
ever, very probable that all the leaves were not put out
of activity in transpiration simultaneously. Thus, I have
observed, with Cytisus laburnum, that the old leaves
begin to show collapse by losing their glossy surface, and
rolling back from the edges at a pressure of 6 to 7 atmo-
spheres, while the young, small leaves, which are com-
posed of growing tissues, remain stiff and turgescent, even
at 16 atmospheres. 1

A preliminary experiment on Cytisus laburnum showed
that the leaves of this plant flagged markedly after an
exposure of five to ten minutes to a pressure of 16 atmo-
spheres. The flagging in this case is indicated by the
folding down of a leaf from the base of its petiole, and the
folding back of its leaflets, so that the whole leaf has the
appearance of the leaf of a sensitive plant {Mimosa pudica)

1 This phenomenon is probably correlated with the relative sizes of the
vacuoles in the old and the young cells ; for it will appear later that the
osmotic pressure of the sap of the young is less than that of the older
leaves.

L 2

148 TRANSPIRATION AND ASCENT OF SAP ch.

which has been stimulated. Besides these motions, the
surface of the leaf loses its gloss and becomes dried-looking,
the edges of the leaf roll up, and the expanded portion
becomes crumpled. The general appearance of the leaves
after twenty minutes' exposure to 16 atmospheres is that
of a leaf which has been exposed to a high temperature
and afterwards dried. Microscopic examination of the
cells of these leaves shows the protoplasm contracted from
the cell- wall just as it is in plasmolysed cells. This appear-
ance is probably brought about by the cell-wall being
pressed in on the protoplasm, and causing the latter to
force out its watery contents. When the pressure is
relieved, the cell-wall, by virtue of its elasticity, recovers
its form, while the protoplasm remains contracted within.
The space included by the cell-walls does not, however,
attain the dimensions it possessed when the cell was turge-
scent, as in that case it was distended by internal pressure
and consequently the leaf formed of such collapsed cells
is flaccid.

After obtaining this result, I set about to determine
the critical pressure for this plant, i.e., the pressure at
which the cells of the leaf would be forced to collapse,
and water would be driven back from them into the stem.

(1) In the first experiment, a small branch of this tree
carrying 9 leaves was fixed in the apparatus. The pres-
sure was maintained at 16 atmospheres. During one hour
of diffuse light, while the conditions within the apparatus
were kept favourable to transpiration, i.e., the space was
dried by calcium chloride, 0950 gr. was forced from the
leaves through the stem into the flask below. During
the first ten minutes of this experiment the leaves began
to flag, and soon showed all the appearances described
above.

(2) A branch of the same tree, carrying 12 leaves, some
old and some young, was submitted to a pressure of
8 atmospheres. After one hour of bright sunshine the

vii OSMOTIC PRESSURES OF LEAF CELLS 149

vessel into which the branch dipped was found to have
gained 0'400 gr. During this time the old leaves had
become flaccid, while the young leaves remained turgid.
Even the old leaves did not become markedly flaccid during
the first forty minutes of the experiment.

(3) A branch with 8 leaves was exposed to a pressure
of 6 atmospheres during one hour mostly of bright sun-
shine. During this time the leaves showed no signs of
becoming flaccid, but the surface lost some of its gloss.
On weighing, it was found that the vessel below had lost
0*007 gr. of water. This amount, however, comes within
the limits of error of the experiment, and, consequently,
we may assume that neither upward nor downward motion
of water occurs in these branches when the leaves are
exposed to a pressure of 6 atmospheres. In this experi-
ment, when the pressure was removed, the leaves recovered
their gloss.

(4) Against 4 atmospheres, the same branch, in inter-
mittent sunshine, transpired 0*622 gr. in one hour and
twenty minutes, while all the leaves remained quite turgid.

At the conclusion of the series on this branch the amount
it transpired at normal pressures still surrounded with
carbon dioxide gas was measured, and was found to be
1 *244 gr. in one hour and ten minutes. In air at normal
pressure the same branch transpired in one hour 0*966 gr.
During these last two experiments, the leaves were slightly
faded.

The decrease in the rate of transpiration with the in-
crease of pressure which is indicated by these results is,
doubtless, more marked than here appears, as it is well
known that the rate of transpiration of a branch falls off
rapidly from the time of cutting it. In Experiment 3,
at 6 atmospheres, which was the second to be made with
this branch, this decrease would have been small, but in
the succeeding experiments would have become more
exaggerated.

i 5 o TRANSPIRATION AND ASCENT OF SAP ch.

As it appeared quite possible that different examples
of the same species might have different osmotic pressures
in their leaves, these branches were all taken from the
same individual, and from a height of about six feet from
the ground.

In this series of experiments there are two sources of
error tending to make the critical pressure appear lower than
it is in reality : First, there is the mechanical crushing of
the conduits themselves owing to the external pressure.
When the osmotic cells experience the pressure, they may,
without themselves suffering any collapse, move in on the
conducting tissues, which, although they are specially
adapted to resist external pressure as well as internal
tension, are elastic to some extent, and consequently will
become somewhat contracted. This will expel a certain
quantity of water from them into the vessel beneath :
and, as the vessel was taken away immediately after the
pressure in the glass cylinder was lowered, the conducting
tissues may not have had time to reassume their former
volume. By this means a quantity of water would be
forced back into the vessel and remain there, and would
tend to counteract the loss due to transpiration. As the
greatest amount of water I have observed forced back
in this way from a branch, which was larger than the
branch used in these experiments, was about 01 gr., as
will be seen later, we may place the critical pressure of
the branch of Cytisus laburnum at 6 to 8 atmospheres.

Effect of carbon dioxide on the osmotic pres-
sure. The second source of error is more difficult to
allow for. The presence of the carbon dioxide surrounding
the leaves undoubtedly acts injuriously on the cells of
the leaf, so that a leaf which has been surrounded with
carbon dioxide for several hours sometimes shows a
darkened appearance, and collapses at a lower pressure
than one which has been put in fresh into the apparatus.
With this plant (Cytisus laburnum), however, the injurious

vii OSMOTIC PRESSURES OF LEAF CELLS 151

effects of carbon dioxide are not so marked nor so rapid
in their manifestation as in others.

Tables 17 and 18 embodying the experiments on Tilia
americana, which was found very sensitive to this gas,
illustrate how carbon dioxide affects the transpiration and
turgescence of the leaves.

Table 17.
Til in (tmr }(((( na in Carbon Dioxide.

Pres-
Experi- sures in
ment. Atmo-
spheres.

Conditions
of Light.

Duration
of Experi-
ment.

Amount of
Water

forced from
Branch .

2-284 gr.

Remarks.

A

15-16

Dull.

60 minutes.

After 15 minutes col-
lapse of leaves ap-
parent. Finally all
leaves were shriv-
elled.

B

10

Dull.
Dull.

40 minutes.

0-988 gr.

Slight collapse at end
of experiment.

C

7-10

60 minutes.

1 71 gr.

The leaves became
flaccid immedi-
ately.

D

7-8

Dull.

60 minutes.

1-452 gr.

Collapse slight.
Leaves rolled at
edges.

E

6

Sunshine.

60 minutes.

0-659 gr.

No loss of tur-
gescence.

F

4

Bright dif-
fuse light.

45 minutes.

0-287 gr.

After 30 minutes
some leaves
slightly crumpled.

G

4

Sunshine

and bright

light.

60 minutes.

0-182 gr.

Remained quite tur-
gescent.

H

3

Sunshine

and bright

light,

60 minutes.

- 0-506 gr.

Remained quite tur-
gescent.

In experiments A, B, D, E, G, and H fresh branches
with 8-11 leaves were used; in C and F the branches

1 52 TRANSPIRATION AND ASCENT OF SAP ch.

already used in A and E were observed. Consequently
they had already been exposed for some time to carbon
dioxide.

At the end of experiment D, when the pressure had
been reduced to normal for about 10 min., the margins
of the leaves unrolled and their usual appearance was
reassumed.

Table 18.
Tilia americana in Air.

Experi-
ment.

Pressure
in Atmo-
spheres.

6

Conditions
of Light.

Duration
of Experi-
ment.

Amount
of Water
Trans-
pired.

Remarks.

A

Dull light.

15 minutes.

+ 0-029 gr.

Fresh branch with
4 large leaves.

B

5

Diffuse
light.

60 minutes.

+ 0-070 gr.

Fresh branch with
9 leaves. No loss
of turgescence ap-
parent.

C 4

Diffuse
light.

00 minutes.

+ 0-111 gr.

No loss of turges-
cence.

Experiment C in Table 18 is subject to a correction
for the elasticity of the branches' conduits. In deter-
mining the amount of the water transpired, the vessel
beneath was placed in position before the pressure was
raised in the glass cylinder and removed for its second
weighing, while the pressure was still maintained. Con-
sequently, some water was squeezed back from the con-
duits, owing to their elastic yield, and remained in
the vessel, diminishing the amount of transpiration
observed. In order to estimate how much ought to
be allowed for this, an experiment was made in which
the same branch was raised to a pressure of 6 atmos-
pheres for ten minutes. While this was maintained, a

vii OSMOTIC PRESSURES OF LEAF CELLS 153

weighed vessel containing water was supplied to its
protruding end, and then the pressure was lowered
to that of the atmosphere. After ten minutes, the
vessel was re- weighed and was found to have lost 0T08 gr.
due to the elastic recovery of the conduits. When this
allowance is made in experiment C, Table 18, the amount
transpired becomes 0*219 gr. instead of 0T11 gr.

In order to determine whether this elastic contraction
of the conduits occurred chiefly in the conduits of the
stem or leaf, experiments were made in which a branch
was first exposed to a pressure of 6 atmospheres for
ten minutes, and while this was still maintained, a
weighed quantity of water was supplied to its lower end
which protruded from the high-pressure apparatus. The
pressure was then immediately lowered, and the branch
was left to draw up water from below for ten minutes by
means of its elasticity, the amount drawn up being
measured by a second weighing. When this amount is
compared with the amount drawn up in a similar experi-
ment with the same branch when all the blades of the
leaves are removed, it is found that the former is very
much greater than the latter quantity. Thus, with a
branch of Tilia americana bearing 11 leaves, the first
amount was 0T08 gr., while the latter was only 0*02 gr.,
a quantity which approaches the limits of error of the
experiment. From this we may conclude that the elastic
contraction takes place chiefly in the conduits of the
leaves.

Determinations with compressed air. At a sub-
sequent date, when it was possible for me to generate
higher air pressures, it was found that the leaves were able
to withstand considerably higher pressures when not
exposed to the harmful effects of carbon dioxide. Thus
the leaves of Helianthus multiflorus in air do not collapse
until a pressure of 20 atmospheres is applied to them,
while those of Cytisus laburnum and Tilia americana did

154 TRANSPIRATION AND ASCENT OF SAP ch. vii

not lose their lustrous appearance or roll at the edges
till pressures between 26 and 38 atmospheres were
applied. Hence, under normal conditions we may believe
that the leaves of these plants will not show signs of
fading till tensions equivalent to 20-30 atmospheres are
generated.

By means of this method useful results were obtained,
but danger attended the determinations. Despite the
strength of the glass cylinders used, two explosions occurred,
fortunately attended by delay in the work only, so that
after a comparatively small number of observations, a
more suitable method was looked for.

Literature.

Dixon, H. H., "On the Osmotic Pressures in the Cells of Leaves," Proc.
Roy. Irish Acad. vol. iv (Ser. 3), p. 61, and Notes from the Botanical School
Trinity College, Dublin, vol. i, p. 44.

Id. "On the Physics of the Transpiration Current," Notes from the
Botanical School, Trinity College, Dublin, vol. i, p. 57.

Id. " A Transpiration Model," Proc. Bog Dublin Soc. 1903, vol. x (N. S.),
p. 114, and Notes from the Botanical School, Trinity College, Dublin, vol. i,

p. 217.

Dixon, H. H.,and Joly, J., "On the Ascent of Sap," Phil. Trans. Roy.
Soc. London, 1895, vol. 186 B, p. 563.

CHAPTER VIII

THE THERMO-ELECTRIC METHOD OF CRYOSCOPY

Relation between osmotic pressure and freez-
ing-point. The method finally adopted for determining
the osmotic pressures in leaves is an indirect one.
As is well known, a relationship exists between the
freezing-point of a solution and the osmotic pressure
it can exert against a semi-permeable membrane. Hence,
if we can determine the freezing-point of the sap
in the vacuoles we shall have a measure of the
osmotic pressure. This cryoscopic method has been
applied to determine the osmotic pressures of various
fluids of the animal body by numerous investigators.
In these cases, comparatively large quantities of the
fluids are available, and consequently Beckmann's method
for determining freezing-points is suitable.

Beckmann's method of cryoscopy. This involves
the use of a thermometer with a large bulb which
must be immersed in the solution the freezing-point
of which is to be determined. The size of the bulb
necessitates the use of a considerable quantity of the
solution, viz., 12 to 15 c.c. as a minimum. 1 Such large
requirements seemed to preclude the application of the
method to the determination of the osmotic pressures of
the sap of transpiring organs, of which but small quantities

1 Beckmann thermometers with diminutive bulbs have been introduced
recently for dealing with small quantities of liquid.

155

156 TRANSPIRATION AND ASCENT OF SAP ch.

can be conveniently obtained. This objection applies all
the more strongly to the other more elaborate methods
of determining the freezing-points of solutions.

Use of thermocouples. In order to circumvent
this difficulty, it was decided to replace the mercurial
thermometer by a thermocouple, and to compare directly
the freezing-point of water with that of the solution.

It seems surprising that thermocouples have not been
in general use for determining the freezing-points of solu-
tions. In the first place, it is evidently possible to make
the thermo-electric method a differential one, viz., com-
parative of the freezing-point of the solution to be examined
with that of pure water under the same conditions, and so
it would seem most of the corrections necessary in the
thermometric method would be rendered needless. Thermo-
couples have great possibilities of sensitiveness, e.g., it is
by no means difficult to obtain by their use a deflection
of the light-spot on a galvanometer scale amounting to
1 mm. for a difference of 0*0001 . With this sensitive-
ness they can be made with a very small heat-capacity,
and will consequently take up the temperature of their
surroundings quickly. Their minute size and ease of
manipulation contrast very favourably with the bulki-
ness of the ordinary freezing-point thermometers. Absence
of parallax in reading the scale and the ease with which
couples having various ranges may be constructed will
also occur as advantages.

Notwithstanding these attractions, I have not been able
to find that thermocouples had been used previously in
cryoscopy, although in many researches their properties
would be of value. This is probably to be explained by
the erratic behaviour they exhibit when set up without
special precautions. When a sufficiently sensitive galvano-
meter is used to give a good deflection for small temperature-
differences, it is found also to be deflected by temperature-
differences acting on accidental junctions in the circuit.

viii THERMO-ELECTRIC CRYOSCOPY 157

Such junctions are usually formed at the binding-screws
between brass and copper or between two different samples
of copper. Strained places even in the copper leads may
also act as thermo- junctions. Another source of trouble is
strains in the galvanometer suspension, which lead to con-
tinual changes in the position of the zero on the scale.
The slowness of the galvanometer needle to take up its
final position may also be mentioned as introducing uncer-
tainty in deciding on the true magnitude of the deflection.

In view of these sources of error, it is evidently of great
importance to have as few connections in the circuit
as possible, and, where the latter are unavoidable, to
secure that they are balanced by similar connections kept
at the same temperature.

Construction of the thermocouples. In order
to eliminate one usual set of connections from the
circuit, i.e., that between the thermocouple and the leads,
it was arranged to utilise the ends of the copper leads
themselves as one pair of elements in the junctions.
These leads, which had a diameter of 0*1 7 mm., extended
right from the junctions to the reversing key (to be
described later). The other pair of elements of the
couple were formed of the ends of a continuous iron,
nickel, german silver, " constantin," or " eureka " wire.

For the work in hand, the eureka-copper junctions were
found most suitable. The eureka alloy has a high thermo-
electric value when forming a junction with copper, and
so is capable of giving a large deflection for a small tem-
perature difference. Its comparatively great resistance
enables one to adjust very conveniently the sensitiveness
by increasing or diminishing the length of the eureka in
the couple. Its low coefficient of variation of resistance
with temperature secures that this convenient resistance
introduces practically no error ; and when, as in the
apparatus to be described, it is enclosed in the freezing-
chamber, the error is so small that it may be disregarded.

i 5 8 TRANSPIRATION AND ASCENT OF SAP ch.

The specific resistance of eureka at is given as

47*4 microhms; its variation at 20 per 1 as - 0048 per cent.

The construction of the eureka-copper thermocouple is

simplicity itself. To each end of the silk-covered piece of

eureka wire, about 1 m. long, a con-
venient length of the copper lead is
soldered. The eureka wire I made
use of had a diameter of 01 9 mm.,
and a resistance of about 16 5 ohms
per metre. The soldered junctions
between the eureka and copper may
be'neatly made by stripping a few
millimetres of the ends of each from
their silk coverings and dipping the
bared tips into a solution of resin in
spirit. After this treatment, if the
ends in contact with one another
are immersed in a tiny drop of
molten solder, a very compact and
good junction is made.
Description of the apparatus.
To accommodate the couple to the
apparatus, the eureka wire before
soldering was wound on a cork
support (Fig. 25 s), leaving some
20 cm. of each end free. This cork
support forms a connecting-piece
between two drawn pine rods (p and
r, Fig. 25) which are destined to carry the junctions
and to keep them in position, one in a test-tube {a) con-
taining the fluid to be examined, and the other in a similar
tube (6) containing distilled water. 1

1 If a finer wire is used, the resistance may be disposed of by winding
it round the lower end of the rod p, so that it remains immersed in the
freezing distilled water. This eliminates any change in the resistance due
to temperature fluctuation.

Fi;

c,

viii THERMO-ELECTRIC CRYOSCOPY 159

The two test-tubes, each about 1 cm. in diameter, are
supported in a perforated cork bung (c), which fits
loosely in an outer large test-tube, which in turn is im-
mersed in the freezing-bath, and forms the freezing-
chamber (/). The perforated bung is held about the
middle of the large test-tube by a metal rod (m) a piece
of stout brass wire fixed into it and passing through
another bung which closes the mouth of the outer test
tube. The rod is prolonged above the second bung (d),
and forms a handle by which both bungs may be removed
simultaneously from the freezing-chamber carrying the
small test-tubes in the lower bung.

The cork connecting-piece (s) carrying the eureka wire
of the couple is furnished with a wide median vertical
perforation parallel to the two pine rods. When the pine
supports are placed in the small test-tubes, the metal rod
is passed through the perforation in the connecting-piece
and works loosely in it. Before fixing in the connecting -
piece the pine supports are thoroughly impregnated with
paraffin-wax by keeping them submerged for some time
in melted wax near its boiling-point. The junctions and
the wires coming from them are laid along the supporting
rods thus prepared and fixed in the connecting-piece, and
are bound to the rods, and the whole is waterproofed and
insulated with several coats of collodion varnish. The
supporting rods are continued above the connecting-piece
and are produced through corresponding perforations in
the upper bung (d), in which they fit loosely. It is
convenient that some kind of easily detached stop (q)
should be fixed on one of the rods above the upper bung
to prevent the rods slipping out of this bung when the
test-tubes are removed. The copper leads (I) emerge
from the freezing-chamber along one of the supporting
rods.

This arrangement, which will be easily understood by
reference to Fig. 25, allows the junctions in each of the

160 TRANSPIRATION AND ASCENT OF SAP ch.

smaller test-tubes to be moved simultaneously by raising
and lowering the upper ends of the pine supports, when
the upper bung is in position and the freezing-chamber
is closed. The double lead may be easily introduced, or
withdrawn, from the perforation in the upper bung by
means of a narrow slit opening into that perforation from
the side of the bung.

From these arrangements it will be seen that the method
has been rendered a comparative and differential one, and
consequently the corrections necessary for the thermo-
metric methods may be partly or wholly dispensed with
here. Both test-tubes gain the same amount of heat from
the stirring. With regard to the loss of heat to the
freezing-bath, the water will tend to lose heat more
rapidly, owing to its higher temperature. This difference
is rendered negligible by the way in which the water
freezes. A continuous layer of ice always separates out
against the wall of the test-tube, and forms a screen
between the bath and the water in which the junction
moves.

Again, the velocity with which the ice and the liquids
in the tubes come into equilibrium, depends on the
amount of ice present, its surface, its fineness of division,
and the energy of the stirring.

In the solution less ice will separate than in the water
for a given temperature of the freezing-bath ; but, at
the same time, it is, in practice, found to be more
finely divided. These two differences will act in opposite
directions.

The calibration-curve given in Fig. 26, which is sensibly
a straight line, shows that these errors practically neutralise
each other, and that in the working of the method the
galvanometer-deflection is proportional to the true depres-
sion of freezing-point of the solution examined.

Reversing key. Bearing in mind the desirability
of eliminating all needless junctions from the circuit,

VIII

THERMO-ELECTRIC CRYOSCOPY

161

one would like to connect the leads coming from
the junctions on the pine supports directly with
the terminals of the galvanometer. The importance
of reversing the current through the galvanometer owing
to shifts in the zero position of the mirror, and the
advantages of being able to disconnect the couple readily
from the galvanometer during various manipulations,
render a key of some form or other necessary in the circuit.

Of/,

ression of Freezing-po
50 100

nt

in hundredths
150

of degree

centigrade
200

250

*

4-0

s.

4
*

4io

.11

= 3-5

*
*

fc

*
*

CO

5 3

,

>

*

&

*

2 2-5

*

,g

*

|20

*

y

*
*

o

,'

*

I 1-5

s

' V

CO

S

3 i-o

s

*

t

5

/

s /

^0-5

*

^

50

100 150

Deflection in millimetres

Fk!. 26.

200

250

Such a break in the continuity of the leads involves a
pair of junctions. Experience shows that even when the
junctions are between two pieces of the same wire, thermo-
electric effects are produced if they are not at the same
temperature. In order to keep the two junctions as closely
as possible at the same temperature, the following arrange-
ment was adopted :

The leads coming from the couple are disposed so that

M

1 62 TRANSPIRATION AND ASCENT OF SAP ch.

their naked ends are exposed on opposite sides of a flat
vertical support. To effect this they were passed
several times through a piece of thin cardboard in
such a manner that when the card was bent and folded
across the support the stitches made of the two wires lay
on opposite sides.

Fig. 27 shows the arrangement. T is an H -shaped piece
of tinned iron about 5 5 cm. long. The cross-piece of

the H is represented by
a broad band about 3 cm.
wide. It is covered by
a thin piece of cardboard
C about 1*2 cm. by 5 cm.
This card carries three
stitches of the ends of
the leads on each side
of its middle line. The
ends of the card are
folded round the cross-
piece of the H, and the
iron is folded in the
middle along the dotted
line (Fig. 27A), so that
the ends of the card are
nipped within the fold.
Then the four ends of
the H are bent out at
right-angles to the folded middle-piece, so as to form a
stand to support this in a vertical position (Fig. 27B).
To prevent the ends of the leads making contact with the
iron, two plates of mica (M, Fig. 27B) are slipped between
the leads and the iron one on each side of the vertical
portion. The mica plates are held in position by the
cardboard.

Connection between the ends of the leads exposed on
this support and those coming from the galvanometer

,M

T\

111

i

V/

B

Fig. i

-'7

.

vim THERMO-ELECTRIC CRYOSCOPY 163

was made in the following way : The bare ends of the
galvanometer leads were fixed on the inner surfaces of
the jaws of a spring wooden clip. When the clip was
closed upon the vertical support of the thermocouple
leads, connection was established between the two pairs
of leads, and the circuit was complete. By releasing the
clip and rotating it round a vertical axis through 180,
and clamping it again on the support, the current from
the couple may be reversed through the galvanometer.

In this form of reversing key, the junctions being of
the same metal and if desired made of the same piece
of metal, thermo-electric effects set up by temperature-
difference at the junctions are reduced to a minimum.
Notwithstanding this, it was found that these differences
of temperature were a source of error. To maintain the
junctions on the opposite sides of the support at the same
temperature and so eliminate the error, these connections
were made underneath liquid petroleum, contained in a
beaker, on the bottom of which rested the support of the
thermocouple leads. The petroleum was kept stirred
during observations.

Arrangements for galvanometer. It is found
convenient to have the galvanometer leads a con-
siderable length, so as to allow a suitable distribution
of the parts of the apparatus ; consequently, it is
essential that they should have a sufficiently large cross-
section, so as to offer but a small resistance ; otherwise
changes in temperature, from which it is impossible to
shield them, will alter the sensitiveness of the apparatus.
With the key described, there is no objection to having the
galvanometer leads of different copper wire and heavier
than those coming from the junctions.

Some special precaution is also needed to secure that
the junctions at the binding-screws and those in the
galvanometer are at the same temperature. In the case
of these connections it is all the more necessary, because

m 2

164 TRANSPIRATION AND ASCENT OF SAP ch.

elements of the junctions are of different materials viz.,
brass and copper. It was found that the different tem-
peratures of the opposite sides of the galvanometer in an
ordinary laboratory could cause quite an appreciable
deflection. To remove this, the galvanometer was placed
in a thermostat, arranged to maintain a temperature of
about 21 (Fig. 28, T). For this purpose one of Hearson's
incubators was used. A hole was cut in the wooden door.
Through this a beam of light illuminated the galvanometer-

Fi.i. 28

mirror, and was reflected back to the scale (S). The inner
glass door was found not to injure the sharpness of the image
of the cross- wire sufficiently to be objectionable. It was
necessary to stand the galvanometer (Fig. 28, G) on a stout
glass plate on the copper floor of the thermostat, which
otherwise slowly sagged under the pressure of its feet. The
thermostat during observations, extending over a year, was
maintained at temperatures which varied very slowly
between 20*2 and 21 5, so that at any moment the parts
of the galvanometer must have been very closely at the
same temperature.

This constancy of temperature was probably also ad-

viii THERMO-ELECTRIC CRYOSCOPY 165

vantageous in maintaining the resistance of the galvano-
meter itself constant.

The galvanometer employed was one of the Ayrton-
Mather pattern, manufactured by the Cambridge Scientific
Instrument Company. Its resistance was 20*7 ohms. The
deflection of the spot of light for one micro-volt, when
the screen was 1 metre distant from the mirror, was 10 mm.,
and for one micro-ampere 206 mm. A translucent screen
was used to receive the spot of light from the galvano-
meter-mirror, which was illuminated with a Nernst-lamp
(Fig. 28, L).

Where one observer is using the apparatus, it is
convenient to have the galvanometer leads so long that
the petroleum key may be placed close to the screen, while
the freezing-bath and thermocouple, etc. (F), may stand
at a level 50 cm. below the screen and somewhat nearer the
observer. This disposition brings the key (K), the supports
of the thermocouple, and the stirrer of the freezing-bath
close to the observer, and he is in a convenient position
for reading the position of the image of the cross-wire.

The apparatus should be set up, and the thermostat
and galvanometer adjusted, on the day before an observa-
tion is made. Once set up, no readjustment should be
necessary.

The freezing-bath is contained in the large cylindrical
glass vessel, shown in Fig. 25, H, with thick walls. To
prepare the bath the vessel is about a quarter rilled with
salt solution, and then finely divided ice is added till the
vessel is filled up to within about 3 cm. of the brim. A
stout brass wire stirrer of the usual form is used to mix
the brine and ice. Salt is added till the desired tempera-
ture is attained. This should be about T5 below the
freezing-point of the solution to be examined. If the
proportion of ice to the liquid is large, this temperature
may be maintained constant by occasionally adding a
little salt. A brass lid is fitted to the freezing-vessel, and

1 66 TRANSPIRATION AND ASCENT OF SAP ch.

supports the large test-tube which forms the freezing-
chamber. It is also perforated to admit a thermometer
into the freezing-bath and to allow the stirrer in the bath
to project from it.

Details of procedure. To make an observation
with the apparatus say, to determine the freezing-
point of a solution the procedure is as follows :
the leads of the thermocouple are slipped through
the slit and the pine supports through the holes in
the upper cork bung, and the stop is fixed on one of
the supports above the cork, to prevent them falling
down. The freezing- chamber is then closed with the cork.
Meanwhile, two small test-tubes, one containing about
2 c.c. of the solution, and the other the same quantity of
distilled water, are being cooled by supporting them in
the freezing-bath, making use of the perforation in the lid
through which the stirrer works.

When it is judged that they have reached their freezing-
point, a little hoar-frost is detached on a cooled platinum
needle from the outside of the freezing-chamber and intro-
duced into the distilled water. Ice crystals are immedi-
ately formed, and some adhere to the needle, which is
then transferred to the salt solution. Crystallisation is
instantaneously started in this, and the needle is with-
drawn. The two test-tubes are now put into the holes
of the smaller cork, and this is fixed on to the lower end of
the wire handle which passes down through the upper
cork, which has been removed from the freezing-
chamber momentarily for the purpose. The junctions on
the lower ends of the pine supports are now immersed in
the freezing liquids in the test-tubes. Thus arranged, the
whole, test-tubes and thermocouple, is put into the
freezing-chamber and the upper cork tightly adjusted.

Stirring of the contents of the test-tubes is immediately
commenced by moving the pine rods up and down. As
these are rigidly connected together, the two test-tubes

vni THERMO-ELECTRIC CRYOSCOPY 167

are subjected to precisely similar conditions in this respect.
The freezing-bath is also kept stirred. The galvanometer
may now be put in circuit with the thermocouple by
fixing the clip on the support in the petroleum key ; and
the petroleum is occasionally stirred. Immediately on
making the contact the spot of light travels from zero.
At first its motion is rapid, but becomes slower and slower
till at last it moves with an almost imperceptible creep.
It comes to rest about 60 sees, after contact is made. It
will be found convenient to allow 75 sees, to elapse before
making a reading. During this time the stirring of the
test-tubes is actively kept up ; for it is surprising how.
quickly the ice rising in the solution allows the lower
layers round the junction to become supercooled. In the
other test-tube the same does not occur, as the ice soon
forms a lining lying against the wall of the test-tube, and
the junction is supported in water surrounded by ice.

When the first reading is made, the clip is disconnected
and the galvanometer mirror swings free. Reversed con-
nection is made when the spot of light is at the limit of
its swing on the side on which the first deflection was
recorded. In this way the suspension of the galvanometer
is kept from any sudden strain which might be produced
by suddenly checking its movement. After 75 sees.,
during which the same active stirring is kept up, a second
reading is made. This first observation after putting in
the solution should be regarded merely as a preliminary
one ; but still, if too much ice has not been present, it
will give the freezing-point within a couple of hundredths
of a degree.

The test-tubes are now raised from the freezing-chamber,
and the one containing the solution momentarily touched
by the finger to give it a little heat. When the upper
cork is readjusted and stirring recommenced, it will be
noticed that the spot of light retires towards zero. If all
the ice is not melted, it will quickly recover its former

1 68 TRANSPIRATION AND ASCENT OF SAP ch.

position ; and the test-tube should again be touched.
When it is certain that almost all the ice is melted in the
solution, it is left in the freezing-chamber and allowed
to cool. Meanwhile connection is broken by removing the
clip from the support in the petroleum key.

When it is judged that radiation has cooled the solution
nearly to its freezing-point, connection is again made by
the clip, and stirring is recommenced. The spot of light
then travels to near its previous resting-place, or possibly be-
yond it. Supercooling may proceed, and the spot of light
will slowly travel indefinitely beyond its previous position,
or crystallisation may supervene, and the spot will return
somewhat on its path and tend to take up a steady position.
In the latter case connection is broken at the clip, and
the mirror allowed to swing free. Connection is again
made, and, after 75 sees., during which vigorous stirring
is kept up, a reading is made. The current is then reversed,
and at the end of 75 sees., another reading is made on
the other side of the zero point. If, however, super-
cooling proceeds, and crystallisation does not automatic-
ally occur, it is necessary to inoculate the solution with
a little hoar-frost. The inoculation should be carried out
when the spot of light has definitely passed the limit of
the first deflection. If it is allowed to cool too far,
much ice will separate, and the concentration of the
solution left over will cause too large a depression ; if,
on the other hand, it is inoculated just at its freezing-
point, so little ice separates that the solution in parts
may continue for some time supercooled, and we may
get too great a deflection. Experience shows that the
smallest depression is obtained if the solution is allowed
to cool 0T to 0*2 below its freezing-point before in-
oculation.

It will often be found that the mean of the second pair
of readings indicates a larger deflection than that of
the first pair by about 1 per cent. This seems to be

viii THERMO-ELECTRIC CRYOSCOPY 169

due to the slow cooling of the support of the junction
in the solution. It will be found that readings after the
junctions have been in the freezing-chamber about 15 min.
do not tend to be greater than the preceding ones. In
the natural routine the second pair of readings are made
about 15 mins. after the junctions have been put in posi-
tion. A third pair of readings made in a similar manner
will plainly show whether the apparatus has reached a
steady state. If the observations have been satisfactory,
they should not diverge from one another by more than
one-half per cent., and with care greater accuracy may
be obtained.

To calibrate the apparatus, sodium-chloride solutions of
known strength are introduced into one of the test-tubes.
The deflection produced by the depression of the freezing-
point of each is observed. These depressions being known
by the work of Raoult, Loomis, Nernst, and Abegg, we
obtain the value of a millimetre deflection of the light-
spot in degrees.

Table 19 (p. 170) exhibits the figures of one of these
calibrations. The individual readings are recorded to
give some idea of the accuracy of the arrange-
ment.

The scale reads continuously from left to right : 250 mm.
marks its middle point.

In this table are recorded the two positions of the spider-
line in the spot of light on the scale for three successive
observations of the freezing-point of each solution. The
deflection corresponding to this freezing-point is obtained
by subtracting the second from the first, and halving the
result. It will be seen that deflections obtained in this
way diverge only slightly from the mean, which is given
in the last column but one. The greatest divergence is not
one-half per cent. In the final column are given the actual
freezing-points of the solutions derived from Raoult's
results quoted from Hamburger.

iyo

TRANSPIRATION AND ASCENT OF SAP ch.

Table 19.

Calibration of Thermocouple.

No. of
Sol.

8"

f-l

o

O

-' -

go*
3^ &

a

o

CD
ft

I.

II.

III.

IV.

V.

VI.

gill.

3*500

3-000
2 500
2-000
1-500
1-000

Observation I.

-p.2

o

fa

479-0

. fa_

23 1

6

CD

s:

2510

De-
flection.

227-9

451

57 3

2541

196-8

417

92-5

254-7

162-2

382-8

124-0

253-4

1294

349-0

154

2515

97-5

317-0

187-4

252-2

64-8

Observation II.

J,

A

.

o

T3 2

o

m 2

/ -

pi .-a

"M a:

Zei

fi'tS

O

CD

fa

fa

q=l

480
449-3
417
383-0
349-0

22-8

58-3

92-8

124

153-3

317-0 187-3 252-1 64

251 4

253-8
254-9
253-5
252-1

228
J 05
162
129
97

No. of
Sol.

I.

II.
III.
IV.

V.
VI.

u .
S

- j +a

So ?

PjrH ?

ft

Observation III.

gm.

500

ooo

500
000
500
000

s.

O

fa

480-5
449 -0
418-0
383-6
340-0
317-0

Ol ' 'xi

o

fa

56

93

124

154

187

(I
5

3

CD

n:

CD -*

CD

qa

CD w
CD

o 'g

fa o

iffa

^

CD ?

251
252
255
253
251
252

22S
196
162
129
H7
65

228-4
196 2
162-2
129 6
97 6
64-9

2060=
1-768 3
1-474=
1181
0-886=
0-596 3

It will be noticed from the numbers recorded in the
table that the position of the zero shifted considerably
during the observations. In the first series, i.e., those on
Solution I, the zero point lay about 251 on the scale,
while during the first observation on Solution II it was
near 254, These shifts of zero show the importance of
being able to reverse the current, and of obtaining the
deflection by two readings, one on each side of the zero
position.

A graph of these observations is given in Fig. 26 (p. 161).
The ordinates correspond to the concentrations of the solu-
tions, and the abscissa? to the measures of the deflections
in mm., caused by the difference in temperature of the
freezing-points of the solutions and that of water. The

viii THERMO-ELECTRIC CRYOSCOPY 171

dotted line is a similar graph of Raoult's freezing-point
determinations. The concentrations are plotted against
the depressions of the freezing-point. In the second graph
the abscissae correspond to hundredths of a degree.

The couple on which these observations were made had
a length of 126 cm. interposed between the junctions, as
it was desired that it should give about 1 mm. deflection
for a temperature difference of 01. The actual
deflection was found to be 110*9 mm. for TOO .

For some time after being made, the thermocouple
used in these observations changed its constant consider-
ably, owing probably to some progressive change in the
metals of the junction and circuit. After nine months,
when the constant was re-determined, it gave a deflection
of 130*4 mm. per 1. It had then become nearly stable,
and observations during the next three months showed that
its deflection varied between 129 *2 mm. and 133 mm.
per degree. The smaller fluctuations are possibly due to
changes in the resistance of the circuit connected with
changes of temperature. They show, the need of re-
determining the constant of the couple during each series
of observations, just as the zero change of the Beckmann
thermometer necessitates a control-experiment in the
thermometric method.

With regard to the temperature of the freezing-bath,
it would at first sight appear of little importance in this
differential method, as, no matter what its temperature is,
it might be thought that it affects each test-tube similarly.
It has, however, been found to have an appreciable effect
on the magnitude of the deflection corresponding to the
freezing-point of a given solution, as will be seen from the
table below, in which are recorded the deflections corres-
ponding to the freezing-point of a solution of T5 gm.
sodium chloride in 100 gm. of water, having a freezing-
point of 0*886, when surrounded with a freezing-bath of
different temperatures.

172 TRANSPIRATION AND ASCENT OF SAP ch.

Temperature of Bath.

Deflection

-10

113 6 mm.

-1-5

115-2

-2

1169

-2-5

117-7

-3

118-1

-4-0

118-8 ,,

From these figures it appeared that when the freezing-
bath is less than T2 below the freezing-point of the
solution under examination, a small alteration in the
temperature produces a greater effect on the deflection
than when the bath is about T5 below the solution. It
consequently seemed best to adjust the freezing-bath
about 1 *5 below the suspected temperature of the freezing-
point.

As has already been pointed out, the influence of the
temperature of the freezing-bath on the apparent freezing-
point deflection is due to the difference in the behaviour
of water and salt solutions on freezing. In the latter
the crystals remain separate and the ice is finely divided.
The difference of density between it and the solution
causes it to rise up somewhat more rapidly, tends to aggre-
gate it at the upper surface, and so permits the lower
layers of the solution to supercool to a small extent. In
the distilled water, on the other hand, the ice adheres to
the walls of the tube, and forms a lining to it, so that
supercooling of the lower layers is less favoured.

The convergence temperature and the velocity of heat
exchange between the freezing-bath and the contents of
the tubes are dependent largely on the heat-capacity of
the solution, and consequently it is of importance that
the two test-tubes should contain approximately the
same amount of liquid ; otherwise the rate of exchange
of heat between the solution and the freezing-bath and
the water and the freezing-bath will be different. Thus
the deflection due to the depression of freezing-point
of a solution containing 15 gm. sodium chloride in
100 gm. of water was found to be 115 6 mm., when the

vin THERMO-ELECTRIC CRYOSCOPY 173

solution and the water stood at a level of 3*3 cm. in
similar tubes ; it was reduced to 115*4 mm., when the
depth of the salt solution was increased to 5 cm. Of
course it is easy to arrange that both tubes should contain
the same amount, and so have practically the same heat-
capacity.

Change of resistance of the circuit due to temperature
changes is guarded against by completely immersing the
eureka or nickel of the couple in the freezing-chamber,
while the resistance of the galvanometer is kept constant
by its being enclosed in the thermostat. The complete
immersion of the connecting-piece of the couple in the
freezing-chamber also secures the elimination of thermo-
electric effects due to want of uniformity in this wire.

From what has been said, it will appear that the thermo-
electric method is capable of considerable accuracy, even
when only two junctions are employed. Of course if it
were desired to work to greater accuracy, there would be
no reason why the number of junctions should not be
increased, thus greatly increasing the galvanometer deflec-
tion for the same temperature interval. In the work in
which we were engaged, however, this would have been
undesirable, as a comparatively large range was required.

But even with a pair of junctions, the hundredth of a
degree could be measured with certainty. With this
accuracy very small quantities of fluid may be dealt with.
The small quantities required render the method particu-
larly suitable to physiological work. Its differential charac-
ter might also be applied with advantage to comparing the
freezing-points of different fluids ; for example, in a com-
parison of jugular and carotid blood.

Literature.

Dixon, II. H., ' Observations on the Temperature of the Subterranean
Organs of Plants," Trans. Boy. Irish. Acad. 1903, vol. 32 B, p. 145.

Id. "A Thermo-electric Method of Cryoscopy,'' Proc. Boy. Dublin Soc.

i 7 4 TRANSPIRATION AND ASCENT OF SAP ch.viii

1911, vol. xiii (N. S.), p. 49, and Notes from the Botanical School of Trinity
College, Dublin, vol. ii, p. 121.

Dixon, H. H., and Atkins, W. R. G., "On Osmotic Pressures in Plants ;
and on a Thermo-electric Method of determining Freezing Points," Proc.
Boy. Dublin Sac. 1910, vol. xii (N. S.), p. 275, and Notes from the Botanical
School of Trinity College, Dublin, vol. ii, p. 47.

Hamburger, H. J., " Osmotischer Druck und Ionenlehre (Wiesbaden,
1902).

CHAPTER IX

METHODS OF EXTRACTINO SAP FOR CRYOSCOPIC

OBSERVATIONS

In the earlier experiments with the thermo-electric
method, the sap employed was obtained by simply crushing
the tissues in linen till they yielded the necessary amount
of liquid. In a few cases at the beginning of the investi-
gation, when a difficulty of obtaining sufficient sap was
anticipated, a modification of this method was used. A
weighed quantity of leaves was broken up to a nearly
uniform pulp in a mortar and a measured quantity of
water was added. When thorough mixing was effected,
the diluted sap was squeezed from the sludge and its
freezing-point determined. The correction for dilution was
obtained by a separate determination of the percentage of
water in the sample of leaves used.

Method of simple pressure. The method of
extraction by simple pressure had always been adopted
hitherto in cryoscopic determinations both for animal
and vegetable juices.

The sap so obtained has been regarded as a fairly average
sample of the sap of the organ pressed. This seemed a
reasonable view to take, inasmuch as the pressures applied
so completely crushed the cells of the tissues that the sap
expressed contained large quantities of protoplasmic frag-
ments, which in the case of green organs were particularly

17:.

176 TRANSPIRATION AND ASCENT OF SAP

CH.

noticeable, owing to the presence of chlorophyll cor-
puscles embedded in them. It seemed allowable to assume
that, where the component cells are so completely disin-
tegrated as is indicated by this observation, all the sap
of their vacuoles must be shed into the expressed fluid ;
or at least there would be no reason to suspect a difference
in composition between the latter and the sap which
remained behind in the organ.

Fairly early in this research, however, observations were
made which, in the light of subsequent work, might have
borne a different interpretation. For example, when leaves
were exposed to the vapour of chloroform, it was found
that the sap was pressed out with much greater ease, and
its freezing-point was very much lower, than that of the
sap coming from the untreated leaves.

This may be illustrated by the experiments made on
the sap of leaves of Hedera helix, shown in Table 20.

Table 20.
Sap ikom Chloroformed and Fkesh Leaves of Hedera helix.

No. of
Expt.

Description of Sap.

227
229

232

233

Pressed from untreated leaves on gathering .
Same sap as in 227 to which a few drops of

chloroform had been added, kept 24 hours . .
From leaves similar to those used in 227 after

they had been 24 hours in the dark

From leaves similar to those used in 227 after

they had been 24 hours in the dark and in

chloroform vapour

A comparison of experiment 227 with 229 shows the
increase of the depression of freezing-point we may expect
from the saturation of the sap with chloroform. Experi-
ment 232 is added by way of comparison to indicate the
change in freezing-point which is experienced by the sap
of untreated leaves when kept for twenty-four hours in
the dark. The depression of the sap pressed from the

IX

METHODS OF EXTRACTING SAP

177

chloroformed leaves is evidently much greater than can
be assigned to the action of the chloroform on the sap, or
to the spontaneous changes in the cells of the leaves,
which appear in Experiments 229 and 232 respectively.

Another result which could be interpreted in the same
sense was furnished by two experiments on the sap of
leaves of Ilex aqui folium. In these it was found that, if
the leaves were killed by heat in a saturated atmosphere,
they yielded a sap having a much greater depression of
freezing-point than that pressed from similar leaves which
had not been heated.

Table 21.
Sap from Fresh and Heated Leaves of Ilex aquifolium.

No. of
Expt,

430
431

Description of Sap.

Sap pressed from fresh leaves

Sap from leaves heated to 97 C. for 30 minutes

Again, it was found that, if a weighed quantity of leaves
be desiccated, reduced to powder, and again made up to
the original weight with water, the sap pressed from the
mass will have a much greater depression than that pressed
from the fresh leaves without passing thiough this treat-
ment. This point is borne out by the following experi-
ments :

Table 22.
Sap from Fresh and Desiccated Leaves of Hedera helix.

No. of
Expt.

434
435
436
437

Description of Sap.

A.

0728
1-031
0-869
1-177

N

178 TRANSPIRATION AND ASCENT OF SAP ch.

In Experiments 436 and 437 the specific electrical con-
ductivities of the saps at C. were also determined, and
were found to be respectively 0*00485 and 0*00623. This
shows that the quantity of electrolytes in the sap pressed
from the desiccated leaves has increased approximately
proportionally with the other dissolved substances.

These observations were made primarily with other
objects in view. But even then the possibility that the
sap pressed from the untreated leaves was not so concen-
trated as that remaining behind in them presented itself.
However, it seemed more probable that the greater con-
centration of the sap derived from the chloroformed,
heated, and desiccated leaves was attributable to changes
due to the treatment in each case, and the investigation
of the disci epancy was deferred to a later date.

Progressive concentration of pressed sap.
Finally a short paper by Marie and (latin again sug-
gested the necessity of investigating this point. These
writers when investigating the cryoscopic value of the sap
of Alpine plants note that the sap expressed first from a
plant-organ has a smaller depression of freezing-point than
that pressed subsequently. They contented themselves
however, with adding the successive samples together, and
take the freezing-point of the mixture as the freezing-point
of the sap of the plant.

This progressive concentration of the sap pressed from
plant-organs had been, it was found, very convincingly
established some years previously by Andre, who claimed
to show by exhaustive chemical analyses of the plant-
organs which he examined that, while the concentration
of the sap expressed by increasing pressures rose, the
proportion of the constituents remained the same.

The following experiments illustrate this progressive con-
centration of successive pressings from the same leaves.
The leaves experimented upon were made up into a pellet,
wrapped in two folds of fine linen and pressed in the jaws

IX

METHODS OF EXTRACTING SAP

179

of a vice. As the vice was screwed up five or six drops
of sap were pressed out and caught in a capsule ; then
the vice was opened and the same leaves re-arranged and
pressed again. The sap exuding on this occasion was
collected and kept separate from the first sample : similarly
a third sample was prepared. Successive pellets of leaves
were dealt with in the same manner ; and so, from the same
set of leaves, three samples of sap were obtained. These
were called 1st, 2nd, and 3rd pressings. For each the
depression of freezing-point A and, in some cases, the
electrical conductivity C, were determined. The latter
measurements were always made at 0.

Table 23.

Iledeva helix: leaves.

Concentration of Sap in Successive Pressings.

No. of Experiment.

1st Pressing.

2nd Pressing.

3rd Pressing.

A.

A.

0-00496

A.

L-579

0-888"

C.
0-00513

458, 459, 460
462, 463, 464

0-998
0-694

1110"

0-782"

These figures show very plainly the increase of concen-
tration in the later samples, and by inference the still
higher concentration of the sap remaining behind in the
pressed leaves. Hence, the concentration of the expressed
sap may be expected, in all cases, to be less than the
average concentration in the vacuoles of the tissues before
the application of pressure.

The explanation of this increased concentration is not
hard to find. When the pressure is first applied, almost
pure water is extruded from the intact cells, for the proto-
plasmic membranes are sensibly semi-permeable, permit-
ting water to pass out under pressure, but resisting, more
or less completely, the passage of dissolved substances.

N 2

i8o TRANSPIRATION AND ASCENT OF SAP ch.

Even in the first pressing many of the cells are usually
burst, and their sap passes out with, and is diluted by,
the much more dilute sap coming from the uninjured
cells. Subsequent pressings contain the sap of a larger
proportion of burst cells, and those which are now burst
have had their sap concentrated by the former application
of pressure. Hence, later samples must be more con-
centrated.

Necessity of rendering the protoplasm perme-
able. From this consideration it appears that the problem
of obtaining an average sample of the sap of a plant-tissue
by pressure resolves itself into the problem of rendering
the cell-membranes permeable, so that the application of
pressure will force out solvent and solutes alike. It need
scarcely be said that the method adopted for rendering
the membranes permeable must not itself alter the con-
centration.

Exposure to toluene vapour first suggested itself as a
means for rendering the protoplasm permeable. Owing to
its extremely small solubility in water, it was hoped that
it would not appreciably alter the freezing-point. By
experiment it was found that A for water saturated with
toluene is approximately 0*024, so that the correction
for its vapour going into solution would not be a serious
one.

Effect of toluene vapour on protoplasmic per-
meability. To test the efficiency of toluene vapour in
making the protoplasm permeable, a sample of leaves of
Hedi ra helix was gathered ; each leaf was halved, and two
lots (A and B) were made, each containing a half of every
leaf These two lots were then kept under the same con-
ditions of moisture and darkness in closed glass vessels, the
only difference being that in the vessel enclosing lot A an
open capsule containing cotton wool soaked in toluene was
placed. After 48 hours the freezing-point and the electrical
conductivity of the sap pressed from the two lots were

IX

METHODS OF EXTRACTING SAP

181

examined. In order to see if the increasing concentration
which is characteristic of the sap pressed from untreated
leaves occurs in the case of the sap pressed from the leaves
exposed to toluene vapour, the sap from this lot was
divided into first, second, and third pressings.

Table 24.
Hedera hcl /. : leaves.

A.

C.

0-00521
00521
0-00560

Lot A (exposed to toluene), 1st pressing ....

,, ,, 2nd pressing . . .

,, ,, 3rd pressing . . .
Lot B (control)

1-865
1-875
1-856
0-868

These results show that, with an exposure to toluene
vapour of 48 hours, the protoplasm has become permeable,
and no longer tends to keep back the dissolved substances
of the vacuoles.

Of course, such prolonged exposure has the objection
that during this process enzymes in the cells may consider-
ably alter the nature of the dissolved substances, and so
lead to a change in the concentration and constitution
of the sap. Accordingly, experiments were made to
determine if shorter exposures would be sufficient.

By means of these experiments it was found that shorter
exposures, e.g., 1 to 5 hours, caused a marked concentra-
tion of the sap expressed when compared with that from
the same leaves untreated ; but much longer exposures
were needed to render all the cells permeable, and so
allow the sap obtained to be a fair sample of that of the
uninjured leaf. The prolongation of the exposure makes
the method objectionable. Accordingly, it was abandoned
as unsatisfactory.

Protoplasm rendered permeable by intense
cold. -The possibility that the protoplasmic membranes

1 82 TRANSPIRATION AND ASCENT OV SAP ch,

might be rendered permeable by exposure to low tem-
perature then suggested itself. At the same time, it was
apparent that the low temperature would have the advan-
tage of arresting changes taking place in the tissues experi-
mented upon. To this end the experimental tissues were
immersed in liquid air. Tissues thus treated immediately
become frozen hard. From the liquid air they were
without delay transferred to a stoppered vessel to prevent
the condensation of moisture on them from the atmos-
phere. When they had assumed the temperature of the
surroundings, they were pressed in the usual manner.

Table 25.
Comparison of Saps Extracted by Various Methods.

No. of
Expt.

Sap from

A.

CxlO-.

C x 10 :1

A

472

Hedera helix, leaves untreated

0-767

403

5

2

473

1-255

605

4

8

476

Part of same sample 19 hours

in toluene vapour

1-444

536

3

8

477

Hedera helix, leaves frozen . . .

1-239

558

4

5

478

Same leaves as 477 in toluene

vapour, 2 hours

0-747

422

5

6

483

Ilex aquifolium, roots untreated

0-531

563

10

6

484

Same sample as 483 frozen .

0-682

029

9

o

486

Ilex aquifolium, leaves untreated

0651

433

6

487

Part of sample 486 frozen . . .

1130

619

5

4

494

Iris germanica, rhizome untreated

0-450

128

2

8

495

Same rhizome as 494 frozen . .

O-82'.t

335

4

510

Pyrus malus, fruit untreated . .

1-507

171

1

1

511

Same fruit as in 510 frozen . .

1-919

161

8

512

Citrus aurantium, fruit untreated

l-044 c

513

Same fruit frozen

1-206

208

1

7

518

Citrus limonum, fruit untreated .

1033

291

o

8

519

Same fruit frozen

1-089

345

3

2

514

Solarium tuberosum, tuber untreated

0-523

555

11

515

Same tuber frozen

0-588

583

9

9

516

1'itis vinifera, fruit untreated . .

2- 51)7

132

5

517

3 185

112

3

538

Chamaerops hum His, leaf untreated

0-365

298

8

1

539

1-529

752

4

9

552

< 'hamaerops humilis, leaf untreated

0-599

508

8

5

554

Same leaf frozen

1-517

926

6

1

549

Beta vulgaris, root untreated . .

1-473

r,7o

3

9

551

1 -761

555

3-2

ix METHODS OF EXTRACTING SAP 183

It is generally found that after liquid air treatment com-
paratively small pressure is needed to obtain the sap,
which flows easily from the tissues without requiring the
disruption of the cells. At the same time the sap is much
freer from the debris of broken cells than that from an
untreated leaf. This sap has always given a greater
depression of freezing-point and usually a higher con-
ductivity than that from the same tissues untreated.
Furthermore, these determinations differed from similar
measurements made on sap of the same tissues exposed
to toluene vapour. The results are tabulated on page 1 82.

The numbers in that table show conclusively that the
concentration of the sap pressed from the untreated tissues
seldom approximates to the concentration of that obtained
from the same tissues after freezing. It is hard to see how
freezing could be supposed to alter the concentration of
the sap, whereas, as we have already seen, it is certain
that the sap pressed from living tissues may be consider-
ably less concentrated than that which remains behind,
and consequently less concentrated than that which was
originally in the cells of the tissue before the pressure
was applied.

Sap unaltered by liquid air. It is well known
that chemical changes are arrested at such low tempera-
tures as that of liquid air ; however, it seemed just
possible that changes might take place in the proteids or
in the protoplasm just as the cold was being applied, and
that these changes might lead to an increase in the
quantities of dissolved substances in the sap.

To set this doubt at rest, determinations were made
of the freezing-points of the sap pressed from the untreated
roots of Beta vulgaris and from the leaves (also untreated)
of Chamaerops humilis before and after freezing in liquid
air ; also of the fluids of an egg and of bull's blood under
the same conditions. These liquids were not cleared in
any way of the matter suspended in them, so it is certain

1 84 TRANSPIRATION AND ASCENT OF SAP ch,

that they contained ample amounts of proteids and of
protoplasm to test the point. The results were as follows :

Table 26.
Effect of Liquid Air <>n Vegetable and Animal Fluids.

No. of
Bxpt.

549

550
552
553
479

480
481

482

C x 10 5

Untreated sap of root of Beta . .
Same sap frozen in liquid air . . .
Untreated sap of leaf of Chamaerops
Same sap frozen in liquid air . . .

White of egg untreated

White of egg frozen in liquid air . .

Bull's blood untreated

Bull's blood frozen in liquid air . .

1-473

570

1-474

574

599

508

0-575 c

502

0-445 :

0-445

0-616

0-584

In no case was a sensibly greater depression detected
after exposure to liquid air. The diminution in the depres-
sion observed in the experiments 553 and 482 appears to
be due to the expulsion of dissolved gases. The frothing
of the sap of Chamaerops on thawing after treatment
with liquid air was very marked. This was not looked
for in the case of the bull's blood.

Hence it appears that there is no reason to believe that
the application of liquid air leads to a concentration in
solutions in contact with proteids and protoplasm.

Again, the sap extracted from plant-organs after exposure
to liquid air does not cause plasmolysis of the cells in these
organs. This was demonstrated both for the sap of the
root of Beta and for the leaf sap of Chamaerops. In the
case of the latter the demonstration is particularly con-
vincing. Sap from the frozen leaf was found to have a
depression of T517, while the value of A for that of the
untreated organ was 0"599. Yet the former caused no
plasmolysis in the cells of a section of the leaf mounted in
it, even after twenty minutes. The difference in concen-
tration indicated by these two freezing-points would of

ix METHODS OF EXTRACTING SAP 185

course rapidly produce plasmolysis. This clearly shows
that no appreciable concentration has been effected by
the treatment, and that the sap pressed from the untreated
organ is not isotonic with that in the vacuoles of its cells.

Of course the application of liquid air cannot stop
changes taking place while the sap is being pressed, as is
evidenced by the production of colour in the sap of many
tissues during the process.

The cells treated with liquid air seem to be rendered
completely permeable. This appears from the fact that
the sap is so easily pressed from the tissues after the
exposure, often without any disruption of the cells. Also
the concentration of successive pressings from these frozen
tissues remains sensibly the same, as is shown in Table 27 :

Table 27.
Hedera helix: leaves.

No. of
Expt.

A.

C x 10\

473

474

Exposed to liquid air, 1st pressing .
,, ,, 2nd pressing .

1-255 606
1-261 597

Hence we may assume that the sap so obtained is a fair
sample of that of the uninjured tissues.

It will be noticed that in most instances the difference
in conductivity between the sap of organs treated with
liquid ah* and that of those untreated is not so marked as
the difference in freezing-point. Comparison of the ratio

for the pairs of experiments will make this clear.

This, perhaps, may be largely attributed to the greater
permeability of the protoplasm to electrolytes, so that
the sap pressed from the untreated organs is relatively
richer in them.

1 86 TRANSPIRATION AND ASCENT OF SAP ch.

The result, however, was not anticipated, as from
Andre's work it appeared that the proportions of the
solutes present in the sap were not altered by their
passage out of the organs under pressure. Hence it was
to be expected that the ratio of the electrolytes to the
other solutes would remain sensibly the same for the sap
pressed from the living tissues and for that from tissues
rendered permeable by liquid air.

The results for the rhizome of Iris germanica and for
the fruit of Citrus limonum are exceptions, for in their
case the sap extracted after freezing appears to contain
a larger proportion of electrolytes to other solutes. This
may very probably be assigned to actual differences in the
sap from two apparently similar portions of the same
massive organ. It is also possible that part of this effect
is due to the greater viscosity of the sap from the treated
organ.

These two factors probably also account for the anoma-
lous fall in conductivity noticed in the sap of the fruit
Pyrus malus and of Vitis vinifera obtained by means of
liquid air.

It is certain that a much less extreme cold than that of
liquid air would render the protoplasm permeable ; but
where this is available, it has the advantage of being very
rapid in its application, and reduces the chances of change
in the sap to a minimum.

Use of heat and chloroform vapour for the
extraction. A few experiments were made with the
object of finding out if the application of heat in a
saturated atmosphere, or the exposure to chloroform
vapour, might be used as a substitute for exposure to
liquid air.

First, with regard to the application of heat, a quantity
of leaves of Ilex aquifolium were divided down the midrib,
and two samples, A and B, were formed, each containing
half of every leaf used. A was wrapped in moist bibulous

IX

METHODS OK EXTRACTING SAP

187

paper, enclosed in a metal box, and placed for ten minutes
in a water-oven at 95. The half-leaves were then
cooled on ice and pressed, the sap flowing out easily.
Sample B was immersed in liquid air, and then pressed.
The results of two pairs of comparative experiments on
different sets of leaves were as follows :

Table 28.
Ilex aquifolium : leaves.

No. of
Expt.

A.

C x 10"'.

500
501
502
503

Sample A from heated half-leaves .
Sample B from frozen half-leaves
Sample A from heated half-leaves
Sample B from frozen half-leaves

1-152"
1-244"

0-81(1
1-305"

677
696
504
844

From these numbers it is evident that ten minutes'
exposure to 95 is not sufficient to render the membranes
permeable with certainty. Owing to the likelihood of
serious changes taking place in the sap, it would not be
feasible to expose the leaves for longer to so high a
temperature.

A similar objection was found to apply to the use of
chloroform. For this test, samples A and B were prepared
in the same way as in the foregoing experiment. A was
then exposed to the vapour of chloroform for thirty
minutes, pressed, and to the sap obtained a few drops of
chloroform were added to ensure saturation. The freezing-
point was then determined in the usual way, except that
the control-tube of the apparatus which usually contained
pure water was charged with distilled water, saturated
with chloroform. This change was, of course, not made
when examining sample B, which before pressing had
been immersed in liquid air. The conductivities were
determined in the usual manner.

1 88 TRANSPIRATION AND ASCENT OF SAP ch.

Table 29.
Hedera helix: leaves from S. aspect.

No. of
Expt.

A.

C x 10 5 .

485
562

492
493

Sample A, half-leaves exposed to chloroform
Sample B, half-leaves exposed to liquid air

1063'
1-315

Here again it appears that the exposure to chloroform
vapour has not been sufficient, and it is evidently inad-
visable to prolong the opportunity for spontaneous changes
beyond thirty minutes.

It appears from the experimental work detailed in this
chapter that the liquid pressed from untreated vegetable
tissues is not an average sample of the sap contained in
the cells of those tissues. However, exposure to intense
cold, as may be conveniently effected by treatment with
liquid air, renders the cells permeable and secures that
the sap yielded up under pressure is not altered in concen-
tration during its passage from the cells. Exposure to
heat, toluene or chloroform vapour cannot be recom-
mended for securing this end.

Literature.

Andre, (i., " Sur la composition des liquides qui circulent dans le vegetal,"
Gompt. rend., 1906, 142, p. 106.

Id., "Sur la composition des sues vegetaux extraifcs des racines," Gompt.
rend., 1906, 143, p. 972.

Id., "Sur la composition des sues vegetaux extraits des tiges et des
feuilles," Gompt. rend., 1907, 144, p. 276.

Id., " Sur la migration des principes solubles dans le vegetal," Gompt.
rend., 1907, 144, p. 383.

Atkins. YV. R. <;., " Cryoscopic Determinations of the Osmotic Pressures
of some Plant Organs," Proc. U(j. Dublin Soc, 1910, vol. xii. (N. S.), p. 463,
and Notes from the Botanical School of Trinity College, Dublin, vol. ii, p. 84.

Dixon, II 11., and Atkins W. R. G., " On < >smotic Pressures in Plants ; and
on a Thermo-electric Method of Determining Freezing Points,'* Proc. Roy.
Dublin Soc, 1910, vol. xii. (N. S.), p. 275.

Id., "Osmotic Pressures in Plants. I. Methods of Extracting Sap from

ix METHODS OF EXTRACTING SAP 189

Plant Organs," Proc. Boy. Dublin Soc. 1913, vol. xiii. (N. S.), p. 422, and
Xotrsfram the Botanical School of Trinity College, Dublin, vol. ii, p. 152.

Id., "Osmotic Pressures in Plants. II. Cryoscopic and Conductivity
Measurements of some Vegetable Saps," Proc. Roy. Dublin Soc, 1913, vol.
xiii. (N. S.), p. 434, and Notes from the Botanical School of Trinity College,
Dublin, vol. ii. p. 173.

Marie, C. H., and Gatin, C. L., " Determinations cryos.opiques effectuees
sur les sues vegetaux." (1912).

Maximow, N. A., " Chemische Schutzmittelder Pflanzen gegen Erfrieren,"
Ber. d. Deutsch. Bot. Gesell., 1912. Pd. 30 s. 52, 293, and 504.

CHAPTER X

OSMOTIC PRESSURES IN PLANTS

Having determined the freezing-point of the unaltered
sap, it remains to calculate from it the osmotic pressure
developed in the cells. The relation connecting these
two quantities has been determined experimentally and
also deduced theoretically. Nernst gives the equation :

A x 12 03 = osmotic pressure in atmospheres at 0.

In Table 30 (p. 192), the depression of freezing-point
and the calculated osmotic pressures are given under
A and P respectively.

Factors tending to lower results. It should
here be mentioned that the osmotic pressure recorded
is probably in no case so high as that obtaining in
the cells of the leaves at the moment when they were
collected ; for when the leaves are plucked the tension
in the water columns is destroyed and the cells of the
leaves are free to gorge themselves with water from the
trachea?. The dilution may be considerable if the cells
were not fully distended when the leaves were gathered,
and if they were severed from the branch with a consider-
able mass of the conducting tracts.

Even when this dilution may be neglected the osmotic
pressure of the sap of the fresh leaf does not represent the
maximum tension the osmotic substances in the cells can

190

ch. x OSMOTIC PRESSURES IN PLANTS 191

exert ; ! for the fully distended cells of a leaf may be
considerably reduced in size before wilting takes place.
The variation in pressure which may be occasioned by
this change of volume can be judged by comparing the
osmotic pressure of the sap extracted from freshly plucked
turgid leaves with that of the sap from part of the same
sample when the first signs of wilting are apparent. An
experiment of this nature was made on a sample of leaves
of Fraxinus oxyphyUa. The sap from the fresh leaves
had an osmotic pressure of 15*15 atm., while that from the
same sample just beginning to wilt had a pressure of
24 - 09 atm. A similar experiment with the leaves of
Wistaria sinensis gave 6*61 atm. and 11 '38 atm. as the
pressure in the turgid and wilting leaves respectively.

Again the osmotic pressures, calculated from the
freezing points of the saps, should, when applied to the
leaves at air temperatures be raised 10-15 per cent.

We may conclude, then, that the tensions developed
on the upper ends of the water columns are usually at least
as great as the pressures recorded in the table ; but that
inasmuch as fresh leaves were always examined, the
pressures indicated in the table are by no means equal
to the maximum tensions which may be exerted on the
rising sap.

Determinations. The majority of the determina-
tions have been made on sap extracted by the liquid-air
method, while some, which are marked in the table
' untreated," have been made on sap simply pressed
from the tissues. It is highly probable that, if the
samples which yielded these measurements had been
treated with liquid-air before pressing, higher values
for the osmotic pressures would have been obtained.
A similar criticism is applicable in a less degree to those
marked ' crushed and diluted." In these an extract

1 It may be pointed out that in low-growing plants high tensions are
probably only developed when the supply of water to the roots is restricted.

192 TRANSPIRATION AND ASCENT OF SAP ch.

of the crushed leaves was examined and the osmotic
pressure of the original sap was calculated, allowing for
dilution, as before described.

Although the determinations made on sap extracted by
these last two methods can only be regarded as minor
limits, it appears that they also have a relative signi-
ficance. Thus the sap yielded without treatment from
three samples of leaves plucked simultaneously from three
separate branches of Syringa vulgaris growing under
similar conditions had freezing-points respectively lo70,
1'581, and 1576. Furthermore, the increase in con-
centration of the sap of the parenchymatous tissues from
the root upwards to the leaf, which is always observable in
the saps extracted by the liquid-air method, is paralleled
by a similar gradient from below upwards in the saps
extracted from these tissues without treatment.

Table 30.
Osmotic Pressures in Plant Organs.

V

No.

of

Expt,

648

;.<;;,
566
574
575
2

3
520
533
528

-.;,

661

5:52

679

11

17

18

Acer pseudoplatanus leaves

Agave amerieana ,,

Aim' plicatile ,, ......

Anthumum andreanum, leaves

,, cristallinum ,, .....

Catalpa bignonioides, leaves (crushed and

diluted) ....
, , speciosa, leaves (crushed and diluted)

Cerasus laurocerasus, leaves

('haiitarrops humilis ,,

( 'mdyline australis ,,

Equisetum telmateia, lateral branches (un-
treated)

,, ,, main stems ....

Eucalyptus globulus, horiz. leaves ....

,, ,, roots (untreated) . .

Vagus silvatica var. purpurea, leaves . . .

Fraxinus excelsior (untreated)

,, oxyphylla, fresh leaves (untreated)

,, ,, leaves beginning to

wilt (untreated) . .

Date of
Obs.

26 s i:;
11/1/13
11/1 13

27 1 L3

27/1/13

30/8/09

30 8/09

28/11/13

29 11/13

28 11/13

12/9/09

10/9/13

29/11/12

23/10/09

4/10 13

2/9/09

11 '.Mill

11/9/09

A.

207
840
292
623

727

905 :

724 r

522

598'

116

946

878

970

433

119

097

230

2 003

P.

14

10

3

7
8

22
20
18
19
13

11
10
11
5
13
25
15

52
11
52
49
73

92
73
31

22
43

38
56
68
33
45
22
15

24-09

OSMOTIC PRESSURES IN PLANTS

J93

Table 30 (contd.).
Osmotic Pressures in Plant Organs (contd.).

No.

of

Expt.

014
615

673
617
660

498
495
496

7

548
546
545
677
573
576
651
652
653
654
1

96

547

572

634

5

39

24

27

84

665

?>
) >

7>

>1

Hedera helix, N. aspect, leaves

,, ,, S. aspect, leaves

Helianthus multiflorus ,,

Ilex aquifolium, mature ultimate leaves

,, ,.. roots

Iris germanica, leaves

rhizome

roots .

Magnolia acuminata, leaves (crushed and

diluted)

Monstera deliciosa, leaves

Mnsa sapientum ,,

Passijiora quadrangular is, leaves ....
Pinus laricio , , ....

Platycerium alcicome ,, ....

Poltjpodium irioides ,, ....

Populus alba, spring leaves

,, ,, summer leaves

,, ,, bark at 40 ft. level ....

,, ,, ,, of root

,, balsamifera, leaves (crushed and

diluted)

Pteris aquilina, leaves (untreated) ....

Saccharum officinarum, leaves

Selaginella mertensii, leaves and stems . .

Syringa vulgaris, leaves

(crushed and diluted)
(untreated) ....
Ulmus campestris, leaves (untreated) . , .

Vitis veitchii, leaves (untreated)

Wistaria sinensis, leaves beginning to wilt

(untreated)

,, ,, leaves fresh

Date of
Obs.

A.

P.

19/3/13

1-468

17 66

19/3/13

1

555 3

18-70

2/10/13

764

9 18

19/3 L3

1

572

18-91

10/9 13

1

156

13 91

1

085

1305

15/11/12

829

9-97

15/11/12

764

9-20

1/9/09

1

858

22-34

10/12/12

552

6 64

10/12/12

785

9-44

10/12/12

1

162

13-98

3/10/13

1

174

14-13

27/1/13

625

751

27/1/13

1

886=

1065

28/8/13

1

326 c

15-95

28/8/13

1

487

17-88

28/8/13

1

215

14-62

28/8/13

1

101

1323

30/8/09

1

639

1972

23/9/09

619

7 44

10/12/12

484

5-83

27/1/13

845

10 16

22/8/13

2

119

25 50

31/8/09

2

234

26-87

13/9 09

2

135

25-68

6/9/09

L

550

18-64

6/9/09

783

9-34

10/9/09

946

11-38

30/9/13

0-709

8-52

Tension in tracheae indicated by osmotic
pressures in leaves. In the foregoing table, as
a rule, only the maximum pressure observed in the
leaves or the roots of each species is recorded. As
a matter of fact, wherever sought, great variations
were found between different individuals of the same
species, and in the same individuals in different posi-
tions and under different conditions. In every case,

o

194 TRANSPIRATION AND ASCENT OF SAP ch.

the osmotic pressure of the leaf cells (ranging as it does
from 5 to 27 atm.) was such that it was well able to sustain
the tension necessary to lift the water current in the plant
and to keep the cells turgid during normal transpiration.
Renner's figures obtained by a different method are in
accordance with this result. He compared the flow due to
the tension set up by the leaves, with that caused by a
known difference of pressure, in branches in which the
resistance had been artificially increased. The tension
thus indicated varied between 10 and 20 atmospheres.

Variations in osmotic pressure. With regard to
the causes of the variations in pressure much still has to
be found out, but the results obtained up to the present
may be of interest.

In the first place, a series of experiments was designed
to test the possibility that the osmotic pressure of the sap
of the leaves on any region of the branches is defined by
the resistance which has to be overcome in drawing the
transpiration current from the roots to that part. Ewart
had previously looked for such a difference by means of
the plasmolytic method, but he seems to have encountered
difficulties and left the question undecided. In the following
experiments sap was pressed from leaves taken at a con-
siderable height above the ground ; and its freezing-point
was compared with that of leaves from near the ground -
level. The results of pairs of experiments bearing on
this question are shown in Table 31 (p. 195).

It there appears that, on the whole, taking the experi-
ments in pairs, the leaves at the lower level contained sap
with a lower (sometimes considerably lower) osmotic
pressure than that of higher leaves. But experiments
are far from satisfactorily bearing out this view ; for it
has been noted that the osmotic pressures of the sap from
leaves at the same level, but at different times and under
different conditions, by no means correspond in each case,
although they are often higher than those of leaves at a lower

OSMOTIC PRESSURE IN PLANTS

'95

level. The reverse, however, is sometimes found, as in
Experiments 6 and 7, where the pressure in the lower is
much greater than in the higher leaves.

Table 31.
Osmotic Pressure and Height above Ground Coiutared.

No of

Description of Sample. Sap from untreated

A.

P.

Expt.

leaves in every case.

6]

Magnolia acuminata, leaves from 38 ft. level .

1-628

19 58

7)

11 11 )5 ) J "* *" 11

l-858 c

22-34

19)

JJ 11 11 11 oo it. ,,

1-373

16-51

20J

)) ) ) n J5 * f" 11

1142

13-74

in

Fraxinus excelsior, leaves from 20 ft. level . .

2-097

25-22

12J

,, ,, shaded leaves from 3 ft. level

1-020

12-27

13)

,, ,, exposed ,, ,, 43 ft. ,,

1-380

1660

14j

,, ,, shaded ,, ,, 2 ft. ,,

1-000

12-03

151
16/

,, ,, exposed ,, ,, 43 ft. ,,

1094

1316

,, ,, shaded ,, ,, 2 ft. ,,

936

11-26

25)

Vitis veitchii, leaves from 1 ft. level ....

0'816

9-81

26 J

94. ft

,, ,, ,, . , ^-x IU. ,, ....

0-653

7-85

27)

)) 11 11 11 * Ik. , , ....

0-783

9 34

28 J

11 5 5 11 11 ^ Ik' 11 ....

0-519

6 24

The possibility that these discrepancies might be due
to resistance in the conducting tracts apart from that
offered by the hydrostatic head had to be examined, and
Experiments 80, 81, 82, 83 on Wistaria sinensis and 21,
22, 23, 24 on JJlm.us campestris (recorded in Tables 32
and 33 respectively) were carried out.

Table 32.
Wistaria sinensis : sap from untreated leaves.

No. of
Expt.

Description of Sample.

A.

P.

M. 1

80)
81/

82)
83/

Shaded, from 3 ft. level on basal shoot . .
Exposed, from 3 ft. level at distal end of

horizontal branch 65 ft. long

Exposed, from 27 ft. level

Shaded, from 3 ft. level

412

0-437
0-550
443

4-95

5-25
6-61
5-53

149

169
162
169

Under M are given the mean molecular weights of the solutes calculated

from the dry weight and freezing-point of the solutions.

o 2

196 TRANSPIRATION AND ASCENT OF SAP ch.

Experiments 80 and 81 were made on sap from the leaves
of an old Wistaria trained on a low wall. One sample of
leaves was gathered from short branches near the base of
the main stem. The leaves were about three feet over
the ground. The second sample of leaves was taken from
the terminal branches of a stem running 65 feet approxi-
mately horizontally along the wall at a level of about
three feet. Here again we find a slight difference in
pressure in favour of the distal leaves.

Table 33.
Ulmus campestris : sap from untreated leaves.

No. of
Expt.

Description of Sample.

A.

P.

M.

21
22
23
24

From short shoots on top of arched
branch in shady position at 18 ft. level .

From short shoots at base of trunk in
shady position at 1 ft. level

From short shoots at outer end of arched
branch in shady position at 10 ft. level .

From short shoots on trunk in sunny

0-888
0-703
1030
1-550

10-68

9-18

12 39

18-64

152
148
165
155

The numbers in Table 33 show the real meaning of the
results which had apparently indicated that the level was
the controlling factor in determining the osmotic pressure.
If the hydrostatic head defined the pressure of the leaves,
it is evident that the pressure in the leaves examined
in Experiment 21 should have been the greatest ; if the
resistance of the water-tracts were the controlling factor,
those of Experiment 23 should have had the maximum
pressure, which should have been much greater than those of
22 and 24. The actual order is 24, 23, 21, 22. From this
it is clear that the resistance of the water-tracts was not
the controlling factor of the pressure ; accordingly some
other cause for its variation must be sought. This cause
seems to be principally the fluctuations in the sugar-
content of the leaves due to difference in illumination.

x OSMOTIC PRESSURE IN PLANTS 197

In Experiments 21 and 22, the leaves of which are from
shaded positions, smaller pressures are found than in
Experiment 24, which was performed on sap from leaves
in a sunny position. Experiment 23, on leaves coming
from the outside of the crown facing a clear north sky,
and being consequently better illuminated than the other
two samples, used in 21 and 22, reveals a higher pressure
than they, though this pressure is considerably lower than
that found in 24.

This effect of illumination in raising the osmotic pressure
in the cells of leaves appeared in a great number of
experiments. Conversely it was also found that the
pressure gradually falls when they are cut off from light.
The difference thus produced may amount to 141 1 atm.

These experiments, taken in conjunction with observa-
tions on the mean molecular weight of the solutes, amply
show that conditions favouring the formation of soluble
carbohydrates are those which raise the osmotic pressure ;
and vice versa, the osmotic pressure falls when conditions
are unfavourable to the formation of these substances, while
not preventing their consumption. It was only to be
expected then that the cells in organs where these sub-
stances are normally stored should have high osmotic
pressures. This point is well illustrated by the high
pressures found in the fruit of Vitis vinifera (38*3 atm.
A=3T85), and of Pyrus malus (23T atm. A = 1-919),
and in the root of Beta vulgaris (2T2 atm. A = T761 ). 1

1 Another point of interest with regard to the distribution of osmotic
pressure in plants may be mentioned.

In almost every case it was found that the older leaves, ccnterls paribus,
had a higher osmotic pressure than the younger leaves on the same plant.
This was observed in Syringa vulgaris, Vitis veitchii, Eucalyptus globulus,
Hedera helix, and especially in Ilex aquifolium. The leaves of the last named
evergreen persist through four or five periods of growth, and it is generally
found at any time that the osmotic pressure of the sap of the leaves of each
successive growth is lower than that of those which precede it. Thus taking
the mean of many observations made throughout the year on the sap from
untreated leaves, the osmotic pressure of the sap of the ultimate three

198 TRANSPIRATION AND ASCENT OF SAP ch.

Inasmuch, as the conditions favouring the formation
and storage of soluble carbohydrates are generally best
realised in the higher levels of trees, it is not surprising
to find that the greater pressures are often developed
there, while smaller pressures usually occur in leaves, etc.,
near the ground level. But as in every case the pressure
observed is much greater than the force required to over-
come the resistance of the water-tracts ; it is evidently
in no way defined by that resistance.

Relation of osmotic pressure to the breaking
stress of the cell-walls. The presence of these high
pressures, amounting to about 30 atm. in the mesophyll,
naturally raises the question as to how far the tensile
strength of the cell- walls is taxed by their distending action.

When these cells are inadequately supplied with water
it is evident that their walls may be barely distended
and there will be no tension in them. On other occasions
when the supply is sufficient their walls may be stretched
by nearly the whole osmotic pressure of their vacuoles.

The actual tenacity of cell-walls has been determined
by various investigators. The following determinations
are quoted from PfefTer. Schwendener found that
sclerenchymatous fibres could support from 15 to 25 kilo,
per sq. mm. before breaking, Weinzierl estimated the
tenacity of the fibrous cells in the leaves of Phormium
tenax at 20'33 kilo, per sq. mm., and those of Allium porro
at 14*71 kilo, per sq. mm. Ambronn gives the breaking
strain for the walls of collenchymatous cells at 8 to 12
kilo, per sq. mm.

More recently the author estimated the tenacity of

leaves on the last growth was 8 - 88 atm. (A = 0'738) ; of the penultimate
three leaves on the last growth was 926 atm. (A = - 770 ) ; of the leaves on
the penultimate growth, 10 - 61 atm. (A = 0'882~), and of the leaves on the
antepenultimate growth, 11 "76 atm. (A = - 978'). If these observations had
been made on sap extracted by the liquid-air method considerably higher
pressures would have been obtained. But observations made by this latter
method fully confirm the rise in pressure with age.

x OSMOTIC PRESSURE IN PLANTS 199

cellulose by loading very gradually a single fibre, detached
from the seed of Gossypium, until it broke. By observing
the breaking weight and the area of the cross-section where
the break occurred, the tenacity was obtained.

Different fibres gave tenacities of 37 kilo, to
60 kilo, per sq. mm. The cell-walls of these fibres are
of pure cellulose, and having been taken from fresh seeds
and soaked in water, were in the imbibed condition, and
consequently resembled in their properties the imbibed
cellulose walls of the mesophyll cells. As in no case can
the fibre support a stress greater than its tenacity we
must regard the lower results as due to flaws in the fibres,
and the highest figures as giving the actual tenacity of
cellulose.

The cells of the leaf approximate in form more or less
to cylinders ; they are seldom, if ever, spheres. There-
fore the stress in the cellulose wall will never exceed the
internal pressure P acting over an area -n-r 2 divided by
the sectional area of the cell-wall 2 -n-rt, where t is the
thickness of the wall, and r the radius of the cylindrical
or spherical cell.

Pnr*

Stress per sq. mm. of cellulose =

2nrt

In Cytisus laburnum, for example, the palisade cells
are approximately 0'06 mm. long, 0-0175 mm. in diameter,
and their cell-walls are 0001 mm. thick. The osmotic
pressure may rise to 30 atm., or about 300 gr. per sq. mm.

300 x 0-0087 ,, ft .
s = ' 2 x 0001 = gr " per q '

The observations quoted above show that the breaking
stress of cellulose is more than 50,000 gr. per sq. mm.
Consequently an osmotic pressure of 30 atm., even if
entirely borne by the cell-wall, will not tax it to near
its limit.

Probably the greatest value for the diameter of any of

200 TRANSPIRATION AND ASCENT OF SAP ch.x

the cylindrical cells of the leaves is under 0*05 mm., but
even in cells having this diameter and having a thickness
of wall equal to 0*001 mm. only, it would require a pressure
of 100 atm. to tax the cell- wall to its breaking point.
Consequently we may take it that the osmotic pressures
of the leaves never seriously tax the tensile strength of
the cell- walls.

Literature.

Dixon, H. H., "The tensile Strength of Cell-walls," Ann. of Botany,
L897, vol. xi, p. 585.

Id., "On the Physics of the Transpiration Current," Notes from the
Botanical School of Trinity College, Dublin, vol. i, p. 57.

Id., "Transpiration and the Ascent of Sap," Progresses Rei Botanicae, 1909,
Bd. iii, s. 1.

Dixon, H. H., and Atkins, W. R. G., "On Osmotic Pressures in Plants ;
and on a Thermo-electric Method of Determining Freezing Points," Proc,
Roy. Dublin Soc, 1910, vol. xii (N.S.), p. 275.

Id., " Changes in the Osmotic Pressure of the Sap of the Developing
Leaves of Syringa vulgaris " Proc. Boy. Dublin Soc, 1912, vol. xiii (N.S.),
p. 219, and Notes from the Botanical School of Trinity College, Dublin, vol. ii,
p. 90.

Id., "Variations in the Osmotic Pressure of the Sap of Ilex aquifolium,"
Proc. Boy. DuMin Soc, 1912, vol. xiii (N.S.), p. 229, and Notes from, the
Botanical School of Trinity College, Dublin, vol. ii, p. 111.

Id., "Variations in the Osmotic Pressure of the Sap of the Leaves of Heclera
helix," Proc Boy. Dublin Soc, 1912, vol. xiii (N.S.), p. 239, and Notes from
the Botanical School of Trinity College Dublin, vol. ii, p. 103.

Id., " Osmotic Pressures in Plants. I. Methods of Extracting Sap from
Plant Organs. II. Cryoscopic and Conductivity Measurements on Some
Vegetable Saps," Proc. Boy. Dublin Soc, 1913, vol. xiii (N.S.), pp. 422, 434,
and Notes from the Botanical School of Trinity College, Dublin, vol. ii, pp. 152,
173.

Pfeffer, W., "Physiology of Plants," Translated by A. J. Ewart (Oxford
1903.)

Renner, <>., " Exi>erimentelle Beitrage zur Kenntniss der Wasserbewe-
gung," Flora, 1911, Bd. 103, Hft. 3, 171.

CHAPTER XI

ENERGY AVAILABLE FOR RAISING THE SAP

Energy available for secretion. It has been
pointed out earlier that, under normal conditions of
transpiration, water is probably extracted from the
mesophyll cells and exposed on the outside of these cells
to evaporation by a secretory action. Hence the lifting
force of the transpiration current in these cases may be
attributed to the expenditure of energy by the proto-
plasm of the leaf- cells. For this the energy entering the
cells at the moment, and that stored as energetic com-
pounds in the protoplasm, are available.

Energy entering the leaf. With regard to the
former the data determined experimentally by Brown and
Escombe are applicable.

By use of a Callendar radiometer they found that the
maximum amount of energy incident on leaves of plants
in Kew in full sunshine amounted to TO to 0*5 cal. per
sq. cm. per minute. Had the observations been made
in a higher position, and in one free from the veil of smoke
hanging over Kew, this amount would have been greater.
The coefficient of absorption of the leaves experimented
upon averaged about 0*7. Consequently each square
centimetre should absorb at least 0'5 x 0*7 cal. per
minute of radiant energy. Of this it was found about
0*25 cal. might be required for the vaporisation of the water
given off (allowing 42 x 10" 5 gr. to be transpired per sq.

202 TRANSPIRATION AND ASCENT OF SAP ch.

cm. per minute), while less than 1 per cent, or 0*0035 cal.
was used in photosynthesis, leaving more than 0*0965 cal.
available for carrying out other processes in the leaf
and for raising its temperature above the surroundings.

External energy of this nature may not be available
for secretion when the temperature is high and the sun-
shine is veiled ; for then vaporisation lowers the tem-
perature of the evaporating surfaces, and these possibly
absorb the energy before available. Thus in an experi-
ment on a leaf of Helianthus annum under such conditions
Brown and Escombe found :

Total radiation falling on the leaf per sq. cm. per min. 02746 cal.

absorbed by the leaf ,, ,, 0'1884 ,,

Energy expended in photosynthesis ,, ,, 0033 ,,

,, ,, vaporisation ,, ,, 0*3668 ,,

In this case 0*1817 cal. must have been derived from
the surroundings, and hence very probably the cells of
the leaves drawing forward the water lost rather than
gained energy in the process.

Energy set free by respiration. With regard to
the stored energy set free by respiration and at least
partially available for secretory processes, precise figures
are not to hand ; but we may infer from certain obser-
vations the order of the amounts available.

Aubert records that 1 gr. of leaves of Hedera helix rises in
respiration 252*1 c.mm. of oxygen per hour. Assuming a
hexose is oxidised according to the equation

C 6 H 12 6 + 60 2 = 6CO, + 6H 2 + 677 2 x 10 3 cal.

we find that the respiration of 1 gr. of these leaves generates
1*27 cal. per hour. 1 gr. of leaf has about 45 sq. cm. under
surface. Therefore the respiration of 1 sq. cm. of Hedera
leaf would generate 0*0282 cal. per hour. The amount
of water vapour exhaled from 1 sq. cm. will probably not
exceed, under normal conditions, 0*0252 gr. per hour.
Consequently for every gramme of water vapour given off
there may be as much as a calorie available for raising it

xi ENERGY AVAILABLE FOR RAISING SAP 203

through the water tracts and secreting it on the outside
of the leaf -cells.

No energy need be spent in separating the water from
the solution in the cells, inasmuch as there is a constant
supply of water on the inner side bordering on the tracheae,
and the water passing out to the seat of evaporation is
immediately replaced by that coming in from the vascular
bundles.

Hence, taking into account only respiratory energy, we
have available for raising the water in a tree and exposing
it for evaporation something approximating to one calorie
for every gramme of water given off.

It has previously been shown that the resistance to be
overcome in moving the transpiration current through the
stems of trees is not much more than equivalent to a head
of water equal to the length of the stem. Hence, as each
cubic centimetre of water given ofr from the leaves of a
tree 100 m. high requires an expenditure of work to the
extent 100 x 100 gr. cm. to lift it, we must add the same
quantity of work to overcome the resistance of the con-
ducting tracts ; and the total work for raising a cubic
centimetre in the tree will be about 2 x 10 4 gr. cm., or in

2 x 10 4
calories 77^ 77^, i.e., about 0'5 calorie.
428 x 10"

As we have seen, the respiratory energy of the leaf
supplies something of the order of one calorie for each
cubic centimetre of water given off and hence would be
quite adequate to do the raising of the sap from the root
to the leaves.

Raising of sap by evaporation. We have already
seen that under certain conditions, e.g., when evapora-
tion from the transpiring cells removes water faster than
their secretory powers can provide it, the menisci formed
in the substance of their walls must support the
tensile columns of water in the plant. Evaporation from
these menisci must provide the traction to raise the

r ..(...

204 TRANSPIRATION AND ASCENT OF SAP ch.

water. The tension is transmitted downwards through
the roots to the absorbing cells. In these cells the actions
which occur must be the converse of those occurring in
the mesophyll. At the root the entry of water depends
on the gradient of pressure on passing from the outside
of the root to the inside of the tracheaa. The fall of pressure
due to the tension in the water is continuous all the way
up the stem to the leaf. Thus we may regard the flow of
water up the highest tree as due to the evaporation and
condensation produced by the difference between the
vapour pressure in the soil spaces and that obtaining round
the leaves. The column of tensile water flows
under the action of this difference from end
to end of the plant.

Model. The relations in these respects
of the leaves to the roots may be illustrated
by two porous pots connected hermetically
by a glass tube about a metre long, the
pots and the tube being completely filled with
water (Fig. 29). If one is immersed in damp
earth and the other supported above it, the
difference in the state of saturation of the
spaces surrounding each will be sufficient to
cause condensation to take place on the
surface of the lower pot and evaporation
to proceed from the surface of the upper one.
Motion of the water upwards may be demon-
strated by the introduction of a mercurial
index into the tube. If evaporation from the
upper pot eliminates more water than condensation on
the lower pot supplies, and if the liquid in the apparatus
is in a state capable of standing tension, the stress developed
by the reduction of the volume of water will drag in the
menisci in the interstices of the walls of both porous pots,
and make them more concave. This will have a two-
fold effect. The rendering of the upper menisci more

Fi<;. 29.

xi ENERGY AVAILABLE FOR RAISING SAP 205

concave tends to bring them into equilibrium with a lower
vapour pressure, and consequently reduces evaporation,
whilst the increased concavity of the lower menisci for the
same reason renders them more ready to condense water
vapour.

When this model has been in action for some time it
will be noticed that the lower pot is at a higher temperature
than the upper one. This difference of temperature is
to be attributed to the " sorting demon " action progres-
sing at the two surfaces. The lower menisci form a trap
for more energetic water vapour molecules in the soil
spaces, the upper retain the less energetic, while those
which are more energetic escape into the surrounding
space. Hence there is a gain of heat to the lower menisci,
and a loss of heat by molecular convection from the
upper. This cooling of the upper menisci maintains a
regular flow of heat into the evaporating surfaces, which
is constantly being abstracted again by the escaping
molecules. As they escape, fresh molecules are drawn into
their places from the water beneath by the attractions
of those remaining in the menisci, and these mutual
attractions find expression in the tensile strength of the
liquid which joins the whole column to the evaporating
menisci. Thus the loss of molecules from the menisci,
kept up by the inflow of heat, is able, by calling into play
the mutual attractions of the water molecules, to set up
a stress in the water which may be transmitted to the
lower menisci.

If we suppose, in order to imitate the conditions in the
plant more closely, the outer surfaces of the two pots in
the model to be covered with osmotic cells, we can readily
see that the conditions are not essentially altered. The
osmotic pressure in these cells need not change the gradients
of pressure in the water. The osmotic pressure is the
pressure which the dissolved substance exerts against the
membranes of the cells, while the tension is in the solvent

2o6 TRANSPIRATION AND ASCENT OF SAP ch.

and is transmitted unaltered across the space in which
the pressure of the solutes is also exerted.

In this respect the osmotic pressure acts just in the same
way as a number of internal supports, keeping the cell
turgid and preventing it from collapsing under the tension
of the solvent which drags the water across the cell.

An evaporation engine. The suitability of evapo-
rating menisci as a mechanism for doing work may be

illustrated by a model
designed by Dr. Joly
(see Fig. 30). A light
fly-wheel is delicately
hung in an air-tight
chamber (a). The short
limb of a J -shaped
glass tube (b) enters
the upper part of this
chamber from above,
and its end, which is
drawn to a fine
nozzle, is vertically
over the edge of the
wheel. The longer
limb of the tube dips
into a well of water
formed by a glass
vessel (c) beneath the
chamber. In the lower part of the chamber, beneath the
wheel, is a small cistern (d) also containing water, and from
this cistern leads a branching tube (e) which distributes the
water to twelve porous pots (/), some of which are seen in
the background.

Evaporation from the surface of the pots removes the
water from the cistern. This diminishes the air pressure
in the air-tight chamber, and the atmosphere, pressing
on the surface of the water in the well beneath, urges

Ft.;. 30.

xi ENERGY AVAILABLE FOR RAISING SAP 207

water up the glass tube and through the fine nozzle. A
series of drops is thus delivered on the edge of the wheel,
and keeps it in constant rotation so long as there is water
in the well beneath. In order to prevent the drops run-
ning round the edge of the wheel, this latter is covered
with a thin coating of paraffin wax. By this means the
drops remain fixed on the edge of the wheel just like the
bucketfuls of water on the rim of an overshot wheel. Lest
the drops should not readily detach themselves immediately
on reaching their lowest position on the edge of the wheel,
a camel's-hair brush, projecting out of the water in the
cistern, is brought sufficiently close to the wheel to be
able to drink the drops off its edge.

The arrangements adopted in setting up this model are
shown in the accompanying illustration. The pots used
were 16 cm. in length and 5 cm. in diameter, and thus
the twelve expose a total effective evaporating surface of
3200 sq. cm. With these arrangements the wheel might
be kept in rotation apparently for an indefinite time,
were it not that the paraffin surface on the edge of the
wheel ceases to preserve its water-repellent character after
a certain amount of wear, and then the drops from the
nozzle run down round the wheel without causing its
rotation. Notwithstanding this, the wheel may be easily
kept in rotation for several days.

When the pots are replaced by a leafy branch the wheel
is kept in lively rotation by the evaporation from the
leaves. Variations in the speed of rotation of the wheel
mark the variations in the amount of evaporation from
the leaves. Thus, exposure to a draught or to a higher
temperature, as in the case of the pots, accelerates the rate
of rotation. In addition to this, the influence of light upon
the stomata may be observed in the increased rate by day;
while conversely darkness diminishes the rate of rotation.

It is interesting to trace in this model how the molar
work of maintaining the wheel in rotation is derived from

208 TRANSPIRATION AND ASCENT OF SAP ch.

the molecular action at the evaporating surface. The
molecules of the liquid with most vis viva emerge from the
superficial layers of the liquid and, escaping by reason
of their momentum from the attractions of their fellows,
dash into the adjacent unsaturated space. Those with
less energy cannot free themselves from the bonds of
their neighbours and, if they get beyond the surface at
all, they must needs fall back again into the body of the
liquid. In this manner, from the surface of an evaporating
liquid there is a constant sorting out of these molecules
which possess the greatest amount of energy, while those
with less remain behind. In consequence of this selective
action, the unevaporated liquid, being composed of those
which possess least energy, is maintained at a lower tem-
perature, and therefore heat continues to flow into it from
its surroundings. It becomes, in fact, a sink of energy.
The heat, which is continually entering at the evaporating
surface, prevents the liquid under ordinary conditions
falling to a temperature much below that of the sur-
rounding objects, and so increases the vis viva of the re-
maining molecules and enables evaporation to continue.
This inflow of heat at the seat of evaporation is the ultimate
source of the energy which raises the water to turn the
wheel. This is true both for the evaporating surfaces of
the porous pots and those of the leaves when secretion
is not taking place.

In the model, the evaporating menisci do work raising
the water, which in its passage turns the wheel ; in an
intact plant the work is done against the weight of the
tensile transpiration stream and the resistance of the
conducting tracts. The capillary forces of the menisci
serve to hold the upper surface of the water in position
whilst the inflowing heat, by disconnecting and removing
molecules in these menisci, tends to shorten the water
columns. The bonds cast off by the escaping molecules
are transferred to others within the liquid, which are thus

xi ENERGY AVAILABLE FOR RAISING SAP 209

drawn to the menisci, and the tension is generated and
maintained in the liquid. It is clear that the capillary
forces of the menisci must be sufficiently great to support
the tension needed to raise the water ; hence the fine
grain of the walls of the transpiring cells forms an essential
link in the mechanism which utilises the energy entering
at the evaporating surfaces in the leaves.

Effect of tension in sap on evaporation.
When the heat entering at the evaporating surfaces is,
as just described, the source of the energy which raises
the transpiration current, it is evident that less water
will be evaporated from these surfaces than from similar
surfaces under similar conditions, but relieved of the
work of drawing the water through the conducting
channels. Accordingly it is of interest to inquire what
proportion the work of raising the sap bears to that of
evaporation, or, in other words, how will the tension in
the sap retard evaporation from the transpiring cells.

As was shown previously, each cubic centimetre of
water given off from the leaves of a tree 100 m. high
requires an expenditure of work amounting to 0'5 cal.
for transporting in the conducting tracts one cubic centi-
metre from the roots to the leaves.

To evaporate a cubic centimetre of water at 20 C.
requires 592'5 cal. Therefore the work done in transport-
ing the water from the roots to the leaves of a 100 m.
tree will not require more than the one-thousandth part
of the energy required for the evaporation of the water.
Hence to obtain the energy needed to raise the water
in a 100 m. tree the amount evaporated will only be
diminished by one-thousandth part. Even taking the
highest and certainly excessive estimate of the resistance,
the amount evaporated will only be reduced by one-thirtieth,
if, in addition to evaporation, the energy absorbed by the
leaf has to do the work of transporting the water from
the roots.

210 TRANSPIRATION AND ASCENT OF SAP ch.

Summary. In bringing to a close this description of
researches on Transpiration and Ascent of Sap, it seems
suitable to summarise br'efly the principal conclusions
which they have established.

The transpiration stream is raised by secretory actions
taking place in the leaf cells, or by evaporation and capil-
larity (imbibition) at their surfaces drawing water from
the tracheae. The state of saturation surrounding these
cells determines which of these agencies is effective.

The configuration, physical properties, and structure of
the wood render the conducting tracts of plants highly
inefficient if regarded as a system for conveying water
urged upwards by pressure or drawn upwards in the
substance of the woody walls. The distribution of living
cells in these tracts is such that their actions cannot
account for the rise of water observed, and there is no
reason to believe that the elimination of these activities,
if attended by no secondary changes in the conducting
tracts or transpiring leaves, will arrest the transpiration
stream.

While thus structural and physiological evidence pre-
vent us from accepting any of the previous physical or
vital theories, the same Configuration, physical properties,
and structure of the wood compel us to admit that the
water in the conducting tracts, when not acted upon by
a vis a tergo, must pass into a state of tension. This state
is necessitated by the physical properties of water when
contained in a completely wetted, rigid and permeable
substance which is divided into compartments. There-
fore when root pressure is not acting and when the leaves
of trees are transpiring, the cohesion of their sap ex-
plains fully the transmission of the tension downwards,
and consequently explains the rise of the sap.

Resistance to a current of water moving through wood
at the velocity of the transpiration stream is approxi-
mately equivalent to a head of water equal in length to

xi ENERGY AVAILABLE FOR RAISING SAP 211

the wood traversed. Hence the tension applied to the
upper end of the water columns, which will be able to
raise the transpiration stream in a tree, must equal the
pressure produced by a head of water twice the height
of the tree. In a tree 100 m. high, therefore a tension
of 20 atm. must be produced.

The cohesion of sap amounting, as it does, to at least
200 atm. is in no way taxed by this tension.

The transpiring cells of the mesophyll normally remain
turgid during transpiration ; accordingly we would expect,
if our line of reasoning is correct, that in high trees the
osmotic pressure keeping them distended must correspond
in magnitude to the tensions necessary to raise the sap.

This surmise has been confirmed by determinations of
the osmotic pressures of the saps of various leaves. These
pressures have always been found adequate to resist the
transpiration tension ; but in many cases other factors
enter in, and the pressures developed are much in excess
of those demanded by transpiration.

Finally, it has been shown that the stored energy set
free by respiration in leaves is quite sufficient to do the
work of secretion against the resistance of the transpira-
tion stream ; while, when the vapour pressure of water
in the surrounding space is low, and when evaporation is
doing the work of raising the sap, the expenditure of
energy in this process will reduce the quantity of water
evaporated only by an imperceptible amount.

Literature.

Aubert, E., " Recherches sur la respiration et l'assimilation des plantes
grasses," Revue general de Botanique, 1892, Tom. 4, p. 373.

Brown, H. T., and Escombe, F., "On the Physiological Processes of
Green Leaves," Proc. Roy. Sac. London, 1905, vol. 76 B, p. 29.

Dixon, H. H., "On the Physics of the Transpiration Current," Notes from
the Botanical School of Trinity College, Dublin, 1897, vol. 1, p. 57.

Id. "Transpiration and the Ascent of Sap," Progressus Rei Botanicae,
1909, Bd. iii, s. 1.

INDEX

Ace and osmotic pressure, 197
Air dissolved in tensile water, 103 ff
Air in conducting tracts, 91 ff
Air-pressure balancingosmotic pressure,

153
Air-pressure in tracheae, 48, 49
Ambronn, 198

Anaesthetics and transpiration, 9 ff
Andre, 178, 186, 188
Atkins, 52, 59, 188
Aubert, 202, 211

Beckmann, 155

Berthelot, 103, 114, 117

Biot, 50, 79

Boehm, 28, 29, 45, 95, 99

Bordered pits, Mechanism of, 98

Boucherie, 49, 79

Brown and Escombe, 2, 25, 201, 211

Bubbles in tensile water, 87, 90, 91, 92,

93
Bubbles in trachea?, 91

Capillarity and transpiration, 4 ff

91, 203 ff
Carbon dioxide, Effect of, on leaves,
144 ff
Influence of, on osmotic pressure,

150
Influence of, on transpiration, 11 ff
Cell-wall, Breaking strength of, 198

Menisci in, 4, 203, 204, 205
Centrifuge for extracting wood-sap, 44,

57
Chloroform, Influence of, on transpira-
tion, 11 ff
Clogging of cut surface, Elimination of,
134

falsities resistance, 125, 126

Cohesion of soap film, 106

of water, 84 ft', 101 ff

containing dissolved air, 103 ff

Theory, 87 ff
Collapse of protoxylem, 97
Colocasiaantiquorum, Secretion by, 8, 9
Concentration of sap by pressure, 178 ff

183
Concentration of wood -sap, 39, 45, 58
Condensation of water on roots, 204
Conducting tracts, Changes in, due to
heat, 54 ff
Structure of, 51, 52
Conductivity of sap, Electrical, 178,

179, 185
Contamination of sap due to heat, 55 ff

removed, 65 ff
Copeland's suction theory, 28, 45
Cotter, 106, 108
Cryoscopy, Beckmann's method, 155

Thermo-electric method of, 156 ff

Darwin, 123, 138

Dead stems, Rise of water through, 50 ff
Transmission of water through,
53 ff
Diffusion through stomata, 2, 3
Dissolved air in tensile water, 103 ff
Dixon, 25, 45, 79, 99, 100, 114, 138,

154, 173, 200, 211
Dixon and Atkins. 174, 188, 189
Dixon and Joly, 26, 45, 114, 154
Donny, 101, 114
Drop experiment, 49

Electrical conductivity of sap, 178,
179, 185
of secreted fluid, 9
Theory, 83

213

214

INDEX

Electrolytes in sap, 178, 186
Elfving, 29, 45, 95, 97
Energy absorbed by leaf, 201

available for raising sap, 201 ff

secretion, 201

relations of leaves and roots, 204

set free by respiration, 202
Errera, 29, 46
Ether, Influence of, on transpiration,

9ff
Evaporation and tension in sap, 209

and transpiration compared, 1

engine, 206

Function of, in transpiration, 4 ft'
in raising sap, 25, 203
Evaporation into various gases, 15 ft

Raising sap, 25, 203
Ewart, 52, 69, 70, 79, 94, 100, 114, 116,

118, 119, 122, 125, 127, 131, 138
Extraction, see sap-extraction

Fading of leaves indicating a change
in water tracts, 54, 56, 57, 62 ff

Freezing point and osmotic pressure,
relation of, 155, 190.

Gas generated in lumina, 37

pressure balanced against osmotic
pressure, 142 ff

Pressure theory, 28
Gases, Transpiration into various, 9 ft'
Gelatine, Penetration of, into walls,

29 ff

Plugging of lumina by, 29 ff
Godlewski, 47, 79
Gravitational theory, 81

Hales, 27, 46

Hamburger, 174

Hartig, 28, 46, 48, 49, 79, 93, 100

Heat causing changes in conducting

tracts, 54 ff, 62 ff
Heat causing changes in sap of stems,
53 ff, 62 ff

used for extraction of sap of leaves,
177, 186
Henslow, 16, 26

Ice in lumina, effect of, 38 ff
Illumination and osmotic pressure, 196
Imbibition, Function of, in transpira-
tion, 4 ff
Theory, 29 ff 84

Jamin's Chain, 28

Janse, 48, 51, 52, 54, 79, 102, 117, 118

Johonnot, 106, 114

Joly, 81, 87, 100, 119, 206

Kammerling, 90, 100
Key, Reversing, for thermocouples,
160

Laplace, 101, 115
Larmor, 2, 44, 46

Leaf-cells, Osmotic pressure of, 139
Pressure and tension in, 140
Tensile strength of walls of,
198 ff
Liquid air for sap-extraction, 181 ff
Lumina blocked with gas, 37
gelatine, 29 ff
ice, 38 ff
paraffin, 32 ff
water vapour, 41 ff
of cells and Tracheae compared, 52
Transmission in, 29 ff, 53

Marie and Gatin, 178, 189
Maximow, 189
Medullary rays, 47, 48, 51
Model of transpiring plant, 204

Osmosis, Function of, in transpiration,

4ff
Osmotic pressure and freezing point,
155, 190

in plant organs, 192
of leaf cells, 4, 139

a measure of tension in

tracheae, 141, 193
and age, 197
height, 195
illumination, 196
resistance, 195
tenacity of cellulose, 198
Effect of carbon dioxide on,

150
measured by external gas

pressure, 142 ff
measured by freezing point,
190 ff

plasmolysis, 142
Variations in, 194
Osmotic theory of Larmor, 44
Overton, 80
Oxydase in sap, 58

Oxygen, Influence of, on transpiration,
9 ff

Pappenheim, 98, 100
Paraffin casts of trachea?, 35

for plugging trachea?, 32 ff
Pfeffer, 198

Physical Theories, 27 ff, 81 ff
Pits, Bordered, Mechanism of, 98
Plasmolytic method for measuring

osmotic pressure, 142

INDEX

215

Poisonous substances in conducting
tracts after heating, 56 ff, 60 ff

Poynting and Thomson, 115

Pressure concentrates sap of tissues,
178 ff
for sap extraction, 175
Velocity of flow in stems, proportional
to, 135 ff

Protoplasm, Semi-permeability of, 4,
179 ff

Protoplasmic streaming and transmis-
sion of water in stems, 52, 53

Protoxylem, Collapse of, 97

Quincke's Theory, 83

Secretion and transpiration, 7 ff

of water, 8 ft'

raising sap, 201, 210
Semipermeability of protoplasm, 4,

179
Soap-film, cohesion of, 106
Stephan, 2
Stomata, area of, 1

Diffusion through, 2, 3
Strasburger, 6, 26, 29, 46, 50, 80, 81,

93, 94, 95, 100, 117, 138
Structure of wood and cohesion theory,

91 ff, 210

vital theory, 51, 210
Subdivision of conducting tracts, 91 ff,

210

RENNER, 194, 200

Resistance and osmotic pressure, 195
of walls to flow, 42, 95
of water-tracts, 116 ff', 124 ff, 195
Effect of temperature on, 70
Ewart's estimates of, 116 ff
falsified by clogging, 125, 126
Respiration, Energy set free by, 202
necessary for transpiration under
water, 24
Root-Pressure, Function of, 95
Roots, Condensation of water on, 204
Osmotic pressures of, 45, 182, 184,
192, 193

Sachs, von, 29, 46, 49, 83, 100
Sap-extraction, 175 ff
by centrifuge, 44, 57
chloroform, 176, 186
desiccation, 177
heat, 177, 186
liquid air, 181 ft'
pressure, 175
toluene, 180
Sap of conducting tracts changed by
heat, 55 ff

concentration of, 39, 45, 58
contaminated by heat, 56 ft"
contamination of, removed,

65 ff
extracted by centrifuge, 44, 57
oxydase in, 58
raised by evaporation, 203

secretion, 7 ff, 210
sugars in, 59

tensile strength of, 110 ff.
tissues, concentrated by pressure,

178 ff
leaves, osmotic pressure of, 139,
190 ff
Schultz, 50, 80
Schwendener, 49, 80. 93, 198

Tensile Film Theory, 83
strength of sap, 110 ff

water, 103 ff
water, Bubbles in, 87, 89
rupture of, 86, 113
Tension in sap, measured by osmotic
pressure in leaf-cells, 141, 193
by flow, 194

retards transpiration, 209
Tension theory, 87 ff
Thermo-couples, Calibration of, 169
Change in constant of, 171
Construction of, 157
Use in Cryoscopy of,- 156 ff
Thermo-electric method of cryoscopy,

156 ff
Trachea?, Contents of, 91 ff'

plugged by effects of heating, 54,

60
Structure of, 96 ff

without semi-permeable membrane,
4, 44
Transmission of water in walls,
29 ff

through dead stems, 53 ff
stems as vapour, 42

proportional to pressure, 135
Transpiring plant, model of, 204
Transpiration, a physical phenomenon,
6
and anaesthetics, 9 ff

evaporation, compared, 1
secretion, 7 ff.
turgor, 6
controlled by supply, 122
-current, velocity of, 52, 132
Function of evaporation, osmosis and

imbibition in, 3 ff
Function of living cells in, 7 ff, 201
influenced b} 7 solutes, 5
into saturated spaces, 17 ff
into various gases, 9 ff
under water, 23

2l6

INDEX

Unger, 1, 26

Ursprung, 54, 5C, 80, 115

Vapour blocking lumina, 41 ff
Influence on transpiration of, 9 ff
Water transmitted as, 42 ff
Vesque, 29, 46, 54, 56, 80
Vital actions in transpiration, 7 ff

lifting water in stems, Assumed,

47 tf
lifting water in stems, looked for,
69 ff
Vital theories, 47, 54

Evidence from structure regarding,
51

Walls of cells, Breaking strength of,
198 ff

of tracheae, Resistance of, to moving
water, 95, 96

of trachea, thickenings on, 96
Weber, 54, 55, 80
Weinzierl, 198
Weslermaier, 48, 80
Wilting of leaves indicating changes in

water tracts, 54, 56, 57, 62 ff
Wolff, 27
Wood-sap, Concentration of, 39, 45, 58

extracted by centrifuge, 44, 57
Wood, Section of elements in, 52
Worthington, 115

R. CLAT AND SONS, LTD., BRUNSWICK ST., STAMFORD ST., S.E., AND BUNGAY, SUFFOLK.

Macmillan's Science Monographs

' I V HE aim of these volumes is to provide a medium through
*- which investigators who have made substantial contributions
to the advance of Science in particular directions may bring
together their results and conclusions and discuss them in connexion
with the related work of others.

Scientific research of to-day is essentially specialised ; and
though investigators may make excursions into various parts of
the field of natural knowledge, their names are usually associated
most closely with studies of specific areas. Surveys thus carried
on are described in papers presented to Scientific Societies during
a period which may extend over a number of years ; so that the
student or the original researcher who desires to know the
position of a subject as represented by the work of a leading
authority upon it, has to refer to many volumes of Transactions
or Proceedings of possibly different Scientific Societies. Occasionally,
in an address or an article in a scientific journal, an investigator
gives a sketch of the outstanding points of his studies of a subject;
but the limitations of space prevent him from doing justice to
himself or his work.

Macmillan's Science Monographs afford to authorities upon
definite aspects of science a means by which an adequate statement
of their work may be made available to the scientific world
within a volume of reasonable dimensions and at a moderate
price. The monographs are not intended to be exhaustive records
of all the researches that have been carried out in particular
subjects, but the expression of the original work of the individual
authors, with such consideration of related contributions by others
as is demanded of a scientific publication. Each volume is,
therefore, unique ; and the series forms a collection of authoritative
works which claims a place in every scientific library and is of
prime value both to the student and the original investigator.

i

Macmillan's Science Monographs

8vo.

VOLUMES ALREADY PUBLISHED

CRYSTALLINE STRUCTURE AND CHEMICAL

CONSTITUTION. By A. E. H. TUTTON, D.Sc, M.A. (Oxon.),
F.R.S., A.R.C.Sc. (Lond.), President of the Mineralogical Society; Member
of the Councils of the Chemical Society and the British Association for the
Advancement of Science. 5.?. net.

STABILITY IN AVIATION. An Introduction to

Dynamical Stability as applied to the Motions of Aeroplanes. By
G. H. BRYAN, Sc.D., F.R.S., Professor of Pure and Applied Mathematics
in the University College of North Wales, formerly Fellow of Peterhouse,
Cambridge. $s. net.

STUDIES IN TERRESTRIAL MAGNETISM.

By C. CHREE, M.A., F.R.S., Sc.D. (Camb.), LL.D. (Aberdeen), Super-
intendent of Kew Observatory ; Late Fellow of King's College, Cambridge ;
Ex-President Physical Society of London. 55. net.

STUDIES IN RADIOACTIVITY. By W. H. BRAGG,

M.A., F.R.S., Cavendish Professor of Physics in the University of Leeds;
formerly Elder Professor of Mathematics and Physics in the University of

Adelaide. $s. net.

THE COTTON PLANT IN EGYPT : STUDIES

IN PHYSIOLOGY AND GENETICS. By W. LAWRENCE BALLS,

M.A., Fellow of St. John's College, Cambridge ; Membre de l'Institut Egyptien ;
Botanist to the Department of Agriculture, Egyptian Government. $s. net.

RESEARCHES IN MAGNETO - OPTICS. With

Special Reference to the Magnetic Resolution of Spectrum Lines. By
P. ZEEMAN, Sc.D., Ph.D., D.Sc. Nobel Laureate ; Professor of Experimental
Physics in the University of Amsterdam. 6s. net.

STUDIES IN WATER SUPPLY. By A. C.

HOUSTON, D.Sc, M.B., CM., Director of Water Examination, Metro-
politan Water Poard. $s. net.

STELLAR MOVEMENTS AND THE STRUC-
TURE OF THE UNIVERSE. By A. S. EDDINGTON, M.A. (Cantab.),
M.Sc (Manchester), B.Sc (Lond.), F.R.S., Plumian Professor of Astronomy,
University of Cambridge. 6s. net.

TRANSPIRATION AND THE ASCENT OF SAP

IN PLANTS. By HENRY H. DIXON, Sc.D., F.R.S., University
Professor of Botany in Trinity College, Dublin ; Director of Trinity College
Botanic Gardens.

IN THE PRESS
THE MUTATION FACTOR IN EVOLUTION;

WITH PARTICULAR REFERENCE TO CENOTHERA. By Dr. R. R.
GATES, Lecturer in Biology, St. Thomas's Hospital.

LONDON : MACMILLAN AND CO., LTD.