Another fact which seems very difficult to reconcile with the

hypothesis of an “ideal typical vertebra,” is the not infrequent

presence of some of the typical elements in duplicate. Not only, as

we have seen, may they severally be absent, but they may severally be

present in greater number than they should be. When we see, in the

ideal diagram, one centrum, two neurapophyses, two pleurapophyses,

two hæmapophyses, one neural spine, and one hæmal spine, we naturally

expect to find them always bearing to each other these numerical

relations. Though we may not be greatly surprised by the absence

of some of them, we are hardly prepared to find others multiplied.

Yet such cases are common. Thus the neural spine “is double in the

anterior vertebræ of some fishes” (p. 98). Again, in the abdominal

region of extinct saurians, and in crocodiles, “the freely-suspended

hæmapophyses are compounded of two or more overlapping bony pieces”

(p. 100). Yet again, at p. 99, we read--“I have observed some of the

expanded pleurapophyses in the great _Testudo elephantopus_ ossified

from two centres, and the resulting divisions continuing distinct, but

united by suture.” Once more “the neurapophyses, which do not advance

beyond the cartilaginous stage in the sturgeon, consist in that fish

of two distinct pieces of cartilage; and the anterior pleurapophyses

also consist of two or more cartilages, set end on end” (p. 91). And

elsewhere referring to this structure, he says:--

“Vegetative repetition of perivertebral parts not only manifests

itself in the composite neurapophyses and pleurapophyses, but

in a small accessory (interneural) cartilage, at the fore and

back part of the base of the neurapophysis; and by a similar

(interhæmal) one at the fore and back part of most of the

parapophyses” (p. 87).

Thus the neural and hæmal spines, the neurapophyses, the

pleurapophyses, the hæmapophyses, may severally consist of two or more

pieces. This is not all: the like is true even of the centrums.

“In _Heptanchus_ (_Squalus cinereus_) the vertebral centres are

feebly and vegetatively marked out by numerous slender rings

of hard cartilage in the notochordal capsule, the number of

vertebræ being more definitely indicated by the neurapophyses

and parapophyses.... In the piked dog-fish (_Acanthias_) and the

spotted dog-fish (_Scyllium_) the vertebral centres coincide in

number with the neural arches” (p. 87).

Is it not strange that the pattern-vertebra should be so little adhered

to, that each of its single typical pieces may be transformed into two

or three?

But there are still more startling departures from the alleged type.

The numerical relations of the elements vary not only in this way, but

in the opposite way. A given part may be present not only in greater

number than it should be, but also in less. In the tails of homocercal

fishes, the centrums “are rendered by centripetal shortening and bony

confluence fewer in number than the persistent, neural, and hæmal

arches of that part”--that is, there is only a fraction of a centrum to

each vertebra. Nay, even this is not the most heteroclite structure.

Paradoxical as it may seem, there are cases in which the same vertebral

element is, considered under different aspects, at once in excess and

defect. Speaking of the hæmal spine, Professor Owen says:--

“The horizontal extension of this vertebral element is

sometimes accompanied by a median division, or in other

words, it is ossified from two lateral centres; this is seen

in the development of parts of the human sternum; the same

vegetative character is constant in the broader thoracic

hæmal spines of birds; though, sometimes, as _e.g._, in the

struthionidæ, _ossification extends from the same lateral centre

lengthwise--i.e., forwards and backwards, calcifying the connate

cartilaginous homologues of halves of four or five hæmal spines,

before these finally coalesce with their fellows at the median

line_” (p. 101).

So that the sternum of the ostrich, which according to the hypothesis,

should, in its cartilaginous stage, have consisted of _four or five

transverse_ pieces, answering to the vertebral segments, and should

have been ossified from four or five centres, one to each cartilaginous

piece, shows not a trace of this structure; but instead, consists

of _two longitudinal_ pieces of cartilage, each ossified from one

centre, and finally coalescing on the median line. These four or five

hæmal spines have at the same time doubled their individualities

transversely, and entirely lost them longitudinally!

* * * * *

There still remains to be considered the test of relative position.

It might be held that, spite of all the foregoing anomalies, if the

typical parts of the vertebræ always stood towards each other in the

same relations--always preserved the same connexions, something like a

case would be made out. Doubtless, relative position is an important

point; and it is one on which Professor Owen manifestly places great

dependence. In his discussion of “moot cases of special homology,” it

is the general test to which he appeals. The typical natures of the

alisphenoid, the mastoid, the orbito-sphenoid, the prefrontal, the

malar, the squamosal, &c. he determines almost wholly by reference to

the adjacent nerve-perforations and the articulations with neighbouring

bones (see pp. 19 to 72): the general form of the argument being--This

bone is to be classed as such or such, _because_ it is connected

thus and thus with these others, which are so and so. Moreover, by

putting forth an “ideal typical vertebra,” consisting of a number of

elements standing towards each other in certain definite arrangement,

this persistency of relative position is manifestly alleged. The

essential attribute of this group of bones, considered as a typical

group, is the constancy in the connexions of its parts: change the

connexions, and the type is changed. But the constancy of relative

position thus tacitly asserted, and appealed to as a conclusive test

in “moot cases of special homology,” is clearly negatived by Professor

Owen’s own facts. For instance, in the “ideal typical vertebra,” the

hæmal arch is represented as formed by the two hæmapophyses and the

hæmal spine; but at p. 91 we are told that

“The contracted hæmal arch in the caudal region of the body may

be formed by different elements of the typical vertebra: _e.g._,

by the parapophyses (fishes generally); by the pleurapophyses

(lepidosiren); by both parapophyses and pleurapophyses (_Sudis_,

_Lepidosteus_), and by hæmapophyses, shortened and directly

articulated with the centrums (reptiles and mammals).”

And further, in the thorax of reptiles, birds, and mammals, “the

hæmapophyses are removed from the centrum, and are articulated to

the distal ends of the pleurapophyses; the bony hoop being completed

by the intercalation of the hæmal spine” (p. 82). So that there are

_five_ different ways in which the hæmal arch may be formed--four

modes of attachment of the parts different from that shown in the

typical diagram! Nor is this all. The pleurapophyses “may be quite

detached from their proper segment, and suspended to the hæmal arch of

another vertebra;” as we have already seen, the entire hæmal arch may

be detached and removed to a distance, sometimes reaching the length

of twenty-seven vertebræ; and, even more remarkable, the ventral fins

of some fishes, which theoretically belong to the pelvic arch, are so

much advanced forward as to be articulated to the scapular arch--“the

ischium elongating to join the coracoid.” With these admissions it

seems to us that relative position and connexions cannot be appealed to

as tests of homology, nor as evidence of any original type of vertebra.

In no class of facts, then, do we find a good foundation for the

hypothesis of an “ideal typical vertebra.” There is no one conceivable

attribute of this archetypal form which is habitually realised by

actual vertebræ. The alleged group of true vertebral elements is not

distinguished in any specified way from bones not included in it. Its

members have various degrees of inconstancy; are rarely all present

together; and no one of them is essential. They are severally developed

in no uniform way: each of them may arise either out of a separate

piece of cartilage, or out of a piece continuous with that of some

other element; and each may be ossified from many independent points,

from one, or from none. Not only may their respective individualities

be lost by absence, or by confluence with others; but they may be

doubled, or tripled, or halved, or may be multiplied in one direction

and lost in another. The entire group of typical elements may coalesce

into one simple bone representing the whole vertebra; and even, as in

the terminal piece of a bird’s tail, half-a-dozen vertebræ, with all

their many elements, may become entirely lost in a single mass. Lastly,

the respective elements, when present, have no fixity of relative

position: sundry of them are found articulated to various others than

those with which they are typically connected; they are frequently

displaced and attached to neighbouring vertebræ; and they are even

removed to quite remote parts of the skeleton. It seems to us that if

this want of congruity with the facts does not disprove the hypothesis,

no such hypothesis admits of disproof.

* * * * *

Unsatisfactory as is the evidence in the case of the trunk and tail

vertebræ, to which we have hitherto confined ourselves, it is far worse

in the case of the alleged cranial vertebræ. The mere fact that those

who have contended for the vertebrate structure of the skull, have

differed so astonishingly in their special interpretations of it, is

enough to warrant great doubt as to the general truth of their theory.

From Professor Owen’s history of the doctrine of general homology,

we gather that Duméril wrote upon “la tête considérée comme _une_

vertèbre;” that Kielmeyer, “instead of calling the skull a vertebra,

said each vertebra might be called a skull;” that Oken recognized in

the skull _three_ vertebræ and a rudiment; that Professor Owen himself

makes out _four_ vertebræ; that Goethe’s idea, adopted and developed

by Carus, was, that the skull is composed of _six_ vertebræ; and that

Geoffrey St. Hilaire divided it into _seven_. Does not the fact that

different comparative anatomists have arranged the same group of bones

into _one_, _three_, _four_, _six_, and _seven_ vertebral segments,

show that the mode of determination is arbitrary, and the conclusions

arrived at fanciful? May we not properly entertain great doubts as to

any one scheme being more valid than the others? And if out of these

conflicting schemes we are asked to accept one, ought we not to accept

it only on the production of some thoroughly conclusive proof--some

rigorous test showing irrefragably that the others must be wrong and

this alone right? Evidently where such contradictory opinions have been

formed by so many competent judges, we ought, before deciding in favour

of one of them, to have a clearness of demonstration much exceeding

that required in any ordinary case. Let us see whether Professor Owen

supplies us with any such clearness of demonstration.

To bring the first or occipital segment of the skull into

correspondence with the “ideal typical vertebra,” Professor Owen

argues, in the case of the fish, that the parapophyses are _displaced_,

and wedged between the neurapophyses and the neural spine--removed

from the hæmal arch and built into the upper part of the neural arch.

Further, he considers that the pleurapophyses are _teleologically

compound_. And then, in all the higher vertebrata, he alleges that

the hæmal arch is _separated_ from its centrum, taken to a distance,

and transformed into the scapular arch. Add to which, he says that

in mammals the displaced parapophyses are mere processes of the

neurapophyses (p. 133): these vertebral elements, typically belonging

to the lower part of the centrum, and in nearly all cases confluent

with it, are not only removed to the far ends of elements placed above

the centrum, but have become exogenous parts of them!

Conformity of the second or parietal segment of the cranium with the

pattern-vertebra, is produced thus:--The petrosals are _excluded_

as being partially-ossified sense-capsules, not forming parts of the

true vertebral system, but belonging to the “splanchno-skeleton.” A

centrum is _artificially_ obtained by sawing in two the bone which

serves in common as centrum to this and the preceding segment; and

this though it is admitted that in fishes, where their individualities

ought to be best seen, these two hypothetical centrums are not simply

coalescent, but connate. Next, a similar _arbitrary bisection_ is

made of certain elements of the hæmal arches. And then, “the principle

of _vegetative repetition_ is still more manifest in this arch

than in the occipital one:” each pleurapophysis is double; each

hæmapophysis is double; and the hæmal spine consists of six pieces!

The interpretation of the third and fourth segments being of the same

general character, need not be detailed. The only point calling for

remark being, that in addition to the above various modes of getting

over anomalies, we find certain bones referred to the _dermo-skeleton_.

Now it seems to us, that even supposing no antagonist interpretations

had been given, an hypothesis reconcilable with the facts only by

the aid of so many questionable devices, could not be considered

satisfactory; and that when, as in this case, various comparative

anatomists have contended for other interpretations, the character

of this one is certainly not of a kind to warrant the rejection of

the others in its favour; but rather of a kind to make us doubt

the possibility of all such interpretations. The question which

naturally arises is, whether by proceeding after this fashion,

groups of bones might not be arranged into endless typical forms.

If, when a given element was not in its place, we were at liberty

to consider it as _suppressed_, or _connate_ with some neighbouring

element, or _removed_ to some more or less distant position;--if,

on finding a bone in excess, we might consider it, now as part of

the _dermo-skeleton_, now as part of the _splanchno-skeleton_, now

as _transplanted_ from its typical position, now as resulting from

_vegetative repetition_, and now as a bone _teleologically compound_

(for these last two are intrinsically different, though often used by

Professor Owen as equivalents);--if, in other cases, a bone might be

regarded as _spurious_ (p. 91), or again as having _usurped_ the place

of another;--if, we say, these various liberties were allowed us, we

should not despair of reconciling the facts with various diagrammatic

types besides that adopted by Professor Owen.

When, in 1851, we attended a course of Professor Owen’s lectures

on Comparative Osteology, beginning though we did in the attitude

of discipleship, our scepticism grew as we listened, and reached

its climax when we came to the skull; the reduction of which to the

vertebrate structure, reminded us very much of the interpretation of

prophecy. The delivery, at the Royal Society, of the Croonian Lecture

for 1858, in which Professor Huxley, confirming the statements of

several German anatomists, has shown that the facts of embryology do

not countenance Professor Owen’s views respecting the formation of the

cranium, has induced us to reconsider the vertebral theory as a whole.

Closer examination of Professor Owen’s doctrines, as set forth in his

works, has certainly not removed the scepticism generated years ago

by his lectures. On the contrary, that scepticism has deepened into

disbelief. And we venture to think that the evidence above cited shows

this disbelief to be warranted.

* * * * *

There remains the question--What general views are we to take

respecting the vertebrate structure? If the hypothesis of an “ideal

typical vertebra” is not justified by the facts, how are we to

understand that degree of similarity which vertebræ display?

We believe the explanation is not far to seek. All that our space will

here allow, is a brief indication of what seems to us the natural view

of the matter.

Professor Owen, in common with other comparative anatomists, regards

the divergences of individual vertebræ from the average form, as due

to adaptive modifications. If here one vertebral element is largely

developed, while elsewhere it is small--if now the form, now the

position, now the degree of coalescence, of a given part varies; it

is that the local requirements have involved this change. The entire

teaching of comparative osteology implies that differences in the

conditions of the respective vertebræ necessitate differences in their

structures.

Now, it seems to us that the first step towards a right conception

of the phenomena, is to recognize this general law in its converse

application. If vertebræ are unlike in proportion to the unlikeness

of their circumstances, then, by implication, they will be like in

proportion to the likeness of their circumstances. While successive

segments of the same skeleton, and of different skeletons, are all in

some respects more or less differently acted on by incident forces,

and are therefore required to be more or less different; they are all,

in other respects, similarly acted on by incident forces, and are

therefore required to be more or less similar. It is impossible to

deny that if differences in the mechanical functions of the vertebræ

involve differences in their forms; then, community in their mechanical

functions, must involve community in their forms. And as we know that

throughout the _Vertebrata_ generally, and in each vertebrate animal,

the vertebræ, amid all their varying circumstances, _have_ a certain

community of function, it follows necessarily that they will have a

certain general resemblance--there will recur that average shape which

has suggested the notion of a pattern-vertebra.

A glance at the facts at once shows their harmony with this conclusion.

In an eel or a snake, where the bodily actions are such as to involve

great homogeneity in the mechanical conditions of the vertebræ, the

series of them is comparatively homogeneous. On the contrary, in a

mammal or a bird, where there is considerable heterogeneity in their

circumstances, their similarity is no longer so great. And if, instead

of comparing the vertebral columns of different animals, we compare

the successive vertebræ of any one animal, we recognize the same law.

In the segments of an individual spine, where is there the greatest

divergence from the common mechanical conditions? and where may we

therefore expect to find the widest departure from the average form?

Obviously at the two extremities. And accordingly it is at the two

extremities that the ordinary structure is lost.

Still clearer becomes the truth of this view, when we consider the

genesis of the vertebral column as displayed throughout the ascending

grades of the _Vertebrata_. In its first embryonic stage, the spine

is an undivided column of flexible substance. In the early fishes,

while some of the peripheral elements of the vertebræ were marked out,

the central axis was still a continuous unossified cord. And thus we

have good reason for thinking that in the primitive vertebrate animal,

as in the existing _Amphioxus_, the notochord was persistent. The

production of a higher, more powerful, more active creature of the same

type, by whatever method it is conceived to have taken place, involved

a change in the notochordal structure. Greater muscular endowments

presupposed a firmer internal fulcrum --a less yielding central axis.

On the other hand, for the central axis to have become firmer while

remaining continuous, would have entailed a stiffness incompatible

with the creature’s movements. Hence, increasing density of the

central axis necessarily went hand in hand with its segmentation: for

strength, ossification was required; for flexibility, division into

parts. The production of vertebræ resulting thus, there obviously

would arise among them a general likeness, due to the similarity in

their mechanical conditions, and more especially the muscular forces

bearing on them. And then observe, lastly, that where, as in the head,

the terminal position and the less space for development of muscles,

entailed smaller lateral bendings, the segmentation would naturally

be less decided, less regular, and would be lost as we approached the

front of the head.

But, it may be replied, this hypothesis does not explain all the

facts. It does not tell us why a bone whose function in a given animal

requires it to be solid, is formed not of a single piece, but by the

coalescence of several pieces, which in other creatures are separate;

it does not account for the frequent manifestations of unity of plan

in defiance of teleological requirements. This is quite true. But it

is not true, as Professor Owen argues respecting such cases, that

“if the principle of special adaptation fails to explain them, and

we reject the idea that these correspondences are manifestations of

some archetypal exemplar, on which it has pleased the Creator to frame

certain of his living creatures, there remains only the alternative

that the organic atoms have concurred fortuitously to produce such

harmony.” This is not the only alternative: there is another, which

Professor Owen has overlooked. It is a perfectly tenable supposition

that all higher vertebrate forms have arisen by _the superposing of

adaptations upon adaptations_. Either of the two antagonist cosmogonies

consists with this supposition. If, on the one hand, we conceive

species to have resulted from acts of special creation; then it is

quite a fair assumption that to produce a higher vertebrate animal,

the Creator did not begin afresh, but took a lower vertebrate animal,

and so far modified its pre-existing parts as to fit them for the

new requirements; in which case the original structure would show

itself through the superposed modifications. If, on the other hand, we

conceive species to have resulted by gradual differentiations under the

influence of changed conditions; then, it would manifestly follow that

the higher, heterogeneous forms, would bear traces of the lower and

more homogeneous forms from which they were evolved.

Thus, besides finding that the hypothesis of an “ideal typical

vertebra” is irreconcilable with the facts, we find that the facts are

interpretable without gratuitous assumptions. The average community of

form which vertebræ display, is explicable as resulting from natural

causes. And those typical similarities which are traceable under

adaptive modifications, must obviously exist if, throughout creation in

general, there has gone on that continuous superposing of modifications

upon modifications which goes on in every unfolding organism.

* * * * *

[I might with propriety have added to the foregoing criticisms, the

remark that Professor Owen has indirectly conferred a great benefit by

the elaborate investigations he has made with the view of establishing

his hypothesis. He has himself very conclusively proved that the

teleological interpretation is quite irreconcilable with the facts.

In gathering together evidence in support of his own conception of

archetypal forms, he has disclosed adverse evidence which I think shows

his conception to be untenable. The result is that the field is left

clear for the hypothesis of Evolution as the only tenable one.]

APPENDIX C.

[_From the_ TRANSACTIONS OF THE LINNEAN SOCIETY, VOL. XXV.]

XV. _On Circulation and the Formation of Wood in Plants. By_

HERBERT SPENCER, _Esq. Communicated by_ GEORGE BUSK, _Esq.,

F.R.S., Sec. L.S._

Read March 1st, 1866.

Opinions respecting the functions of the vascular tissues in plants

appear to make but little progress towards agreement. The supposition

that these vessels and strings of partially-united cells, lined with

spiral, annular, reticulated, or other frameworks, are carriers of the

plant-juices, is objected to on the ground that they often contain

air: as the presence of air arrests the movement of blood through

arteries and veins, its presence in the ducts of stems and petioles

is assumed to unfit them as channels for sap. On the other hand, that

these structures have a respiratory office, as some have thought, is

certainly not more tenable, since, if the presence of air in them

negatives the belief that their function is to distribute liquid, the

presence of liquid in them equally negatives the belief that their

function is to distribute air. Nor can any better defence be made for

the hypothesis which I find propounded, that these parts serve “to give

strength to the parenchyma.” Tubes with fenestrated and reticulated

internal skeletons have, indeed, some power of supporting the tissue

through which they pass; but tubes lined with spiral threads can yield

extremely little support, while tubes lined with annuli, or spirals

alternating with annuli, can yield no support whatever. Though all

these types of internal framework are more or less efficient for

preventing closure by lateral pressure, they are some of them quite

useless for holding up the mass through which the vessels pass; and the

best of them are for this purpose mechanically inferior to the simple

cylinder. The same quantity of matter made into a continuous tube would

be more effective in giving stiffness to the cellular tissue around it.

In the absence of any feasible alternative, the hypothesis that these

vessels are distributors of sap claims reconsideration. The objections

are not, I think, so serious as they seem. The habitual presence of

air in the ducts that traverse wood, can scarcely be held anomalous if

when the wood is formed their function ceases. The canals which ramify

through a Stag’s horn, contain air after the Stag’s horn is fully

developed; but it is not thereby rendered doubtful whether it is the

function of arteries to convey blood. Again, that air should frequently

be found even in the vessels of petioles and leaves, will not appear

remarkable when we call to mind the conditions to which a leaf is

subject. Evaporation is going on from it. The thinner liquids pass

by osmose out of the vessels into the tissues containing the liquids

thickened by evaporation. And as the vessels are thus continually

drained, a draught is made upon the liquid contained in the stem and

roots. Suppose that this draught is unusually great, or suppose that

around the roots there exists no adequate supply of moisture. A state

of capillary tension must result--a tendency of the liquid to pass into

the leaves resisted below by liquid cohesion. Now, had the vessels

impermeable coats, only their upper extremities would under these

conditions be slowly emptied. But their coats, in common with all the

surrounding tissues, are permeable by air. Hence, under this state

of capillary tension, air will enter; and as the upper ends of the

tubes, being both smaller in diameter and less porous than the lower,

will retain the liquids with greater tenacity, the air will enter the

wider and more porous tubes below--the ducts of the stem and branches.

Thus the entrance of air no more proves that these ducts are not

sap-carriers, than does the emptiness of tropical river-beds in the dry

season prove that they are not channels for water. There is, however, a

difficulty which seems more serious. It is said that air, when present

in these minute canals, must be a great obstacle to the movement of sap

through them. The investigations of Jamin have shown that bubbles in a

capillary tube resist the passage of liquid, and that their resistance

becomes very great when the bubbles are numerous--reaching, in some

experiments, as much as three atmospheres. Nevertheless the inference

that any such resistance is offered by the air-bubbles in the vessels

of a plant, is, I think, an erroneous one. What happens in a capillary

tube having impervious sides, with which these experiments were made,

will by no means happen in a capillary tube having pervious sides.

Any pressure brought to bear on the column of liquid contained in the

porous duct of a plant, must quickly cause the expulsion of a contained

air-bubble through the minute openings in the coats of the duct. The

greater molecular mobility of gases than liquids, implies that air will

pass out far more readily than sap. Whilst, therefore, a slight tension

on the column of sap will cause it to part and the air to enter, a

slight pressure upon it will force out the air and reunite the divided

parts of the column.

To obtain data for an opinion on this vexed question, I have lately

been experimenting on the absorption of dyes by plants. So far as I can

learn, experiments of this kind have most, if not all of them, been

made on stems, and, as it would seem from the results, on stems so far

developed as to contain all their characteristic structures. The first

experiments I made myself were on such parts, and yielded evidence that

served but little to elucidate matters. It was only after trying like

experiments with leaves of different ages and different characters,

and with undeveloped axes, as well as with axes of special kinds, that

comprehensible results were reached; and it then became manifest that

the appearances presented by ordinary stems when thus tested, are in a

great degree misleading. Let me briefly indicate the differences.

If an adult shoot of a tree or shrub be cut off, and have its lower

end placed in an alumed decoction of logwood or a dilute solution of

magenta,[69] the dye will, in the course of a few hours, ascend to a

distance varying according to the rate of evaporation from the leaves.

On making longitudinal sections of the part traversed by it, the dye is

found to have penetrated extensive tracts of the woody tissue; and on

making transverse sections, the openings of the ducts appear as empty

spaces in the midst of a deeply-coloured prosenchyma. It would thus

seem that the liquid is carried up the denser parts of the vascular

bundles; neglecting the cambium layer, neglecting the central pith,

and neglecting the spiral vessels of the medullary sheath. Apparently

the substance of the wood has afforded the readiest channel. When,

however, we examine these appearances critically, we find reasons for

doubting this conclusion. If a transverse section of the lower part,

into which the dye passed first and has remained longest, be compared

with a transverse section of the part which the dye has but just

reached, a marked difference is visible. In the one case the whole of

the dense tissue is stained; in the other case it is not. This uneven

distribution of stain in the part which the dye has incompletely

permeated is not at random; it admits of definite description. A

tolerably regular continuous ring of colour distinguishes the outer

part of the wood from the inner mass, implying a passage of liquid

up the elongated cells next the cambium layer. And the inner mass is

coloured more round the mouths of the pitted ducts than elsewhere: the

dense tissue is darkest close to the edges of these ducts; the colour

fades away gradually on receding from their edges; there is most colour

where there are several ducts together; and the dense tissue which

is fully dyed for some space, is that which lies between two or more

ducts. These are indications that while the layer of pitted cells

next the cambium has served as a channel for part of the liquid, the

rest has ascended the pitted ducts, and oozed out of these into the

prosenchyma around. And this conclusion is confirmed by the contrast

between the appearances of the lowest part of a shoot under different

conditions. For if, instead of allowing the dye time for oozing through

the prosenchyma, the end of the shoot be just dipped into the dye and

taken out again, we find, on making transverse sections of the part

into which the dye has been rapidly taken up, that, though it has

diffused to some distance round the ducts, it has left tracts of wood

between the ducts uncoloured--a difference which would not exist had

the ascent been through the substance of the wood. Even still stronger

is the confirmation obtained by using one dye after another. If a shoot

that has absorbed magenta for an hour be placed for five minutes in the

logwood decoction, transverse sections of it taken at a short distance

from its end show the mouths of the ducts surrounded by dark stains in

the midst of the much wider red stains.

Based on these comparisons only, the inference pointed out has little

weight; but its weight is increased by the results of experiments on

quite young shoots, and shoots that develope very little wood. The

behaviour of these corresponds perfectly with the expectation that a

liquid will ascend capillary tubes in preference to simple cellular

tissue or tissue not differentiated into continuous canals. The

vascular bundles of the medullary sheath are here the only channels

which the coloured liquid takes. In sections of the parts up to which

the dye has but just reached, the spiral, fenestrated, scalariform,

or other vessels contained in these bundles are alone coloured, and

lower down it is only after some hours that such an exudation of dye

takes place as suffices partially to colour the other substances of the

bundle. Further, it is to be noted that at the terminations of shoots,

where the vessels are but incompletely formed out of irregularly-joined

fibrous cells which still retain their original shapes, the dye runs up

the incipient vessels and does not colour in the smallest degree the

surrounding tissue.

Experiments with leaves bring out parallel facts. On placing in a dye

a petiole of an adult leaf of a tree, and putting it before the fire

to accelerate evaporation, the dye will be found to ascend the mid-rib

and veins at various rates, up even to a foot per hour. At first it

is confined to the vessels; but by the time it has reached the point

of the leaf, it will commonly be seen that at the lower part it has

diffused itself into the sheaths of the vessels. In a quite young leaf

from the same shoot, we find a much more rigorous restriction of the

dye to the vessels. On making oblique sections of its petiole, mid-rib,

and veins, the vessels have the appearance of groups of sharply

defined coloured rods imbedded in the green prosenchyma; and this

marked contrast continues with scarcely an appreciable change after

plenty of time has been allowed for exudation.

The facts thus grouped and thus contrasted seem, at first sight, to

imply that while they are young the coats of these ramifying canals

lined with spiral or allied structures are not readily permeable,

but that, becoming porous as they grow old, they allow the liquids

they carry to escape with increasing facility; and hence a possible

interpretation of the fact that, in the older parts, the staining of

the tissue around the vessels is so rapid as to suggest that the dye

has ascended directly through this tissue, whereas in the younger

parts the reverse appearance necessitates the reverse conclusion. But

now, is this difference determined by difference of age, or is it

otherwise determined? The evidence as presented in ordinary stems and

leaves shows us that the parts of the vascular system at which there

is a rapid escape of dye are not simply older parts, but are parts

where a deposit of woody matter is taking place. Is it, then, that

the increasing permeability of the ducts, instead of being directly

associated with their increasing age, is directly associated with the

increasing deposit of dense substance around them?

To get proof that this last connexion is the true one, we have but to

take a class of cases in which wood is formed only to a small extent.

In such cases experiments show us a far more general and continued

limitation of the dye to the vessels. Ordinary herbs and vegetables,

when contrasted with shrubs and trees, illustrate this; as instance

the petioles of Celery, or of the common Dock, and the leaves of

Cabbages or Turnips. And then in very succulent plants, such as

_Bryophyllum calycinum_, _Kalanchoë rotundifolia_, the various species

of _Crassula_, _Cotyledon_, _Kleinia_, and others of like habit, the

ducts of old and young leaves alike retain the dye very persistently:

the concomitant in these cases being the small amount of prosenchyma

around the ducts, or the small amount of deposit in it, or both. More

conclusive yet is the evidence which meets us when we turn from very

succulent leaves to very succulent axes. The tender young shoots of

_Kleinia ante-euphorbium_, or _Euphorbia Mauritanica_, which for many

inches of their lengths have scarcely any ligneous fibres, show us

scarcely any escape of the coloured liquid from the vessels of the

medullary sheath. So, too, is it with _Stapelia Buffonia_, a plant of

another order, having soft swollen axes. And then we have a repetition

of the like connexion of facts throughout the _Cactaceæ_: the most

succulent showing us the smallest permeability of the vessels. In

two species of _Rhipsalis_, in two species of _Cereus_, and in two

species of _Mammillaria_, which I have tried, I have found this so.

_Mammillaria gracilis_ may be named as exemplifying the relation under

its extreme form. Into one of these small spheroidal masses, the dye

ascends through the large bundles of spiral or annular ducts, or cells

partially united into such ducts, colouring them deeply, and leaving

the feebly-marked sheath of prosenchyma, together with the surrounding

watery cellular tissue, perfectly uncoloured.

The most conclusive evidence, however, is furnished by those _Cactaceæ_

in which the transition from succulent to dense tissue takes place

variably, according as local circumstances determine. _Opuntia_ yields

good examples. If a piece of it including one of the joints at which

wood is beginning to form, be allowed to absorb a coloured liquid,

the liquid, running up the irregular bundles of vessels and into many

of their minute ramifications, is restricted to these where they pass

through the parenchyma forming the mass of the stem; but near the

joints the hardened tissue around the vessels is coloured. In one of

these fleshy growths we get clear evidence that the escape of the

dye has no immediate dependence on the age of the vessels, since, in

parts of the stem that are alike in age, some of the vessels retain

their contents while others do not. Nay, we even find that the younger

vessels are more pervious than the older ones, if round the younger

ones there is a formation of wood.

Thus, then, is confirmed the inference before drawn, that in ordinary

stems the staining of the wood by an ascending coloured liquid is due,

not to the passage of the coloured liquid up the substance of the wood,

but to the permeability of its ducts and such of its pitted cells as

are united into irregular canals. And the facts showing this, at the

same time indicate with tolerable clearness the process by which wood

is formed. What in these cases is seen to take place with a dye, may

be fairly presumed to take place with sap. Where the dye exudes but

slowly, we may infer that the sap exudes but slowly; and it is a fair

inference that where the dye leaks rapidly out of the vessels, the sap

does the same. Inferring, thus, that where-ever there is a considerable

formation of wood there is a considerable escape of the sap, we see

in the one the result of the other. The thickening of the prosenchyma

is proportionate to the quantity of nutritive liquid passing into it;

and this nutritive liquid passes into it from the vessels, ducts, and

irregular canals it surrounds.

But an objection is made to such experiments as the foregoing, and to

all the inferences drawn from them. It is said that portions of plants

cut off and thus treated, have their physiological actions arrested,

or so changed as may render the results misleading; and it is said

that when detached shoots and leaves have their cut ends placed in

solutions, the open mouths of their vessels and ducts are directly

presented with the liquids to be absorbed, which does not happen in

their natural states. Further, making these objections look serious, it

is alleged that when solutions are absorbed through the roots, quite

different results are obtained: the absorbed matters are found in the

tissues and not in the vessels. Clearly, were the experiments yielding

these adverse results conducted in unobjectionable ways, the conclusion

implied by them would negative the conclusions above drawn. But these

experiments are no less objectionable than those to which they are

opposed. Such mineral matters as salts of iron, solutions of which

have in some cases been supplied to the roots for their absorption,

are obviously so unlike the matters ordinarily absorbed, that they

are likely to interfere fatally with the physiological actions. If

experiments of this kind are made by immersing the roots in a dye,

there is, besides the difficulty that the mineral mordant contained

by the dye is injurious to the plant, the further difficulty that the

colouring matter, being seized by the substances for which it has an

affinity, is left behind in the first layers of root-tissues passed

through, and that the decolorized water passing up into the plant is

not traceable. To be conclusive, then, an experiment on absorption

through roots must be made with some solution which will not seriously

interfere with the plant’s vital processes, and which will not have

its distinctive element left behind. To fulfil these requirements I

adopted the following method. Having imbedded a well-soaked broad-bean

in moist sand, contained in an inverted cone of cardboard with its

apex cut off for the radicle to come through--having placed this in

a wide-mouthed dwarf bottle, partly filled with water, so that the

protruding radicle dipped into the water--and having waited until the

young bean had a shoot some three or more inches high, and a cluster

of secondary rootlets from an inch to an inch and a-half long--I

supplied for its absorption a simple decoction of logwood, which,

being a vegetal matter, was not likely to do it much harm, and which,

being without a mordant, would not leave its suspended colour in the

first tissues passed through. To avoid any possible injury, I did not

remove the plant from the bottle, but slightly raising the cone out of

its neck, I poured away the water through the crevice and then poured

in the logwood decoction; so that there could have been no broken end

or abraded surface of a rootlet through which the decoction might

enter. Being prepared with some chloride of tin as a mordant, I cut

off, after some three hours, one of the lowest leaves, expecting that

the application of the mordant to the cut surface would bring out

the characteristic colour if the logwood decoction had risen to that

height. I got no reaction, however. But after eight hours I found, on

cutting off another leaf, that the vessels of its petiole were made

visible as dark streaks by the colour with which they were charged--a

colour differing, as was to be expected, from that of the logwood

decoction, which spontaneously changes even by simple exposure. It was

then too late in the day to pursue the observations; but next morning

the vessels of the whole plant, as far as the petioles of its highest

unfolded leaves, were full of the colouring matter; and on applying

chloride of tin to the cut surfaces, the vessels assumed that purplish

red which this mordant produces when directly mixed with the logwood

decoction. Subsequently, when one of the cotyledons was cut open

by Prof. Oliver, to whom, in company with Dr. Hooker, I showed the

specimen, we found that the whole of its vascular system was filled

with the decoction, which everywhere gave the characteristic reaction.

And it became manifest that the liquid absorbed through the rootlets,

in the central vessels of which it was similarly traceable, had part

of it passed directly up the vessels of the axis, while part of it

had passed through other vessels into the cotyledon, out of which, no

doubt, the liquid ordinarily so carried returns charged with a supply

of the stored nutriment. I have since obtained a verification by

varying the method. Digging up some young plants (Marigolds happened to

afford the best choice) with large masses of soil round them, placing

them in water, so as gradually to detach the soil without injuring the

rootlets, planting them afresh in a flower-pot full of washed sand,

and then, after a few days, watering them with a logwood decoction, I

found, as before, that in less than twenty-four hours the colouring

matter had run up into the vessels of the leaves. Though the reaction

produced by the mordant was not so strong as before, it was marked

enough to be quite unquestionable.

As these experiments were so conducted that there was no access to

the vessels except through the natural channels, and as the vital

actions of the plants were so little interfered with that at the end

of twenty-four hours they showed no traces of disturbance, I think the

results must be held conclusive.

Taking it, then, as a fact that in plants possessing them the vessels

and ducts are the channels through which sap is distributed, we come

now to the further question--What determines the varying permeability

of the walls of the vessels and ducts, and the consequent varying

formation of wood? To this question I believe the true reply is--The

exposure of the parts to intermittent mechanical strains, actual or

potential, or both. By actual strains I of course mean those which the

plant experiences in the course of its individual life. By potential

strains I mean those which the form, attitude, and circumstances

common to its kind involve, and which its inherited structure is

adapted to meet. In plants with stems, petioles, and leaves, having

tolerably constant attitudes, the increasing porosity of the tubes and

consequent deposit of dense tissue takes place in anticipation of the

strains to which the parts of the individual are liable, but takes

place at parts which have been habitually subject to such strains

in ancestral individuals. But though in such plants the tendency to

repeat that distribution of dense tissue caused by mechanical actions

on past generations, goes on irrespective of the mechanical actions

to which the developing individual is subject, these direct actions,

while they greatly aid the assumption of the typical structure, are

the sole causes of those deviations in the relative thickenings of

parts which distinguish the individual from others of its kind. And

then, in certain irregularly growing plants, such as Cactuses and

Euphorbias, where the strains fall on parts that do not correspond in

successive individuals, we distinctly trace a direct relation between

the degrees of strain and the rates of these changes which result in

dense tissue. I will not occupy space in detailing the evidence of this

relation, which is conspicuous in the orders named, but will pass to

the question--What are the physical processes by which intermittent

mechanical strains produce this deposit of resistant substance at

places where it is needed to meet the strains? We have not to seek far

for an answer. If a trunk, a bough, a shoot, or a petiole, is bent

by a gust of wind, the substance of its convex side is subject to

longitudinal tension: the substance of its concave side being at the

same time compressed. This is the primary mechanical effect. There is,

however, a secondary mechanical effect, which here chiefly concerns

us. That bend by which the tissues of the convex side are stretched,

also produces lateral compression of them. Buttoning on a tight

glove and then closing the hand, will make this necessity clear: the

leather, while it is strained along the backs of the fingers, presses

with considerable force on the knuckles. It is demonstrable that the

tensions of the outer layer of a mass made convex by bending, must,

by composition of forces, produce at every point a resultant at right

angles to the layer beneath it; that, similarly, the joint tensions

of these two layers must throw a pressure on the next deeper layer;

and so on. Hence, if at some little distance beneath the surface of a

stem, twig, or leaf-stalk, there exist longitudinal tubes, these tubes

must be squeezed each time the side of the branch they are placed on

becomes convex. Modifying the illustration just drawn from the clenched

hand will make this clear. When, on forcibly grasping something, the

skin is drawn tightly over the back of the hand, the whitening of

the knuckles shows how the blood is expelled from the vessels below

the surface by the pressure of the tightened skin. If, then, the

sap-vessels must be thus compressed, what will happen to the liquid

they contain? It will move away along the lines of least resistance.

Part, and probably the greater part, will escape lengthways from the

place of greatest pressure: some of it being expelled downwards, and

some of it upwards. But, at the same time, part of it will be likely

to ooze through the walls of the tubes. If these walls are so perfect

as to permit the passage of liquid only by osmose, it may still be

inferred that the osmose will increase under pressure; and probably,

under recurrent pressure, the places at which the osmotic current

passes most readily will become more and more permeable, until they

eventually form pores. At any rate it is manifest that where pores

and slits exist, whether thus formed or formed in any other way, the

escape of sap into the adjacent tissue at each bend will become easy

and rapid. What further must happen? When the branch or shoot recoils,

the vessels on the side that was convex, being relieved from pressure,

will tend to resume their previous diameters; and will be helped to do

this by the elasticity of the surrounding tissue, as well as by those

spiral, annular, and allied structures which they contain. But this

resumption of their previous diameters must cause an immediate rush

of sap back into them. Whence will it come? Not to any considerable

extent from the surrounding tissues into which part of it has been

squeezed, seeing that the resistance to the return of liquid through

small pores will be greater than the resistance to its return along the

vessels themselves. Manifestly the sap which was thrust up and down

the vessels from the place of compression will return--the quantities

returning from above and from below varying, as we shall hereafter

see, according to circumstances. But this is not all. From some side a

greater quantity must come back than was sent away; for the amount that

has escaped out of the tube into the prosenchyma has to be replaced.

Thus during the time when the side of the branch or twig becomes

concave, more sap returns from above or below than was expelled upwards

or downwards during the previous compression. The refilled vessels,

when the next bend renders their side convex, again have part of their

contents forced through their parietes, and are again refilled in the

same way. There is thus set up a draught of sap to the place where

these intermittent strains are going on, an exudation proportionate

to the frequency and intensity of the strains, and a proportionate

nutrition or thickening of the wood-cells, fitting them to resist the

strains. A rude idea of this action may be obtained by grasping in one

hand a damp sponge, having its lower end in water, while holding a

piece of blotting-paper in contact with its upper end, and then giving

the sponge repeated squeezes. At each squeeze some of the water will

be sent into the blotting-paper; at each relaxation the sponge will

refill from below, to give another portion of its contents to the

blotting-paper when again squeezed.

But how does this explanation apply to roots? If the formation of

wood is due to intermittent transverse strains, such as are produced

in the aërial parts of upright plants by the wind, how does it happen

that woody matter is deposited in roots, where there are no lateral

oscillations, no transverse strains? The answer is, that longitudinal

strains also are capable of causing the effects described. It is

true that perfectly straight fibres united into a bundle and pulled

lengthways would not exert on one another any lateral pressure, and

would not laterally compress any similarly-straight canals running

along with them. But if the fibres united into a bundle are variously

bent or twisted, they cannot be longitudinally strained without

compressing one another and structures imbedded in them. It needs but

to watch a wet rope drawn tight by a capstan, to see that an action

like that which squeezes the water out of its strands, will squeeze the

sap out of the vessels of a root into the surrounding tissue, as often

as the root is pulled by the swaying of the plant it belongs to. Here,

too, as before, the vessels will refill when the pull intermits; and

so, in the roots as in the branches, this rude pumping process will

produce a growth of hard tissue proportionate to the stress to be borne.

These conclusions are supported by the evidence which exceptional

cases supply. If intermittent mechanical strains thus cause the

formation of wood where wood is found, then where it is not found,

there should be an absence of intermittent mechanical strains. There

is such an absence. Vascular plants characterized by little or no

deposit of dense substance, are those having vessels so conditioned

that no considerable pressures are borne by them. The more succulent a

petiole or leaf becomes, the more do the effects of transverse strains

fall on its outer layers of cells. Its mechanical support is chiefly

derived from the ability of these minute vesicles, full of liquid, to

resist bursting and tearing under the compressions and tensions they

are exposed to. And just as fast as this change from a thin leaf or

foot-stalk to a thick one entails increasing stress on the superficial

tissue, so fast does it diminish the stress on the internally-seated

vascular tissue. The succulent leaf cannot be swayed about by the wind

as much as an ordinary leaf; and such small bends as can be given to

it and its foot-stalk are prevented from affecting in any considerable

degree the tubes running through its interior. Hence the retentiveness

of the vessels in these fleshy leaves, as shown by the small exudation

of dye; and hence the small thickening of their surrounding prosenchyma

by woody deposit. Still more conspicuously is this connexion of facts

shown when, from the soft thick leaves before named and such others as

those of _Echeveria_, _Rochea_, _Pereskia_, we turn to the thick leaves

that have strong exo-skeletons. _Gasteria_ serves as an illustration.

The leathery or horny skin here evidently bears the entire weight of

the leaf, and is so stiff as to prevent any oscillation. Here, then,

the vessels running inside are protected from all mechanical stress;

and accordingly we find that the cells surrounding them are not

appreciably thickened.

Equally clear, and more striking because more obviously exceptional,

is the evidence given by succulent stems which are leafless. _Stapelia

Buffonia_, having soft procumbent axes not liable to be bent backwards

and forwards in any considerable degree by the wind, has, ramifying

through its tissue, vessels that allow but an extremely slow escape

of dye and have unthickened sheaths. Such of the Euphorbias as have

acquired the fleshy character while retaining the arborescent growth,

like _Euphorbia Canariensis_, teach us the same truth in another way.

In them the formation of wood around the vessels is inconspicuous

where the intermittent strains are but slight; but it is conspicuous

at those joints on which lateral oscillations of the attached branches

throw great extensions and compressions of tissue. Throughout

the _Cactaceæ_ we find varied examples of the alleged relation.

_Mammillaria_ furnishes a very marked one. The substance of one of

these globular masses, resting on the ground, admits of no bending from

side to side; and accordingly its large bundles of spiral and annular

vessels, or partially-united cells, have very feebly-marked sheaths

not at all thickened. In such types as _Cereus_ and _Opuntia_ we see,

as in the Euphorbias, that where little stress falls on the vessels,

little deposit takes place around them; while there is much deposit

where there is much stress. Here let me add a confirmation obtained

since writing the above. After observing among the Cactuses the very

manifest relation between strain and the formation of wood, I inquired

of Mr. Croucher, the intelligent foreman of the Cactus-house at Kew,

whether he found this relation a constant one. He replied that he did,

and that he had frequently tested it by artificially subjecting parts

of them to strains. Neglecting at the time to inquire how he had done

this, it afterwards occurred to me that if he had so done it as to

cause constant strains, the observed result would not tell in favour of

the foregoing interpretation. Subsequently, however, I learned that he

had produced the strains by placing the plants in inclined attitudes--a

method which, by permitting oscillations of the strained joints,

allowed the strains to intermit. And then, making the proof conclusive,

Mr. Croucher volunteered the statement that where he had produced

constant strains by tying, no formation of wood took place.

Aberrant growths of another class display the same relations of

phenomena. Take first the underground stems, such as the Potato and

the Artichoke. The vessels which run through these, slowly take up

the dye without letting it pass to any considerable extent into the

surrounding tissues.[70] Only after an interval of many hours does the

prosenchyma become stained in some places. Here, as before, an absence

of rapid exudation accompanies an absence of woody deposit; and both

these go along with the absence of intermittent strains. Take again the

fleshy roots. The Turnip, the Carrot, and the Beetroot, have vessels

that retain very persistently the coloured liquids they take up. And

differing in this, as these roots do, from ordinary roots, we see

that they also differ from them in not being woody, and in not being

appreciably subject to the usual mechanical actions. In these cases,

as in the others, parts that ordinarily become dense, deviate from

this typical character when they are not exposed to those forces which

produce dense tissue by increasing the extravasation of sap.

To complete the proof that such a relation exists, let me add the

results of some experiments on equal and similarly-developed parts,

kept respectively at rest and in motion. I have tested the effects on

large petioles, on herbaceous shoots, and on woody shoots. If two such

petioles as those of Rhubarb, with their leaves attached, have their

cut ends inserted in bottles of dye, and the one be bent backwards

and forwards while the other remains motionless, there arises, after

the lapse of an hour, scarcely any difference in the states of their

vessels: a certain proportion of these are in both cases charged with

the dye, and little exudation has been produced by the motion. Here,

however, it is to be observed that the causes of exudation are scarcely

operative; the vascular bundles are distributed all through the mass

of the petiole, which is formed of soft watery tissue; and they are,

therefore, not so circumstanced as to be effectually compressed by the

bends. In herbaceous stems, such as those of the Jerusalem Artichoke

and of the Foxglove, an effect scarcely more decided is produced; and

here, too, when we seek a reason, we find it in the non-fulfilment of

the mechanical conditions; for the vascular bundles are not so seated

between a tough layer of bark and a solid core as to be compressed at

each bend. When, however, we come to experiment upon woody shoots, we

meet with conspicuous effects, though by no means uniformly. In some

cases oscillations produce immense amounts of exudation--parallel

transverse sections of the compared shoots showing that where, in the

one that has been at rest, there are spots of colour round but a few

pitted ducts, in the one that has been kept in motion the substance of

the wood is soaked almost uniformly through with dye. In other cases,

especially where there is much undifferentiated tissue remaining, the

exudation is not very marked. The difference appears to depend on

the quantity of liquid contained in the shoot. If its substance is

relatively dry, the exudation is great; but it is comparatively small

if all the tissues are fully charged with sap. This contrast of results

is one which contemplation of the mechanical actions will lead us to

expect.

And now, with these facts to aid our interpretation, let us return to

ordinary stems. If the upper end of a growing shoot, the prosenchyma

of which is but little thickened, be allowed to imbibe the dye, the

vessels of its medullary sheath alone become charged; and from them

there takes place but a slow oozing. If a like experiment be tried

with a lower part of the shoot, where the wood in course of formation

has its inner boundary marked but not its outer boundary, we find

that the pitted ducts, and more especially the inner ones, come into

play. And then lower still, where the wood has its periphery defined

and its histological characters decided, the appearances show that

the tissue forming its outer surface begins to take a leading part

in the transmission of liquid. What now is the explanation of these

changes, mechanically considered? In the young soft part of the shoot,

as in all normal and abnormal growths that have not formed wood, the

channels for the passage of sap are the spiral, annular, fenestrated,

or reticulated vessels. These vessels, here included in the bundles

of the medullary sheath, are, in common with the tissues around

them, subject, by the bendings of the shoot, to slight intermittent

compressions, and, especially the outermost of them, are thus forced

to give the prosenchyma an extra supply of nutritive liquid. The

thickening of the prosenchyma, spreading laterally as well as outwards

from each bundle of the medullary sheath, goes on until it meets the

thickenings that spread from the other bundles; and there is so formed

an irregular cylinder of hardened tissue, surrounding the medulla and

the vascular bundles of its sheath. As soon as this happens, these

vascular bundles become, to a considerable extent, shielded from the

effects of transverse strains, since the tensions and compressions

chiefly fall on the developing wood outside of them. Clearly, too, the

greatest stress must be felt by the outer layer of the developing wood:

being further removed from the neutral axis, it must be subject to

severer strains at each bend; and lying between the bark and the layer

of wood first formed, it must be most exposed to lateral compressions.

Among the elongated cells of this outer layer, some unite to form the

pitted ducts. Being, as we see, better circumstanced mechanically,

they become greater carriers of sap than the original vessels, and,

in consequence of this, as well as in consequence of their relative

proximity, become the sources of nutrition to the still more external

layers of wood-cells. The same causes and the same effects hold with

each new indurated coat deposited round the previously indurated coats.

This description may be thought to go far towards justifying the

current views respecting the course taken by the sap. But the

justification is more apparent than real. In the first place, the

implication here is that the sap-carrying function is at first

discharged entirely by the vessels of the medullary sheath, and

that they cease to discharge this function only as fast as they are

relatively incapacitated by their mechanical circumstances. And the

second implication is, that it is not the wood itself, but the more

or less continuous canals formed in it, which are the subsequent

sap-distributors. This, though readily made clear by microscopic

examination of the large pitted ducts in a partially lignified shoot

that has absorbed the dye, is less manifestly true of the peripheral

layer of sap-carrying tissue finally formed. But it is really true

here. For this layer, though nominally a layer of wood, is practically

a layer of inosculating vessels. It is formed out of irregular lines

and networks of elongated pitted cells, obliquely united by their

ends. Examination of them after absorption of a dye, shows that it is

only along the continuous channels they unite to form that the current

has passed. But the essentially vascular character of this outer and

latest-formed layer of the alburnum is best seen in the fact that

the vascular systems of new axes take their rise from it, and form

with it continuous canals. If a shoot of last year in which growth

is recommencing, be cut lengthways after it has imbibed a dye, clear

proof is obtained that the passage of the dye into a lateral bud takes

place from this outermost layer of pitted cells, and that the channels

taken by the dye through the new tissue are composed of cells that pass

through modified forms into the spiral vessels of the new medullary

sheath. This transition may be still more clearly traced in a terminal

bud that continues the line of last year’s shoot. A longitudinal

section of this shows that the vessels of the new medullary sheath do

not obtain their sap from the vessels of last year’s sheath (which, as

shown by the non-absorption of dye, have become inactive), but that

their supplies are obtained from those inosculating canals formed out

of last year’s outermost layer of prosenchyma, and that between the

component cells of this and those of the new vascular system there are

all gradations of structure.[71]

It is not the aim of the foregoing reasoning to show that mechanical

actions are the sole causes of the formation of dense tissue in

plants. Dense tissue is in many cases formed where no such causes

have come into play--as, for example, in thorns and in the shells

of nuts. Here the natural selection of variations can alone have

operated. It is manifest, too, that even those supporting structures

the building up of which is above ascribed to intermittent strains,

may, in the individual plant of a species that ordinarily has them, be

developed to a great extent when intermittent strains are prevented.

We see this in trees that are artificially supported by nailing to

walls; and we also see a kindred fact in natural climbers. Though in

these cases the formation of wood is obviously less than it would be

were the stem and branches habitually moved about by the wind, it

nevertheless goes on. Clearly the tendency of the plant to repeat the

structure of its type (in the one case the structure of its species;

and in the other case that of the order from which it has diverged in

becoming a climber) is here almost the sole cause of wood formation.

But though in plants so circumstanced intermittent mechanical strains

have little or no direct share, it may still be true, and I believe

is true, that intermittent mechanical strains are the original cause;

for, as before hinted, the typical structure which the individual

thus repeats irrespective of its own conditions, is interpretable

as a typical structure that is itself the product of these actions

and reactions between the plant and its environment. Grant the

inheritance of functionally-produced modifications; grant that natural

selection will always co-operate in such way as to favour those

individuals and families in which functionally-produced modifications

have progressed most advantageously; and it will follow that this

mechanically-caused formation of dense substance, accumulating from

generation to generation by the survival of the fittest, will result

in an organic habit of forming dense tissue at the required places.

The deposit arising from exudation at the places of greatest strain,

recurring from generation to generation at the same places, will come

to be reproduced in anticipation of strain, and will continue to

be reproduced for a long time after a changed habit of the species

prevents the strain--eventually, however, decreasing, both through

functional inactivity and natural selection, to the point at which it

is in equilibrium with the requirement.

Another side of the general question may now be considered. We have

seen how, by intermittent pressures on capillary vessels and ducts

and inosculating canals, there must be produced a draught of sap

towards the point of compression to replace the sap squeezed out. But

we have still to inquire what will be the effect on the distribution

of sap throughout the plant as a whole. It was concluded that out of

the compressed vessels the greater part of the liquid would escape

longitudinally--the longitudinal resistance to movement being least.

In every case the probabilities are infinity to one against the

resistances being equal upwards and downwards. Always, then, more sap

will be expelled in one direction than in the other. But in whichever

direction least sap is expelled, from that same direction most sap will

return when the vessels are relieved from pressure--the force which is

powerful in arresting the back current in that direction being the same

force which is powerful in producing a forward current. Ordinarily, the

more abundant supply of liquid being from below, there will result an

upward current. At each bend a portion of the contents will be squeezed

out through the sides of the vessels--a portion will be squeezed

downwards, reversing the current ascending from the roots, but soon

stopped by its resistance; while a larger portion will be squeezed

upwards towards the extremities of the vessels, where consumption and

loss are most rapid. At each recoil the vessels will be replenished,

chiefly by the repressed upward current; and at the next bend more of

it will be thrust onwards than backwards. Hence we have everywhere in

action a kind of rude force-pump, worked by the wind; and we see how

sap may thus be raised to a height far beyond that to which it could be

raised by capillary action, aided by osmose and evaporation.

Thus far, however, the argument proceeds on the assumption that there

is liquid enough to replenish every time the vessels subject to

this process. But suppose the supply fails--suppose the roots have

exhausted the surrounding stock of moisture. Evidently the vessels thus

repeatedly having their contents squeezed out into the surrounding

tissue, cannot go on refilling themselves from other vessels without

tending to empty the vascular system. On the one hand, evaporation from

the leaves causing a draught on the capillary tubes that end in them,

continually generates a capillary tension upwards; while, on the other

hand, the vessels below, expanding after their sap has been squeezed

out, produce a tension both upwards and downwards towards the point

of loss. Were the limiting membranes of the vessels impermeable, the

movement of sap would, under these conditions, soon be arrested. But

these membranes are permeable; and the surrounding tissues readily

permit the passage of air. This state of tension, then, will cause an

entrance of air into the tubes; the columns of liquid they contain

will be interrupted by bubbles. It seems, indeed, not improbable that

this entrance of air may take place even when there is a good supply

of liquid, if the mechanical strains are so violent and the exudation

so rapid that the currents cannot refill the half-emptied vessels

with sufficient rapidity. And in this case the intruding air may

possibly play the same part as that contained in the air-chamber of a

force-pump--tending, by moderating the violence of the jets, and by

equalizing the strains, to prevent rupture of the apparatus. Of course

when the supply of liquid becomes adequate, and the strains not too

violent, these bubbles will be expelled as readily as they entered.

Here, as before, let me add the conclusive proof furnished by a direct

experiment. To ascertain the amount of this propulsive action, I took

from the same tree, a Laurel, two equal shoots, and placing them in

the same dye, subjected them to conditions that were alike in all

respects save that of motion: while one remained at rest, the other

was bent backwards and forwards, now by switching and now by straining

with the fingers. After the lapse of an hour, I found that the dye had

ascended the oscillating shoot three times as far as it had ascended

the stationary shoot--this result being an average from several trials.

Similar trials brought out similar effects in other structures. The

various petioles and herbaceous shoots experimented upon for the

purpose of ascertaining the amount of exudation produced by transverse

strains, showed also the amount of longitudinal movement. It was

observable that the height ascended by the dye was in all cases greater

where there had been oscillation than where there had been rest--the

difference, however, being much less marked in succulent structures

than in woody ones.

It need scarcely be said that this mechanical action is not here

assigned as the sole cause of circulation, but as a cause co-operating

with others, and helping others to produce effects that could not

otherwise be produced. Trees growing in conservatories afford us

abundant proof that sap is raised to considerable heights by other

forces. Though it is notorious that trees so circumstanced do not

thrive unless, through open sashes, they are frequently subject

to breezes sufficient to make their parts oscillate, yet there is

evidently a circulation that goes on without mechanical aid. The

_causes_ of circulation are those actions only which disturb the liquid

equilibrium in a plant, by permanently abstracting water or sap from

some part of it; and of these the first is the absorption of materials

for the formation of new tissue in growing parts; the second is the

loss by evaporation, mainly through adult leaves; and the third is the

loss by extravasation, through compressed vessels. Only so far as it

produces this last, can mechanical strain be regarded as truly a cause

of circulation. All the other actions concerned must be classed as

_aids_ to circulation--as facilitating that re-distribution of liquid

that continually restores the equilibrium continually disturbed; and

of these capillary action may be named as the first, osmose as the

second, and the propulsive effect of mechanical strains as the third.

The first two of these aids are doubtless capable by themselves of

producing a large part of the observed result--more of the observed

result than is at first sight manifest; for there is an important

indirect effect of osmotic action which appears to be overlooked.

Osmose does not aid circulation only by setting up, within the plant,

exchange currents between the more dense and the less dense solutions

in different parts of it; but it aids circulation much more by

producing distention of the plant as a whole. In consequence of the

average contrast in density between the water outside of the plant and

the sap inside of it, the constant tendency is for the plant to absorb

a quantity in excess of its capacity, and so to produce distention

and erection of its tissues. It is because of this that the drooping

plant raises itself when watered; for capillary action alone could only

refill its tissues without changing their attitudes. And it is because

of this that juicy plants with collapsible structures bleed so rapidly

when cut, not only from the cut surface of the rooted part, but from

the cut surface of the detached part--the elastic tissues tending to

press out the liquid which distends them. And manifestly if osmose

serves thus to maintain a state of distention throughout a plant,

it indirectly furthers circulation; since immediately evaporation

or growth at any part, by abstracting liquid from the neighbouring

tissues, begins to diminish the liquid pressure within such tissues,

the distended structures throughout the rest of the plant thrust

their liquid contents towards the place of diminished pressure. This,

indeed, may very possibly be the most efficient of the agencies at

work. Remembering how great is the distention producible by osmotic

absorption--great enough to burst a bladder--it is clear that the

force with which the distended tissues of a plant urge forward the sap

to places of consumption, is probably very great. We must therefore

regard the aid which mechanical strains give as being one of several.

Oscillations help directly to restore any disturbed liquid equilibrium;

and they also help indirectly, by facilitating the re-distribution

caused by capillary action and the process just described; but in the

absence of oscillations the equilibrium may still be restored, though

less rapidly and within narrower limits of distance.

One half of the problem of the circulation, however, has been left out

of sight. Thus far our inquiry has been, how the ascending current

of sap is produced. There remains the rationale of the descending

current. What forces cause it, and through what tissues it takes place,

are questions to which no satisfactory answers have been given. That

the descent is due to gravitation, as some allege, is difficult to

conceive, since, as gravitation acts equally on all liquid columns

contained in the stem, it is not easy to see why it should produce

downward movements in some while permitting upward movements in

others--unless, indeed, there existed descending tubes too wide

to admit of much capillary action, which there do not. Moreover,

gravitation is clearly inadequate to cause currents towards the roots

out of branches that droop to the ground. Here the gravitation of the

contained liquid columns must nearly balance that of the connected

columns in the stem, leaving no appreciable force to cause motion. Nor

does there seem much probability in the assumption that the route of

the descending sap is through the cambium layer, since experiments on

the absorption of dyes prove that simple cellular tissue is a very bad

conductor of liquids: their movement through it does not take place

with one-fiftieth of the rapidity with which it takes place through

vessels.[72]

Of course the defence for these hypotheses is, that there must be a

downward current, which must have a course and a cause; and the very

natural assumption has been that the course and the cause must be other

than those which produce the ascending current. Nevertheless there

is an alternative supposition to which the foregoing considerations

introduce us. It is quite possible for the same vascular system

to serve as a channel for movement in opposite directions at

different times. We have among animals well-known cases in which the

blood-vessels carry a current first in one direction and then, after a

brief pause, in the reverse direction. And there seems an _à priori_

probability that, lowly-organized as they are, plants are more likely

to have distributing appliances of this imperfect kind than to have

two sets of channels for two simultaneous currents. If, led by this

suspicion, we inquire whether among the forces which unite to produce

movements of sap, there are any variations or intermissions capable of

determining the currents in different directions, we quickly discover

that there are such, and that the hypothesis of an alternating motion

of the sap, now centrifugal and now centripetal, through the same

vessels, has good warrant. What are the several forces at work? First

may be set down that tendency existing in every part of a plant to

expand into its typical form, and to absorb nutritive liquids in doing

this. The resulting competition for sap will, other things being

equal, cause currents towards the most rapidly-growing parts--towards

unfolding shoots and leaves, but not towards adult leaves. Next we

have evaporation, acting more on the adult leaves than on those which

are in the bud, or but partially developed. This evaporation is both

regularly and irregularly intermittent. Depending chiefly on the

action of the sun, it is, in fine weather, greatly checked or wholly

arrested every evening; and in cloudy weather must be much retarded

during the day. Further, every hygrometric variation, as well as every

variation in the movement of the air, must vary the evaporation.

This chief action, therefore, which, by continually emptying the

ends of the capillary tubes, makes upward currents possible, is one

which intermits every night, and every day is strong or feeble as

circumstances determine. Then, in the third place, we have this rude

pumping process above described, going on with greater vigour when the

wind is violent, and with less vigour when it is gentle--drawing liquid

_towards_ different parts according to their degrees of oscillation,

and _from_ different parts according as they can most readily furnish

it. And now let us ask what must result under changing conditions

from these variously-conflicting and conspiring forces. When a warm

sunshine, causing rapid evaporation, is emptying the vessels of the

leaves, the osmotic and capillary actions that refill them will be

continually aided by the pumping action of the swaying petioles,

twigs, and branches, provided their oscillations are moderate. Under

these conditions the current of sap, moving in the direction of least

resistance, will set towards the leaves. But what will happen when the

sun sets? There is now nothing to determine currents either upwards

or downwards, except the relative rates of growth in the parts and

the relative demands set up by the oscillations; and the oscillations

acting alone, will draw sap to the oscillating parts as much from above

as from below. If the resistance to be overcome by a current setting

back from the leaves is less than the resistance to be overcome by a

current setting up from the roots, then a current will set back from

the leaves. Now it is, I think, tolerably manifest that in the swaying

twigs and minor branches, less force will be required to overcome the

inertia of the short columns of liquid between them and the leaves than

to overcome the inertia of the long columns between them and the roots.

Hence during the night, as also at other times when evaporation is not

going on, the sap will be drawn out of the leaves into the adjacent

supporting parts; and their nutrition will be increased. If the wind is

strong enough to produce a swaying of the thicker branches, the back

current will extend to them also; and a further strengthening will

result from their absorption of the elaborated sap. And when the great

branches and the stem are bent backwards and forwards by a gale, they

too will share in the nutrition. It may at first sight seem that these

parts, being nearer to the roots than to the leaves, will draw their

supplies from the roots only. But the quantity which the roots can

furnish is insufficient to meet so great a demand. Under the conditions

described, the exudation of sap from the vessels will be very great,

and the draught of liquid required to refill them, not satisfied by

that which the root-fibres can take in, will extend to the leaves.

Thus sap will flow to the several parts according to their respective

degrees of activity--to the leaves while light and heat enable them to

discharge their functions, and back to the twigs, branches, stem, and

roots when these become active and the leaves inactive, or when their

activity dominates over that of the leaves. And this distribution of

nutriment, varying with the varying activities of the parts, is just

such a distribution as we know must be required to keep up the organic

balance.

To this explanation it may be objected that it does not account for

the downward current of sap in plants that are sheltered. The stem and

roots of a drawing-room Geranium display a thickening which implies

that nutritive matters have descended from the leaves, although there

are none of those oscillations by which the sap is said to be drawn

downwards as well as upwards. The reply is, that the stem and roots

tend to repeat their typical structures, and that the absorption of

sap for the formation of their respective dense tissues, is here the

force which determines the descent. Indeed it must be borne in mind

that the mechanical strains and the pumping process which they keep

up, as well as the distention caused by osmose, do not in themselves

produce a current either upwards or downwards: they simply help to move

the sap towards that place where there is the most rapid abstraction

of it--the place towards which its motion is least resisted. Whether

there is oscillation or whether there is not, the physiological demands

of the different parts of the plant determine the direction of the

current; and all which the oscillations and the distention do is to

facilitate the supply of these demands. Just as much, therefore,

in a plant at rest as in a plant in motion, the current will set

downwards when the function of the leaves is arrested, and when there

is nothing to resist that abstraction of sap caused by the tendency

of the stem- and root-tissues to assume their typical structures. To

which admission, however, it must be added that since this typical

structure assumed, though imperfectly assumed, by the hothouse plant,

is itself interpretable as the inherited effect of external mechanical

actions on its ancestors, we may still consider the current set up by

the assumption of the typical structure to be indirectly due to such

actions.

Interesting evidence of another order here demands notice. In the

course of experiments on the absorption of dyes by leaves, it happened

that in making sections parallel to the plane of a leaf, with the

view of separating its middle layer containing the vessels, I came

upon some structures that were new to me. These structures, where they

are present, form the terminations of the vascular system. They are

masses of irregular and imperfectly united fibrous cells, such as those

out of which vessels are developed; and they are sometimes slender,

sometimes bulky--usually, however, being more or less club-shaped. In

transverse sections of leaves their distinctive characters are not

shown: they are taken for the smaller veins. It is only by carefully

slicing away the surface of a leaf until we come down to that part

which contains them, that we get any idea of their nature. Fig. 1

represents a specimen taken from a leaf of _Euphorbia neriifolia_.

Occupying one of the interspaces of the ultimate venous network, it

consists of a spirally-lined duct or set of ducts, which connects with

the neighbouring vein a cluster of half-reticulated, half-scalariform

cells. These cells have projections, many of them tapering, that insert

themselves into the adjacent intercellular spaces, thus producing

an extensive surface of contact between the organ and the imbedding

tissues. A further trait is, that the ensheathing prosenchyma is

either but little developed or wholly absent; and consequently this

expanded vascular structure, especially at its end, comes immediately

in contact with the tissues concerned in assimilation. The leaf of

_Euphorbia neriifolia_ is a very fleshy one; and in it these organs

are distributed through a compact, though watery, cellular mass. But

in any leaf of the ordinary type which possesses them, they lie in

the network-parenchyma composing its lower layer; and wherever they

occur in this layer its cells unite to enclose them. This arrangement

is shown in fig. 2, representing a sample from the Caoutchouc-leaf,

as seen with the upper part of its envelope removed; and it is shown

still more clearly in a sample from the leaf of _Panax Lessonii_,

fig. 3. Figures 4 and 5 represent, without their sheaths, other such

organs from the leaves of _Panax Lessonii_ and _Clusia flava_. Some

relation seems to exist between their forms and the thicknesses of the

layers in which they lie. Certain very thick leaves, such as those of

_Clusia flava_, have them less abundantly distributed than is usual,

but more massive. Where the parenchyma is developed not to so great an

extreme, though still largely, as in the leaves of Holly, _Aucuba_,

_Camellia_, they are not so bulky; and in thinner leaves, like those

of Privet, Elder, &c., they become longer and less conspicuously

club-shaped. Some adaptations to their respective positions seem

implied by these modifications; and we may naturally expect that in

many thin leaves these free ends, becoming still narrower, lose the

distinctive and suggestive characters possessed by those shown in the

diagrams. Relations of this kind are not regular, however. In various

other genera, members of which I have examined, as _Rhus_, _Viburnum_,

_Griselinia_, _Brexia_, _Botryodendron_, _Pereskia_, the variations in

the bulk and form of these structures are not directly determined by

the spaces which the leaves allow: obviously there are other modifying

causes. It should be added that while these expanded free extremities

graduate into tapering free extremities, not differing from ordinary

vessels, they also pass insensibly into the ordinary inosculations.

Occasionally, along with numerous free endings, there occur loops; and

from such loops there are transitions to the ultimate meshes of the

veins.

These organs are by no means common to all leaves. In many that afford

ample spaces for them they are not to be found. So far as I have

observed, they are absent from the thick leaves of plants which form

very little wood. In _Sempervivum_, in _Echeveria_, in _Bryophyllum_,

they do not appear to exist; and I have been unable to discover them in

_Kalanchoë rotundifolia_, in _Kleinia ante-euphorbium_ and _ficoides_,

in the several species of _Crassula_, and in other succulent plants. It

may be added that they are not absolutely confined to leaves, but occur

in stems that have assumed the functions of leaves. At least I have

found, in the green parenchyma of _Opuntia_, organs that are analogous

though much more rudely and irregularly formed. In other parts, too,

that have usurped the leaf-function, they occur, as in the phyllodes of

the Australian Acacias. These have them abundantly developed; and it

is interesting to observe that here, where the two vertically-placed

surfaces of the flattened-out petiole are equally adapted to the

assimilative function, there exist two layers of these expanded

vascular terminations, one applied to the inner surface of each layer

of parenchyma.

Considering the structures and positions of these organs, as well as

the natures of the plants possessing them, may we not form a shrewd

suspicion respecting their function? Is it not probable that they

facilitate absorption of the juices carried back from the leaf for

the nutrition of the stem and roots? They are admirably adapted for

performing this office. Their component fibrous cells, having angles

insinuated between the cells of the parenchyma, are shaped just as they

should be for taking up its contents; and the absence of sheathing

tissue between them and the parenchyma facilitates the passage of the

elaborated liquids. Moreover there is the fact that they are allied to

organs which obviously have absorbent functions. I am indebted to Dr.

Hooker for pointing out the figures of two such organs in the “Icones

Anatomicæ” of Link. One of them is from the end of a dicotyledonous

root-fibre, and the other is from the prothallus of a young Fern. In

each case a cluster of fibrous cells, seated at a place from which

liquid has to be drawn, is connected by vessels with the parts to which

liquid has to be carried. There can scarcely be a doubt, then, that in

both cases absorption is effected through them. I have met with another

such organ, more elaborately constructed, but evidently adapted to the

same office, in the common Turnip-root. As shown by the end view and

longitudinal section in figs. 6 and 7, this organ consists of rings of

fenestrated cells, arranged with varying degrees of regularity into a

funnel, ordinarily having its apex directed towards the central mass

of the Turnip, with which it has, in some cases at least, a traceable

connexion by a canal. Presenting as it does an external porous surface

terminating one of the branches of the vascular system, each of these

organs is well fitted for taking up with rapidity the nutriment laid

by in the Turnip-root, and used by the plant when it sends up its

flower-stalk. Nor does even this exhaust the analogies. The cotyledons

of the young bean, experimented upon as before described, furnished

other examples of such structures, exactly in the places where, if

they are absorbents, we might expect to find them. Amid the branchings

and inosculations of the vascular layer running through the mass of

nutriment deposited in each cotyledon, there are conspicuous free

terminations that are club-shaped, and prove to be composed, like those

in leaves, of irregularly formed and clustered fibrous cells; and some

of them, diverging from the plane of the vascular layer, dip down into

the mass of starch and albumen which the young plant has to utilize,

and which these structures can have no other function but to take up.

Besides being so well fitted for absorption, and besides being

similar to organs which we cannot doubt are absorbents, these

vascular terminations in leaves afford us yet another evidence of

their functions. They are seated in a tissue so arranged as specially

to facilitate the abstraction of liquid. The centripetal movement

of the sap must be set up by a force that is comparatively feeble,

since, the parietes of the ducts being porous, air will enter if the

tension on the contained columns becomes considerable. Hence it is

needful that the exit of sap from the leaves should meet with very

little resistance. Now were it not for an adjustment presently to

be described, it would meet with great resistance, notwithstanding

the peculiar fitness of these organs to take it in. Liquid cannot be

drawn out of any closed cavity without producing a collapse of the

cavity’s sides; and if its sides are not readily collapsible, there

must be a corresponding resistance to the abstraction of liquid from

it. Clearly the like must happen if the liquid is to be drawn out of

a tissue which cannot either diminish in bulk bodily or allow its

components individually to diminish in bulk. In an ordinary leaf, the

upper layer of parenchyma, formed as it is of closely-packed cells

that are without interspaces, and are everywhere held fast within

their framework of veins, can neither contract easily as a mass, nor

allow its separate cells to do so. Quite otherwise is it with the

network-parenchyma below. The long cells of this, united merely by

their ends and having their flexible sides surrounded by air, may

severally have their contents considerably increased and decreased

without offering appreciable resistances: and the network-tissue

which they form will, at the same time, be capable of undergoing

slight expansions and contractions of its thickness. In this layer

occur these organs that are so obviously fitted for absorption. Here

we find them in direct communication with its system of collapsible

cells. The probability appears to be, that when the current sets into

the leaf, it passes through the vessels and their sheaths chiefly into

the upper layer of cells (this upper layer having a larger surface

of contact with the veins than the lower layer, and being the seat

of more active processes); and that the juices of the upper layer,

enriched by the assimilated matters, pass into the network-parenchyma,

which serves as a reservoir from which they are from time to time

drawn for the nutrition of the rest of the plant, when the actions

determine the downward current. Should it be asked what happens where

the absorbents, instead of being inserted in a network-parenchyma,

are, as in the leaves of _Euphorbia neriifolia_, inserted in a solid

parenchyma, the reply is, that such a parenchyma, though not furnished

with systematically arranged air-chambers, nevertheless contains air

in its intercellular spaces; and that when there occurs a draught upon

its contents, the expansion of this air and the entrance of more from

without, quickly supply the place of the abstracted liquid.

If then, returning to the general argument, we conclude that these

expanded terminations of the vascular system in leaves are absorbent

organs, we find a further confirmation of the views set forth

respecting the alternating movement of the sap along the same channels.

These spongioles of the leaves, like the spongioles of the roots, being

appliances by which liquid is taken up to be carried into the mass

of the plant, we are obliged to regard the vessels that end in these

spongioles of the leaves as being the channels of the down current

whenever it is produced. If the elaborated sap is abstracted from the

leaves by these absorbents, then we have no alternative but to suppose

that, having entered the vascular system, the elaborated sap descends

through it. And seeing how, by the help of these special terminations,

it becomes possible for the same vessels to carry back a quality of sap

unlike that which they bring up, we are enabled to understand tolerably

well how this rhythmical movement produces a downward transfer of

materials for growth.

* * * * *

The several lines of argument may now be brought together; and along

with them may be woven up such evidences as remain. Let me first point

out the variety of questions to which the hypothesis supplies answers.

It is required to account for the ascent of sap to a height beyond that

to which capillary action can raise it. This ascent is accounted for

by the propulsive action of transverse strains, joined with that of

osmotic distention. A cause has to be assigned for that rise of sap

which, in the spring, while yet there is no considerable evaporation to

aid it, goes on with a power which capillarity does not explain. The

co-operation of the same two agencies is assignable for this result

also.[73] The circumstance that vessels and ducts here contain sap and

there contain air, and at the same place contain at different seasons

now air and now sap is a fact calling for explanation. An explanation

is furnished by these mechanical actions which involve the entrance

or expulsion of air according to the supply of liquid. That vessels

and ducts which were originally active sap-carriers go completely out

of use, and have their function discharged by other vessels or ducts,

is an anomaly that has to be solved. Again, we are supplied with a

solution: these deserted vessels and ducts are those which, by the

formation of dense tissue outside of them, become so circumstanced that

they cannot be compressed as they originally were. A channel has to be

found for the downward current of sap, which, on any other hypothesis

than the foregoing, must be a channel separate from that taken by the

upward current; and yet no good evidence of a separate channel has

been pointed out. Here, however, the difficulty disappears, since one

channel suffices for the current alternating upwards and downwards

according to the conditions. Moreover there has to be found a force

producing or facilitating the downward current, capable even of drawing

sap out of drooping branches; and no such force is forthcoming. The

hypothesis set forth dispenses with this necessity; under the recurring

change of conditions, the same distention and oscillation which before

raised the sap to the places of consumption, now bring it down to the

places of consumption. A physical process has to be pointed out by

which the material that forms dense tissue is deposited at the places

where it is wanted, rather than at other places. This physical process

the hypothesis indicates. It is requisite to find an explanation of the

fact that, when plants ordinarily swayed about by the wind are grown

indoors, the formation of wood is so much diminished that they become

abnormally slender. Of this an explanation is supplied. Yet a further

fact to be interpreted is, that in the same individual plant homologous

parts, which, according to the type of the plant, should be equally

woody, become much thicker one than another if subject to greater

mechanical stress. And of this too an interpretation is similarly

afforded.

Now the sufficiency of the assigned actions to account for so many

phenomena not otherwise explained, would be strong evidence that the

rationale is the true one, even were it of a purely hypothetical

kind. How strong, then, becomes the reason for believing it the true

one when we remember that the actions alleged demonstrably go on in

the way asserted. They are ever operating before our eyes; and that

they produce the effects in question is a conclusion deducible from

mechanical principles, a conclusion established by induction, and a

conclusion verified by experiment. These three orders of proof may be

briefly summed up as follows.

That plants which have to raise themselves above the earth’s surface,

and to withstand the actions of the wind, must have a power of

developing supporting structure, is an _à priori_ conclusion which may

be safely drawn. It is an equally safe _à priori_ conclusion, that if

the supporting structure, either as a whole or in any of its parts, has

to adapt itself to the particular strains which the individual plant is

subject to by its particular circumstances, there must be at work some

process by which the strength of the supporting structure is everywhere

brought into equilibrium with the forces it has to bear. Though the

typical distribution of supporting structure in each kind of plant may

be explained teleologically by those whom teleological explanations

satisfy; and though otherwise this typical distribution may be ascribed

to natural selection acting apart from any directly adaptive process;

yet it is manifest that those departures from the typical distribution

which fit the parts of each plant to their special conditions are

explicable neither teleologically nor by natural selection. We are,

therefore, compelled to admit that, if in each plant there goes on

a balancing of the particular strains by the particular strengths,

there must be a physical or physico-chemical process by which the

adjustments of the two are effected. Meanwhile we are equally compelled

to admit, _à priori_, that the mechanical actions to be resisted,

themselves affect the internal tissues in such ways as to further the

increase of that dense substance by which they are resisted. It is

demonstrable that bending the petioles, shoots, and stems must compress

the vessels beneath their surfaces, and increase the exudation of

nutritive matters from them, and must do this actively in proportion

as the bends are great and frequent; so that while, on the one hand,

it is a necessary deduction that, if the parts of each plant are to be

severally strengthened according to the several strains, there must

be some direct connexion between strains and strengths, it is, on the

other hand, a necessary deduction from mechanical principles that the

strains do act in such ways as to aid the increase of the strengths.

How a like correspondence between two _à priori_ arguments holds in

the case of the circulation, needs not to be shown in detail. It will

suffice to remind the reader that while the raising of sap to heights

beyond the limit of capillarity implies some force to effect it, we

have in the osmotic distention and the intermittent compressions caused

by transverse strains, forces which, under the conditions, cannot but

tend to effect it; and similarly with the requirement for a downward

current, and the production of a downward current.

Among the inductive proofs we find a kindred agreement. Different

individuals of the same species, and different parts of the same

individual, do strengthen in different degrees; and there is a clearly

traceable connexion between their strengthenings and the intermittent

strains they are exposed to. This evidence, derived from contrasts

between growths on the same plant or on plants of the same type, is

enforced by evidence derived from contrasts between plants of different

types. The deficiency of woody tissue which we see in plants called

succulent, is accompanied by a bulkiness of the parts which prevents

any considerable oscillations; and this character is also habitually

accompanied by a dwarfed growth. When, leaving these relations as

displayed externally, we examine them internally, we find the facts

uniting to show, by their agreements and differences, that between

the compression of the sap-canals and the production of wood there

is a direct relation. We have the facts, that in each plant, and in

every new part of each plant, the formation of sap-canals precedes

the formation of wood; that the deposit of woody matter, when it

begins, takes place around these sap-canals, and afterwards around

the new sap-canals successively developed; that this formation of

wood around the sap-canals takes place where the coats of the canals

are demonstrably permeable, and that the amount of wood formation is

proportionate to the permeability. And then that the permeability and

extravasation of sap occur wherever, in the individual or in the type,

there are intermittent compressions, is proved alike by ordinary cases

and by exceptional cases. In the one class of cases we see that the

deposit of wood round the vessels begins to take place when they come

into positions that subject them to intermittent compressions, while it

ceases when they become shielded from compressions. And in the other

class of cases, where, from the beginning, the vessels are shielded

from compression by surrounding fleshy tissue, there is a permanent

absence of wood formation.

To which complete agreement between the deductive and inductive

inferences has to be added the direct proof supplied by experiments. It

is put beyond doubt by experiment that the liquids absorbed by plants

are distributed to their different parts through their vessels--at

first by the spiral or allied vessels originally developed, and then by

the better-placed ducts formed later. By experiment it is demonstrated

that the intermittent compressions caused by oscillations urge the

sap along the vessels and ducts. And it is also experimentally proved

that the same intermittent compressions produce exudation of sap from

vessels and ducts into the surrounding tissue.

That the processes here described, acting through all past time, have

sufficed of themselves to develope the supporting and distributing

structures of plants, is not alleged. What share the natural selection

of variations distinguished as spontaneous, has had in establishing

them, is a question which remains to be discussed. Whether acting

alone natural selection would have sufficed to evolve these vascular

and resisting tissues, I do not profess to say. That it has been a

co-operating cause, I take to be self-evident: it must all along

have furthered the action of any other cause, by preserving the

individuals on which such other cause had acted most favourably.

Seeing, however, the conclusive proof which we have that another cause

has been in action--certainly on individuals, and, in all probability,

by inheritance on races--we may most philosophically ascribe the

genesis of these internal structures to this cause, and regard natural

selection as having here played the part of an accelerator.

EXPLANATION OF PLATE.

Fig. 1. Absorbent organ from the leaf of _Euphorbia neriifolia_.

The cluster of fibrous cells forming one of the terminations of the

vascular system is here imbedded in a solid parenchyma.

Fig. 2. A structure of analogous kind from the leaf of _Ficus

elastica_. Here the expanded terminations of the vessels are imbedded

in the network-parenchyma, the cells of which unite to form envelopes

for them.

Fig. 3. Shows on a larger scale one of these absorbents from the leaf

of _Panax Lessonii_. In this figure is clearly seen the way in which

the cells of the network-parenchyma unite into a closely-fitting case

for the spiral cells.

Fig. 4. Represents a much more massive absorbent from the same leaf,

the surrounding tissues being omitted.

Fig. 5. Similarly represents, without its sheath, an absorbent from the

leaf of _Clusia flava_.

Fig. 6. End view of an absorbent organ from the root of a Turnip. It is

taken from the outermost layer of vessels. Its funnel-shaped interior

is drawn as it presents itself when looked at from the outside of this

layer, its narrow end being directed towards the centre of the Turnip.

Fig. 7. A longitudinal section through the axis of another such organ,

showing its annuli of reticulated cells when cut through. The cellular

tissue which fills the interior is supposed to be removed.

Fig. 8. A less developed absorbent, showing its approximate connexion

with a duct. In their simplest forms, these structures consist of only

two fenestrated cells, with their ends bent round so as to meet. Such

types occur in the central mass of the Turnip, where the vascular

system is relatively imperfect. Besides the comparatively regular forms

of these absorbents, there are forms composed of amorphous masses

of fenestrated cells. It should be added that both the regular and

irregular kinds are very variable in their numbers: in some turnips

they are abundant, and in others scarcely to be found. Possibly their

presence depends on the age of the Turnip. Judging from the period

during which my investigations were made, namely winter and early

spring, I suspect that they are developed only in preparation for

sending up the flower-stalk.

[Illustration: Figs.1–8.]

Let me add that experiments on circulation in plants made during the

state of inactivity, when it is to be presumed that the vessels and

tissues contain but little gap, are much more successful than those

made in the summer. It would seem that when the tissues are fully

charged with sap the taking up of dyes is comparatively slow and the

above-described effects are not so easily demonstrable.

* * * * *

[An expert writes concerning this essay:--“I have not attempted to

annotate critically this paper. There is no doubt that many of your

conclusions are perfectly sound, particularly those relating to the

passage of crude sap through the _cavities_ of the elements of the

wood, though the opinion that the actual passage was through the walls

very generally held till about 12 years ago.”]

APPENDIX D.

ON THE ORIGIN OF THE VERTEBRATE TYPE.

[_When studying the development of the vertebrate skeleton,

there occurred to me the following idea respecting the possible

origin of the notochord. I was eventually led to omit the few

pages of Appendix in which I had expressed this idea, because

it was unsupported by developmental evidence. The developmental

evidence recently discovered, however, has led Professor Haeckel

and others to analogous views respecting the affiliation of the_

Vertebrata _on the_ Molluscoida. _Having fortunately preserved a

proof of the suppressed pages, I am able now to add them. With

the omission of a superfluous paragraph, they are reprinted

verbatim from this proof, which dates back to the autumn of

1865, at which time the chapter on “The Shapes of Vertebrate

Skeletons” was written._--December, 1869.]

The general argument contained in Chap. XVI. of Part IV., I have

thought it undesirable to implicate with any conception more

speculative than those essential to it; and to avoid so implicating it,

I transfer to this place an hypothesis respecting the derivation of the

rudimentary vertebrate structure, which appears to me worth considering.

Among those molluscoid animals with which the lowest vertebrate animal

has sundry traits in common, it very generally happens that while the

adult is stationary the larva is locomotive. The locomotion of the

larva is effected by the undulations of a tail. In shape and movement

one of these young Ascidians is not altogether unlike a Tadpole.

And as the tail of the Tadpole disappears when its function comes

to be fulfilled by limbs; so the Ascidian larva’s tail disappears

when fixation of the larva renders it useless. This disappearance of

the tail, however, is not without exception. The _Appendicularia_

is an Ascidian which retains its tail throughout life; and by its

aid continues throughout life to swim about. Now this tail of the

_Appendicularia_ has a very suggestive structure. It is long, tapering

to a point, and flattened. From end to end there runs a mid-rib, which

appears to be an imbedded gelatinous rod, not unlike a notochord.

Extending along the two sides of this mid-rib, are bundles of muscular

fibres; and its top bears a gangliated nervous thread, giving off, at

intervals, branches to the muscular fibres. In the _Appendicularia_

this tail, which is inserted at the lower part of the back, is bent

forwards, so as not to be adapted for propelling the body of the animal

head foremost; but the homologous tails of the larval Ascidians are

directed backwards, so as to produce forward movement. If we suppose

a type like the _Appendicularia_ in the structure and insertion of

its permanent tail, but resembling the larval forms in the direction

of its tail, it is, I think, not difficult to see that functional

adaptation joined with natural selection, might readily produce a

type approximating to that whose origin we are considering. It is a

fair assumption that an habitually-locomotive creature would profit

by increased power of locomotion. This granted, it follows that

such further development of the tail-structures as might arise from

enhanced function, and such better distribution of them as spontaneous

variation might from time to time initiate, would be perpetuated. What

must be the accompanying changes? The more vigorous action of such an

appendage implies a firmer insertion into the body; and this would be

effected by the prolongation forwards of the central axis of the tail

into the creature’s back. As fast as there progressed this fusion of

the increasingly-powerful tail with the body, the body would begin to

partake of its oscillations; and at the same time that the resistant

axis of the tail advanced along the dorsal region, its accompanying

muscular fibres would spread over the sides of the body: gradually

taking such modified directions and insertions as their new conditions

rendered most advantageous. Without further explanation, those who

examine drawings of the structures described, will, I think, see that

in such a way a tail homologous with that of the _Appendicularia_,

would be likely, in the course of that development required for its

greater efficiency, gradually to encroach on the body, until its

mid-rib became the dorsal axis, its gangliated nerve-thread the spinal

chord, and its muscular fibres the myocommata. Such a development of an

appendage into a dominant part of the organism, though at first sight

a startling supposition, is not without plenty of parallels: instance

the way in which the cerebral ganglia, originally mere adjuncts of the

spinal chord, eventually become the great centres of the nervous system

to which the spinal chord is quite subordinate; or instance the way in

which the limbs, small and inconspicuous in fishes, become, in Man,

masses which, taken together, outweigh the trunk. It may be added that

these familiar cases have a further appropriateness; for they exhibit

higher degrees of that same increasing dominance of the organs of

external relation, which the hypothesis itself implies.

Of course, if the rudimentary vertebrate apparatus thus grew into,

and spread over, a molluscoid visceral system, the formation of the

notochord under the action of alternating transverse strains, did not

take place as suggested in § 255; but it does not therefore follow

that its differentiation from surrounding tissues was not mechanically

initiated in the way described. For what was said in that section

respecting the effects of lateral bendings of the body, equally

applies to lateral bendings of the tail; and as fast as the developing

tail encroached on the body, the body would become implicated in the

transverse strains, and the differentiation would advance forwards

under the influences originally alleged. Obviously, too, though the

lateral muscular masses would in this case have a different history;

yet the segmentation of them would be eventually determined by the

assigned causes. For as fast as the strata of contractile fibres,

developing somewhat in advance of the dorsal axis, spread along the

sides, they would come under the influence of the alternate flexions;

and while, by survival of the fittest, their parts became adjusted

in direction, their segmentation would, as before, accompany their

increasing massiveness. The actions and reactions due to lateral

undulations would still, therefore, be the causes of differentiation,

with which natural selection would co-operate.

APPENDIX D 2.

THE ANNULOSE TYPE.

The production of a segmental structure by undulatory movements,

suggested in Appendix D, as also in B (first published in 1858) as

explaining the vertebral column, has been recently suggested by Prof.

Korschelt as the cause of that segmentation of the annulose type which

gives the name to it. He espouses a--

“view which is based upon the assumption that at first an

unsegmented, elongated ancestral form was produced by terminal

growth, whereupon the entire body became separated at once into

a large number of segments by a re-arrangement of the individual

organs. This assumption is supported by the consideration that

with the lateral sinuous movement of the body, and with the

rigidity of the tissues caused by increasing differentiation,

the formation of alternating regions of greater and less

motility was of considerable advantage to the individual, and

rendered possible a further elongation of the body. The first

cause for the appearance of metameric segmentation would then be

sought in the manner of locomotion and in mechanical conditions.

However, this latter view is not supported in any way by

embryology.” (_Embryology of Invertebrates_, Part I, pp. 349–50.)

I venture to think the confession that this view “is not supported in

any way by embryology” should be joined with the confession that it

is at variance with that abstract embryology which comprehends the

process of development in general. The assumption that there took

place “a re-arrangement of the individual organs” of “an unsegmented,

elongated ancestral form,” in such wise that the organs, previously

single, presently became multiple, so that instead of one organ of each

kind there were substituted many organs of each kind, is inconsistent

with the general law of evolution, organic and other--implies not

integration but disintegration. Everywhere the advance is from many

like parts performing like functions to relatively few unlike parts

performing unlike functions. The higher forms of the annulose type

itself show this. Compare a myriapod and a crab. In the one we have

not only a great number of similar segments bearing similar limbs,

but we have in each segment a dilatation of the main blood-vessel--a

rudimentary heart--a swollen portion of the nerve cord--a small

ganglion--and so on; whereas in the other, besides relatively few

segments and few limbs (sundry of them extremely unlike the rest) we

have a vascular system concentrated into a central heart with arteries

and a concentrated nervous system, such that the great ganglia in the

integrated carapace immensely subordinate the ganglia of the remaining

segments; and similarly with the other organs. Now unless it be denied

that these highest decapods have been evolved from low types akin to

myriapods in composition, it must be admitted that the progress has

been from a string of many like segments with similar sets of organs

to a group of relatively-few unlike segments with dissimilar sets of

organs. If so we cannot rationally deny that the progress has been

of this nature up from the lowest annelid, instead of having been,

as Prof. Korschelt’s hypothesis implies, of opposite nature at the

beginning.

In a preceding passage a clear recognition of the normal course of

development occurs. In opposing the view set forth in §§ 205–7 of this

work, Prof. Korschelt says:--

“It seems scarcely favourable to this theory that the degree

of independence which the individual segments present is

comparatively slight. The most important organs (nervous system,

body musculature, blood-vascular system) show themselves to be

single fundaments of the entire body, and are also developed as

such even though they also exhibit evidences of metamerism. Even

the excretory canals may give up their segmental isolation and

become united to one another by means of longitudinal canals.”

(_Ib._ p. 348.)

On turning back to § 206, the reader will, I think, demur to the

assertion that the independence is “comparatively slight”; seeing

that, as in _Ctenodrilus_, a single segment sometimes becomes separate

and reproduces other segments to form a new series. Instead of

admitting that “the most important organs” “show themselves to be

single fundaments of the entire body,” it may be held, contrariwise,

that their original independence in each segment is masked only to

the degree involved by their co-operation as parts of a compound

organism. But chiefly I remark that when it is said that “the excretory

canals may give up their segmental isolation and become united” by

“longitudinal canals,” there is a clear confession that the isolation

of these organs was original and their union superinduced--an

implication that the course of evolution is as I have described it, and

at variance with the course of evolution assumed by Prof. Korschelt.

Yet another incongruity is involved in his interpretation. He writes:--

“Just as in the consideration of the tapeworm chain we were

induced by the comparison with unsegmented forms to refer the

entire chain to an unsegmented individual, and, on the other

hand, to see in the proglottis, not a complete individual, but

only the abstricted hinder portion of the body of the Cestode,

in the same manner, and with much more reason, we adhere to the

individuality of the Annelid body.” (P. 349.)

And then on the preceding page, referring to the composition of the

Annelid body, he says:--“The most natural comparisons are those

with the tapeworm chain and with the strobila of the Scyphomedusæ.”

Now since it is here assumed that the tapeworm and the strobila are

analogous in composition, it is implied that the detached proglottis

and the detached medusa are analogous; and hence if we are to regard

the proglottis as “not a complete individual but only the abstricted

hinder portion of the body of the Cestode,” then we must similarly

regard the medusa as not a complete individual, but only the abstricted

hinder portion of the strobila. This commits us to the strange

conclusion that whereas individuality is ascribed to the original

simple polyp, and by and by to the partially-segmented strobila, though

these are without special senses and with only rudiments of muscular

and nervous systems, individuality is denied to the detached medusa,

which has organs of sense, a distinct nervo-muscular system and a

considerable power of locomotion, as well as a generative system:

traits which in other cases characterize developed individuals. Here

also, then, there seems to be an inversion of the ordinary conception.

This conception of the proglottis and the medusa is, I see, accepted by

some as tenable. But if we accept it we must accept also an analogous

conception, which will I think be regarded as untenable. It is that

supplied by the _Aphides_. From an egg proceeds a series of sexless and

wingless females, and at the end of the series there come winged males

and females with resulting gamic reproduction. If instead of forming

a discrete series the imperfect females formed a concrete series,

the members of which could individually feed without being detached

from one another, as the segments of a tapeworm can, the parallelism

would be complete; and then, according to the view in question,

we should have to regard the perfect males and females eventually

arising, not as individuals but as terminal portions of the series,

containing generative products and having wings for the dispersion of

them--locomotive egg-bearing segments of the chain. Whoever espouses

this view must hold either that the first imperfect female of the

series was the individual or that the entire string of them constituted

the individual (in conformity with a view once propounded by Prof.

Huxley). But he must do more than this. Since the _Aphides_ have

descended from some winged species of the order _Hemiptera_, he must

hold that among those remote ancestors each particular fly, male or

female, was an individual; but that when abundant food and inert life

led to the partheno-genetic habit, and to chains of sexless forms, the

males and females eventually produced at the end of each chain, though,

like their remote ancestors, possessed of procreative organs and wings,

are not individuals.

[Some memoranda bearing on the question here discussed, mislaid at the

time when the chapter dealing with it was revised, have been discovered

in time for utilization in this appendix.]

One of my critics says:--

“You have overstated the case in your favour: the alimentary

canal does not, as you suggest, show a segmentation

corresponding to that of the other organs in Annelids. Either

it is a simple uniform tube, or else its differentiations

(pharynx, œsophagus, crop, intestine) are quite independent of

the repetition of the somites.”

In presence of statements made in works of authority, this objection

greatly surprises me. I meet with the descriptive word “moniliform”

applied to the intestine in some Annelids, and then in the Text Book of

Claus, translated and edited by Sedgwick, it is said, concerning the

alimentary canal in the _Annelida_:--

“This is followed by the gastric region of the gut, which

occupies the greatest portion of the length of the body, and

is either regularly constricted in correspondence with the

segments, or possesses lateral diverticula.” (P. 365.)

And again on p. 369 it is said:--

“The intestine usually preserves the same structure in its

entire length and is divided by regular constrictions into

a number of divisions or chambers, which correspond to the

segments and dilate again into lateral diverticula and cæca.”

The alimentary canal thus presents the segmental character as clearly

as consists with fulfilment of its function. If the successive segments

are co-operating units of a compound animal having but one mouth,

then, necessarily, the gut cannot be completely cut into parts, each

answering to a segment, for there could be, in that case, no passage

for the food. If the portion of the intestine belonging to each segment

has a conspicuous dilatation, or has a cæcum on each side, it exhibits

the segmental character as much as the physical requirements permit. So

far from being at variance with the hypothesis, its structure exhibits

a verification of it.

The next objection runs as follows:--