Marine Biological Laboratory

,.r..,.A J^y 51, 1941

Accession No. '^5622

Gven By "^^^ Hacnlllan Go.

Place He- rork City

THE COURSE OF EVOLUTION

CAMBRIDGE
UNIVERSITY PRESS

LONDON: BENTLEY HOUSE

NEW YORK, TORONTO, BOMBAY
CALCUTTA, madras: MACMILLAN

TOKYO : MARUZEN COMPANY, LTD

All rights reserved

THE

COURSE OF EVOLUTION

BY DIFFERENTIATION OR
DIVERGENT MUTATION RATHER THAN

BY SELECTION

by

J. C. WILLIS

M.A., Sc.D., Hon. Sc.D. (Harvard), F.R.S.

European Correspondent, late Director, Botanic
Gardens, Rio de Janeiro

CAMBRIDGE

AT THE UNIVERSITY PRESS

1940

PRINTED IN GREAT BRITAIN

CONTENTS

Preface

page vii

Chapter I. The Coming of the Darwinian Theory of

Natural Selection 1

II. Contacts with Darwinism (i). The Podo-

stemaceae g

III. Contacts with Darwinism (ii). Endemism,

Age and Area 24

IV. The Hollow Curve 33
V. Contacts with Darwinism (iii). Mutation 43

VI. Contacts with Darwinism (iv). Adaptation 52

VII. Isolation q\

VIII. Differentiation 65

IX. Divergence of Variation 74

X. Some Test Cases between the Rival Theories.

A. Numerical gg

XI. Some Test Cases between the Rival Theories.

B. Morphological I03

XII. Some Test Cases between the Rival Theories.

C. Taxonomic 132

XIII. Some Test Cases between the Rival Theories.

D. Geographical Distribution 142

XIV. General Discussion lg4
XV. Final Summary of Conclusions 191

Appendices I94

References to Literature 200

Index

203

OJ622

LIBRARY

PREFACE

/Vn accident in 1905, and the nature of my official occupation,
forced me to work that could be done in spare time with the
aid of a pen and a library, and since then I have largely devoted
myself to the study of geographical distribution. The dictionary
for which I was responsible emphasised in my mind the enormous
variety in sizes and distribution of families, genera, and species.
All seemed a nearly hopeless confusion. Yet this is not nature's
way; her work is always beautifully planned, as Darwin had
already shown in the wonderful theory of evolution, whose
establishment as a working guide through the intricacies of life
was due to him, and gave hiin his lasting claim to fame. Without
a mechanism to operate it, however, few were prepared to make
so great a break with what had gone before. In natural selection,
Darwin produced an apparently serviceable mechanism, which
was so familiar to every one that it had a great appeal, soon
resulting in the establishment of evolution in an unassailable
position. But during the last fifty years there has always been
an underlying feeling that all was not well -sWth natural selection.
The writer, though brought up in its strictest school, soon began
to feel very doubtful about it, and a few years of experience with
tropical vegetation made him realise that selection could not be
responsible for evolution. From that time onwards he has never
ceased to bring up objections to it, though rarely has any answer
to these been attempted. Selection is now no longer required as
a support for evolution, and must take its proper place, which is
one of great importance, as has been pointed out here and
elsewhere.

The writer then set out, some thirty-five years ago, to find
some definite laws underlying the welter of facts in distribution.
The first thing that really set him upon the track was the
discovery in 1912 of the "hollow curve" formed by the numbers
of species in the genera of the Ceylon flora, a curve which soon
proved to be universal in both floras and faunas. This led to the
development of the theory set out in Age and Area in 1922.
Being, among other things, a flat contradiction of the theory of
gradual adaptation through the agency of natural selection, this
theory of age and area was not accepted, but as the counter

viii PREFACE

arguments brought up mostly assumed that the older theory was
sound, the writer's faith remained unchanged, and he continued
to follow up his beliefs. They are now yielding interesting results,
of which the present book is one, while another, dealing with
distribution, and whichis perhaps even more subversive of current
opinions (used as a shelter for so much in national policies), is
upon the road to completion.

The present book, the logical sequence of Age and Area, has
been greatly delayed by various inconveniences, and by the great
quantity of statistical work required. This was so great a burden
that I can hardly sufficiently express my gratitude to my friend
Mr John Murray, late of the Indian Educational Department,
who undertook a great deal of it, and with his trained mathe-
matical skill was able to do it well and rapidly. I am also deeply
indebted for aid to Dr W. Robyns, Director of the Botanic
Garden at Brussels, Dr B. P. G. Hochreutiner, Director of that
at Geneva, and Sir Arthur Hill, Director of that at Kew, at all
of which places, and especially the first named, I have done
much work. My friend Mr G. Udny Yule has helped me very
greatly with criticism and assistance, and I am also much indebted
for help to Mr J. S. Bliss, Dr C. Balfour Stewart, and many
others.

J. C. WILLIS

LES TERRAGES

AVENUE DES ALPES
MONTREUX

25 March, 1940

tif |LI3RAR^

CHAPTER I

THE COMING OF THE DARWINIAN THEORY
OF NATURAL SELECTION

J\s a recent product of evolution, man must have arrived upon
the scene to find himself in a world that was already well pro-
vided with animals and with plants. Some animals would be
actively hostile and dangerous to him, some would be afraid of
him, some would be indifferent; some plants would be poisonous,
some good to eat or to provide useful materials, some indifferent.
Man would presumably inherit some notion of what to eat, and
how to obtain it, but it is clear that in his early days the struggle
for existence must have been severe, especially if one remembers
the prolonged infancy and helplessness of his offspring. He
probably had greater brain power, and may have had some
leanings towards co-operation, otherwise chiefly showTi by insects.
Whilst failure in the struggle for existence at the very beginning
would probably have meant his complete extermination, the risk
would lessen as he established himself in various different places
removed from that in which he probably began. It is an intriguing
thought that he may owe his first survival to having arisen in
some place not troubled by dangerous animals, or to some other
stroke of what seems like mere luck.

From verv earlv times he must have been struck by the bodily
likenesses of many of the organisms by which he was surrounded.
He would soon recognise the difference between the male and the
female of the same species, and he would distinguish, for example,
between the tiger, the leopard, and the cat, or between the wolf,
the fox, and the dog. He would see the evident likenesses that
run through these triads, and that it was greater between tiger
and cat than between tiger and dog. He would see other like-
nesses between goose, duck, and swan, between owl, eagle and
hawk, or yet again between lizard, snake, and crocodile. But he
would also notice that there were overriding distinctions among
these various animals — that both the cat group and the dog
group could be included in a greater group that we now call the
Mammals, the eagle group and the duck group in the greater
group of Birds, and so on. Thus there would grow up the notion
of groups within groups, which is the essence of all classification.

WED I

2 THE DARWINIAN THEORY [ch. i

The likenesses between plants are often less immediately
obvious, and as compared with animals they seem to have been
less remarked mitil a few centuries ago. One may see this lack of
observation upon the part of mankind in the common names
of plants, which are often old. Thus one often finds such names
as meadow-rue, marsh-marigold, rock-rose, sea-heath, wood-
sorrel, and the like, applied to plants that are in no way closely
related to the rue, the marigold, the rose, the heath, or the sorrel,
though they may have a superficial likeness in the leaves, in the
look of the flowers, in their colour, or in the taste. But at the
same time, one must also notice that many plants belonging to
the same families (as now recognised) have similar names. Thus
many Cruciferae, with their cress-like taste (cress itself is a
member of the family), have names like bitter-, penny-, rock-,
thale-, wart-, water-, winter-, and yellow-cress. The same taste,
however, occurring in the seeds of the garden Tropaeolum, that
plant used to be known as Indian cress, though it belonged to a
totally different family. This also illustrates the now familiar fact
that to place an organism in its proper relationships one must not
rely upon a single character only. The name vetch is common
among the British Leguminosae, and grass among the Gramineae,
though here again one finds members of other families, often
unrelated to the grasses, known as arrow-, cotton-, eel-, goose-,
knot-, scorpion-, scurvy-, and whitlow-grass, because of some
resemblance in habit, leaves, or other things.

Gradually the true likenesses of plants began to be recognised
to such an extent that they were grouped into species and
genera within families, and these again within larger groups,
especially by the work of Tournefort, Linnaeus, Jussieu, Brown,
Endlicher, and many others of more recent date, so that now
we have what is probably a reasonably good classification of
them.

Till about a century ago, the universally accepted view of the
origin of plants and animals was that they had been specially
created, each species in the form in which it now appears upon the
earth, whilst their varieties were formed later, as the areas
occupied by the species became larger or more varied. But it was
clear that though one might group together the buttercup family,
or the cat family, special creation would not explain, though it
made the need of explanation greater, why they should possess
such likenesses as caused them to be thus grouped together.
Since the time of Aristotle vague ideas had been floating about.

CH. I] OF NATURAL SELECTION 3

that such groups might owe their origin and their likeness to
descent from some common parent, accompanied by such modi-
fication in different directions that there would arise forms like
the wolf and the dog, or the apple and the pear, showing an
obvious family resemblance though differing in detail. But
owing to the lack of any mechanism that seemed in any way
capable of bringing it about, this idea of "evolution" was not
seriously taken up, except by a few like Lamarck and St Hilaire,
and never became what one may term practical politics until the
coming in 1859 of Charles Darwin's famous book ^'The Origin of
Species by means of Natural Selection, or the Preservation of
Favoured Races in the Struggle for Life", preceded, on 1 July
1858, by a joint paper by Darwin and by Alfred Russel Wallace,
an independent discoverer, read at the Linnean Society. Both
writers had been more or less inspired by reading Malthus (30)
to realise the struggle for existence that must always be going on
wherever living beings occur, a struggle which becomes the
fiercer the more that thev are crowded together, as for instance
at the birth of young, or of germination of seeds, for it is well
known that both animals and plants tend to produce more off-
spring than there is room for. Though by the aid of wind, water,
animals, etc. the seed may be scattered to some extent, the chief
crowd will always tend to be near together, and the great struggle
will be among the seedlings, rather than between them and the
parent, against which they will usually have but little chance.
As there will generallv be too manv seedlincrs for the available
space, the struggle will be severe, even if the competitors be
connected with the parent by an offshoot or runner. The survivors
will largely be chosen by chance, for early arrival on the spot, a
less shady or better watered position, a better or softer patch of
soil, and so on, will all be of greater advantage to the young
seedling than any advantage that it may carry in itself as com-
pared with its competitors of the same species, just as in the
human struggle for existence parental advantage, the right school
tie, etc. are of value. If it finds itself late in germination, upon
poor soil, in a place with insufficient water, and so on, natural
selection or competition will kill it out, inasmuch as it is unsuited
to the conditions with which it has met, even though it may be
suited well enough to what one may call the normal conditions of
the place. It may also be killed out if it be the offspring of parents
that have been used to somewhat different conditions, for it will
probably carry with it their suitability to conditions. The more

1-2

4 THE DARWINIAN THEORY [ch. i

like those from which it came that the conditions are, the better
chance will the young plant have, whereas if it come from some
distance, where the conditions are likely to be somewhat dif-
ferent, it will be more a matter of good luck should it succeed in
establishing itself in the new locality.

As competitors begin to decrease, the struggle for existence
will probably become somewhat less intense. When mature, the
struggle will be largely that to secure the most of any small space
for expansion of roots or of leaves that may become vacant.

Seeing a struggle like this, it seems natural to suppose that if
any of the youngsters possessed any character that might give it
any advantage against the rest, however slight, it would tend to
win in the struggle more often than not. It is a remarkable thing
that inasmuch as evolution is only clearly shown in structural
characters, and natural selection was trying to explain evolution,
it ignored the functional characters, and tried to explain the
structural ones. But of course if the functional characters had
been the only ones that were acted upon, there would have been
little to show that any evolution had gone on at all. There would
obviously be no need for all the structural differences.

Assuming that the advantageous character were inherited,
another plant might win in the next generation, and so on, the
character perhaps (another assumption) becoming more and
more marked in each generation until at last, when taken
together with other characters that had also varied (whether in
correlation with the first, or also under the influence of selection),
a specific difference was arrived at, and a new species would have
been formed. As this would have been formed by a definite
adjustment to the local conditions, it would be what is usually
called adapted to them ; this type of adaptation we shall call in
future structural adaptation, as it was in structure that the
changes were supposed to show that had brought the advantages
with them. As it would tend to be the unimproved offspring of
the old species, which retained its specific characters, that would
be defeated in the struggle for existence, the old species would
thus tend to decrease in numbers, being gradually reduced to the
rank of a small and local group of plants, which might be looked
upon as a relic of former vegetation, and which in time would die
out altogether. And, supposing the original species to be found
upon a considerable area, where there might be differences in
conditions between different parts, then it might vary in two or
more directions, giving rise to two or more species. In this case

CH. i] OF NATURAL SELECTION 5

the old species would tend to become discontinuous in its distri-
bution by being replaced in some of its area by the new ones.

On the face of it, this suggested mechanism for the carrying on
of evolution, to which Darwin gave the name of Natural Selection
("or the preservation of favoured races in the struggle for life")
seemed eminentlv reasonable, and one that could do the work
required. But the struggle was necessarily of each individual of a
species for itself alone, and if one individual showed a favourable
variation while its neighbours did not, the variation would soon
tend to be lost by crossing. This was shown by Fleeming
Jenkin (21) in a criticism which Darwin considered as the best
that was ever made of his work. It therefore became necessary to
stipulate for the same variation to appear in many more or less
adjacent individuals of the species, scattered as a rule over a
considerable area. Crossing would then be useful, rather than
injurious. This in turn meant that the variation must probably
have been controlled, directlv or indirectlv. bv the external con-
ditions, and these would most likely be those of climate or of soil,
for the biological conditions largely depend upon which particular
plants may happen to surround the individual concerned at any
given place.

Instead of an external force, there might of course have been
some compelling internal force which made a whole lot of indi-
viduals vary in the same way, and in this case one would certainly
expect all to vary. Whether the force were external or internal,
unless all varied alike over a considerable area, the advantage
would be lost by crossing. In either case, it is a little difficult to
see where natural selection got any leverage, for there would be
no competition between the new and the old, except at the margin
between them, where the new would in any case tend to be lost
by crossing. When Darwin gave way, as he was forced to do, to
this criticism of Fleeming Jenkin, the freedom of the natural
selection theory was really lost.

The struggle for existence, felt as it was in every community
and family, was such a commonplace of everyday life, that the
principle had a very great psychological appeal, and was soon
taken up on all hands. The long neglected theory of evolution
rose "in the attitude of claimant to the throne of the world of
thought, from the limbo of hated, and as many hoped, of for-
gotten things " (66). Rarely has any h\^othesis met with greater
success than did natural selection. A mechanism familiar to
everyone seemed able to operate the long wished for process of

6 THE DARWINIAN THEORY [ch. i

evolution. Every man felt, as Mrs Arber has said, that he was
one of those picked out by it, and so he felt it his duty to support
the theory. Though Darwin's immortal service was really the
establishment of evolution, the name Darwinism became
attached rather to the theory of natural selection, which became
a cult, and which now exercises enormous influence in the world
at large, even national policies being in some instances largely
tinged with it. This is another instance of the influence of the
dead hand, so well brought out by Woolf in After the Deluge,
chap. I.

Evolution itself is now so well established that it has no longer
any need whatever for any assistance or support from the hypo-
thesis of natural selection, and whether the latter be true or not
matters little or nothing. What we have to do is to follow up the
theory of evolution, and find out something more about its
working.

Natural selection was a new theory that was a complete
reversal of the old. Instead of being created suddenly, so that
at once thev showed all their differences, which are often con-
siderable, and usually more or less discontinuous, living beings
were formed gradually by the selection and accumulation of small
diff'erences that gave some advantage to their possessors in the
struggle for existence that was a daily commonplace of life.
Creation in its usual sense was replaced by evolution, and the
appearance of larger differences by the accumulation of smaller.
The family resemblances that were mentioned above were now
explained, thus removing to a period immensely farther back the
conception that the phenomena of the life of animals and plants
were pre-ordained, and throwing open to research a vast field of
knowledge.

With natural selection itself, time has dealt less kindly. It
acquired an immense prestige by its success in establishing
evolution, but has not proved so useful in the further advance of
science as was expected. It contains too many assumptions, and
has required too many supplementary hypotheses to enable it to
offer sure ground upon which to build, and is ceasing to be in-
voked as it used to be. It was at one time known as the doctrine
of "nature red in tooth and claw", and as such has become
largely incorporated into the theory of life that underlies the
general policies of the world.

At the time of its greatest success, a rival, pre-Darwinian,
system of evolution, known by the name of Differentiation, was

CH. i] OF NATURAL SELECTION 7

in process of development, and was being pushed by the famous
zoologists Owen and Mivart. It must be admitted that against
the psychological appeal of "Darwinism" it had no chance, but
at the same time, there was even then much truth underlying it,
and as time has gone on people are becoming more and more
inclined to think that in some respects at any rate it will give a
closer approach to the truth than will selection, the absolute need
for which as a support for evolution has now passed by. Special
creation went too far in one direction, natural selection in the
other, and differentiation may be called a kind of compromise.

CHAPTER II

CONTACTS WITH DARWINISM.
THE PODOSTEMACEAE

i-T is not intended here to write a history of the movement known
as Darwinism, but rather to sketch the author's contacts with it,
which have lasted for fifty years.

The pubh cation of the Origin of Species created a revolution in
the world of science, but like most great changes in ways of
thought it was very unwelcome to the older men, who rarely
came round so far as to accept it in any whole-hearted way. In
the next few vears there was a flood of ant i -Darwinian literature,
and many incisive criticisms were made upon natural selection
(rather than upon evolution) from one of which we quote the
next sentence: "It follows, therefore, that if we accept the
Evolutionists' view, every specialised chemical compound met
with in some living beings only must fulfil the condition, that
every approximation to the complete compound must have
been of advantage to the being in which it was produced in the
struggle for life. . .unless these very substances existed in, and
formed points of difference between, Mr Darwin's few original
forms" (29, p. 134). Maclaren also points out that change of
climate does not change the chemistry of a plant, so that there
is no opening for natural selection in a change of conditions.

It was clear that there must be discontinuity in evolution,
and this was difficult to harmonise with the view that it had
proceeded by gradual accumulation of minute steps. Chemical
substances of differing nature could not be formed from one
another by slow and gradual steps, nor could gradual steps in the
formation of such a substance as the green colouring matter of
plants (chlorophyll), for example, be of value. Yet this, probably
one of the early formed organic substances, providing the food
for plants and animals alike, ranks with water and protoplasm
among the most important chemical substances in the world.

The writer has been chiefly occupied with economic botany for
over forty years, and to him these considerations have long been
a fatal objection to the current theory of evolution — the gradual
passage, by reason of improving structural adaptation to the
surrounding conditions of life, from small variations through

CH. II] THE PODOSTEMACEAE 9

larger to well-marked varieties, to species, and to higher forms.
There is no inherent reason why economic botany should remain
what it now is, an ever-increasing mass of facts with little or no
co-ordination. What little of this there is, as may be seen at once
by consulting Wiesner's standard treatise, is very largely
confined to such observations as that a and h, belonging to the
same family, produce similar economic products. This alone
shows that the facts of economic botany must be explicable upon
evolutionary lines. Yet, with the exception of the theory that
poisonous plants have evolved the poison as a protection against
animals, natural selection has never attempted to explain any-
thing in the realm of economic botany, which ought by this time
to be a properly classified scientific discipline, with general
principles running through it. One chemical fact must follow
from another.

Something the same may be said of geographical distribution,
which has been a favourite study of the author for the last thirty-
five years. This again consists of a stupendous mass of facts,
connected together by little more than a tissue of speculation.
Sir Joseph Hooker, its great leader of former days, wrote: "All
seem to dread the making Botanical Geography too exact a
science; they find it far easier to speculate than to employ the
inductive process", and the position is not so very different
even yet. It has always been admitted that any theory of the
mechanism of evolution must stand or fall according to whether
it can or cannot interpret the facts of distribution. The two are
obviously and inextricably bound together and to them should
be added the facts of economic botany.

At first natural selection seemed to offer an explanation of
these geographical facts, indeed so promising an explanation
that Hooker became one of Darwin's chief lieutenants, never
following out to their conclusions some of the lines of work upon
which he had begun. Gradually, however, it was discovered
that the employment of natural selection was not leading to real
advance, and the first enthusiasm died away, leaving distribution
in the Cinderella-like position that it still occupies. Those who
had leanings in the direction of distributional study turned more
and more to the rising science of ecology, known as natural history
of plants when the author taught the beginnings of it under Sir
Francis Darwin in 1891-4. But though ecology is all-important
for the details of local distribution, it cannot answer the wide
questions which are the province of geographical distribution

10 CONTACTS WITH DARWINISM I [ch. ii

properly so-called. Some new theoretical background is required,
other than natural selection, which has proved a very broken
reed upon which to lean.

Those who have tried to make evolution work upon Darwinian
lines, i.e. in the "upward " direction from minute variety through
variety and species, and so on, have met with continually
increasing difficulties, with some of which we now propose to deal.

For the variations that were ultimately to form the basis of
new species, Darwin relied principally upon the "infinitesimal"
or continuous variation that was well kno\Mi always to be going
on in every possible character. Thus, supposing one measured
the length of 500 leaves from similar plants of the same species,
one might find the average to be 25 mm. The greatest single
number would probably be found to show this length, but there
would be almost, if not quite, as many measuring 24 or 26 mm.,
somewhat fewer for 23 and 27 mm., and falling away more and
more quickly, but at about the same rate on either side. Investi-
gation gradually showed that there were definite limits to this
kind of variation. It follows the ordinary curve of frequency
distribution. If one cross two individuals both having a very
high degree of the character, the average of their offspring does
not retain that high level, but falls back, or regresses. The high
level can only be maintained by strenuous selection in each
generation. Further, it is also found by experience that one
cannot, by means of selection, pass a certain maximum. This
kind of variation, in other words, is not fully hereditary nor is it
irreversible, like the differences that characterise species, and
cannot be indefinitely added up without some external aid. The
experiences of sugar beet and other breeding show this well
enough; never can one go beyond a certain point unless, by
hybridisation or in other ways, one introduces new factors. In
the struggle for existence, mere chance has much too large a share
in determining the victors to allow even the maximum to be
reached. Thus, on this ground alone, this type of variation was
disqualified as forming an essential part of the evolutionary
mechanism.

But this is not the only difficulty that arises in trying to use
this kind of variation, which is always linear, or up-and-down.
A leaf may vary infinitesimally in length, or in breadth, in the
depth of its incisions, or in the degree of number and length of its
hairs, but it does not vary except in sudden steps in such a
direction as that from alternate to opposite, from simple to com-

CH. II] THE PODOSTEMACEAE 11

pound, from pinnate to palmate, from dorsiventral (facing
upwards, with different anatomy on the two sides) to isobilateral
(facing sideways, hke Gladiolus leaves, with the same anatomy
on both sides), from parallel veined to net veined, or in other ways
that could be mentioned. Now variations in length and breadth
are rarely of much importance for distinction of species, unless so
great that there is a wide difference between the averages in the
two cases, while the other characters that have just been men-
tioned will be seen at once to be such as are of great importance
in distinction between one species and another. This is another
fatal objection to the use of this kind of variation as part of the
mechanism of evolution. Some kind of variation was required
that was not only inherited and irreversible, but also differen-
tiating and not merely linear, or up-and-down.

Another serious difficulty was the fact that species were very
rarely distinguished from one another by a single character only.
Usually there were from two to six characters marking them off
from one another, some of them more variable than the rest, and
more liable to overlap from one species to the other, so that one
had to examine a great number of specimens of each of the
species to be sure that their overlap was not due simply to lack
of real difference. Thus in Cornus, to take the first genus that
comes to hand, C. kousa and C. capitata are closely allied species.
The key division is that the former has the calyx truncate, the
latter 4-lobed, and the involucral bracts more or less ovate as
against obovate. But there are so many minor points of difference
that the description of either takes up nearly twenty lines (49).
None of the characters afford any opening for natural selection
to work upon, so far as can be seen, but supposing that it had
worked upon one, were all the rest simply correlations? One could
hardly imagine it working upon one at a time, for what would
ensure that a should be followed by b, which was unconnected
with it (as are the two distinguishing characters above quoted)?
Nor could one imagine it picking out a variation that included a
little of each of a, b, c, d, etc., when these were unconnected,
unless they were in some way correlated.^ But if correlation were
to be invoked to this extent, it must be the principal, though
perhaps only passive, factor in evolution as sho^\Ti by the
characters that distinguish its finished product. Nearly all the

1 Cf. Origin of Species, chap, vii, first few pages, for remarks upon this
subject. Incidentally Darwin there suggests the "somewhere" which has
proved such a useful refuge to the defender of natural selection.

12 CONTACTS WITH DARWINISM I [ch. ii

characters must be correlations. And why did one not find, in the
fossil records, any species that had been fossilised before this
complicated process had been completed?

It is clear that, on the theory of gradual adaptation, a very
long time must be allowed to get from one species to another.
This means that the change of conditions must go on for a long
time also, for if a small change in structure enabled the species
growing in one locality to survive there, there would be no urgent
reason why they should continue to vary in the same direction,
unless the conditions also continued to vary in the same direction
as that in which they had begun to do so.

Another difficulty was to understand why variations of this
kind should usually go so far as to pass what one may call the
rough-and-ready line of distinction between species — that they
should be, mutuallv, more or less sterile.

One does not find to anv serious extent in the fossil record,
species which represent real intermediates between existing or
fossil species. One finds rather examples of species that have
some of the characters of one, some of another. But one does not
find species (as from the constant occurrence of the few characters
side by side in existing species one might expect to do) that show
intermediate characters between alternate and opposite leaves,
between palmate and pinnate leaves, between erect and climbing
stems, between racemose and cymose inflorescences, between
flowers with and without a cyclic perianth, between isomerous
and heteromerous flowers, between imbricate, valvate, and con-
volute aestivation, between flowers with the odd sepal posterior
and with it anterior, between stamens in one and in more whorls,
between anthers opening by splitting or by teeth, valves, or pores,
between 3-locular and 4-locular ovarv, between ventral and dorsal
raphe, between loculicidal and septicidal fruits, and so on through
all the important structural characters.

All these were very serious difficulties, while it had also to be
remembered that in any case evolution could only go on if the
needful variations in the right direction should appear, for, unless
this should happen, it was evident that natural selection could
do nothing. One could not imagine the "mixed" variation of
characters a, &, c, etc., above-mentioned appearing at all, unless
most of it was simply correlation, and if the differences had to
appear one by one, the chance of all appearing was but small, and
the time required would be enormous. Forty years ago it was
clear to the writer that some form of sudden and irreversible

CH. II] THE PODOSTEMACEAE 13

variation was required, such as was supplied by de Vries' theory
of mutation (48).

Evolution by gradual variation thus has many difficulties in
its path, which in the first enthusiasm of natural selection were
passed over with little notice. Under the influence of the criticism
of Fleeming Jenkin, it had to be admitted that all the new plants
of a considerable area must vary more or less in the same direc-
tion to prevent the new variation from being lost by crossing. It
would be lost at the edge of its territory, but would presumably
survive in the middle. The area of the parent species would thus
tend to become more or less discontinuous. It had to be assumed
that the parent did not vary in a favourable direction also, but
as all variation was assumed to be structural (it could hardly
be otherwise, as natural selection was trying to explain a
structural evolution), it was easy to suppose that the parent could
not vary in such a way. It also had to be assumed that the con-
ditions continued to change for a very long time, to such an
extent anyhow as to pass the sterility line, or a new species could
not be formed. This new species would evidently be well adapted
to the new conditions whose existence was responsible for its
coming into being, but it had also to be assumed that when
formed, or partly formed, it would then prove so suited to the
region in which the parent was still supreme as to kill out the
latter there also. This was a pure assumption, but was necessary
in order to explain the spread of the newer and better-adapted
species, which in turn was to explain their wide distribution. We
have shown in Age and Area, p. 34, that the older species will
probably gain continually upon the younger in rate of dispersal,
supposing, which seems to be the case, that there is no reason
(when they are taken in groups) why one should spread more
rapidly than another nearly related to it. If the area to which the
new species was ultimately to reach were very large, it was really
rather absurd to talk of it as adapted to the whole area. It must
have been just a case of luck that it proved so sufficiently suited
to far-away places as to be able to establish itself there, though
once arrived it would begin to suit itself in detail to the local
conditions. And it must not be forgotten that early species would
have the best chance both of rapid travel and easy settlement.

Finally, among the difficulties of Darwinism, it was evident
that the variations must be such that natural selection could
work upon them when they did appear, and as to that we have
but little evidence.

14 CONTACTS WITH DARWINISM I [ch. ii

The hypothesis of evolution by small variation has never, so to
speak, been officially abandoned, but it has been so altered by
supplementary hypotheses that it is hardly recognisable, and the
theory of mutation, brought up by de Vries, has largely taken its
place. A mutation, which when obvious is often called a sport,
at once produces a morphological or structural character or
characters that are definitely distinct from those which were
found in the parent form, and not only that, but which have come
to stay, and are (practically) irreversible. It is always possible,
of course, though not very probable, that some later mutation
may change them, or some of them, back again, or to something
else. Here, then, was a hypothesis that surmounted the chief
difficulties mentioned above, and provided hereditary variations
that were differentiating and (practically) irreversible.

Mutation was taken up, though slowly, as people gradually
realised the fatal nature of the objections to linear and infinitesi-
mal variations. Unfortunately for its speedy success, some doubt
was thrown upon the genuinely mutational nature of the pheno-
mena upon which it based. Some, at any rate, appeared to have
been due to hybridisation. But in spite of this setback, mutation
had come to stay, and we shall trace some of its history below.
People say that a sport is not capable of succeeding by itself, but
we do not know what would happen if it were really viable, and
plenty of time were allowed.

Natural selection was, of course, essentially a theory of gradual,
progressive, and more or less continuous adaptation to sur-
rounding conditions. It is evident that living things are suited
to them, for if they were not they would soon be killed out in the
struggle for existence. Some theory that will explain adaptation
is, therefore, very desirable. It was largely because it seemed so
capable of doing this that natural selection was so enthusiastically
taken up.

Each new species was formed, according to Darwin, because it
was an adaptational improvement upon its immediate ancestor.
Once this was fully realised, there was a great rush into the study
of adaptation. It was taken for granted (it could hardly be
otherwise) that as natural selection was trying to explain evolu-
tion, which showed itself mainly in external structural characters,
these characters must also, of necessity, be the means of expres-
sion of adaptation. Evolution has undoubtedly gone on in
morphological change, but as yet we are practically without any
proof that the change also represents the adaptation that may

CH. II] THE PODOSTEMACEAE 15

have gone on. What natural selection undoubtedly does is to
work with the individual, and to kill out, upon the whole, those
individuals that are below the average in any species— man or
animal or plant — but we have no proof that it works in the same
way with species as a whole or as units, killing out one species or
variety to make room for another, unless in particular conditions
which are more or less local. A species a may be killed out in one
place, because of unsuitable local conditions, whilst its rival h
may be killed out in another, for the same reason. If structural
differences go for anything, there must be a great adaptational
difference between the Dicotyledons and the Monocotyledons,
yet both grow intermingled almost everywhere, and in much the
same proportions. There is no " monocotyledonous " mode of life
that suits a Monocotyledon better than a Dicotyledon, yet there
are very great structural differences between them.

During this period, the possibility of internal, functional, or
physiological adaptation was ignored. Yet adaptation has far
more to do with the physiological than with the morphological
characters, if indeed it has anything to do with the great bulk of
these. There are very few external characters to which one can
point as definitely physiological. The leaves, roots, stems,
flowers, and fruit are so to a great extent, but not differences in
these (such as palmate or pinnate leaves, or drupes and berries),
except rarely. Adaptation to climate, which is a physiological
difference between one form and another, is primarily a purely
internal adaptation. To have any chance of survival, a species
must be suited to a greater range of climate than that with which
it perhaps began. As it migrates into new territory, it will
probably begin to become adapted to the slight changes with
which it may meet as it moves with (usually) very great slowness
into slightly differing conditions.

A vast amount of energy was put into the study of adaptation
during the last quarter of last century, and the imagination was
pushed to the extreme limit to find some kind of adaptational
value in even the least important features of plants, such as a
few hairs in the mouth of a corolla, an unpleasant smell (to some
human beings), and innumerable other characters (cf. books of
this period, such as 23). Unfortunately for the adaptationist and
for the theory of natural selection, which was founded upon
adaptation, no one was ever able to show that the important
morphological features of plants, which showed so conspicuously
in the characters that marked families, tribes, genera, and most

16 CONTACTS WITH DARWINISM I [ch. ii

often also the species, had any adaptational value whatever, and
the higher that one went in the scale, from species upwards, the
more difficult was it to find such a value. This, when one comes to
think it over, is really a very puzzling fact — why should the
differences become larger the higher one goes? Is the struggle for
existence greater among the higher groups, between two families
for example, than between two species, and between these than
between two individuals? A glance at the table of family
characters, given as Appendix i, will illustrate this.

This list of the important characters that distinguish families
from one another is after all not so very large. Each family has
something to show under most heads. In any pair of allied
families that may be taken, there will be mutual agreement in
many characters, but a contrasting difference in others, one
character of a pair being taken in preference to the other, and
that character tending to be shown right through the family,
though there are nearly always exceptions in the larger families,
the number of exceptions tending to rise with the size of the
family. Most of the pairs of characters that are given are such
that they do not admit of intermediates, and this divergence of
variation, as it is called, is constantly to be found in nature
between organisms that are so alike in most of their characters
that they are evidently allied in descent. Divergent differences
may show between one species and the next, between one genus
and the next, as with the berry-fruited Cucubalus in Caryo-
phyllaceae, between one tribe, sub-family, or family and the
next. As one goes downwards in the scale from family characters,
one finds more and more characters coming into use, but they can
still very often be arranged in divergent pairs.

One does not find (usually it is impossible) intermediates be-
tween the two characters of a pair, except in a few like superior
and inferior ovary, where semi-inferior is possible. But to imagine
intermediates between alternate and opposite leaves, or between
most of the pairs given, is to ask too much of selection. These
characters must, one would imagine, be the result of some sudden
change, which would give one or the other.

The individual characters are so divergent from one another in
each pair that it is clear, as in fact has long been well enough
known, that variation is definitely divergent. This was always, as
Guppy has said, a worry to Darwin, for it was extraordinarily
difficult to understand how an evolution, working "upwards"
through the variety and species, could drop out at each stage the

CH. II]. THE PODOSTEMACEAE 17

organisms necessary to make the divergence show more and more
as one went up in the scale. As Guppy points out in Age and
Area, p. 104, Hooker was definitely considering the idea or
nucleus of a theory of differentiation (19, ii, 306) but "no induc-
tive process based on Darwin's lines could have found its goal in
a theory of centrifugal variation. . . . Huxley was in the same case.
For he held views of the general differentiation of types, and his
road that would lead to the discovery of the causes of evolution
started from the Darwinian position. That road was barred to
him."

One cannot conceive of any of these family differences being
formed under the influence of natural selection. One cannot even
suggest, in any single case, which of the two characters is the
earlier, or what advantage can be gained by one as against the
other, or as against any possible intermediate, if such a thing
could exist at all. One must also remember, in dealing with
natural selection, that there must have been an enormous de-
struction of intermediates, of which we find no fossil record of
any note.

The supporters of natural selection mostly (at present, that is,
for they are apt to change over to the reverse explanation, that of
local adaptation) look upon the small and local genera and species
which occur in such great numbers, as being the losers in the
struggle for existence, i.e. the relics of a former vegetation, now
upon the way to extinction. A very remarkable thing about these
relics, which they do not attempt to explain, is that they do not
occur, except very rarely, in two or more different localities, with
a wide separation. For example, there are hardly any cases
known where they occur in two different continents, and few
where thev are found in the interior of two different countries on
one continent. Nor do the great majority of them belong to small
and isolated genera, but to the large genera (cf. p. 26), which
natural selection regards as the successes. The "relics" therefore
must have belonged to ancestral species which must have been
widely distributed to give rise to their present descendants. Why
then, when in one or more regions of slightly different conditions
new species were developed, did not the old species become dis-
continuous in its distribution, leaving relics in several different
places?

Under the natural selection theory, the large genera in the big
families, like Senecio, Ranunculus, or Poa, are supposed to have
been the best adapted and therefore the most successful. But

WED 2

18 CONTACTS WITH DARWINISM I [ch. ii

they are worldwide in their distribution, which must therefore
have gone on in early times. Has natural selection been gradually
diminishing in its effects?

The characters given in the " family " list are very important in
the distinctions between families, but they also appear very
frequently in distinctions between tribes, less often between
allied genera, and still less often between two allied species. It is
evident, therefore, that they can be rapidly produced, and do not
necessarily need a long and gradual evolution from species up-
wards. It is difficult to see how this can be so, unless they can be
the subject of single sudden changes, which as they are usually
divergent is not difficult to imagine.

It is very difficult to apply the Darwinian explanation, that
distribution is due to superior adaptation, to a genus like Senecio,
for most of its species are, compared to the genus, quite local. If
there be any marvellous adaptation, then, to account for the
enormous distribution, it must be generic, and no one has ever
been able to make even a suggestion as to what it may be, or
wherein it is shown. The generic characters are purely morpho-
logical, with less functional adaptation even than the specific.

The writer, as personal assistant to that best and kindest of
men. Sir Francis Darwin, who had helped his father in so much of
his work, was, of course, brought up in the arcana of natural
selection, and accepted it with enthusiasm. His first research was
upon adaptational lines, but he was not satisfied with the adap-
tational explanation of things, and when soon afterwards, in
1896, he went to Ceylon to succeed Dr Trimen, his views under-
went a complete change. The leisure time of the first six years was
devoted to a detailed study in both Ceylon and India of that
remarkable family of water plants the Podostemaceae (51-55),
containing about forty genera with 160 species, found in all the
tropics, with overlap into cooler regions. All live upon the same
substratum of water-worn rock (or anything firm, like timber,
that may be caught in the rock) in rapidly flowing water. They
are annuals, flowering immediately that the spathe comes above
water in the dry season, and then dying. If accidentally laid bare
by an unusual fall of water in the vegetative season, they soon
die without flowering. All the food comes from the water, and
they have no competition for place, except among themselves.
Enormous quantities of minute seed are produced, which have no
adaptation at all (except in Farmeria) for clinging to their place
in the swift current. At most one in a thousand or two may be

CH. II] THE PODOSTEMACEAE 19

caught in some fragment of old plant, or in some other place
where it can germinate.

At the period when this study was undertaken, the Podo-
stemaceae, with their strange look of lichens or seaweeds, their
peculiar mode of growth, their great variety of form, were looked
upon as obviously showing adaptation in the highest degree, and
it was for this reason that the work was undertaken. But among
the conclusions drawn from it was this, that apart from those
adaptations which they showed in common with all water plants,
such as the lack of strengthening tissues and of stomata, there
was in them little evidence of any special adaptation whatever.
The conditions under which they lived were the most uniform
that it was possible to conceive — the same mode of life, no com-
petition with other forms of life, the same substratum, the same
light (varying from day to day with the depth of the water), the
same temperatures, the same food, everything the same. Yet in
spite of this, the plants showed an enormous variety of form,
greater than that of any other family of flowering plants what-
soever, while water plants as a rule show little variety in form,
and have but few genera and species. Still more remarkable was
it that their morphology diff'ered for each continent, flattened
roots in the Old World, flattened shoots in the New, so that it was
usually possible to say by a simple inspection what was the
probable habitat of a species never seen before. It was hard to
believe that natural selection, working upon structural modifica-
tions that have never been shown to have any functional value,
could do this. The linking genus, Podostemon itself, covers an
immense area, including that of many of the smaller genera, and
is less dorsiventral than they are, though all show a highly dorsi-
ventral flower, which stands erect, and is commonly wind
pollinated, an unexpected combination of characters for the
selectionist to explain.

Once these very remarkable facts were fully realised, the
explanation that seemed much the most probable was that on the
whole the highly dorsiventral genera were descended from Podo-
stemon or from some form like it. It could not be the other way
about, for the flowering plants as a rule are not dorsiventral,
except in the structure of the leaf, and very often in the flower.
Nor could there have been some intermediate form, for that
would have had to be more dorsiventral than are the flowering
plants in general. One only of the local genera, Willisia in the
Anamalai mountains in South India, shows as much ordinary

2-2

20 CONTACTS WITH DARWINISM I [ch. ii

symmetry in its shoots as Podostemon, and as in that genus they
grow adventitiously from a creeping root.

The plants of this family grow in conditions of uniformity that
can hardly be matched in any other flowering plants, but
amongst them is included the uniform action of a force which
cannot be escaped. Growing as they do, always upon smooth
water-worn rock, they cannot send their roots into the substratum,
so that the normal polarity of the young plant, which sends its
root down and its shoot up, is completely disturbed. By no con-
tortions can the plants grow normally, though the rock may be of
any kind of slope.

There was no evidence to be found that would show that
natural selection had anything to do with the multiplicity of form
in these plants, for all were growing under the same conditions ;
but there was always this inescapable force urging dorsiventrality .
Under these circumstances, though he had started out with great
faith in adaptation and natural selection, the author became a
convert to the theory of mutational origin of species, adopting
from the very first the view that mutations or sudden steps might
at times be large enough to form species at one stroke. There were
no signs of real intermediates, yet surely here if anywhere they
might have been expected. An ordinary plant of another family,
growing more or less vertically upwards, would not usually come
under the continual influence of any powerful agent which would
tend to make its mutations go in any particular direction, but
with the Podostemaceae they were always being pushed in the
direction of dorsiventrality by the maximum force that nature
was capable of exercising in that direction. The mutations of
ordinary plants would give rise to specific differences in which
one could see no result of any particular directing force — there
was little or nothing to choose between them, and they were
morphological differences, with no adaptational value. In the
Podostemaceae, on the other hand, the mutations showed the
result of the continual force that was acting upon them, in a
dorsiventrality that on the whole tended to be continually more
and more marked the more local the genus might be. But it was
only an adaptation in the sense that moving restlessly in bed
might be described as adapting oneself to wearing pyjamas of
the fabric of which hairshirts were made. The dorsiventrality
was simply a morphological feature which had been forced upon
the plants. Upon this view, the difference in morphology between
the American and the Asiatic forms was also easily accounted for

CH. II] THE PODOSTEMACEAE 21

by some small difference in morphology between the first parents
in the two countries, which had the effect of urging the first
mutations in somewhat different directions. It is, of course, true
that natural selection might do the same with the same start, but
it is not quite so easy to imagine.

Any member of the family seems to be able to live without anv
great difficulty where any other member can live (53, p. 535)
though probably they have some preference as to speed of water,
and one must remember that in any case this varies with the
level of the river, being usually faster the higher the level.
People who came with me to look at the Podostemaceae growing
in the river near Peradeniya, when they saw the flat, closely
adherent Lazvias or Hydrohryums, used to say "obvious adapta-
tions to escape being carried away by the fast water". But in
Brazil the comparatively enormous Moureras and other forms,
3 or 4 ft. long, yet attached only at one end, lived in water that
was going at twice the speed of that in Ceylon.

Much or most of the evolution that had gone on, therefore,
seemed to be completely de luxe, for there was no need for the
new forms, nor was there any adaptational niche that would suit
one form onlv. and not also many others. It would almost seem
as if, in cases like this, if not perhaps in most, evolution must go
on, whether there be any adaptational reason for it, or not.

The explanation of the distribution of the Podostemaceae,
as given by current theories based upon natural selection, en-
counters some awkward difficulties. The most highly dorsiventral
forms are the most local, i.e. they are "the relics of previous
vegetation, defeated by the more widely distributed ones". In
other words, the family began with extreme dorsiventrality, and
then, so to speak, repented of it to some extent. But to become
less dorsiventral under the constant and utmost influence of a
force that is urging movement in the opposite direction, can
hardly be looked upon as likely to happen under the influence of
natural selection and the whole situation becomes an impasse.

The phenomena sho^^^l by the Podostemaceae are almost
exactlv matched in the allied family Tristichaceae, which has
much the same distribution, and are also matched by the pheno-
mena shown by the most completely parasitic plants, such as
Rafflesiaceae or many fungi, which, though they grow in mar-
vellously uniform conditions, none the less show important
structural differences.

The universality of this type of distribution, with the more

22 CONTACTS WITH DARWINISM I [ch. ii

primitive genera the more widely distributed, and the most
highly modified the most local, taken together with other features
shown by the Podostemaceae, made the writer realise that in
trying to work evolution from the variety — which upon the
theory of natural selection was an incipient species — upwards to
species and further, we were trying to work it backwards. Once
this fact had been fully grasped, as it was about thirty-five years
ago, the theory of natural selection became for him a theory
which in its youth had done a marvellous piece of work, but had
exhausted itself in that effort, and was not likely to lead to any
further serious advances, as indeed had already been shown in its
breakdown in the study of adaptation in the last quarter of the
nineteenth century.

During the six years that this work occupied, the writer had
frequent opportunities of visiting the tropical forest, and soon
realised that the struggle for existence was mainly among the
seedlings that tried to commence life upon any small spot upon
which, owing to fall of a tree, the breaking off of a branch, or for
other reason, there was rather more light than usual. But most
of the seedlings were of differing species, and commonly also of
different genera. And as never twice would the same assortment
of seedlings have to be encountered, and never twice the same
conditions of weather, it was impossible to see how slight varia-
tions towards adaptational advantage could be of any use. Mere
chance, as we have already pointed out (p. 3), must evidently be
the chief factor in determining the survivors. Ecological adapta-
tion to slight climatic and other changes must evidently be
internal rather than external. It was possible, as Harland has
suggested, that slight changes of this kind might entail some genie
change, and these, when added up over long periods, might give
rise to morphological mutations. But this has little or nothing
to do with the straightforward natural selection that was
normally accepted, and in any case is working downwards from
above, as does differentiation.

The fiercest struggle for existence that a plant is ever likely
to encounter is that into which it must be thrown at its birth,
when it will have to compete with other seedlings upon land
already very fully occupied. Any form that is not adapted to the
conditions in which it finds itself at that time will be remorselessly
killed out, unless the time is short, hy reason of its unsuitahility ^
and that is what natural selection really means. Anything that
is in any way handicapped — by unsuitability to the conditions,

CH. II] THE PODOSTEMACEAE 23

by anj^thing unfavourable in the spot upon which it is trying to
grow, by mere late arrival as compared with its competitors, and
so on, cannot survive in such a struggle, unless the handicap
imposed by one thing is compensated by a start in some other.
The actual winner or winners will be mainly picked out by chance,
and will in all probability be derived from parents that are
already living somewhere close by, and which may therefore be
looked upon as already adapted to the climate and other condi-
tions. In all probability this adaptation will be to a reasonably
large range of temperature and other climatic conditions, for
unless this were so, survival would be very improbable in most
places. There is also reason to suppose, that if it be done slowly
enough, a species may, as it moves slowly about the world,
become slowly acclimatised to other conditions, for the range of
some species is so enormous, and includes such varied condi-
tions, that without some possibility of this kind it is difficult to
understand.

CHAPTER III

CONTACTS WITH DARWINISM, continued.
ENDEMISM, AGE AND AREA

JtIaving by this time (1902) completely thrown over natural
selection as the chief mechanism of evolution, the author's next
piece of work was a study of the remarkable flora of Ritigala
mountain, lying isolated in the flat "dry" zone of Ce^don, in
which little or no rain falls for the almost six months of
the southwest monsoon. A note on the flora had already been
published by Trimen (45). The mountain, over 2500 ft. high, falls
with a steep cliff to face the south-west wind, and the summit,
of but a few acres, receives rain during that monsoon, thus
forming an outlier of the "wet" zone flora, which otherwise only
begins upon the mountains about 40 miles away to the south.

The flora of Ritigala summit, of over 100 species, contains one
or two which are quite local to it, or endemic, in the botanical
sense. The rest of the plants are largely to be found in the wet
zone, but not in the intermediate country, which is at a much
lower elevation, and is shown by geological evidence probably to
have been dry since the Tertiary period.

Endemism, about which the writer has published a good deal of
work, is, it is hardly too much to say, a crucial feature upon whose
proper explanation largely hangs niuch of the whole matter of
evolution and of geographical distribution. The best known
endemic of Ritigala is Coleus elongatus Trim. (46, and Plate 74),
easily distinguished by having a calyx of five equal sepals instead
of one of two lips, and by having a pendulous cymose inflorescence
of five stalked flowers, in place of the sessile bunch of five flowers
that is the usual thing in Labiatae. There also occurs upon the
summit the closely related C. barbatns, widely distributed in
tropical Asia and Africa, and upon the natural selection theory
the most "successful" of all the Colei, but here growing together
with C. elongatus the most "unsuccessful", and in the same way,
upon open rocky places. Why was this so, upon the hypothesis of
natural selection? No satisfactory answer could be given by its
supporters, and they were obliged to bring in two supplementary
hypotheses, which were mutually contradictory. Some said that

CH. Ill] ENDEMISM, AGE AND AREA 25

C. elongatus was a local adaptation, i.e. a success, but if so, why-
did it not have a different habitat from C. harhatust Others
offered the reverse explanation, and said that it was a relic of
previous vegetation, i.e. a failure. But again, why did it continue
to grow in the same places as C. harhatus, the most widespread
and successful of the Coleil Why was it not killed out? And why
was it morphologically distinct from all other Colei, with a few
exceptions in Africa? Had a pendulous inflorescence with stalked
flowers given rise to a normal Labiate one, which otherwise
characterises much of this genus of 150 species of tropical Asia
and Africa ? And how did the calyx change from five equal teeth
to two lips, one presenting four teeth, one one? Regular variation
in a calyx would always affect the teeth equally; a two-lipped
condition could only be the result of some sudden change. The
final refuge of the natural selectionist is usually to say that the
peculiarities must have been useful at some other time, or at
some other place. But the conditions upon the summit of Riti-
gala, and in all probability in the country between it and the wet
zone, had not altered since the Tertiary, and there was no sign of
C elongatus anywhere else, while its most successful and closely
related rival, C. harhatus^ was upon the same summit, in similar
places, and about equally common. Neither of the diametrically
opposed solutions offered by the natural selectionists would hold
water, especially as no adaptational value could possibly be read
into either inflorescence or calyx, whereas the problem was easily
solved by imagining C. elongatus to have arisen by a single
mutation from C. harhatus. And why was there another endemic
in the mountain mass of the wet zone also? Was it a case of
isolation resulting in a new species upon Ritigala? This was the
only probable explanation other than that of mutation which has
been offered, and as the wet-zone endemic has neither the equal
sepals nor the pendulous inflorescence, marked mutation must
have gone on. There was no opening for natural selection, even
could it have produced such differences among the few dozen
plants of both species upon the summit of Ritigala. It was also
clear that upon the principles of natural selection, as altered by
Darwin after the destructive criticisms of Fleeming Jenkin (21),
there was not room enough upon the summit of Ritigala to allow
of the development of even one endemic, to say nothing of two or
three, or of the surprising fact that the most common and wide-
spread species of Coleus was also living there with the local
endemic, and in the same or similar places.

26 CONTACTS WITH DARWINISM II [ch. m

This work fully confirmed the author's doubts concerning the
efficacy of natural selection, and the weakness of the explanations
that were put forward in its name. He also became interested in
endemics for their own sake, for it was becoming evident that
upon a correct explanation of them depended much of the proper
understanding of what had gone on in the course of the evolution
and geographical distribution of plants that had occurred in the
earth's past history.

Work upon endemism has been continued ever since the first
experience upon Ritigala, and has led to many interesting results,
many of which were published in a book upon Age and Area in
1922, and others of which it is hoped to publish in another book
dealing with Geographical Distribution only. One of the first
interesting points to come out was the very great number that
were confined each to one (or more rarely to two or more) of the
mountain summits of Ceylon (57). It was shown that over a
hundred species were confined to one or more hill-tops. Thus the
large tropical genus Eugenia showed E. Fergusonii and E. aprica
in the mountains north-east of Kandy, E. cyclophylla and a
variety of E. Fergusonii upon Adam's Peak, E. phillyraeoides
upon Kalupahanakanda, E. pedunculatus in the Rangala moun-
tains, and E. rotundifolia and E. sclerophylla upon the peaks
above 6000 ft. The mountains, all rising from a plateau, thus had
eight peculiar Eugenias, which one could not figure as being
refugees from the plains by way of the plateau (an explanation
sometimes advanced). They also contained six endemic Hedyotis,
ten Strobilanthes, four Atiaphalis, and so on. Plants like this are
usually supposed to be relics of previous vegetation and it was of
special interest to notice here what in fact is generally the case
throughout the warmer parts of the world. The nineteen genera
that show more than one mountain endemic are represented in
Ceylon by 268 species, or 14 species per genus against an average
representation of 2-7 species, and in the world as a whole these
genera contain 4095 species, or 215 per genus, against an average
of about 13. They are thus not only very large genera but also
genera that make up nearly 10 per cent of the whole flora of
Ceylon, and 2 per cent of that of the world. And this is the
general rule with regard to endemics, wherever they may occur.

It looked as if there must be some definite reason for the
commonness of endemics upon mountain tops, and I suggested
cosmic rays, though mere isolation might be sufficient.

The mountains of Ceylon thus behaved, in regard to endemism,

CH. Ill] ENDEMISM, AGE AND AREA 27

just like the separate islands of an archipelago, where again the
endemic species behave in this manner, belonging to large genera,
with a distinct tendency to differ among themselves upon the
different islands. It was, therefore, concluded that there was
nothing peculiar in the existence of an oceanic island that should
give rise to endemics, other than the qualities that it shares with
mountain tops, which show like islands in their possession of local
species. "Of these the most obvious is isolation, and we may,
I think, justly draw the conclusion that has often been put
forward, and say that isolation, as isolation, favours the produc-
tion of new forms" (57).

The study of endemism begun in Ceylon was recommenced at
Rio de Janeiro early in 1912, and soon led to the hypothesis of
age and area about which many papers and a book (66) were
published in the following ten years. By the courtesy of the
Editor of the Annals of Botany I am allowed to quote, with
modification and omission, from a paper of 1921 (65) a short
summary then written:

Examining on many occasions, from 1896 onwards, the...
Flora of Ceylon (46),... I gradually found, somewhat to my
surprise, that the strictly local species confined to that island, or
endemic species, as we usually call them, which are very numerous
in Ceylon, showed on the average the smallest areas of distribu-
tion there, whether in the grand total or in individual families
(cf. 70, p. 12). On the older view of the meaning of endemic
species, which I then held, this seemed a very remarkable thing —
that species which were generally looked upon as having been
specially evolved to suit the local conditions should be so rare in
those very conditions. If these species were specially adapted to
Ceylon, therefore, it could not be to the general conditions of the
island, but must be to strictly local conditions within its area.
There was clearly no difference between island endemics and
those of the mainland. Accordingly, still more remarkable did it
seem when I came to study in detail the local distribution of
these endemic species in Ceylon, and found that, as a rule, they
were not confined each to one spot or small region characterised
by some special local peculiarity in conditions, to suit which they
might have been supposed to have evolved. Not only so, but
such spots were frequently to be found with no local species upon
them. Only about a quarter of the whole number were confined
to single spots, and more than half of those were restricted to the
tops of single mountains (57). The remaining three-quarters
occupied areas of larger and larger size,* and in diminishing
numbers as one went up the scale. . . . The very rare species are
as a rule well localised, but the rare and rather rare . . . cover areas

28 CONTACTS WITH DARWINISM II [ch. iii

that overlap one another like the rings in a shirt of chain mail.
Now a little consideration will soon show that from the point of
view of evolution to suit local conditions this is a verv remarkable
state of affairs. If A and B grow in overlapping areas, both must
be growing in the coincident portion, and what keeps A from
growing into the rest of 5's territory, B into ^'s? In reality the
case is more complex, for if all the species were entered, there
would be. . .a dozen overlapping at any one point. It is all but
inconceivable that local adaptation should be so minute as this,
with soil essentially the same throughout, and the rainfall, etc.
varying much from year to year. The species would have to be
adapted to wide range in rainfall, and to very slight in a com-
bination of other factors. It was clear that the old ideas of
particular adaptation were quite untenable.

Nor would the other popular theory, which equally survives
to-day, satisf}^ the knowledge that I now had about local distri-
bution. How could species be dying out in this remarkable chain-
mail pattern, and why were there so many with small areas?
Had one perhaps arrived in Ceylon just in time to see the
dying out of a considerable flora? And why did so many choose
mountain tops as a last resort? If they had climbed from below,
they must have plenty of adaptive capacity, and should be able
to compete with the new-comers. Still more, why did each one or
two choose a diff"erent mountain?. . .It was difficult to believe
that the plains were once inhabited by diff^erent species at every
few miles, whilst many mountains with endemics did not even
rise direct from the plains, but from a high plateau.

Counting up all the species of the Ceylon flora, and dividing
them into three groups — those endemic to Ceylon, those found
onlv in Cevlon and South India, and those with a wider distribu-
tion abroad than this (which I termed zvides for short) — I found
(59) the endemics to be graduated downwards from few of
large distribution area to many of small (e.g. common 90, rare
192), and the wides in the other direction (e.g. common 462,
rare 159), with the Ceylon-South India species intermediate. In
other words, the average area occupied by an endemic was small,
that by a Cejdon-South India species larger, and that by a wide
the largest of all. A cursory examination of other floras showed
me that their endemic species also behaved in the same way, . . .
and I Avas at last furnished with what seemed to me to be a much
more feasible explanation of the distribution of species in general,
and endemics in particular.

Having disposed, to my own satisfaction, of the notion that
endemics were moribund species, I adopted the view that in
Ceylon the mdes were the first species {o7i the whole'^) to arrive,
and had therefore on the whole occupied the largest areas. The
Ceylon-South India species, on my view, must have arisen from

^ I.e. in any genus the wide would usually be the first to arrive.

CH. Ill] EXDEMISM, AGE AND AREA 29

them at points in general south of the middle of the Indian
peninsula, and would on the whole be younger in Ceylon than the
wides, and therefore occupy lesser areas on the average. The
Ceylon endemics would arise from the wides (or Ceylon-South
Indians) in Ceylon, and would be the youngest, and on the
average occupy the least areas. All the figures of course must be
worked in averages, for an endemic of one group might be
occupying a large area when the first wide of another arrived.

Confirmatory evidence was soon obtained from the floras of
New Zealand, Jamaica, Australia, and the Hawaiian Islands.
The figures for New Zealand are as follows:

mge in N.Z.

Wides

Endemics

881-1080 miles

201

112

641- 880

77

120

401- 640

53

184

161- 400

38

190

1- 160

30*

296

* Largely undoubted introductions of recent years.

Facts like these, which are universal, cannot be the result of a
selection, but must have some more mechanical explanation. The
only one that to the writer seemed at all satisfactory was simply
age, as was explained in Age and Area, though of course age in
itself was not exactly a factor in distribution. There are very
many factors that may affect dispersal, but if one suppose
factor a to produce an effect in distribution in a long time x that
may be represented by 1, one may reasonably expect that in
time 2x it will produce an effect 2. If the effects of all the factors
be added up, the total effect in time x may be represented by m,
and in time 2x by 2m. Obviously there will be great individual
differences between species, so the proviso was made that com-
parison (with a view to determining questions relating to age)
must only be made between allied forms, which were most likely
to behave in an approximately similar way under similar circum-
stances. The quickly reproducing, herbaceous Compositae must
only be compared with other Compositae, not with the slowly
reproducing trees of the Dipterocarpaceae or the Conifers; and
so on. One form might even occupy in a decade what might take
the other several centuries to occupy. And not only must this
precaution be taken, but closely allied species, even, must be
taken in tens, to allow of averaging the effects of the many factors
that might take part in their distribution. But, bearing these
things in mind, one might say that large area of distribution

30 CONTACTS WITH DARWINISM II [ch. iii

meant considerable age, small area small (each set of plants com-
pared being taken from the same circle of affinity).

And also, one must always remember that the distribution of
plants is very largely controlled and determined by the presence
or absence of barriers, which may be of many kinds. There may
be simple physical barriers like the sea, or a mountain chain;
there may be the barrier of a climatic change from warm to cold,
or from dry to wet, and so on; there may be ecological barriers
imposed by the habit or other peculiarities of the plant itself, and
so on. The whole quesliion is discussed in detail in chap, v (p. 32)
of Age and Area.

So axiomatic did all this seem, that the author was somewhat
surprised by the vehement opposition that it encountered. The
explanation of this perhaps lies in the fact that geographical
distribution would thus be transferred to a more mechanical
sphere than had hitherto been allotted to it. No longer, especially
in view of the regular arithmetical arrangement, could the natural
selection theory supply a full explanation of the facts of evolution
into genera and species, and no longer, in face of the fact of
increase in number do^\Tiwards in the case of endemics, upwards
in case of wides (table on p. 29), could it supply a full
explanation of the facts of distribution, or of the nature of
endemics. Sooner or later, it seemed to the author, these new
discoveries meant that natural selection, in its present form at
any rate, would cease to be so important a factor in evolution,
and with evolution of course went distribution and many other
branches of biological science.

One of the most important things that would necessarily
follow from the acceptance of age and area was the replacement
that it asked of the long-cherished notion that endemics in general
were either relic forms, or local adaptations, by the supposition
that when they occurred in very small areas they were mostly
young beginners as species, that had not yet had time to occupy
larger areas. In many cases of course barriers (especially barriers
due to climatic or soil conditions) that would in any event obstruct
or prevent further spread were so close that only small areas
could be covered, even though the species might be very old.
Other species, again, of very limited distribution, and that more
especially in the north within reach of the effects of the cold of
the last glacial periods, were evidently relics. Sinnott (Age and
Area, p. 86) gives, as examples of this class in North America,
Carya, Platiera, Madura, Garrya, Sassafras, Xanthorhiza,

CH. Ill] ENDEMISM, AGE AND AREA 31

Baptisia, Nemopanthus, Ceanothus, Dirca, Dionaea, Hudsonia^
Rliexia, Ptelea, Decodon, Houstonia, Symphoricarpus, etc., many
of which are fossil in the Old World. As they also include most
of the woody endemics of North America, and as each of them
belongs to a different family, it is highly probable, if not certain,
that they are relics. But, as already pointed out, they are lost in
the crowd when considered in connection with their own families,
especially as most of them are but small genera. And though
they may be relics of a previously more woody vegetation of
North America, we have no reason to suppose that they are being
killed out by superior species — they have probably been much
reduced by change of climate and are not quite so well suited to
the conditions that now exist. In warmer countries one com-
paratively rarely finds endemics of this kind; the endemics, as
has already been pointed out (Age and Area, pp. 91, 165; and
p. 26 above), occur chiefly in the large and "successful" genera,
like Ranunculus or Poa in New Zealand, or Eugenia in Ceylon or
in Brazil.

Among these just quoted relics there occurs Ceanothus, with
forty species in North America only, a genus that must be
counted as large for that country. In a recent discussion, Arto-
carpus, the jak and breadfruit genus, which is the third largest
genus in the large family of the Moraceae, and has over sixty
species scattered over Indo-Malaya and China, was quoted as a
relic, on the ground of the occurrence of fossils outside its present
area. This kind of definition of relic seems to the writer something
of a begging of the question. We can no longer be sure that any
plant is not a relic. The whole British flora must evidently consist
of relics, except perhaps the very local species farthest from the
land that has been submerged, and yet the flora is in reality a
very young one in its present position. If a change of conditions
affect a country, it is in the highest degree improbable, except in
a case like the coming of the ice, that it will kill out all the former
flora — it will be gradually and partly replaced by newcomers
that better suit the newer conditions, and if the conditions change
back again, these may be in turn replaced by the older flora, and
gradually things may become much as they were before the first
change.

One reason, perhaps, for the unpopularity of age and area was
the realisation that it was incompatible with the current view of
the way in which evolution had gone on. If we follow it to its
logical conclusion, it is clear that as the family in general occupies

32 CONTACTS WITH DARWINISM II [ch. iii

a larger area than the genus, the genus than the species, the
family must be the oldest, or (where, as is often the case, one
genus covers the family area) as old as its oldest genus. This turns
the Darwinian theory upside dowTi, for upon it the family is a
later appearance. There is, however, no evidence for this. How-
ever far back we go in the geological record, we always find
families that are identical with some of those of the present day.
They are also usually widely separated, so that even at that early
period it is clear that if evolution followed the Darwinian plan, it
must already have travelled far, though we find no evidence
whatever of any intermediate stages upon the way.

CHAPTER IV

THE HOLLOW CURVE

J. H E chief result of the work upon Age and Area, perhaps, was
the discovery of the " hollow curve of distribution " (cf. chap, xviii
of Age and Area, p. 195), a curve which shows in all cases of dis-
tribution that I have yet examined, whether of animate or even
of inanimate things. My opponents have gone to great trouble to
show that it holds, for example, with the names in a telephone
book, or even with the distribution by size and shape of a pile of
gravel, in other words that distribution is in general what one
may call very largely accidental, and not determined by adapta-
tion in so far as concerns general distribution about the world,
which is exactly what I wished to prove.

The curve was first noticed in 1912 in regard to the flora of
Ceylon, which consisted of 573/1 (573 genera each with one species
in Ceylon), 176/2, 85/3, 49/4, 36/5, 20/6 and so on. If one take the
first few numbers, one finds that the numbers to right and left
of any single number (e.g. of 176/2) add up to more than twice as
many (573/1 -f- 85/3 = 658) as itself, so that the curve must be
hollow as shown in the figures below. It turns the corner between
3 and 5, and as the numbers get small it becomes more or less
irregular.

The curve was also found to show, but not in such detail, with
the areas covered by species. If one divide the species of a genus
or family into those of large, small, and medium areas, one finds
that if one add together the numbers in the large and the small,
they make more than twice as many as those in the medium, or
in other words they make a hollow curve, like those shown in the
illustrations.

Now not only does this hollow curve show with the distribution
of species by areas, but it also shows with the distribution of
genera in a family by the number of species that they contain.
We must always remember that statistics must only be applied
to numbers and to related forms, which as a general rule will
behave in much the same way. Take, for example, the family
Monimiaceae, of 33 genera and 337 species. The two largest
genera, Siparuna with 107, and Mollinedia with 75 species, range

WED ^

34

THE HOLLOW CURVE

[CH. IV

SIPARUNA

(Monimioceoe) PR.

The familLj consists of SiparunaClO/j,
Mollinedia(7''), and genera with
31,25.15.15,11,7.5.5.4,4,4,3,3,3,3.
2,2.2,2,1,1,1,1,1, 1,1,1,1, tend

6. Mexico. 24. Dominica.

ll.fGuaremala and 25. St. Vincent.
Nicaragua. 38.Costa Rico.

19- St. Vincent.
2 O.Nicaragua
21. Mexico.
22.Mexico.

41. C America

42. Mexico (frequent)
51 Costa Rica.
54Costa Rica.

Pig. 1. Distribution of Sipariina in South America. The local species in the
Andes etc. are simply massed together, not shown each in its own place. The
numbered list shows the localities of northern species.

CH. IV] THE HOLLOW CURVE 35

from Mexico to Rio de Janeiro or south of it, the smaller genera
over less districts. Siparuna has one species that covers the whole
South American area of the genus, some of intermediate areas,
and a great many of very small areas. Mollinedia, on the other
hand, though its total area is much the same, has only one species
that even ranges as far as from Rio de Janeiro to Monte Video;
most of its species are quite local, and over 95 per cent are so local
as to count as relics under the natural selection conceptions. Is it
a failure because of the small areas occupied by the individual
species, or a success because of their number, and the area
occupied by the genus as a whole? What is selection doing in
these two cases? And still more, what is it doing or going to do
with the rest of the family, where the genera contain 30, 25, 15,
15, 11, 7, 6, 5, 4, 4, 4, 3, 3, 3, 3, 2, 2, 2, 2, 1, 1, 1, 1, 1, 1, 1, 1, 1,
1, 1, 1 species respectively? One cannot draw a line in a curve like
this, to separate the sheep from the goats. Relics would not be
made in steadily diminishing numbers, nor would local adapta-
tion be neatly graduated like this. All families of reasonable size
show the same curve, as seen in fig. 2, which gives the fifteen
largest families of flowering plants. Close together though they
are, the curves never touch. When turned into logarithmic curves,
as in the next figures (3, 4), they all give approximations to
straight lines, i.e. they have the same mathematical form, and
must be the expression of some definite law which is behind
evolution and distribution, and does not agree with current
views about these subjects. Many distributional subjects show
the same form of curve, as may be seen in fig. 5, which shows
families of plants and animals, lists of endemics, floras, fossils,
and areas occupied, all mixed up. The curve shows in the names
in the telephone book, where the very common names are few,
the very uncommon many. It shows in the list of numbers of
hotels in towns in the advertisements in Bradshaw, where (in the
one examined) only London and Bournemouth had large num-
bers, while a great many had only one each, and there were a
few in the intermediate numbers.

This similarity interested me very much, and I have lately
completed a study of the distribution in Canton Vaud (Switzer-
land), where I live, of the surnames of farmers, a class who move
about less than others. Vaud is about the size of Gloucestershire,
but divided into valleys often separated by very high mountains,
which make intercourse between the vallevs difficult. After a
day on the farm, a young man is not going to cross a high moun-

3-2

36

THE HOLLOW CURVE

[CH. IV

tain range to see his best girl, but marries in his own valley. The
result has been a very interesting distribution of surnames.

1^

FAniLICS in ORDtR Of SIZE
SHOWmC MUMbERS or CLOEPA

WITH DirrtRcnT nunBLRi or iPLCiLb.

00

0\

6135

cnera

Fig. 2. Hollow curves exhibited by the grouping into sizes of the genera in
the first 15 largest families of flowering plants. Each curve is diagonally
above the preceding one, as indicated by the heavy black dots (points of
origin). Note that the curve almost always turns the comer between the
point marking the number of genera with 3 species, and that marking the
number with 5 (indicated by the dotted lines). The number after the name
of the family shows the number of genera in it.

In a great proportion of the villages in the canton, some
hundreds in number, there are local names found only (i.e. en-
demic) in one village each, sometimes on one farm only, some-
times on two or more. Sometimes the names occur in two or

CH. IV]

THE HOLLOW CURVE

37

N2 of species

10

(00

■2 -4 -6 -8 10 1-2 1-4

log (N? Of species)

Fig. 3. Logarithm curve for Rubiaceae (from WiUis, Dictionary).
(By courtesy of the Editor of Nature.)

2

.

\

o^V.

^ ^

>o,

1

0

'■'^^s

c
a

■^o 1

Number
o

o^ —

^^v. ""

■QO

— ^

V.

^^——^

0
1

•2

■4

■ ' -6
5

■a

log(
N

it

Number 0

IC

jmberofi

"^species)
pecies

2 ' 1
20

V ' i-e

30

i;8

2
(00

Fig. 4. Logarithm curve for Chrysomelid beetles (from old Catalogue).
(By courtesy of the Editor of Nature.)

38

THE HOLLOW CURVE

[CH. IV

more villages, but always in diminishing numbers as one goes
upwards, just as with the plants. Most often the villages are in
the same neighbourhood, but at times they are as far off as a

Monospecific Genera at this end of curve

• C

ft) ^
. «■ "^ j-

/June of Genera
ospp.

Number of species (or size of area.J

Fig. 5. Mixed curves, to show the close agreement of the hollow curves,
whether derived from families of plants grouped by sizes of genera (Com-
positae, Hymenomycetineae, Simarubaceae), families of animals (Chryso-
melidae, Amphipodous Crustacea, Lizards), endemic genera grouped by sizes
(Islands, Brazil, New Caledonia), local floras grouped by (local) sizes of genera
(Ceylon, Cambridgeshire, Italy), local faunas (Birds of British India, British
Echinoderms), Tertiary fossils by sizes of genera, or Endemic Compositae
of the Galapagos by area. [By courtesy of the Editor of Nature.]

man can walk in one or two days, their distances and the tracks
on which they lie showing the directions of emigration of young
fellows in search of work.

The distribution of species of plants that occurred in Ceylon,

CH. IV] THE HOLLOW CURVE 39

for example, outside the island, was found to go, on the average,
with their distribution inside the island,^ but natural selection
could not adapt a plant that was to come, say from West Africa,
to suit Ceylon better than a plant that had only come, say from
Bombay. If anything, one would expect the latter to suit Ceylon
the better. And the same thing showed with the names of the
farmers. Rochat is a very common name in the valley of Joux,
and has spread farthest into the country round, while the names
that are less common have spread less. It is impossible to main-
tain that the possession of the name Rochat gives any advantage
in the struggle for existence as against the name Capt, which is
less common in Joux, and has not spread so far beyond it (fig. 6).
Natural selection can have nothing to do with the distribution of
surnames, which behave just like species of plants.

All these various curves match, and must be determined by
the same rules. There would seem to be a necessity to reconsider
the idea that distribution is determined by natural selection, as
indeed we have already seen. Adaptation can only be to the con-
ditions that exist round about the plant, and it is absurd to
suppose that the bulrush or the silverweed, for example, that
(in the same specific form) occurs in New Zealand as well as in
Europe could have become, in Europe let us say, adapted to
New Zealand conditions. That it suits them is simply due to luck,
and to local adaptation, as it slowly moved from place to place.
But in any place where it was not fairly well suited, it would
usually be killed out remorselessly and promptly by natural
selection.

Many other cases might be brought up, but the fact that distri-
bution shows these hollow curves, which cannot be explained by
aid of the theory of natural selection, will suffice to show that
that theory in its turn is meeting with almost insuperable diffi-
culties. It was difficulties like this which made my friend
Dr Guppy, who had devoted most of his life to the study of
distribution, adopt in 1906 the theory of evolution by diff'eren-
tiation, whilst, as the result of completely independent investiga-
tions upon different lines, I myself adopted it in 1907. The theory
itself is pre-Darwinian. The idea that underlies it, as formulated
by Guppy, is that in the early days of the flowering plants the
climates of the world were damper and more uniform. The world
as a whole seems to have become drier since that time, so that the

^ I.e. a plant widely distributed in Ceylon was on the average widely
distributed outside it.

40

THE HOLLOW CURVE

[CH. IV

climates must have become more differentiated into damper and
drier, warmer and colder, etc., than they once were. With them
the plants have become differentiated, and it was commonly
supposed that this was done, as in the theory of natural selection,

CANTON VAUD

Distabution of name THEVENAZ

•agriculturist ♦commercial or professional -5" Crotx»V' ^'*'

^ C6.S0O mK ) ♦ ^f

BalUt

Kilometres

Crandson

j!/ (<2,000lnh

el

Orkei
(J.SOO inh)

CANTON VAUD.

DlstrLbutlon of name ROC HAT

& po9sibl< varittUi ROJARD, R055AT. RUCHAT
• agriculturist. ♦ commercial or professiooa

o I 1 a fc St T » < la

Kilometres.

IMONTREUX

Fig. 6. Distribution of two surnames, Thevenaz and Rochat,

in French Switzerland.

more or less in suitability to the various climates, but that the
whole specific or other difference appeared, not by gradual
adaptation, but at one step. The writer, however, is not prepared
to admit that these things are necessarily connected, without
further evidence.

CH. IV] THE HOLLOW CURVE 41

To this, I have added the facts of the hollow curves, which are
universal, not only in the distribution of plants and animals, but
in other things, as we have just seen in the case of that of the
farmers' surnames in Vaud, which matches to a nicety the distri-
bution of plants. It seems to me impossible to reconcile these
curves with the theory of natural selection, and there are other
very serious objections to this latter theory. To reconcile the
theory of differentiation with the hollow curves, I have added to
it the supposition that the evolution that goes on, and which is
shown in the morphological characters of plants, has little or
nothing to do, directly, with adaptation, and certainly not with
direct adaptation. The characters do not necessarily indicate
adaptation at all. Every now and then a character appears, like,
for example, climbing habit, of which natural selection can make
use, and which is therefore retained, but natural selection was
not the direct cause of its (complete) appearance, nor was its
appearance, in all probability, as accidental as that theory would
involve. It appeared full-fledged, and was advantageous, or at
any rate not harmful. And if it had no necessary adaptational
value behind it, there was no particular reason why a species
showing it should spread at a different speed from other species
of the same or closely allied genera, a supposition which at once
made the hollow curves a normal feature of distribution. And
there was also no reason why the many new species that have
appeared should not appear at a rate that was in any case not
determined by the necessities of adaptation, as we have seen in
the case of the Podostemaceae. A species that reached a sand
dune, for example, if it were reasonably suited to it upon arrival,
would gradually adapt itself in more detail to some definite local
conditions there, by physiological adaptation controlled by
natural selection.

What the mechanism was by which this evolution was carried
on, we do not know. I suggested in 1907 that "a group of allied
species represents so many more or less stable positions of equi-
librium in cell division".

The occurrence of the hollow curve for distribution of plants
by areas, or for distribution of genera by numbers of species,
shows that neither in geographical distribution (strictly so-called)
nor in evolution can natural selection be invoked as it once was,
as the principal factor. Any influence that it has must either be
very small, or else exerted in an indirect way. One cannot, upon
such a curve, draw any dividing line, and say that those upon

42 THE HOLLOW CURVE [ch. iv

one side are to be regarded as successes, those upon the other as
failures. Nor can one picture to oneself a system in which the
number of failures that die out (and in all families more or less
alike) is at first to be about 38 per cent (the monotypic genera, of
one species), then only about 12 per cent (the genera of two
species) and so on in decreasing numbers.

CHAPTER V

CONTACTS WITH DARWINISM, continued.

MUTATION

X HE coming of mutation was mentioned above (p. 14) and it
was pointed out that it seemed to get over some, at any rate, of
the difficulties inherent in the employment of gradual variation.
In particular, as the new form was qualitatively, and not merely
quantitatively different from the old, the change was differen-
tiating. Further, it was practically irreversible, and might also
be hereditary.

But it was gradually realised that its employment brought in
its train other difficulties which were almost as great, so long, at
any rate, as one adhered to natural selection as the driving force
in evolution. This adhesion definitely handicapped the theory,
preventing it from giving its proper stimulus to biological progress.
Since people wished to combine it with natural selection, they
had to stipulate that mutations must be very small. It was very
hard to see how it could work with large mutations that might
effect such differences as distinguish the Monocotyledons from
the Dicotyledons, or even those that divide one family or genus
from another, and which might change the whole character of the
plant. If these were to be allowed, one could no longer imagine
progress by small, gradual, and progressive adaptation, and this,
determined in everj^ detail by natural selection, was still the
ruling principle invoked in evolution. If we remove direct
advantage from the list of factors that mav be immediatelv
operative in causing evolution to go on, it is evident that the
structural mutations that distinguish one form from another
need not, perhaps even cannot, proceed in gradual stages, unless
there be, as of course is by no means impossible, some at present
inscrutable law that guides them. But fossil evidence gives but
little support to this conception. Real intermediates are rare;
what are commonly called intermediates are usually things that
combine some of the characters of one with some of the other. If
one find a plant showing, to give an imaginary case, four of the
characters of Ranunculaceae to three of the Berberidaceae, it is
sure to give rise to discussion and dispute.

44 CONTACTS WITH DARWINISM III [ch. v

More than thirty years ago the writer published a paper upon
the distribution of the Dilleniaceae (72), in which he adopted the
notion that mutation might at times be so large that there might
appear in one step a new species, or perhaps even a new genus.
Intermediate stages were not considered to be necessary, though
it was pointed out that in one or two cases intermediate forms,
perhaps hybrids, were found living side by side.

By that time the author had completely discarded the theory
of natural selection as the chief driving force in evolution, re-
garding it primarily as a means of getting rid, promptly, of
anything that was seriously unsuited to the conditions under
which it had to live. There was, of course, no definite reason why
selection should not at times, under favourable circumstances,
produce new forms, but it seemed unlikely that such production
was at all common, or that it should produce forms of specific
rank. It could not be looked upon as operative in regard to the
bulk of the morphological characters which show us that evolution
has gone on, and which in consequence have always tended to be
regarded as in some way showing progressive adaptation. The
author had also abandoned the idea that there was such wonderful
morphological or structural adaptation in the flowering plants.
Each, of course, must be fairly well suited to the place in which it
grew, for if it were not, natural selection would soon dispose of it ;
but that was all, in most cases. Real adaptation was largely
internal as was clearly indicated (1) by the enormous range of
many species without any serious morphological change from one
region, or one set of conditions, to another; (2) by the great
numbers of plants that were to be found in the same conditions
(as nearly as made but little difference) and yet showed such great
morphological differences that they could be classified into many
different families and genera, though they might all come into
one ecological category, like the Podostemaceae or the plants of
a moor or a sand dune. The common plants of a moor in Britain,
for example, include Betula, Calluna, Carex, Cornus, Empetrum,
Erica, Kobresia, Listera, Molinia, Nardus, Potentilla, Scirpus,
and Vaccinium, covering a great range of the flowering plants, as
can be seen at a glance. Another indication (3) was the great
numbers of species of one genus that might at times be found in
similar conditions, like Mesemhryanthemums in South Africa,
while single species of other genera ranged over great differences
in conditions.

The structural differences that showed in plants to such an

CH. v] MUTATION 45

extent were often so clear cut, and so distinct, that it seemed to
the writer quite evident that they must in general have been
formed by sudden change, or mutation. Gradual change, picking
out advantageous variation, would be very unlikely indeed
always to produce tlie same structural character, such, for example,
as is sho^vn by a berry or a drupe, or by opposite leaves. Why
should berries be most often found in the near (systematic)
neighbourhood of capsules, drupes in that of achenes or nuts?
Why should selection pick out leaves that were exactly opposite,
ovules with the raphe exactly dorsal or ventral, or why such
clearly marked and exactly formed fruits as capsules, berries, etc. ?
Selection would obviously act with decreasing force as the leaves
came nearer and nearer to being opposite (or alternate, for then
they show a definite phyllotaxy or arrangement), or the raphe
to being dorsal or ventral, etc. In actual fact, between many of
these characters, intermediate stages were not possible. One
could only take the one or the other side of a very divergent
variation, such as alternate or opposite leaves, dorsal or ventral
raphe, etc. The mutation, in so far as the characters themselves
were concerned, paid no attention to functional or adaptational
requirements. It was impossible to conceive of any adaptational
need that would ensure that all Monocotyledons should have a
single cotyledon, together with a parallel-veined leaf, a trimerous
flower, and a peculiar anatomy. There is not even a "mono-
cotyledonous " mode of life to which this great morphological
change might be supposed to adapt them. For that matter,
there is not even a " ranunculaceous " or a " thalictroid " mode of
life. The larger a genus, family, or other group of plants is, the
greater is the variety in the conditions of life in which it is found
(comparing, as usual, only related forms), and also the larger the
area covered by single individual species of the family or genus.
To return to the Dilleniaceae; assuming that mutations could
be of generic size, the author drew up a scheme according to
which the whole tree of the family could be looked upon as
derived by descent from a genus so comparatively simple in
structure, and so widely distributed, as Tetracera. A sketch was
drawn of a suggested manner in which the evolution might have
proceeded, showing all the existing genera, which might even be
the whole tree of the family. Of course some geological or other
catastrophe might have killed out some more or less local genera,
though it would be unlikely to have done so to any genus that
was already very widespread. Incidentally Tetracera is not the

46 CONTACTS WITH DARWINISM III [ch. v

largest genus in the family, but it is the most widespread, and
when these two characters do not agree in pointing out what is
probably the oldest genus, the author considers that distribution
rather than size should be regarded as more important.

In thus supposing that a genus could appear at one stroke, and
that one genus could, directly, give rise to another, the author
was definitely going beyond mutation pure and simple, and
adopting the theory of differentiation, in which, as the changes
were large, the idea that the morphological differences represented
adaptational improvements was discarded. In other words,
though evolution was unquestionably going on, and was on the
whole, though more notably in the animal world, producing
higher and higher types, there was no need to suppose that there
was necessarily any adaptational reason in the innumerable
structural changes that showed themselves in the course of that
evolution, and indeed showed that evolution was going on at all.
It seemed much more probable that most of those features which
we were accustomed to call adaptational improvements had
appeared already full-fledged. If new features that thus appeared
were really harmful, or met with ill-luck, they were promptly
removed by the action of natural selection. If they were bene-
ficial, or not harmful, and met with average luck, they were
retained.

The sketch showed the way in which it was suggested that
evolution had proceeded, from the large and widespread genera
down to the small and local, but there was then about as much
chance to prove this kind of mutation as to prove that natural
selection could do what was required to form a new species, for
it must not be forgotten that this has not yet been done. There is
some reason to suppose that it can produce new varieties, but
no proof that it can cross the line of mutual sterility that usually
lies between species. Both theories derive varieties from a parent
species, but selection derives them from a parent which is at an
earlier stage of development, and perhaps fated to die out, the
varieties being considered as on the way to species. Differentia-
tion does not admit this, but regards them as later stages of
development by mutation than the species that gave rise to
them, and with which they are not necessarily in competition,
though perhaps they may sometimes go on, by further mutation,
to become new species. The essential point at present, for dif-
ferentiation, is to prove that evolution proceeded in the direction
from family to species, and not the reverse.

CH. v] MUTATION 47

When one looks at the great differences that exist, for example,
between the Dicotyledons and the Monocotyledons, and these in
several different points, it seems to be an unnecessary handicap
to accept the idea that mutations must necessarily be small,
especially when we have no facts to prove that this must be the
case. The characters are so completely unrelated to anything in
the way of adaptation that it becomes very difficult to conceive
of them as having been gradually acquired, especially when one
remembers that intermediates between them are all but im-
possible, and could not in any case have any adaptational value,
so that unless there is some recondite law in the background
that can force things to proceed in such a manner, there seems no
reason for it. There might for example, be (probably is) some
physical or chemical law that at present we do not know, com-
pelling genes or chromosomes to behave in a certain way.^ But
as one sees the phenomena at present, how can one pass by
gradual stages from two cotyledons to one (or vice versa), from
net veining to parallel, from a 5-merous to a 3-merous flower,
from the one kind of anatomy to the other? The only reasonable
way to account for it is to suppose that the characters of Mono-
and Dicotyledons were handed down as the lines of descent
resulting from a mutation in very early times which split off the
one from the other. No adaptational difference can be found, nor
is there any " monocotyledonous " mode of life. As one comes up
the scale from species, the plants are found to grow in greater and
greater variety of conditions, and to belong to more and more of
the various ecological groups. If monocotism suit a grass or a
bamboo better than dicotism, why does it also suit a tulip, a
Zostera, a Potamogeton, an iris, or an orchid? And why, if there is
any adaptational difference betw^een the two great groups, do
they occur with such regularity in almost every part of the world
in the proportion of one to four? There are small places where
these figures vary very much, but the only large ones are usually
near the limits of vegetation, a fact which suggests that there are
differences in age between the two groups. Hooker pointed out
this numerical relationship in 1888 (18), and it remains one of the
many problems in geographical distribution which are com-
pletely inexplicable upon the hypothesis of natural selection, and
which are left unmentioned by its supporters.

Another direction in which the theory of mutation makes

^ My friend Dr C. Balfour Stewart suggests that it is probably electrical,
as is probably the spUtting of the chromosomes in reproduction.

48 CONTACTS WITH DARWINISM III [ch. v

things much easier to understand is the widespread correlation
of characters, for which natural selection can offer no explanation.
Why is the possession of tendrils, or of hooked leaves or stems,
always accompanied by a weak and flexible stem? Why has a
dorsiventral leaf, such as is possessed by a vast number of plants,
always a layer of palisade tissue towards the upper side, for
making the best use of the light that falls upon it? Why, in the
Compositae, have the heads of flowers an involucre of bracts,
why has the style two stigmas, why is the ovary unilocular, why
is there only one ovule and that erect, and why is there no endo-
sperm? And why do all these characters go together in practically
every instance in a family of 18,000 species? The same sort of
questions may be asked for any other family, whilst they would
be absurd in the case of adaptational characters. Nothing but
descent from a common ancestor (or ancestors) will explain
them, and evolution upwards from individuals and varieties will
not do it; it must have been the other way, as differentiation
would have it. Evolution apparently must go on, at any rate if
the appropriate stimuli are present, but there is no necessary
adaptational reason for much of it, at any rate, and we find
practically no gradual stages in the fossil record. To accept
mutation, and that of any necessary size, would seem to be the
simplest theory upon which to work until something better
turn up.

An objection often brought up is that no such mutations —
large, viable, not recessive, and not lethal — have been seen. But
no one has ever seen a species formed by natural selection. Yule
has estimated that one such mutation in fifteen to thirty years,
upon any small spot of the earth's surface, would be sufficient to
account for all the flowering plants that exist. The chance of
seeing such a mutation is all but non-existent, and if the result
were found at present, people would at once put it down as
another relic and leave it at that. Until we can control mutation
— and signs are not wanting that we may be able to do so at some
future time — we can hardly hope to get proof for this proposition.

One must not forget that the mutations that have been studied
have, as a rule, been mutations that have occurred in cultivated
plants, or otherwise in unnatural conditions, conditions which in
themselves perhaps stimulated a greater mutability than usual.
We have not properly considered the case of mutations under
completely natural conditions, which are well kno^vn to be much
less common. If a mutation appear in a seedling of some tree in

CH. v] MUTATION 49

the jungle, the chances are that it will inherit the suitability of
its parents to the local conditions, and that if the mutation be not
seriously harmful, it will not be interfered with in any way by
natural selection, and will be allowed to survive, and in time, if
hereditary, to propagate itself. May it not be that something of
this kind is an explanation of the great majority of the innu-
merable structural differences that we see in plants, and which
so often only appear when the serious struggle for existence is
over, or practically over? One cannot imagine that it can have
any importance in the struggle for existence whether a plant
have or have not one or two cotyledons, a parallel- veined or a
net-veined leaf, a 3-merous or a 5-merous flower, and so on. To
the vast majority of the characters upon which we base our
classifications natural selection is probably completely indif-
ferent. It is well known, incidentally, that most of those
characters which we consider as usually of family rank (App. I)
may at times appear as generic, or even specific, so that it is
evidently quite easy for them to be acquired, while at the same
time the structural agreement between them is amazing. Nothing
but sudden mutation will easily account for such phenomena.

A case in which mutation of this kind looks as it might have
happened in nature is that of the columbine {Aquilegia), which
looks as if it might have arisen from the larkspur {Delphinium),
the latter having a dorsiventral flower with one spur, the former
a regular flower with five spurs. Nothing but mutation can cross
the (numerical) gap between these genera, and one actually sees
an almost exactly similar mutation happening frequently in the
toad-flax.

A good illustration (and dozens similar to this could be given)
of the very great probability of large mutation is that afforded
by the three families Centrolepidaceae, Eriocaulaceae, and
Restionaceae, all of which, independently, split into two sections,
Diplantherae with dithecous anthers, and Haplantherae with
monothecous. One cannot conceive of this by anything but a
direct mutation, which would produce morphological similarity
in all.

Some quotations bearing upon this subject may be made from
a paper now of some age (57), in which attempts were made to
show that local or endemic species were usually separated from
the widely distributed, and usually fairly closely related, species
that accompanied them by differences which could only be
passed over by mutations, often "large".

WED A

50 CONTACTS WITH DARWINISM III [ch. v

Ranunculus sagittifolius, confined to the high mountain region
about Nuwara Eliya (Ceylon), differs widely from the only other
Ceylon buttercup, R. Wallichianus (South Indian also), which
occurs side by side with it, though in drier and sunnier places,
but is closely allied to R. reniformis of the mountains of the
western Indian peninsula, differing mainly in the petals, which
are five in the Ceylon species, twelve to fifteen in the Indian one.
. . .Are we to suppose the conditions of life so different in the
Ceylon and Indian mountains that a five-petalled flower will suit
the one, a twelve-petalled the other? Or how is the one to pass
into the other, or both to arise from a common ancestor, except
by discontinuous variation? Can it be supposed that the simple
obovate-lanceolate leaf of Acrotrema intermedium fits it for the
Kitulgala district (Ceylon), while the pinnate leaf with linear-
lanceolate segments of A. Thwaitesii fits that species for the
Dolosbage district, but a few miles away, a trifle higher up, and
in a similar climate?. . .A. lyratum, characterised by very long
peduncles, is found only on the summit of Nillowekanda, an
isolated precipitous rock. . .is it to be supposed that the long
peduncles are any advantage . . . ? W^hat advantage can the two
ovules of Polyalthia Moonii and P. persicifolia be against the one
of the other species? P. rufescens, another species with two ovules,
and closely allied to both, occupies the Cochin district of South
India, and why should there be three species in so similar a
country . . . ? And how did the one form arise from the other, or
both arise from a common ancestor, except by mutation? Similar
queries might be asked 800 times for the 800 endemics ... in the
Ceylon flora.

The only possible explanation to my mind was that provided
by the "parent and child" theory, that parent and child might,
and very often did, exist side by side.

The general principle on which India and Ceylon have been
peopled with the many species which they contain would seem to
be that one very common species has spread widely, and, so to
speak, shed local endemic species at different points, or in other
cases that one species has spread, changing at almost every point
into a local endemic species, which has again changed on reaching
new localities.

A very good proof for mutation, and indeed for differentiation
also, is provided by the work done by Mr G. Udny Yule and the
writer upon the statistics of evolution (76). We showed that the
evolution of new genera out of old followed with very great close-
ness the rule of compound interest. After some time one genus
becomes two, and so on. But if genera are formed like this it is
hard to believe that they can have been formed by gradual steps,

CH. v] MUTATION 51

and it would also show that the larger must be in general the
ancestors of the smaller. Evolution seems to have proceeded
upon a definite plan; "the manner in which it has unfolded itself
has been relatively little affected by the various vital and other
factors, these only causing deviations this way and that from the
dominant plan".

4-2

CHAPTER VI

CONTACTS WITH DARWINISM, continued,

ADAPTATION

jT^daptation, or suitableness, with an implied meaning of
having been suited by some particular agency, is a subject that
has been as much discussed as any in biology, and especially
since the publication of the theory of natural selection, which is
essentially based upon it. Under that theory a new organism
only comes into existence because it is an adaptational improve-
ment upon that from which it is derived. In other words, im-
provement in adaptation is the only reason for which new organisms
are evolved. But the only thing that shows that they are new
organisms is a structural or morphological difference between
them and other forms, even if the latter be obviously closely
related to them. It was, therefore, taken for granted (it could
hardly be otherwise) that the 7norphological or structural characters
were the expression of the adaptation that had gone on, and therefore
had, themselves, a greater or less adaptational value.

Once this was fully realised, there was a great rush into the
study of adaptation, especially during the 'eighties and early
'nineties of last century. But in spite of all the work that was
put into it, no one ever succeeded in showing that even a small
percentage of the structural characters, that were the reason why
plants were divided into so many families, genera and species,
had any adaptational meaning or value whatever. No value
could be attributed to opposite as against alternate leaves (or vice
versa), to dorsal against ventral raphe, to opening of anthers by
pores or by slits, and so on.

In the characters of the plants of average moist climates (often
called mesophytes, as occupying the middle position), it was very
difficult to find anything that could be called in any way adap-
tive, except those general characters which are common to most
of the higher plants and occur in almost every kind of conditions,
such as roots (which are adapted to taking up food), leaves
(adapted to forming food by aid of the energy of light taken in),
flowers, fruits, seeds, etc. But as one went outwards to either
extreme, to the water plants (hydrophytes) on the one side, or to

CH. VI] ADAPTATION 53

the plants growing in dry climates (xerophytes) on the other, one
began to find characters more or less individual to the species,
that had something definite to do with the mode of life of the
plants, and which therefore might be called adaptive characters.
On the one side one found the somewhat negative characters of
absence of strengthening tissue and absence of stomata, with
diminution or absence of the roots ; on the other side one found
the more positive characters such as sinking of the stomata in
pits, hairy or waxy leaves, and in the most extreme cases, such as
the cacti, of storage of water in the tissues. But few of all these
characters, of whichever group, though they might make great
changes in the general look of the plants, were of great importance
in the separation of plants into species, or into genera, and still
less into families. There is little evidence that even such great
adaptations as are involved in the development of hj^drophytes
or of xerophytes can cause such great morphological or structural
differences as actually exist between plants. A mere glance at the
composition of any ecological group of plants that are suited to
any given situation is sufficient to show the truth of this. Take,
for example, the plants that occur in boggy places in Britain,
of which lists may be found in Tansley (44) or Bonnier. There
are about twenty genera represented, of which eight are Mono-
cotyledons, whereas the average proportion of Monocotyledons
is only one in five. Among woodland plants they are one in three,
whereas upon cliffs they are absent. Though the differences
between them and the Dicotyledons are about the most important
structural differences that occur, there is no evidence to show that
they have any adaptational value whatever. The bog plants also
show both alternate and opposite leaves, superior and inferior
ovaries, capsules that are septicidal and loculicidal, and that
open by lids, that are divided into several loculi or have only
one, whilst there are also achenes, follicles, berries, drupes, and
schizocarps among the fruits. The twelve genera of Dicotyledons
belong to ten different families, including both Polypetalae and
Sympetalae, and so on. In Ericaceae, where two genera occur,
one has a berry, the other a capsule. Nowhere is there any indi-
cation that the supposed structural adaptation had anything to
do with the fact that they all live in bogs, and must therefore be
adapted, or suited at any rate, to that mode of life. Other
British ecological groups of plants — those of chalk-downs, moun-
tains, and dunes, etc. — will show similar results. Everywhere one
finds that there are plants showing the important characters of

54 CONTACTS WITH DARWINISM IV [ch. vi

classification and distinction, and even showing, in many cases,
both members of the contrasting pairs that are given in the list of
family characters (Appendix I). These characters show no relation
whatever to any of the ecological features that may give the
character to the locality. Almost any family or genus, if large
(i.e. old, upon the theory of Age and Area, with its subsidiary
Size and Space) may be found in almost any kind of locality,
represented by some of its species. For example, in the bog flora
just mentioned, there occurs Sedum villosum, a member of a large
genus of 450 species usualh^ xerophytic. And not only so, but it
is a hairy species, bearing a character usually specially associated
with xerophytism. Morphologists have long maintained that
structural characters have nothing to do, directly, with the life
or functions of the plant, and it would appear that they are right
in this contention, which violently contradicts the supposition of
selection as a chief cause in evolution. The evolution that has
produced more than 12,000 genera and 180,000 species has not
been, primarily, an adaptational evolution, as the writer tried to
show twenty-five years ago in the case of the Podostemaceae.

The agency by which plants were to become adapted to the
conditions in which they were found was, of course, that of
natural selection, for the competition upon which it is based,
which we call the struggle for existence, will evidently kill out
those that are in any way seriously unsuited to the conditions.
It may also kill out some or many of those that are well suited, if
they be in any way handicapped, as by too shady a position when
in the seedling stage, by a poor water supply, or by many other
things. But in itself this killing out would not produce any
advance in complexity of structure or function, the things that
we regard as showing that evolution has gone on. Certain assump-
tions were therefore needed. Only advantageous changes could
be picked out, and it was therefore supposed that (usually) when
a gradual change of local conditions began, some of the offspring
varied in such a direction as to give them an advantage. It had
also to be assumed that the parent did not vary in this way. The
process being repeated in every generation (another assumption),
the improved forms always winning, the difference from the
parent ultimately became specific, showing as a rule more or less
sterility when crossed with the parent. The parent was supposed
not to adapt itself (yet another assumption), but to become a
relic and gradually to die out for want of offspring viable in the
new conditions.

CH. VI] ADAPTATION 55

For such a process to be successful, there are other assump-
tions that we must make. We have (1) to assume — which goes
against much or most of the evidence — that a morphological
change has some adaptational value, (2) that such a variation
will appear at the time when it is wanted (for otherwise there will
be nothing for natural selection to work upon), (3) that the
conditions will continue to vary in the same direction long enough
to permit of the adding up of small variations until the specific
(sterility) line is passed, (4) that the operation is so strenuous
that at some point upon the way the sterility line will be safely
passed, (5) that at some point when the species is fully embarked
upon the change, a better variation, but working in another
direction, is not offered to it by nature, thus confusing the result,
(6) that when one variation has achieved its full result, it shall be
followed by another, often in a completely different direction
(for one species usually differs from another in several characters)
without interfering with the mutual sterility, and (7) that the
variation is so eminently desirable that it will be followed up
until the new structural feature, for instance alternate (or oppo-
site) leaves, palmate, pinnate, peltate, stipulate or exstipulate,
gland dotted, or other type of leaf, anther opening by slits, valves,
or pores, dorsal or ventral raphe, achene, follicle, pod, nut,
schizocarp, berry, drupe, etc., is fully perfected.

The whole thing is largely based upon the third assumption
given above. For example, the climate (not the weather) must
change gradually in the direction of warmer or cooler, wetter or
drier. But these changes are well known to be so slow that they
can only be detected in averages of a century or more — a period
longer than the life of most plants, except many trees — whilst
weather is continually changeable. Suppose a plant to have
begun to vary in the direction of suitability to increased drought,
and then there comes, as so commonly happens, a cycle of wetter
years; what is going to happen then? Botanists have somewhat
neglected weather effects, when compared with agriculturists. In
the RepoH of the Sudan Agricultural Research Service for 1937,
which I have lately reviewed, it is stated that the average good
yields of the whole Gezira, in which the weather conditions were
as stated, were reflected on the Government Farm, where the
yields were much the same; "and once again we get an illustra-
tion of the comparatively small effects which local conditions
may have". This is familiar to all who have to do with crops, and
puts considerable difficulty in the way of anyone who imagines

56 CONTACTS WITH DARWINISM IV [ch. vi

local adaptation to local needs, except upon very large areas.
Would adaptation be likely steadily to follow a line based only
upon averages, in such circumstances ? It would hardly seem likely.

U 4 cotton, one of the great successes of cotton breeding, was
locally bred at Barberton in the Transvaal for certain needs, and
has proved to be a superior cotton for an immense area. But a
Darwinian species would almost certainly be a species produced
upon a local area, and if it began to spread about in its early
stages it would be lost (as Fleeming Jenkin showed) by crossing
with its neighbours, a fate from which U4 was, of course, carefully
protected.

Conditions other than those of climate or soil are hardly likely
to change continuously in one direction, except upon broad
general lines, such as a change from forest to grassland or vice
versa, and even this is probably determined by climatic change.

There is another type of adaptation, which we may call
adaptation to movable conditions. A climbing plant will remain
adapted to climbing almost anywhere that there are erect plants,
so long as it is suited to the climate and other general conditions.
A water plant can travel over an immense area, finding suitable
conditions in innumerable places. The American pitcher plant,
Sarracenia, is now quite happily established in a bog near to
Montreux, and so on.

Geographical distribution was also explained by the selec-
tionists as based upon adaptation. The better adapted species
were those that spread the furthest. But how did a species
become adapted, let us say in Asia Minor, to the conditions that
occur in New Zealand? It must be just a case of luck. If the
species were old, so that it had plenty of time to adapt itself
wherever necessary, and as in this case it would probably have a
good deal of capacity to withstand extremes, or adaptability, it
would probably be able to find places whose existence would
enable it to get across the vast distances. When at last it reached
New Zealand it would probably soon find places in which the
conditions were sufficiently like those just left to enable it to live
there. One would, perhaps, expect those plants that were
evolved in regions where there was great variety of conditions to
be those most likely to spread widely ; it may be so, but we have
at present no evidence to go upon.

In dealing with the adaptation of a plant to changed conditions
man always tends to be in too great a hurry. When Europeans
first went to the tropics, they tried to acclimatise there the plants

CH. VI] ADAPTATION 57

of Europe, with no success except in the high mountains, where
many herbaceous, but rarely arboreous, things have taken a hold
upon ground from which the original plant associations had been
removed. In the same way they tried to acclimatise in Europe
tropical things like the dahlia or the potato, but even after the
lapse of centuries these plants remain "half-hardy". In both
these cases the change of conditions was too great to allow of
physiological adaptation, which might perhaps have taken place
in a gradual acclimatisation over a very long period of time with
only very slight alteration in conditions at each step. Or it may
have been only that the range of capacity to withstand conditions
was not sufficient even after the utmost had been done in accli-
matisation. Time and gradual progression are the most essential
things in acclimatisation.

A very great difficulty in the path of acceptance of natural
selection as a cause for gradual adaptation is the fact that so
many of what look like real morphological adaptations require so
much correlation. Climbing plants come into this group, though
they are obviously well suited to climbing. The habit cannot be
difficult to acquire, for there are so many cases of the closest
relatives, one climbing, one erect. A climber also needs a support,
which is usually an erect plant, so that erect plants must have
been the earlier. But one cannot imagine natural selection
picking out the beginnings of weak and flexible stems, whether
by gradual change or by small mutations. And when at last they
were formed, as obviously there would be no value in developing
tendrils or other means of climbing until the stems were weak,
they would collapse into the darker lower levels of vegetation,
and would have to undergo physiological adaptation to living in
greater darkness. Then they would have to learn to form climbing
organs, and finally, learning to climb, they would once more have
to adapt themselves to life in greater light. And what use would
the beginnings of tendrils or other climbing organs be? And why,
after having learnt to live in greater darkness, should the plant
want to grow up into the light once more? Yet it would be dragged
up by the tendrils, and would probably suff'er from the excess of
light. There is too much, and too complicated internal adaptation
required, to say nothing of the external. One must look with
great suspicion upon such an easy interpretation of such struc-
tural features as climbing stems as being simply adaptations. If
they were gradually formed, the work was too complicated for
natural selection to perform.

58 CONTACTS WITH DARWINISM IV [ch. vi

It is clear that in adaptation to climbing a large part of the
adaptation, if not perhaps all, must be internal and physiological,
and we are inclined to think that it is to such adaptation that
the name should be practically confined, while such things as
climbing plants might be called suited to climbing. If a plant, as
will usually be the case, move only a very small distance from the
parent, it is probable that it will not need more than the minimum
of physiological adaptation to suit it to the new place, and so on
at every move. But such adaptation will not necessarily show
any morphological changes visible to the outside. If one look at
the distribution of such a widespread plant as Hydrocotyle
asiatica, which ranges from the plains of Ceylon, with a tem-
perature range of 70-90° F., to the south of New Zealand with
winter snow and frost and a weak sun, one finds it to be essentially
the same plant throughout. The Ceylon plants are suited to the
Ceylon conditions, the New Zealand to those of New Zealand.
But it is customary to speak of it as "adapted" to both. If it
suits them both, it must be just a case of luck, with local adapta-
tion going on as it has moved from one to another. One very
much doubts, after considerable experience with acclimatisation,
if seed from the plains of Ceylon would suit New Zealand without
a lot of previous physiological adaptation, or vice versa.

Liberian coffee was gradually acclimatised to higher levels in
Java by carrying seed a little higher at each generation. In Ceylon,
when we tried to acclimatise the beautiful Cyperus Papyrus with
European seed, we failed, but seed from India was a success.

The whole question of correlation of characters is an extremely
difficult one when looked at from the point of view of natural
selection. If large, it implies that most of the characters con-
cerned have no bearing upon natural selection, and do not
interfere with the results produced by the modification in the
first character, thus further implying that the change in that is
sufficient to carry the new species past the line of mutual sterility
that will usually divide it from the old. The characters of
climbing plants had some evident connection, for all were useful
in climbing, but that does not apply to the characters that one
finds correlated in an ordinary species, which have no apparent
connection of any kind, nor anything to which one can attach
any adaptational value. Their best explanation seems to be that
they have gone together in the apparently purposeless and un-
accountable way in which characters in mutations so often seem

to go.

CH. VI] ADAPTATION 59

The mere fact that the prominent genera that occupy any kind
of marked ecological standpoint, such as a bog, a saltmarsh, a
mountain, a chalk-down, are usually the large and widespread
genera, is enough to show that there was but little selection —
they were the oldest and got there first, and being adaptable they
became functionally modified to suit their new surroundings.

What has been said about gradual adaptation applies equally
to the view at present rather in favour that mutations were
small, and that selection presently resulted in another small step,
and so on. But what is to ensure that a small step in one direction
shall be followed by a second, or that conditions shall continue
to change in such a way as to make it worth while for such a
thing to occur?

The balance of probability would seem to be in favour of the
appearance of structural characters by single mutations, and in
that case it seems rather absurd to talk about adaptations in
them. The adaptation is rather the internal and functional
adaptation.

It would seem quite possible that climatic conditions all over
the world have been gradually differentiating and becoming more
varied as time has passed. On the whole, they have almost
certainly become drier, though probably not in such places as
many coastal regions. This would affect newly formed species by
gradual^ restricting their freedom of movement, or even by
forming impassable barriers. To move in a region of more or less
uniform climate would probably require comparatively little of
fresh adaptation to each new habitat, but if the climate were
changing from one place to another, this adaptation would have
to be greater, and would presumably need more time. This would
in turn make the rate of travel slower, and it is quite possible
that the change of climate might, so to speak, pass it upon the
way, and erect a barrier some distance in front, the species
reaching the limit of possible acclimatisation. This would seem
to have happened in Ceylon, for example, where the island is
rather sharply marked out into dry and wet zones. Comparatively
few species are found on both sides of the divide, and really
frequent in both zones. Many genera show a number of species
in the wet zone with few in the dry, others the reverse, whilst of
the genera that are confined to one zone, most occur in the wet.

It is clear that it is somewhat stretching a point to say that
new genera, arising locally, as we have seen will in all probability
be the case, are adapted to wide spread over the world. As only

60 CONTACTS WITH DARWINISM IV [ch. vi

rarely do the really large families show a single genus with the
range of the whole family, a feature very common in small and
frequent in medium-sized families, and as still more rarely does
that genus in a large family show a species with the whole range
of the genus, it is clear that any adaptation responsible for wide
spread must be generic. What is much more probably the case,
inasmuch as these widespread genera are admittedly of simple
rather than of complex type, is that the parent of the genus still
possesses great adaptability, or suitability to a considerable range
of conditions. This will enable it to move far with less difficulty
than usual, and as at the same time its structural evolution, which
has probably little or no relation to adaptation, will be going on,
it will give rise to more and more species. These will probably
inherit their parent's general suitability to conditions, but it is
quite probable that it may all the time be getting less (perhaps at
each mutation), so that each new species may be liable to become
more localised than its predecessor in regard to the total range
possible to it, while at any given time it will of course be more
local on account of its greater youth.

CHAPTER VII

ISOLATION

In a paper on the floras of hill-tops in Ceylon, published in
1908 (57), the author drew attention to the great proportion of
local endemics — one-eighth of the total number of endemic
species — that were to be found upon one only, or upon more than
one, of the mountain tops of the south-west of Ceylon. The
principal massif (the central) is to the south-west, a smaller to
the north-east of it, and there are a number of more or less
isolated peaks separate from them, the most isolated being Riti-
gala in the north of the island (p. 24). The highest summit,
Pedurutalagala, attains 8296 ft. ; Adam's Peak, the best known,
is 7353 ft., and is rather isolated at the south-western edge of the
central mass. There are ninety-seven well-marked Linnean species
endemic upon these mountain tops, with eleven varieties of these
or other species, some of which are usually reckoned as indepen-
dent species. Of the species upon single mountain tops, there are
no fewer than twelve upon Adam's Peak, which is so steep that
its summit does not present any great area of vegetation for the
last 2000 ft.

Since the widely distributed species, those that were not en-
demic to Ceylon, however localised in Ceylon they might be, were
never confined to hill-tops, it was clear that there was, quite
probably, some definite force or influence acting to cause these
local endemics to exist in the places where they occurred. For a
long time, opponents of my views maintained that they were
relics of previous vegetation, and in fact this view is still popular.
As they occur in general at higher levels than other species of their
genera that are found in Ceylon, it was suggested that they had, so
to speak, fled up the hills from their rivals. But if they could do
this, they must have had a good capacity for internal, physio-
logical adaptation, and it seems strange that they could not adapt
themselves to staying where they were. And as most of the
mountains rise from a central plateau, it seems very remarkable
that so many of them should each be upon its own mountain. It
seemed to me very probable that each was an endemic that had
arisen upon the spot where it was found, and at various times, in
conversation and elsewhere, I have suggested that the immediate

62 ISOLATION [ch. vii

mechanism of their formation might be the action of cosmic rays,
which would be more marked at high elevations.

It was clear that these mountain tops showed a distribution of
plants like that which was shown by a group of islands forming
an archipelago. Now the only thing obviously in common be-
tween the two was isolation, and I therefore drew the conclusion
that isolation as isolation favoured the production of new forms.
At that time we knew little or nothing about the genes and
chromosomes, and since then Harland has put forward the likely
suggestion that long continued gene separation may lead to gene
change, which of course in its turn might lead to definite mutation.
Since about three-quarters of this mountain-top flora has no
special adaptation for distribution by wind or by animals, it is
highly probable that individuals of the more widely distributed
species lower down would very rarely reach the higher summits,
whose plants would be, and remain, very isolated. In this con-
nection it is worth special notice that the islands which show con-
siderable local endemism, like the Hawaiian islands, are very
commonlv mountainous.

Whether the mutation w^hich the author considers to have been
the origin of any one of the species was due to one or the other
cause, or to both, there would not be, in either case, any serious
opening for gradual adaptation under the influence of natural
selection. It must also be remembered that the number of indi-
viduals is very small (cf. p. 25). In this connection, I may quote
from Age and Area the footnote on p. 206: "A few days before
I left Rio, Dr Lofgren found, on a little island about three miles
off the coast, a new and very distinct Rhipsalis, of enormous size.
He told me that there were only four examples on the island."
I may also refer to the examples given in the same book on
p. 151.

Sixty-eight of the 108 endemics (including the eleven varieties)
are found upon one mountain only, the other forty upon more
than one. It is a very striking fact that these mountain endemics
belong, not to small and local genera, but chiefly to large and
widespread ones, as was shown on p. 26.

The general conclusion from this piece of work is that isolation
favours the development of new forms, and that local conditions
have but little eff'ect in developing, though they may have much
in determining the survival of, these new forms, and that conse-
quently natural selection, upon adaptational grounds, is unlikely.
It is more than doubtful whether any given species has been

CH. vii] ISOLATION 63

specially adapted to the exact local conditions in which it is
found, except by the internal, physiological adaptation that must
always be going on. It would be killed out at birth if not reason-
ably well suited to the local conditions.

Turrill (47) has shown that in the Balkans one may find pairs
of altitudinally differing species like Bellis longifolia and sylvatica,
a fact which affords further evidence in favour of the author's
view that isolation and elevation, one or both, may lead to the
formation of endemic species upon mountains. In Ceylon, of the
sixty-two genera represented by endemic species upon mountain
tops, forty-three also have endemics at lower elevations, and only
nineteen have not, a fact which makes the supposition that
those of high levels are relics seem a little far-fetched.

The Podostemaceae as a family are very isolated, and they
grow submerged in water, usually at what are only moderate
elevations, yet they have many species. Though isolated from
other plants, they usually cover their own habitat, the rocks,
fairly thickly, so that one hesitates to suggest that they would
have so many species were they really isolated as individuals. It
would seem more likely in their case that they owe their numbers
to the overhead force of plagiotropism that is always at work
upon them. There are probably quite a number of causes that
may lead to the formation of new species.

Lakes formed by elevation of the coast of the Black Sea con-
tain, I am assured, endemic species of cockles, a fact which would
seem to favour isolation, especially as they are close to sea-level.

Siparuna (p. 35) has the great bulk of its local species in the
mountains rather than in the plains, and the same is the case with
many other genera, whilst many of the isolated islands that
contain so many endemics are also mountainous. These facts
might seem in favour of elevation (cosmic rays) rather than isola-
tion, but other plants, such as the Dipterocarps, show many
endemic species in the plains, usually in dense forest. Here
species formation is probably connected with age.

Probably both isolation and elevation may be potent causes
leading to well-marked development of new species. In the
former case mutation is quite probably due to slow gene change,
as Harland has suggested, but this would probably bring about
sudden mutation by the adding up of strains until they became
so strong as to cause some sudden kaleidoscopic change. In the
latter case, if the cause of mutation be some effect of the bom-
bardment of the genes by cosmic rays, one might expect the

64 ISOLATION [ch. vii

mutation to be sudden, and as most of the mountain endemics
are well-marked Linnean species, it was perhaps very definite
also, the principle of divergence of character coming into
operation.

Another well-known series of facts is that families which are
widely distributed chiefly or only in the more broken southern
hemisphere, have rarely any genera that cover the whole of their
area of distribution. In the plants that are more marked in the
northern hemisphere, on the other hand, there is very often a
genus that does cover the whole area, even if that also includes
the southern hemisphere, such for example as Ranunculus^
Senecio, or Solanu7n. Whether this difference has anything to do
with the isolation of so many areas in the south, we do not know,
but the fact is suggestive.

We are still very far indeed from any proper understanding of
the operations that have been concerned in evolution, except
that natural selection must evidently play a less conspicuous, or
at any rate, a less direct part. It looks as if, more especially under
certain circumstances such as elevation or isolation, evolution
must go on, and this supposition is borne out by such things as
the progressive change that shows in such plants as Stratiotes
described by Miss Chandler (3), where a whole series of species
differing in characters of no conceivable functional importance,
have succeeded one another in successive geological horizons. If
these changes had been a little more marked, we should have had
two or more genera succeeding one another, and this point must
always be borne in mind, together with the tendency to diver-
gence, in considering extinct genera.

CHAPTER VIII

DIFFERENTIATION

VV ITH his customary scrupulous fairness, Darwin went out of
his way to draw attention to an axiom of taxonomic botany that
was seriously opposed to the theory of evolution by adaptation
through the agency of natural selection. "Those classes and
families which are the least complex in organisation are the most
widely distributed, that is to say that they contain a larger pro-
portion of widely distributed species. " Incidentally, as the
simpler families must upon the whole be the older, this goes a
good way towards proving the correctness of the theory of age
and area.

Now upon the theory of natural selection, it is clear that the
successful genera must be those that have the largest numbers of
species, or the widest distribution, or both; but as they have been
developed by adaptive selection, they should surely on the whole
be the most complex and specialised, showing the most signs of
adaptation. This has always been a difficulty to the supporters of
natural selection, and one which has been passed over with little
remark. It can be at once explained by the hypothesis brought
forward in Age and Area, for upon that the older forms will be
the more widespread, and by reason of their age they must be
the simpler on the whole, as having been more early formed in the
process of evolution. But age and area is incompatible with the
theory of natural selection.

Age and area leads on directly to the theory which Guppy has
called Differentiation, though a simpler and better descriptive
term might perhaps be found — mutation perhaps, or differential
or divergent mutation, for example, if it were admitted that
mutations might be large. The essential feature of the theory,
originally adumbrated by Geoffroy St Hilaire (41), is that evolu-
tionary change goes downwards from the family towards the
species, not in the opposite direction. A family begins as a family,
and is not graduallv formed bv the destruction of intermediates.
At the same time, of course, when it begins it is also a genus and
a species, which at the start are all-important to the family; if the
species be killed out, the family disappears. As it grows, the single
genera and species become less important to it. The name

WED S

66 DIFFERENTIATION [ch. viii

differentiation was given by Guppy (11), whose concept it was that
in the far back days of damper and more uniform climate most of
what are now the large (or widespread, or both) families were
formed, each at one stroke by well-marked mutations, and they
then slowly began to grow in size by further mutations. As time
went on, and the earth perhaps became drier on the whole, the
variety of climate would increase, and mutations perhaps be
more rapid, but their "size" is supposed to have become less, so
that fewer great divisions, like for example the Monocotyledons,
tended to appear. As differentiation went on in the climates, so
it went on in the living forms. This does not mean that they were
necessarily formed in adaptation to the climates but rather
perhaps that the climatic change gave the stimulus which resulted
in further mutations. Mutations might be of any rank, from
variety up to division, so that any difference might appear at one
stroke. If the newly formed plant could pass through the sieve
of natural selection, and escape the dangers that threatened its
very existence when it first began, it might then begin to spread,
and once established in several places it would be, comparatively
speaking, safe. As the original species thus survived as well as
the offspring, the family must necessarily increase in number in
such a way that when plotted by their numbers of species, its
genera would form the "hollow" curve. It is quite possible
that after a certain lapse of time a species 77iust die out (43),
and it is still more possible that it may change into another by
some simultaneous mutation. We have seen a small instance of
simultaneous mutation in the sudden loss of smell that happened
to all the plants of musk some years ago, and may perhaps see
the results of series of such mutations in the consecutive species
of Stratiotes described by Miss Chandler (3) and other such
series.

While under natural selection new forms only arise as the
result of improvements in adaptation, under differentiation they
arise because evolution must go on, at any rate whenever the
needful stimuli, or conditions, are present, as we have seen in the
case of the Podostemaceae (p. 20). Under natural selection the
small variety becomes a larger one, and so on. It seems to the
writer, as it did to Dr Guppy, that in trying to make evolution
work in this way, people have been trying to work it backwards,
and it is with the object of showing the necessity of proper
revision of the current view that the present book is written.
A number of more or less crucial test cases are given below, all of

CH. viii] DIFFERENTIATION 67

which seem to point to the supposition that differentiation gives
a more correct picture of the direction of movement of evolu-
tionary change than does natural selection, even though we have
no clear vision of the mechanism that was involved in making the
changes that occurred.

The family is supposed to have arisen by some well-marked
and sudden mutation (or conceivably a series of smaller ones,
probably at close intervals), which would at one stroke change
two or more characters and pass the line of mutual sterility that
commonly divides species from one another. As the characters
that divide the families are, after all, not so very numerous
(cf. Appendix I) each family must take a different combination,
sometimes taking one of a given pair, sometimes the other, and
in every kind of mixture. Many families, for example, have
alternate rather than opposite leaves, or superior rather than
inferior ovary, but only the Cruciferae have alternate exstipulate
leaves, bractless racemes of ^ regular flowers, sepals in two whorls
of two, four petals, two short and four long stamens, superior
ovary of two carpels, unilocular with replum, a pod-like fruit,
and exalbuminous seeds. As the characters run in contrasted
pairs (or triads), we have no information as to whether there is
any advantage in one side rather than the other, or in either as
against any possible intermediate, or indeed that any has any
adaptational value. There is thus no evidence to show in which
direction evolution moved, and we are perfectly free to select
that for which we think that the evidence is better. It is this
evidence, or rather, some of it, which we propose to bring forward
below.

Nor can we say with any likelihood of accuracy that the
change indicated in any one pair is larger than that in another.
Is it a greater change from two cotyledons to one than from
alternate to opposite leaves? We do not know; all we have to go
upon is that the latter is much more common. With one or two
rare exceptions, there is no difficulty in supposing all Mono-
cotyledons to have descended from at most a few different
ancestors, whilst one may find alternate and opposite leaves side
by side in many cases of allied genera or species. There is nothing
inherently absurd in the idea that a family might be founded by
a single mutation.

About 1902 the writer became a convert to the theory of
mutation, but it seemed to him completely illogical to insist that
mutation could only be very small, when before us, in every

5-2

68 DIFFERENTIATION [ch. viii

family, there lay so much evidence that species, genera, tribes,
sub-families and families were so continually separated by such
well-marked divergent characters as leaves opposite or alternate,
anthers opening by slits or by pores, ovules l-2-oo in each
loculus, raphe dorsal or ventral, and many more such differences,
which allowed of no intermediate or transition forms upon
which natural selection might operate, which were such that one
could not conceive of natural selection choosing between them,
and which were so constant in their inorphological character — a
feature that one could not expect natural selection to bring forth.
They could only, it would appear, be the result of definite single
mutations, and therefore mutations must at times be large. And
if large in regard to these characters, which are very often of
"family" rank, why not in all cases?

In May 1907, without having seen Dr Guppy's book, the
author published what was essentially the same theory (70),
largely based upon the study of the Podostemaceae, and upon
ten years' experience of tropical vegetation. Both authors were
convinced that the great importance at that time attributed to
adaptation was exaggerated. Natural selection was trying to
construct a tree from the twigs downwards. But though a tree
grows from the ground upwards, it always has young twigs and
leaves (which may be looked upon as representing genera and
species), though each one, when the tree is small, has a much
greater value in proportion to the whole organism than when the
tree is large. It seemed to us clear that in trying to show that
evolution proceeded in the order

Small variety — Large — Species — Genus — etc. ,

people were trying to make it work backwards, and that the
proper order was

Family — Tribe — Genus — Species — Variety.

The relative rank of these groups varied as time went on. When
very young, the family, the genus, and the species were the same,
but as the family grew in size (just as with the tree mentioned
above) the species became of less and less relative rank when
compared to it.

To turn to geographical distribution; upon the theory of
natural selection, the large and widely distributed genera are the
successes, the small and local the failures or relics. The success
was always put down to better adaptation to conditions, though
no one tried to explain how a species that derived its adaptation.

CH. viii] DIFFERENTIATION 69

say, in Europe, was able to spread as far as New Zealand. It
could not become, in Europe, adapted to the conditions of New
Zealand, and its appearance there must be due simply to the
chance that the conditions resembled one another in both places,
and that there were conditions in between that were not dissimilar
at the time that the plant reached them. There can be no doubt
that as a plant moves very slowly about the world, it can become
adapted as it moves to the slightly different conditions that occur
at each move. If it come to a place where the change is too
sudden for it to adapt itself, it will then have come against a
barrier to further spread — an ecological or a climatic barrier, to
be added to the barriers of physical nature, such as high moun-
tains, or open seas, that so often occur. By the formation of one
of these barriers after the plant has passed, the distribution of a
species may become discontinuous.

The success of a species under natural selection means usually
the greater or less extermination of the one from which it
descended, and which was not so well adapted. Under the theory
of differentiation, on the other hand, which goes with that of age
and area, the large and " successful " genera are simply the oldest,
while the small and local are in general the youngest. There is no
special adaptational reason for size or spread, so that, within any
close circle of relationship (which will more or less ensure the
same general reactions to the outside world), the rate of spread of
two or more forms will not usually be widely different. This is the
essence of the theory put forward in Age and Area.

In diagram 7 we have indicated what we imagine to be the
probable general course of evolution under the theory of dif-
ferentiation. The family is represented at the start by a solitary
monospecific genus A, which will throw off new forms by muta-
tions. At first they will probably be produced very slowly indeed,
but as A increases its numbers (and with them its area, thus
probably coming under the stimulus of different conditions) will
probably appear more rapidly. Whether the earliest mutations
will be more often specific or generic we have no idea, but most
probably a second genus B will be "thrown" before so very long
a time. This will again begin slowly, and it will be a long time
before it throws a new genus Bb, whereas A will probably throw
its second new genus C before Bb appears. All this, of course, is
dealing in averages, and we do not know that this particular B
will necessarily be slower than A, though on the average the
second genus will be behind the first throughout. On the average

70

DIFFERENTIATION

[CH. VIII

A, as the oldest genus, should have the greatest area and the
greatest number of species, B the second, C the third, Bb probably
the fourth, and so on, but only on averages. Whilst in Ranuncu-
laceae Ranunculus has 325 species to 250 in Clematis, one would
hesitate, and rightly so, to say that the former was the older, when
one remembers that it is herbaceous, and Clematis shrubby.
As time goes on, it is clear that the rate at which new genera

Fig. 7. Evolution by differentiation. Each genus is supposed to survive

the whole way along the line at right angles to its origin, e.g. A still

survives at H, B at Be, and so on. In order to save complication, the

lines to show the gro^^i:h Bb, Be, &c. are not shown.

are formed will increase. Each genus will begin with one species,
and after a time will form more, so that the few older genera of
the family will contain the greatest numbers of species. The result
will be (cf. 66, p. 185) the gradual formation of the familiar
hollow curve already described, with a few large genera of dif-
ferent sizes at the top, and many monospecific genera at the
bottom, the numbers increasing from top to bottom at an
accelerating rate. As there will rarely, upon this theory, be any
appreciable adaptational difference between species or even
genera, there will be little or no reason why the older ones should
be killed out (as there is under natural selection), and so the
increase in numbers will lead inevitablv to the hollow curve.

cn. viii] DIFFERENTIATION 71

Up to the present, this theory is the only one which can make
any pretence of explaining the hollow curve. The latter is so
universal that it is evidentlv a general law which must be ex-
plained. But it is, of course, in direct contradiction to the theory
of natural selection. With the latter theorv one can make no
predictions as to what may be found in the arrangement and
characters of families, genera, or species. With differentiation
one can make a beginning in this direction, and this alone makes
a strong claim in its favour.

If differentiation be the rule, it is clear that the ultimate result
of the growth of a family from one original genus ^4 to a fair
number of genera should in general be the formation of groups
within groups, like the cat group or the dog group within the
larger group of Mammals. By the principles of differentiation
and of divergence of variation, each genus thrown will tend to be
markedly different from the parent that throws it. If it were
thrown very far back in the history of the famil}^ it will have
had time to throw more (^enera in its turn. These mav, but so far
as one can see, not necessarily must, display the character or
characters that made their parental genus, B for example, dif-
ferent from its own parent A upon the main line of the family.
When these genera upon the second line B had become more or
less numerous, or in any case if the characters of their parent had
included two or more of the characters w^hich we usually rank as
"family" characters, they would form a group Bh, Be, Bd, etc.,
well marked off from the first group which was being formed by
the genera upon the main line. A, B, C, D, etc., and to these two
groups we should probably give the rank of sub-families, or
tribes, according to our conception of the value of their characters.
One of these groups would most probably be larger and perhaps
more widely dispersed than the other, and both would continue
to grow and to spread. Supposing that the family escapes with-
out very great damage all the various accidents that may befall
it, and that all its genera behave fairly closely in the same way,
as would be the case under differentiation, the original parent A
will have the largest number of species (theoretically) and the
largest area of occupation, while the other genera, B, C, etc.,
will be successively smaller in these respects, as we have seen
in the ^Monimiaceae for example. The difficulty in defining what
is or is not a sub-family or a tribe is the same as that of defining
a genus or a species. We have no standard to work by in defining
the value of a certain character, other than the way in which it

72 DIFFERENTIATION [ch. viii

appears in the group under consideration. Ruminate endosperm
being characteristic of all Annonaceae, and of none of the allied
Magnoliaceae, becomes very important in regard to these two
families, while in the palms, etc., it may characterise some only
of the species of a genus. Upon the theory of differentiation this,
of course, simply means that in the one case the ancestor that
showed it was the ancestor of a whole family, in the other only
of a few species. Any of those characters which we usually con-
sider as especially "family" characters may appear at any stage
from family down to species, but on the whole are more common
as one goes upwards in a family from the species.

One thing that is always brought up as an argument against
those who object to the explanation of evolution by natural selec-
tion is that the fossil records show many extinct genera, of
families still existing. The theory of natural selection, based upon
adaptation, with its prompt killing out of less-adapted ancestors,
accounts easily for this, while differentiation, which supposes the
ancestors to live on together with their descendants, cannot do so.
But one is apt to forget that the explanations of the facts of
palaeobotany have for many years been such as could be made to
jfit with the all-powerful theory of selection. One is reminded of
the defence of phrenology in The Professor at the Breakfast Table.
There are a number of things that must be taken into considera-
tion before one can fully explain the fossil records.

In the first place, it seems not impossible, as Small has shown
(43), that there may be a definite limit to the life of species and
genera. In his summary he says: "From this the important de-
duction can be made that species die a normal death, presumably
from the senescent sterility of old age, with, perhaps, a minor
part being played by the progressive restriction of survival condi-
tions for a senescent species . . . the species number in a genus is
shown to follow the series

1-2-4-8-16-32-44-43-41-37-29-13-0.

This gives 24 million years as the normal lifetime of an
ordinary genus."

This is supported by such facts as those brought out by Miss
Chandler (3), who found in different recent horizons a whole
series of fossil species of Stratiotes, differing structurally from one
another, but with nothing to which one could possibly attribute
any adaptational value. The loss of smell by musk (p. 66) shows
that a whole species can undergo a simultaneous change ; a larger

CH. viii] DIFFERENTIATION 73

mutation than this might have changed the whole of it to another
species. Then again, if a geological catastrophe come along, it
may easily destroy a whole species, or even genus, that has not
yet been able to spread far enough to get beyond its range. Unless
a fossil is found to cover such an area that it is unlikely that such
a fate may have overtaken its living representatives, it seems to
the writer in the highest degree unsafe to look upon it as an
ancestral form of existing species. It is more likely to be a lateral
mutation thrown off from the main line, and exterminated as a
genus by some happening.

Lastly, there should be mentioned the all but complete absence
of transition stages in the fossils, a fact which violently disagrees
with the supposition that evolution was gradual and continuous.

CHAPTER IX

DIVERGENCE OF VARIATION

1 T has long been known, though it has excited but little interest^
that there is a great tendency in variation to be divergent. As
Guppy says (66, p. 104) Hooker, in his lecture upon Insular
Floras, "shadowed out a general notion of Centrifugal Variation
operating through countless ages'. It appears almost as a sug-
gestion, but the idea had been evidently floating half-formed in
his mind ever since he wrote his essay on the Tasmanian flora in
the late 'fifties. It was the nucleus of a theory of Divergence or
Differentiation that acquired more definite outlines as time went
on, since it reappears in the intensely interesting account of a talk
with Darwin which is given in a letter to Huxley in 1888 (19,
n, p. 306).

" We can perhaps understand the long intervals of time now.
For the confirmation that such a theorv would have derived
from a line of research instituted on Darwin's lines was denied
to him. The two proved to be incompatible. For no inductive
process based on Darwin's lines could have found its goal in
a theory of centrifugal variation. 'I well remember', Hooker
describes in a letter to Huxlev in 1888, 'the worrv which that
tendency to divergence caused him (Darwin). I believe I first
pointed the defect out to him, at least I insisted from the first
on his entertaining a crude idea which held that variation was
a centrifugal force, whether it resulted in species or not.' Huxley
was in the same case. For he held views of the general differen-
tiation of types, and his road that would lead to the discovery
of the causes of evolution started from the Darwinian position.
That road was barred to him."

There can be no doubt, when one looks at the various characters
that are used in taxonomic distinction between one form and
another, that the bulk of them are divergent, and that the more
so the higher one goes in the tables of characters, upwards from
species to families. Take for example the list of "family"
characters given in Appendix I, and note the great proportion of
distinctions in which there cannot even be an intermediate, by
reason of the marked divergence, and where, in any case, there
can be no functional difference between the intermediate and the

CH. IX] DIVERGENCE OF VARIATION 75

two extremes. For instance, among the characters will be found
the following:

Root tap or adventitious
Stem monopodia! or sympodial
Leaves alternate or opposite
simple or compound
palmate or pinnate
parallel or net veined
Inflorescence racemose or cymose
Flower spiral or cyclic

mon- or di-oecious
iso- or hetero-merous
regular or zygomorphic
Receptacle above or below calyx
Parts of flower in 2s, 3s, 4s, 5s, etc.
Calyx in one or two whorls
Odd sepal anterior or posterior
Corolla free or united

imbricate, valvate, or convolute
alternate with, or superposed to, sepals
Stamens in one, two, or more whorls

diplostemonous or obdiplostemonous
free or united
Anther versatile or not

opening by slits, pores, valves, etc.
Pollen in various patterns of cell wall
Carpels free or united

1 to 00

transverse or anterojDOsterior
Placentation parietal, axile, etc.
Raphe ventral or dorsal
Micropyle up or down
Style basal or terminal
Stigma capitate or lobed

Fruit achene, follicle, capsule, drupe, berry, etc.
Seed with or without endosperm

one, few, or many
Embryo straight, curved, twisted, etc.

and many more equally divergent, whilst in the few cases where
intermediates are possible, no functional value or disadvantage
can be read, either into them or into one of the extreme diver-
gents. In no case, in these family characters, has any functional
value been shown, in a definite and unmistakable manner,
though suggestions have been made in one or two cases.

If now one go on to the characters used in the keys which
determine the genus and species of a plant belonging to one of

76 DIVERGENCE OF VARIATION [ch. ix

these families, one finds the same kind of divergences, more and
more marked on the whole in approaching the top of the list (the
first divisions in the keys), and least marked in the characters
that distinguish one species from another. This fact of increasing
divergence as one gets nearer to the top of the list has always
been a great difficulty in the path of the supporters of natural
selection, and has been left discreetly unmentioned by them.

Opening a volume of Engler, the family displayed is the Cyclan-
thaceae, composed of six genera only. The first and most obvious
division, into Carludoviceae with male flowers in fours, and
Cyclantheae with male and female flowers in alternating rings or
spirals, picks out the two most important genera, one in each of
the groups, though the first group, with five genera and forty-five
species, is much larger than the second, with one and four.
IncidentaUy, how did selection, or gradual adaptation, produce
these two very distinct types of inflorescence, and what was
intermediate between them? Taking first the Carludoviceae, the
genus Ludovia, with two species in Guiana and Amazonas, is
first cut off*, having only a rudimentary perianth in the male
flower (why, on the theory of selection, did it spoil its attractive-
ness to insects?). The four genera left are divided into Carlu-
dovica, which has forty species covering the whole range of the
family in tropical America (north and south) and the West
Indies, and which has a short perianth in the female flower,
against a long one in the other three (again, attractiveness
apparently spoiled), and an inferior ovary against a superior.
Carludovica is by far the largest genus in the family, far out-
numbering all the rest put together, and has a distribution
covering that of the whole family, just as we have seen to be a
general rule (p. 64). Does it owe its "success" to its inferior
ovary, and if so, wherein does the advantage lie, for the flowers
are so crowded that one cannot tell from the outside that the
ovary is inferior? And if there is an advantage there, what about
the reduced perianth?

Evodianthus, next, is distinguished from the two other genera
by having the stamens inserted in the tube of the perianth, while
the others have them on the disk ; and finally Stelestylis has the
stalk of the male perianth flat and hollow, and a pyramidal style,
while Sarcinanthus has the male perianth forming six-sided
pyramids, and no style. All three are small and little dispersed
genera with two species in Costa Rica and the West Indies, one
in eastern Brazil, and one in Costa Rica, respectively.

CH. IX] DIVERGENCE OF VARIATION 77

Cyclantheae has only one genus, Cyclanthus, with four species
in tropical South America from Peru northwards, and in the
West Indies. It is thus the second genus of the family both in
number of species and in dispersal, but as it is so much smaller
than Carludovica, we must suppose that it was only cut off from
that genus rather late. Its dispersal is much smaller than that of
Carludovica.

There is no reason for supposing the small genera to be relics;
it is far simpler to imagine them all split off by large mutations
from Carludovica in its gradual dispersal over the whole area of
the family. The first split probably gave Cyclanthus, which is the
second largest genus, and has a division to itself in the family,
whilst it would seem extremely probable that the mutation that
gave rise to it was an extra "large'" one, for the difference is so
great between it and the rest. The other four genera, smaller, and
with less distribution, count in the same group with Carludovica,
from which they are not so markedly different. The general im-
pression that one gains here, as in almost all cases, is that after
the big mutation which first gave rise to the family, there follow^ed
others which gradually became less and less marked, and which
kept more or less closely wathin the boundaries that were indi-
cated by the first mutation that occurred after the formation of
the family.

The first impulse of many will be to say that Cyclanthaceae
form an exceptional famil}^ and perhaps also to say that the keys
are artificial things. But the exceptional families, and the diver-
gences that are shown in the keys, have both to be explained by
natural selection, or by any other theory of evolution, just as
much as have the ordinary families and the smallest divergences.
Natural selection would be very hardly pressed to find any ex-
planation of the very remarkable differences between Carludovica
and Cyclanthus, especially as it is all but impossible even to
imagine that there can be any intermediate stages, and no use-
value can be put to either of the extremes or to any conceivable
intermediate.

This supposition, that the first mutation, in a family newly
formed by a large change from some ancestral form, may be in
turn large, is well supported by an examination of the keys to the
various families that are given in any general text-book of
systematic botany. In the list in Appendix II, I have extracted
from the keys in my Dictionary, 6th ed. (which are mostly taken
from the Natiirlichen Pflanzenfamilien), the first dichotomy in

78 DIVERGENCE OF VARIATION [ch. ix

each case, omitting a few keys where the first break is into three
or more. These sixty famihes of course were selected for the
Dictionary as being the larger or more important, and we shall
go on to deal with the smaller ones below.

It will be seen at once that these are characters the bulk of
which are of the same rank as the "family" characters given in
Appendix I. To take a few examples, one finds among them such
character-pairs as these, which can all be matched in the family
characters :

1. Leaves opposite — alternate Gentianaceae

2. Leaves in two ranks — not Musaceae

3. Inflorescence racemose — cymose Verbenaceae

4. Flower naked — with perianth Betulaceae

5. Perianth actinomorphic — zygomorphic Campanulaceae

6. Calyx polysepalous — gamosepalous Caryophyllaceae

7. Calyx valvate — imbricate Mimoseae (Legum.)

8. Stamens free — in tube Meliaceae

9. Stamens two — one Orchidaceae

10. Carpels free — united Annonaceae

11. Carpels six to fifteen — three to five Hydrocharitaceae

12. Ovule one per loculus — two Euphorbiaceae

13. Fruit a berry — loculicidal capsule —

septicidal capsule Ericaceae

14. Fruit many-seeded — one-seeded Myrsinaceae

15. Fruit achene — follicle Ranunculaceae

Incidentally, how does natural selection account for, or explain,
the difi'erences shown in 2, 3, 6, 7, 9, 11, 12, 13, 14, 15, to say
nothing of the others?

If one go on to the second dichotomy in a key to a big family,
one finds that there are on the whole fewer, though still some, of
the "family" differences shown, but these become less frequent
in proportion to the total, as one goes down the list.

It might be thought, perhaps, that small families would show
a diff'erence from the larger ones, possibly in having smaller
divergences in their classification into genera. If they were really
relics, as they are often supposed to be, this might be the case,
but in actual fact it is not found, as a glance at Appendix III will
show. This contains the distinguishing characters of the genera in
the families that contain two only. Here again one finds such
distinctions as:

Leaves opposite — alternate Caryocaraceae

Erythoxylaceae
Trigoniaceae

CH. IX] DIVERGENCE OF VARIATION 79

Perianth five — four Achatocarpaceae

Flower 5-nierous — 3-merous Limnanthaceae

K and C alternate — superposed Caricaceae

Corolla valvate — convolute Quiinaceae

Corolla free — united Xyridaceae

Stamens few — oo Salicaceae

Carpels two — three Balanopsidaceae

Capsule — berry Balsaminaceae

Taccaceae

It is clear that in these small families the first split, which is
only into two genera, shows just as important divergences as does
the first split in the large families, which is into two sub-families,
and the two genera of the small family are just as well separated
as are the two (average) largest genera of the big family, which
head its two chief sub-groups. The importance of this fact we shall
better appreciate when we return to its discussion in the Test
Cases (below, p. 112). If small families really consisted of relics,
one would not expect that their genera should be divided by
divergences of any special size, and certainly not that the diver-
gences would be of the size and forin that one expects to find
between the sub-families of large families, or even between the
large families themselves.

If one take a number of monotypic families, or families of one
genus, from the first edition of Engler, and look at the distinc-
tions there given for dividing the species of each of the genera
into two chief groups, one finds these characters to be of some
systematic importance, and often to be characters that are not, or
hardly, capable of having intermediates. It is very hard to see
how characters of such divergence should be those supposed to
be left in the genera that survive of what is supposed to be a
dying family. Here are a few examples :

Monocotyledons .

Typhaceae. Fruit with longitudinal groove, and open-

ing in water; seed not united to fruit
wall.
Fruit without groove, not opening in water;
seed united to fruit wall.

Sparganiaceae. Inflorescence branched.

Inflorescence not branched.

Naiadaceae. Dioecious. Stem and back of leaf spiny.

Testa of manv lavers of cells.
Monoecious. Stem and back of leaf not
spiny. Three layers.

80 DIVERGENCE OF VARIATION [ch. ix

Cannaceae. Three outer staminodes separate.

Two outer united, third free.
Dicotyledons.

Casuarinaceae. Twigs whorled, rarely 4-angled, and then

hairy in fork.
Twigs not whorled, or 4-angled with whorls
of four leaves.

Myricaceae. Female flower with two to four or more

bracteoles, not accrescent to fruit.
Female flower wdth two lateral bracteoles,
accrescent to fruit, making two wings.

Myzodendraceae. Male flower with two stamens.

Male flower with three stamens.

Grubbiaceae. Flowers in threes in axils of foliage leaves.

Fruit hairy.
Flowers in threes in axils of opposite bracts.
Fruit not hairv.

Ceratophyllaceae. Fruit without spines or wings.

Fruit with spines or wrings.

Moringaceae. Seed without wings.

Seed with wings.

Nepenthaceae. Seeds egg-shaped with no appendages.

Seeds with long hairlike coat.

Myrothamnaceae. Two bracteoles. Stamens free.

No bracteoles. Stamens in column.

Platanaceae. Leaves usually 5-nerved.

Leaves usually 3-nerved.

Both in the monotype and the ditype families it will be seen at
once that the characters that distinguish the species in the one
and the genera in the other, are of the " family " type rather than
of the specific or generic type found in large families. And most
often they allow of no intermediates. Nothing but divergent
mutation will explain such things.

It is fairly clear that the larger genera tend to head sub-
families, or groups of whatever rank may be considered appro-
priate in the family concerned. It will therefore be of interest to
study one or two families in greater detail, and the first that
comes up in a random choice is the Ranunculaceae. We shall
expect, upon the theory of diff'erentiation or divergent mutation
which we have been discussing, that the chief division in the key
will usually lead to the two chief groups into which the family is
divided, and that each of these will be headed by one of the two

CH. IX] DIVERGENCE OF VARIATION 81

or three largest genera in the family. Of course, since we cannot
be sure of what is the largest divergence, nor be sure that that
divergence necessarily came first in the mutations, we shall not
expect every family to show such a result with any certainty,
though one may expect it to show more often than not. The im-
portant point is that each sub-group should be headed by a com-
paratively large genus. If the group be small in proportion to the
family, the genus may be small in proportion to some of the
largest genera of the family ; if large, one will expect its leading
genus to be larger.

We shall further expect to find the smaller genera practically
all included in the key that is marked out by the first divergent
mutation. That is to say, that we shall in general expect them to
be grouped as satellites round the big genera, not. as one might
expect if they were relics, in small and comparatively isolated
groups, which need not necessarily be closely related to the big
groups of the present day. We shall, therefore, expect the charac-
ters of these small genera to be less and less marked the smaller
(by the number of species in them) that they are, and to be, so to
speak, squeezed in between the well-marked characters of the
large genera. Real relics, on the other hand, would be more likely
to be distinguished fairly clearly from their relatives of the same
familv, bv characters that might even be as marked as those that
show in the first or second dichotomy of the key.

The Ranunculaceae, a family of medium size, not very much
larger than the average size for all families, have seven genera
that (in comparison with the rest of the family) we may call
large, each one containing at least seventy-five species. There are
nine of intermediate size with ten or more, but none exceeding
twenty (figures some years old), and ten small with nine or less.
This gap between the large and the intermediate genera is not an
uncommon occurrence, especially in families of small and medium
size, and should be well worth further investigation.

The big genera are :

Belonging

Spp.

to group

Aconiium

150

N. temp.

A2

Anemone

130

Cosmop.

B

Aquilegia

75

N. temp.

A2

Clematis

250

Cosmop.

B

Delphinium

175

N. temp.

A2

Ranunculus

325

Cosmop.

B

Thalictrum

75

N. hemisphere

B

Total 1180 or 87 per cent of the family

WED

82 DIVERGENCE OF VARIATION [ch. ix

The intermediate genera are :

Actaea

15

N. temp.

A2

Adonis

10

N. temp. Old World

B

Coptis

10

N. temp, and Arctic

A2

C alt ha

20

Temp.

A2

Ilelleborus

15

Eur., Medit. region

A2

Isopyrum

20

N. temp.

A2

Nigella

16

Eur., Medit. region

A2

Paeonia

15

Eur., Asia, W.N. Amer.

Al

Trollius

12

N. temp, and Arctic

A2

Total 133 or 10 per cent of the family

The small genera are:

Anemonopsis

1

Japan

A2

Callianthemum

5

Mts. of Eur., C. Asia

A2

Er ant his

7

N. temp. Old World

A2

Glaucidium

2

Japan, China

A 1

Ilamadryas

4

Antarctic Amer.

B

Leptopyrum

1

C. Asia

A2

Myosurus

7

Temp.

B

Oxygraphis

9

N. temp. Asia and Amer.

B

Trautvetteria

6

Japan, N. Amer.

B

Xanthorrhiza

1

Atl. N. Amer.

A2

Total 43 or 3 per cent of the family
Grand total 1356 spp. in 26 genera, average

52.

The classification used here is that of Engler and Prantl, in
their first edition — Paeonieae (A 1) and Helleboreae (A 2) being
marked off" from Anemoneae (B); Paeonieae have only two
genera. Of the large genera given above, three belonging to
group A 2 have an average of 133 species per genus, and are only
North temperate in distribution, while four belong to group B,
average 195, and are cosmopolitan in distribution in three cases,
the fourth being only North hemisphere. On the face of it, by the
greater size and greater distribution, B would appear to be an
older group than A. The intermediate genera are intermediate
both in size and in distribution, and the small genera are evidently
the lowest in both respects.

Now the very old and large genera, upon the theory of dif-
ferentiation, must owe their origin to the earliest generic muta-
tions in the family, and upon the principle of divergence of
variation, we shall expect these variations to be, on the whole, the
most divergent that occur in the family. In other words, the
larger genera of a family should be separated by well-marked
divergences, while the smaller will be less so. This is exactly what

CH. IX] DIVERGENCE OF VARIATION 83

we do find. If we draw up a key to the Ranunculaceae, dealing
only with the seven big genera given in the list above, it will be
found to be just such a divergent key, so that to place a species
in its proper genus is a very simple matter. Here is the whole
key:

A. Ovules on both sides of ventral nerve of carpel: follicle —

Flower with 2 or (2) honey-leaves:
Honey-leaves sessile, odd leaf of

perianth spurred, projecting. Delphinium

Honey-leaves stalked, odd leaf

helmet-shaped, erect. Aconitum

Flower with 5 honey-leaves. Aquilegia

B. Ovule solitary at base of ventral nerve: achene —

Ovule with one integument.
Ovule pendulous:

Leaves opposite. Clematis

Leaves alternate (exc. involucre). Anemone

Ovule erect. Ranunculus

Ovule with two integuments. Thalictrum

The key is a very simple affair, with widely divergent charac-
ters at every stage, so that there can be no difficulty whatever in
placing any species in its genus, were these the only genera in
the family. It is only when the smaller genera are included that
anv difficultv is found. With each new one that is added, the
characters that have to be used become more numerous and more
complicated. These seven large genera cover practically the whole
range of variation that is found in the family, to say nothing of
including 87 per cent of the whole, and the rest of the genera
come within, or very close to, the range thus indicated. If one
add to the seven large genera the rest of the family, which con-
sists of small genera not exceeding twenty species, one finds that
the steps which in the above key lead only to Ranunculus lead
also to Myosurus, Oxygraj^his, Trautvetteria, and Hamadryas.
A whole series of new steps in identification is now required, but
the important and interesting point is that all the new additions
come within the original key, or very nearly so. The new additions
that have to be made to the lists of characters are all at the
generic end of the key or close to it, with few exceptions. Instead
of finding that '"ovule erect" leads straight to Ranunculus, we
have to have a supplementary key like the following:

6-2

84 DIVERGENCE OF VARIATION [ch. ix

(Ovule erect)

Flower hermaphrodite.

Fruit with no hard layer in wall.

Ovule ultimately pendulous ; perianth

leaves spurred. Myosurus

Ovule always erect; perianth leaves
not spurred:
With honey-leaves. Oxygraphis

Without honev-leaves. Trautvetteria

Fruit with hard layer in wall. Ranunculus

Flower dioecious. Hamadryas

While almost all of the new and smaller (younger, according to
age and area) genera that have to be added to the key that we
obtained from the large (old) genera are added simply in such a
way that they cluster around some of the big genera, like those
just given cluster around Ranunculus, one finds every now and
then one or more genera (usually clustered) which do not so
obviously represent satellites of the big genera, but have a focal
point of their own. Thus among the intermediate genera in
Ranunculaceae there appears Paeonia, whose characters require
a splitting of the early character of distinction given above and
marked A. Instead of leading directly to Aquilegia, Delphinium,
and Aconitum, as at present, A has now to include Paeonia, which
cannot be easily split off, as was Ranunculus, by extension of the
generic end of the key, but has to be split off as follows :

A: Follicle, etc.

Outer integument of ovule longer than inner; Paeonia

no honev-leaves; ovarv wall fleshv.
Outer integument not longer, sometimes one Aquilegia, etc.,

integument only; honey-leaves or not; as before

ovary wall rarely fleshy.

Passing yet further down the scale of genus-size, Paeonia
becomes accompanied by Glaucidium, with two species in the
mountains of Japan and China (a much smaller distribution than
that of Paeonia, as one would expect upon age and area). As
the separation of Paeonia was so comparatively high up in the
scale, this small group of two genera is evidently of somewhat
different rank from that which surrounds Ranunculus, and is
often regarded as a sub-family; but it is important to notice that
it is hardly of the rank of the other two sub-families. As a key
to the three sub-families, we have

CH. IX] DIVERGENCE OF VARIATION 85

A. Ovules on both sides of ventral nerve; follicle —

(1) Outer integument of ovule Sub-fam. I. Paeonieae

longer.

(2) Outer integument of ovule Sub-fam. II. Helleboreae

not longer.

B. Ovule solitary at base ventral Sub-fam. III. Anemoneae

nerve; achene.

As Paeonia is comparatively small, it is extremely probable that
it is much vouno-er than the Helleboreae, which include three of
the first seven very large genera; and this is confirmed by its
small distribution as compared with them.

It is clear that if we suppose the big genera of a family to be
the first formed, and that by the most divergent variation that
(on the whole) occurs in the family, whilst the intermediate and
smaller genera are younger, we can get a satisfactory picture of
what seems to have gone on. The big genera, formed by early and
divergent variation, mark out the outer limits (or nearly so) of
the familv, the intermediate and small ones, which are on the
whole the younger, coming later and filling in the outline thus
made. In the later stages of the family, the divergences tend to
become smaller and smaller, especially as the possibilities of large
divergences have become somewhat used up. At each stage the
divergence is probably limited by what has already occurred, and
with comparatively few exceptions keeps within the limits thus
marked out. If, as in Annonaceae, the commencing mutation,
which gave rise to the family, includes a berry fruit, then this may
be a family character; if, as in Myrtaceae, it is produced in the
second mutation, the berry may characterise the sub-family
resulting from that. It may even be produced in later and later
mutations, and be the mark only of a tribe, a sub-tribe, a group
of genera, a single genus, or it may even mark only some of the
species in a genus.

The key to a family, if well constructed, in all likelihood gives
a clue to the mutations by which that family evolved into its
present condition. But one must remember that while a group of
the largest genera will doubtless be older than a similar group of
smaller ones in the same family, those that are actually largest,
or those that are the most widely distributed, need not necessarily
be the oldest, for there are so many accidents that may befall
plants in the shape of geological and other changes. Once a genus
becomes so large and important that it has many species and

86 DIVERGENCE OF VARIATION [ch. ix

covers great areas, the chances of its complete disappearance,
unless mere age, or further (probably universal) mutation can do
it, are small. The intermediate genera, on the other hand, may
often have suffered complete extinction, and still more the
smallest genera.

What has been said is also strongly supported by the facts of
distribution. There can be no doubt that in any given family, the
distribution of the genera goes on the whole with their size, as
has been shown in Age and Area, chap, xii, p. 113 (Size and
Space). Age, size of genus, and area occupied by it, all go
together.

It is clear that this analysis of the Ranunculaceae fully sup-
ports the theory of differentiation as against that of natural
selection, upon which no prediction can possibly be made as to
the size or composition of a family.

As another example, let us take the sub-family Silenoideae in
the Caryophyllaceae. It contains eighteen genera, whose numbers
of species, from the latest monograph (35), where the numbers in
the large genera are evidently rounded off, are

400, 300, 90, 80, 30, 30, 25, 10, 8, 7, 5, 5, 4, 4, 4, 1, 1, 1.

The first two genera, Silene (400) and Dianthus (300), which
contain 700 out of the total number of 1005 species in the sub-
family, are instantly picked out (supposing these to be the only
genera in the group) by the very first dichotomy that is given in
the key, which splits the Silenoideae into two tribes. All the
Lychnideae, headed by Silene, show a calyx with commissural
ribs; the Diantheae, headed by Dianthus, not so. The other
Lychnideae contain 80, 10, 8, 7, 5, 5, 4, 1, 1 species, and the other
Diantheae show 90, 30, 30, 25, 4, 4, 1, adding up, the one to 121
species, the other to 184, or in both cases much fewer than in the
big genus at the head of the group (400-121 and 300-184). Each
tribe is headed by a big genus, and the one tribe adds up to 521,
the other to 484, showing a difference just as indicated in Test
Case II, p. 94. The figures seem to indicate that in the Diantheae
there were more genera produced of intermediate size, so that
perhaps the stimulus of genus formation came earlier, and
resulted in the greater number of species shown by the smaller
Diantheae than by the smaller Lychnideae.

As the divergence just considered includes all the Silenoideae
on one side or the other, it is not unlikely that it was the first
mutation to appear after the first formation of the group by the

CH. IX] DIVERGENCE OF VARIATION 87

mutation that produced Silene itself. All later mutations come
within it, in the sense that the effects of this first mutation are
shown in them all. If we now follow only the tribe Lvchnideae,
Pax's key next splits off, by triple (or more probably by two
separate) divergences two genera, Ciicubalus, with one species in
Eurasia, and with berry fruit, and Drypis, with one species in
south-east Europe, and with capsule with lid; but as these are
small and rather local genera, and could evidently be split off
from any genus with a capsule, it is unlikely that they were
formed at this early stage. The next division in the key is more
probably that which split off Melandrium with eighty species in
the northern hemisphere. South Africa, and South America,
which differs from Silene by its fully unilocular capsule as against
a capsule multilocular at the base. In view of the great dispersal
of Melandrium, it is by no means improbable that it may have
been formed even earlier than Dianthns, and having met with
greater vicissitudes, such as the separation of Old and New
Worlds, has lost many more species than either Silene or Dian-
thus. In both Melandrium and Silene the capsule has two teeth
to a carpel, and each has a closely related genus with one tooth
per carpel, which was probably split off later {Viscaria near
Silene, Lychnis near Melandrium). Further mutations might give
the two genera Uebelinia and Agrostemma near to Melandrium,
by changing the relative position of carpels and calyx segments,
which are opposite in Melandrium and alternate in the two small
genera — a change which could only come by some mutation. They
might also give Heliosperma as a mutation from Melandrium, it
having only two rows of papillae on the seed, instead of having
them all over, and Petrocoptis as a mutation from Lychnis, the
latter having the teeth of the carpel twice as many as the styles,
the former once. It will be noticed that this phenomenon appears
(in Silenoideae-Lychnideae) in two places, and must have
appeared independently in these two, though the morphology
or the structural features are the same in each case.

CHAPTER X

SOME TEST CASES BETWEEN THE
RIVAL THEORIES

A. NUMERICAL

JL T is now almost unquestioned that existing plants and animals
have been produced by an evolution that, on the whole, has gone
forward, producing organisms of increasing complexity such as
man and the higher animals and plants. But many of the " lower "
things, the seaweeds, the lichens, the smaller ferns, the insects,
etc., have not been killed out, but have also increased very
greatly in number. This has always been difficult to explain upon
the current theory, but is perhaps more easy of explanation if we
consider that evolution was not altogether a matter of con-
tinuous improvement in adaptation, at any rate as indicated in
external characters, which are almost the only things to show us
that there has been any great evolution at all.

We have seen that a good case can be made out for differen-
tiation, in so far as it implies that a family most probably began
(at one step) as one genus with one species, of family rank, giving
rise later to other genera and species carrying the family characters
(but often with modifications in various directions), and making
in this way a family whose numbers would steadily increase,
inasmuch as there was no necessary reason why any of them should
die out, as there was under natural selection, which killed out the
less well-adapted ancestors. The loss of this first species and
genus would of course exterminate the family, but as it grew in
size, the loss of one genus with one species would matter less and
less, the rank of the genus with reference to the family becoming
continually less, the smaller the genus in proportion to the size of
the family.

The adoption of the theory of differentiation of course turns
the working of the mechanism of evolution the other way round,
and in the opinion of the writer puts events in their proper
sequence. It therefore seems clear that the first thing to be done
is to decide which of the two views is the more correct one to
take. Did evolution go in the direction from variety and species
towards higher forms (Darwinism), or in the reverse way (Dif-
ferentiation)? Did the family begin as a species of family rank,
or was it gradually formed by the destruction of intermediates?

CH. x] A. NUMERICAL 89

We are still far from any understanding of the actual mechanism
of evolution, but if we can feel sure of the direction in which it
worked, we shall have made one step in advance which may open
a way to profitable lines of research.

For example, take the case of economic botany, with its back-
ground of applied organic chemistry. So long as we imagine a
plant A, producing a valuable substance a;, to be descended from
some ancestor unknown, and quite probably unknowable, we are
heavily handicapped in tracing the origin and chemistry of x.
But if the descent, as differentiation would have it, were the
other wav, and the actual ancestors of A mav still be alive, so
that their chemistry may be studied, the work is greatly sim-
plified. Instead of remaining a vast mass of facts with little or
no co-ordination, economic botany may become a definitely
scientific subject, producing knowledge, not merely supplying it
in a dictionarv form, and we shall be able to look to valuable
results as yet quite unforeseen.

Endemic or local plants, again, if they be regarded as usually
the youngest in their own circles of affinity, and therefore as
"the latest thing" in breeding, in chemistry, etc., may become of
great importance, instead of being regarded as practically
negligible relics, as at present.

The writer hopes that the work here described may aid in
putting workers upon the right path towards a discovery of the
actual mechanism of evolution, and it seems to him that it may
be to cytology that we should look for the next step in advance.
As yet, the mutations that have appeared seem usually to be
lethal, recessive, or non-viable, but this is no proof that viable
or dominant mutations cannot appear also. If the result of
such a mutation were to be found growing anywhere, people
would at present say that it was another relic, and leave it at
that. Guppy has pointed out that many of the species that
have been found once only, and have never been seen again in
spite of search, are quite probably the result of such mutations,
which were in the early stages of establishing themselves, and
were perhaps exterminated by collecting specimens, or were not
viable (cf. 66, p. 151).

As the two theories of the direction of evolution are diametri-
cally opposed, it seemed to the writer possible to devise some
crucial tests between them. A number of these have been thought
out from the principles laid down in Age and Area-, these sug-
gested others, which have led to more. This simple fact, that these
principles can be so extensively used for prediction, goes to show

90 TEST CASES [ch. x

their general correctness, for the rival theory of natural selection
cannot be used to make predictions at all. All the evidence ob-
tained seems to point in the same direction, and seems to show
that evolution is moving as an ordered whole, upon lines that
have an arithmetical or mathematical basis. The general mathe-
matical propositions that underlie the theory that is here being
put forward have been worked out fully by Mr G. Udny Yule,
whose paper (75) contains a very readable and simple general
introduction and summarv that should be read bv all who take
any interest in the subject under discussion.

The actual evolution of new genera and new species seems
largely determined by a simple following, differing in speed in
each individual case, of the law of continual doubling, as was
shown by Yule and the author in 1922 (76). Sir James Jeans has
said that "All the pictures which Science now draws of nature,
and which alone seem capable of according with observational
fact, are mathematical pictures." In this he was referring more
especially to the physico-chemical sciences, but the work de-
scribed here, and in Age and Area, gives the impression that
biology will have to be added to them, though not in such a
clearlv-cut condition.

In this and the following chapters, some test cases are de-
scribed, all giving evidence which seems not infrequently conclu-
sive that the theory of differentiation, or divergent mutation,
is a more probable explanation of evolution than is that of
natural selection. The number of cases described may seem
excessive to some, but the writer, w^ho is now growing old, has
tried to make his position as secure as possible, and has therefore
chosen a number of tests from various parts of the subject.

TEST-CASE I. INCREASE IN NUMBER
WITH EVOLUTION

It is admitted that as time has gone on, plants have increased
vastly in number. But how did natural selection, working through
gradual adaptation, produce such an increase? The very name
selection would seem to imply the picking out of some from among
many. One would expect the ultimate result to be a few "super-
plants ", not a vast and increasing number with no evidence to
show that any one was superior to its immediate relatives.

On the theory of natural selection new variations, to have any
chance of persistence, must have been produced so that acci-
dentally or otherwise they suited the conditions, or more com-

CH. x] A. NUMERICAL 91

monly, some difference in the conditions, better than did their
immediate ancestor, which must have been suited to the condi-
tions to survive and reproduce. This would most probably mean
some difference in the physical conditions, especially of climate
or of soil, or in physical differences due to the presence of other
organisms, such as greater shade, greater demand for some
chemical constituent of the soil, or other thing. But ivhy should
a change in soil, or in climate, or in biological surroundings, un-
less perhaps it were very strongly marked, involve any morpho-
logical change? It is very difficult to see any connection between
these things.

Unless by some accidental happening, or in the rare case of a
"pure stand" — a solitary species occupying a large area — the
surroundings made by other plants would be continually variable.
Weather also is changeable, and unless a species were suited
from its birth to this fact, it would have a very poor chance of
survival in any case. Soil varies from one spot to another, and
so on. Unless variation in the conditions went continuously
in the same direction, as for example in a change of climate (not
of weather), it is very difficult to see why variations in the
morphological characters of the plant should go always in the
same direction, as is required if they are to be added up to
make specific differences. And it is difficult to see why, for
example, there should be any need for change at all in a species
that occurs, as do most species, principally in one association of
plants.

But to get increasing numbers of species, one species must (at
any rate very often) give rise to two or more, not simply to one
new one, unless, as on the theory of differentiation, the parent
survive as well as the offspring. But upon the theory of gradual
adaptation, to get two or more species from one without losing
them by intercrossing in the early stages, one must have dif-
ferent conditions in different parts of the range of the same
parent species. In other words, it must occupy a fairly large
area to get into such variety, and this is the basis of the explana-
tion of the local species as relics, though they far outnumber the
widely distributed ones, even in the most "successful" genera.

But if all the local species are failures, where does the increase
in number come from? Even in his own diagram (6, p. 90)
Darwin begins with eleven species, which at the next stage
become reduced to seven, the rest disappearing. At an indefi-
nitely later stage, shown in faint lines, they have increased to

92 TEST CASES [ch. x

fourteen. The relative proportions of widely and of narrowly
distributed species were not well known at that time, nor the
relative proportions of the genera in a family, both shown in the
hollow curves. Nor was it realised that no boundary could be
fixed dividing endemic species or genera from non-endemic. A
mere glance at the hollow curves will show this, or at a contour
map (chap, xiii, case 27). Even the big genera consist largely or
even principally of local or endemic species.

As an actual case, we may take the Monimiaceae, already de-
scribed upon p. 33. There are two large genera and thirty small.
What is selection going to do with these latter, which contain
30, 25, 15, 15, 11, 7, 6, 5, 4, 4, 4, 3, 3, 3, 2, 2, 2, 2, 1, 1, 1, 1, 1,
1, 1, 1, 1, 1, 1, 1, species respectively? The ones will presumably
disappear first on the whole, and the family should logically be
reduced ultimately to the two large genera of 107 and 75 species,
most of which again are relics in the sense that they only occupy
small areas. It is clear that natural selection, working upon the
lines usuallv laid down for it, would result in a tremendous dimi-
nution. And not only do the numbers of species in the genera
follow the law of Age and Area — the hollow curves — but so do
the areas that they occupy. The diameters of areas occupied by
genera of one or two species average about 560 miles, with three
to five species about 830, with six to eleven 1766, with fifteen to
thirty about 2310, and the two large genera about 5500 miles.
One can draw no lines of distinction. If the ultimate end of
natural selection is to be a small number, why begin with so large
a one? Whence did thev all come, and whv were thev evolved at
all? Under differentiation expansion is the rule, for each one may
ultimately give two, and there is no necessary reason for the
older ones to die out as they must under natural selection. Once
established in a small way, if there is no necessary difference in
adaptational value between one morphological form and another
nearly allied to it from which it may even have arisen, a species
may go on indefinitely, though by reason of the presence of
barriers to spread — physical, climatic, ecological, etc. — it may
never be able to expand over very large areas of country.

The same results as are shown by the Monimiaceae are shown
by any other family that one may take, especially if it be of fairly
reasonable size. The Cruciferae, with 350 genera, begin higher up
(with larger genera than the Monimiaceae) and end with 56 twos
and 145 ones. The Compositae end with 148 twos and 446 ones
(old figures).

CH. x] A. NUMERICAL 93

It is difficult to understand, upon the theory of natural selec-
tion, how the long tails of genera that contain only one or very
few species, and that occur in all but the very smallest families
(and are often indicated there), ever came to be evolved at all.
Natural selection looks upon them as the failures, and upon the
large genera with many species as the successes; the latter are
also widely distributed about the world in practically all cases.
But ivhy should a genus with many species occupy a large area?
There must, upon the adaptation theory, have been in it a mar-
vellous generic adaptation. If we take the first hundred genera in
my Dictionary (5th ed.) with fifty or more species, half of them
show a distribution right round the world, and at least half the
remainder cover immense areas. The smallest ranges are those of
Acantholimon (Eastern Mediterranean) and Agathosnia and Aloe
(South Africa). But, with ranges like this, these large genera
must be very old, to have reached so many continents before
communications were broken, and how did thev come to find, in
those early times, so great a variety of conditions as to lead to so
many sjDecies, at a time when conditions are usually supposed to
have been much more uniform than now?

If the small genera of one or a very few species are to be looked
upon as relics, why are there so many of them, and wh}^ do their
numbers increase tow^ards the bottom? It was shown (in 66,
p. 185) that out of 12,571 genera of flowering plants, 4853, or
38-6 per cent, had only one species each, 12-9 per cent had two
species, and 7-4 per cent had three. The numbers diminish up-
wards, following the regular hollow curve, shown not only by the
grand total, but by each individual family down to quite small
ones. The larger the family, the more accurately does it show the
hollow curve, a fact which does not favour the view that the tail
of small genera is composed of relics. Why should a "successful"
family have so many? One cannot draw a line through such a
curve, and say that all on one side of it are to be looked upon as
failures, on the other side as successes. To explain the curves, the
selectionists are thus obliged to admit that natural selection
shows its results in a continual and decreasing diminution of
numbers, as indeed one would to some extent expect from its
name. But if so, why did nature produce so many at first, only
to cut them down later, and where does the increase in number
come from, that is undoubtedly shown by the vegetable king-
dom? Was there no selection in ancient times? Differentiation,
on the other hand, as Yule has shown (75), necessarily results in

94 TEST CASES [ch. x

the production of genera in such a way that the result must be a
hollow curve.

The result of this first test is thus clearly in favour of dif-
ferentiation.

TEST-CASE II. THE SIZE OF THE LARGEST
GENUS IN A FAiMILY

On the theory of natural selection, the parent of a new species
will tend to become a relic, ultimately disappearing, but on that
of differentiation, there is no necessary reason why this should
happen. The parent may survive, probably does, long after the
throwing of offspring that may be specifically or even generically
distinct. As time goes on, the mutations in any one line seem to
tend to become perhaps less marked, so that generic mutations
perhaps become less frequent in proportion. It is possible that at
first, when considerable divergence is more easy, all or most of the
divergences may be what we should consider as generic. But on
the whole, it is evident that in any case the earlier members of a
family should be larger than the later ones — in numbers of
species if genera, in area occupied if species. They started first,
and on the average they should keep in front, so long as one con-
siders only related forms growing in similar conditions, as already
fully explained in Age and Area. The oldest genus in a family,
therefore, should in general tend to be the largest genus in it, and
the older and larger the family, the larger should its largest genus
be. But we have no absolute test of age, and must not try to
make comparisons of age, except between close relatives in
similar conditions. To say that the largest genus in a quickly
reproducing, mainly herbaceous family like the Compositae is older
than, or even as old as, the (far smaller) largest genus in the slowly
growing and reproducing giant trees of the Dipterocarpaceae, is
to make a statement which has nothing whatever to back it. The
latter, though only 5 per cent of the size of the first, may even be
very much the older genus. All kinds of accidents also interfere
with arithmetical regularity in these matters, so that it is really
very astonishing to see how regular the figures are, in spite of all
the geological or climatic changes, or other outside interferences.
None the less, as has already been shown in Age and Area,
p. 188, the supposition that the size of the largest genus goes with
the size of the family (a fact which could not be predicted by the
aid of natural selection) is borne out when one takes averages.

CH. x] A. NUMERICAL 95

The table given there shows this clearly, and some later figures
show it equally well:

ize of family in

genera

Av

erage

of the largest

(not in species)

genera in each (species)

1

12

2-3

43

4-8

94

9-20

129

21-40

153

41-70

195

71-100

313

101-250

330

Over

611

The requirement of differentiation, that the size of the largest
genus of the family shall go up with that of the family itself, is
fully borne out, while no theory of natural selection or of gradual
adaptation can offer any explanation of the facts.

TEST-CASE III. THE RELATIVE
SIZES OF GENERA

We may now consider the relative sizes of the genera in a family
or other group. Upon the theory that they were formed by
gradual adaptation one cannot say more about their probable
relative sizes than that some (the ''successful" ones) will pro-
bably be large, and some (the "failures'" or "relics'") small. Nor
can one give even an indication of what their relative numbers
will be. Further, one will also be inclined to expect to find some
kind of distinction shown between the successes and the failures.
But if differentiation be the more correct view to take, evolution
is no longer of necessity a direct expression of continually im-
proving adaptation, nor is the geographical distribution of plants.
It is clear that if that be so, there would be little reason for one
plant to spread, on the average, faster than its near relatives.
All in a related group would tend to spread at a more or less
uniform speed. But the speed of spread would depend upon
many factors, and to average these out, as already explained in
Age and Area, plants should only be taken in groups of say ten
allied forms, which should only be compared with other tens
allied to the first. Plants of systematic affinities that were widely
different might spread at completely different speeds, or plants
that differed in habit, like trees and herbs, or in speed of repro-
duction or other things. But on averages, with groups of allies
growing in fairly similar conditions, the oldest genus of a family
should be the largest, whilst the others should show a continually

96 TEST CASES [ch. x

decreasing size, but increasing numbers, with decreasing age. The
result would be to give one of the hollow curves which we have
described above. A little thought will soon show that the diminu-
tion in size will not be proportionate to that in age, for the older
that a genus is the more rapidly will it tend to gain upon those
younger than itself (66, p. 34),

As a genus or species (they are the same at the start) increases
in number of individuals and in area occupied, it w411 begin to
"throw" offspring differing from itself, by mutations occurring
at infrequent intervals, sometimes of generic rank, but more often
of specific. The average size of a genus is about fourteen to fifteen
species, but this does not mean, as one is tempted to suppose, that
a generic mutation may occur once in fourteen to fifteen times.
Rather it means that the average age of a genus may be more
or less represented by the average age of those which possess
fourteen to fifteen species. Some of the throws will be undoubted
species, some undoubted genera, some again of doubtful rank.

Supposing, which seems the most probable, that a new species
or genus begins upon a small area, it will probably be a very long
time before it occupies a more considerable space with more
individual representatives. But while it may wait a very long
time for the first throw, it would seem probable that the frequency
of the throws will on the whole increase w4th the number of the
individuals in the species, which in turn will tend to increase more
and more rapidly as time goes on (cf. Age and Area, pp. 33-4).
The first line of descent, that from the original genus (and species,
of course) of the family, will always have the start of the second,
which arises from the first generic throw of the original genus.
But as time goes on, there will be a continually increasing number
of lines of descent with the continual formation of more and more
genera to head them, so that at last we shall get the familiar curve
shown by any table of numbers of species in the genera of any
particular family of reasonable size. Thus a recent enumeration
of the Caryophyllacea^ (35) gives the following figures (bigger
genera obviously rounded to nearest ten or more) :

400, 300, 160, 100, 100, 90, 70, 40, 40, 30, 30, 30, 25, 23, 20,
20, 20, 20, 20, 18, 16, 15, 12, 10, 10, 10, 8, 7, 6, 6, 6, 5, 5, 5,
5, 4, 4, 4, 4, 4, 3, 3, 2, 2, 2, 2, 2, 2, 2, 1, 1, 1, 1, 1, 1, 1, 1, 1,
1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1.

If, then, genera are formed upon this principle — and that this
is quite a probable approximation to what really happens is

CH. x] A. NUMERICAL 97

shown by the universality of the hollow curve — we shall expect
to find that there will be a gap between the numbers of species in
the two largest genera of the family. This gap will obviously be
due to the fact that the two first genera of a family will usually
have been formed the one a good while before the other. During
that interval of time the first one will probably be able upon the
average to throw one or more species before the second genus
appears. It will thus get the start of the latter, and will con-
tinually gain upon it. It follows from this that the larger the
original genus now is, the greater, on the average, should be the
gap between it and the second genus. As we have just seen,
the third genus of a family will, upon the average, be separated
from the second by less time than is the second from the first, and
the time-separations will become less and less as we go downward
to the smaller genera. We shall, therefore, expect the gaps in
numbers of species also to lessen.

Turning to the facts, this is exactly what we do find. Out of
all the families given in my Dictionary — about 240 with two or
more genera — only eleven, mostly very small, show no difference
between first and second genus. The two larger families are
Bignoniaceae and Sapotaceae, where the four top genera are all
given as having 100 species. But on the average of all the families,
the first difference is ninetj-nine, while the second gap is only
thirty-two, the third eleven, the fourth eleven, the fifth six.

The result of this test case, therefore, is in favour of differentia-
tion. In fact, one cannot, with progressive arithmetical links
like these between the genera, consider natural selection or gradual
adaptation as having had much to do with the evolution.

TEST-CASE IV. PROPORTIONS OF SMALL
GENERA IN FAMILIES

If natural selection of gradual adaptation be the moving power
of evolution, and the small genera and local species be the relics
of what must be regarded as the failures, then one would certainly
expect that these ought to be more numerous in proportion in the
small and local families, which are also regarded as relics. If, on
the other hand, differentiation has been the mechanism, one will
expect that the larger a family grows, the more rapid will be its
proportionate production of small genera, for each genus, whether
small or large, may be able to throw new ones, so that the small
genera will be increasing (in number, not in size) more rapidly
than the large, the family following the hollow curve. . pV A m

WED .J\5^ ll J^ /

98 TEST CASES [ch. x

If we test these suppositions upon the facts, we soon find that
the large and "successful" families have many more "relics" in
them than have the small and "unsuccessful". In the Compo-
sitae, the largest of all, the monotypic or one-specied genera form
37-8 per cent of the total, while in the 151 small families containing
not more than ten genera each they are only 29 per cent (figures
twenty years old). The families with eleven to fifty genera have
33 per cent of monotypes, those with fifty-one to one hundred
have 36 per cent and those above have 39 per cent, a result which
agrees well with the theory of differentiation, but not with that
of natural selection. Even with the ditypic genera, their per-
centage in families up to 200 is 12-25, and 12-75 above.

If the small genera of one or two species are to be looked upon
as relics of former floras, why are they so numerous? About
38 per cent of all genera are monotypic, and over 12 per cent
ditypic, so that these groups alone make up half the total number.
Over 80 per cent of all genera have ten or fewer species. The hollow
curve, as we have seen, goes so smoothly and uniformly that
there is no possibility of drawing a line between successes and
failures. The only explanation of these curves upon the theory of
natural selection would seem to be that selection, as indeed one
might expect from its name, is continually picking out fewer and
fewer, so that its effect will be ultimately shown (when the relics
have died completely out) in a vast di7ninutio7i of the numbers of
species and genera. In other words, it is on its way to pick out a
few "super-plants" from among a mass of inferiors. But if so,
why did nature begin with so many? Their evolution cannot be
explained by natural selection. The whole attempt to explain
things upon this theory leads to so many absurdities that it
becomes untenable. The simplest explanation is evidently that
by using the theory of gradual adaptation in structural characters
one is trying to work backwards.

Every formation of a genus of two species (perhaps one may
be enough) increases the number of genera that may be looked
upon as capable of giving new genera of one, and as the larger
genera also may be looked upon as similarly capable, the rate of
production of monospecific genera will increase with the
size of the family. As already explained, the ones, as newcomers,
will be particularly slow at first in establishing themselves, so
that there will always be a time lag between them and the twos.

This test also fully favours the theory of differentiation.

CH. x] A. NUMERICAL 99

TEST-CASE V. THE HOLLOW CURVE

Many years ago it was shown that this curve, which is described
in Age and Area, p. 195, and in Chap, iv above, is a universal
feature of distribution in plants and in animals, both in regard to
the areas occupied, and to the sizes of the genera in families by
number of species contained. When plotted logarithmically, in
the latter case, they give close approximations to straight lines,
showing that they have the same mathematical form, and must
be due to the operation of the same law. The production of such
curves seems to the writer to place an almost insuperable ob-
stacle in the path of those who would explain evolution and
distribution in terms of gradual adaptation by means of natural
selection. Yule has shown (75) that the curve would result from
the continual doubling of the species and genera concerned,
when one supposes the parent to survive as well as the offspring,
as is the case according to the theory of differentiation. The curve
then becomes a normal and necessary feature of the evolution
that is going on, whereas under the theory of natural selection it
is totally inexplicable. Opponents have tried to belittle it by
showing that one can get similar curves from the names in the
telephone book, and such like conglomerations of inanimate
things. I have lately shown that the distribution of family
surnames of farmers in Canton Vaud (69) is just like the distribu-
tion of species, and therefore must follow the same laws, as it
gives the same curve. Natural selection could not determine it,
therefore it cannot be the determinant in the general distribution
of plants.

Nothing but a uniform pressure would ensure that results
could be expressed in hollow curves. Family by family, and
genus by genus, whether in numbers or in areas, all alike obey
the same law. Natural selection could not produce results like
this, and the only cause yet suggested is age, which represents
the resultant of all the forces acting. If they produce an average
result of <r in a long time 1, they will produce 2x in time 2. Age
thus forms a measure of distribution, but one cannot compare
unrelated forms, and must always work in tens of allied species,
to average out the differences that there may be between them.

It is clear that this test gives an unqualified verdict in favour
of differentiation.

7-2

100 TEST CASES [ch. x

TEST-CASE VI. SIZE AND SPACE

The hypothesis of Size and Space is more fully described in Age
and Area, p. 113 ; it follows from that hypothesis. " On the whole,
keeping to the same circle of affinity, the larger families and
genera will be the older, and will therefore occupy the most
space." If adaptational improvement ceases to be the prime (or
even perhaps an important) factor in evolution, there is no special
reason why one species should spread more rapidly, or over a
greater area, than other species closely related to it. As an illustra-
tion the case of distribution of species in Britain was taken, and it
was shown that it increased with the size of the genus.

"A good proof for the general correctness of Size and Space is
that . . . the further out we go among the islands, the larger on the
average do the genera become (in the number of species that they
contain in the world). Whilst the world average for a genus is
12-13 species, the non-endemic genera found in India contain on
the average about 50 species in the world, in New Zealand
about 75, and in the Hawaiian Islands about 100.

"The smaller families usually occupy smaller areas than the
larger, and the question arises whether they should be con-
sidered of equal rank to the latter. Guppy has suggested a
grouping of families into classes based upon these principles, for
which he has suggested the title Rank and Range, and it is clear
that in all future systematic work, the question of area must
occupy some attention" (66).

It is clear that the facts shown under Size and Space cannot
be explained by aid of the hypothesis of natural selection or of
gradual adaptation, and can at present only be easily explained
by that of differentiation.

TEST-CASE VII. "SOME STATISTICS OF EVOLUTION

AND GEOGRAPHICAL DISTRIBUTION,

AND THEIR SIGNIFICANCE"

To give details would simply be to repeat the paper of Mr G.
Udny Yule and the author, in Nature, vol. cix, 9 February 1922,
p. 177, and it will suffice to call attention to it. The general con-
clusion was that: "Inasmuch as all families, both of plants and
animals, show the same type of curve, whether graphic or loga-
rithmic, it would appear that in general the manner in which
evolution has unfolded itself has been relatively little affected by
the various vital and other factors, these only causing deviations
this way and that from the dominant plan." It follows that
evolution must have been by mutation, and that this must, at
times anyway, have been large, as demanded by the theory of
differentiation.

CH. x] A. NUMERICAL 101

TEST-CASE VIII. THE HALVING OF THE
SPECIES IN A FAMILY

We have seen (p. 96) that, in general, there is one genus in
each family which on the average has at the present time nearly
a hundred species more than the second genus ; the difference is
only on the average thirty-two between the latter and the third
genus, and so on. Only when one comes down into the smaller
genera does coincidence in number happen at all seriously, and
it happens more and more the nearer one comes to the bottom
of the list, so that at last, if of any size, the family ends with a
streamer of monotypic genera, or genera of one species each. This
hollow curve, which is always formed, is what is to be expected
upon the theory of differentiation, and natural selection is
helpless to explain it.

The hollow curve is due to the continual doubling of each genus
in turn by the throwing of a new genus so that, as time goes on,
the total number of genera undergoes an increase, which is con-
tinually more and more rapid, as the numbers grow. And as time
goes on, the genera already formed are supposed to increase their
number of species in the same way. We have supposed, as the
simplest solution of the problem for the meanwhile, that each
genus will on the average throw a new genus rather than a new
species once in every so many throws. In the counting that is
being used for this particular paragraph, ^ the total number of
families with more than one genus is 235. Taking the number of
species in each genus of a family, and arranging the genera in
descending order, the total number of species has been counted
for each family, and halved, and a dividing line drawn immediately
to the right of the genera required to make up the full half. This
of course means that the genera on the left may contain the
exact half (this is rare) or slightly or even considerably more ; but
the numbers on the right-hand side of the line never exceed, and
very rarely equal, those on the left. For example, three families
are given :

Aristolochiaceae 300

Basellaceae 14

Elatinaceae 19

60 10 8 1
3 111

19

All of these have the dividing line after the first genus, and this
proves to be the rule when the family is small, but not when it is
large. Out of the 235 families, no fewer than ninety-eight, or
41-7 per cent, have the dividing line after the first genus, as shown

1 The numbers are continually being revised for my Dictionary,

102 TEST CASES [ch. x

above. Fifty of them have two, three, or four genera, the actual
figures for the whole number being 22/2 (twenty-two of two
genera), 21/3, 7/4, 9/5, 8/6, 6/7, 2/8, 5/9, 2/10, 2/11, 1/12, 1/13,
2/15, 2/17, 1/19, 2/21, and one each of 24, 26, 28, 78 (Moraceae)
and ninety-nine genera (Solanaceae). Arranging these ninety-
eight families in order of size of the largest genus in each, one
finds that though the average size of the families in each group of
ten goes down with the average size of the largest genus, there
are nevertheless, in the first ten, four families with less than ten
genera each, but each headed by a very large genus {Begonia,
Oxalis, Piper, Impatiens).

The next lot of families is composed of those where the dividing
line comes after the second genus, as in Primulaceae: 250, 120, 90,
and so on to ten ones, total 651. While the average size of the
ninety-eight families with dividing line after the first genus was
7-9 genera with 201 species, the average size of those with the line
after the second genus is 14-9 genera with 249 species. Going on
in the same way through the whole number, we get the following
table : Average

Families

r

Gen.

>

Spp.

Dividing line after the

First genus

98

7-9

201

Second

50

14-9

249

Third

22

31-6

571

Fourth

20

63

997

Fifth

10

72-9

847

Sixth

6

73-2

1036

Seventh to tenth

13

128

1436

Over

16

387

5050

Figures in italics break the regularity of the table of averages.

The larger, on the average, that the family becomes, the more
is the dividing line pushed to the right, until in the Compositae,
the largest family of all, it only appears after the thirtieth genus.

It is clear that there is some arithmetical reason behind all this,
and the simplest explanation is that it is due to the continual
increase of species in genera other than the original one, when
the latter divides off new genera (which again divide) at average
intervals. In any case, the facts do not agree with any hypothesis
of gradual adaptation working from below upwards.

These numerical tests, to which others might be added, are
thus all in favour of differentiation rather than of natural
selection or of gradual adaptation.

CHAPTER XI

SOME TEST CASES BETWEEN THE
RIVAL THEORIES

B. MORPHOLOGICAL

iN AT URAL selection, being a conmion phenomenon of everyday
experience, has exercised such a fascination that it has to a
notable extent inhibited people from trying properly to think
out how a principle, whose essence is competition with partial
escapes into usually temporary success every now and then by
improved adaptation, can produce the ordered arrangement,
taxonomy, and morphological or structural uniformity with which
we are familiar. Herschel the astronomer, in an early criticism of
the Origin of Species, is said to have called it the "law of
higgledy-piggledy", and when one tries to imagine what mor-
phology would be, under its unrestricted operation, it is difficult
to meet this criticism. Why should natural selection produce such
comparative uniformity in morphological structure? Why should
there be such morphological likeness between the members of
whole families, tribes, genera, or even divisions like the Mono-
cotyledons? Why should the morphology remain the same, and
not improve in later evolutions? Why should the larger (older)
families appear in almost every kind of ecological conditions,
though the members of any one of these families show greater
structural resemblance among themselves than do the plants of
the association that inhabits any given spot? A grass is an un-
mistakable grass, whether in the tropics or in the arctic zone, in
a dry or in a wet climate, in a bog or on a moor. To say that this
is the case because it is a grass, and must retain the morphology
of a grass, is no explanation, but only throws the task of explana-
tion a little further back. Why and how were the grasses, or the
crucifers, or the composites, evolved at all? Why is there nothing
in common, in structural features, between say a grass and a
crucifer growing in the same kind of conditions, and side by side,
on a moor or in a pasture? One would expect natural selection,
working by gradual adaptation to similar conditions, and deter-
mining the structural features (as it must do if it is to be an
explanation of evolution) to produce something of similarity. In

104 TEST CASES [ch. xi

actual fact, however, there is rarely much or any structural like-
ness among the members of a given association of plants, unless
they happen to belong to the extremes of the principal ecological
divisions like xerophytes on the one side and hydrophytes on the
other, or to special ecological groups like climbers or parasites,
which do not, incidentally, grow in any special conditions, or in
associations. Even in these cases, the ecological characters that
mark them are rarely such as have great importance in classi-
fication.

Were it not for the great structural differences that exist, we
could not tell that evolution had gone on to so great and complex
a degree. There might be herbs, shrubs, and trees, water-plants,
epiphytes, climbers, plants of dry climates, bulbs, tubers, and so
on, with other more or less adaptive forms, but there seems no
a priori reason to suppose that we should find such things pro-
duced by an adaptive evolution as the structural differences that
mark whole families like the grasses or crucifers, and distinguish
them from one another. As Went has said (50), we see the mor-
phological differences, and assume that they must have some
physiological explanation. But there is nothing to show that
there is any physiological need for them. What connection can
be shown between the great bulk of the structural features of
plants and their physiological necessities? Man is adapted, region
by region, to almost every kind of conditions that can be found
upon the surface of the earth, yet he is all undoubtedly of one
species, and does not show any great structural differences. And
there are numerous similar cases with plants, though these are
slower in movement, and have not covered so much ground.
Some cover a very large area with no serious structural dif-
ferences, like Hydrocotyle asiatica, Sanicula europea or Hippuris
vulgaris, while in other places where the conditions are very much
alike throughout, a genus may show a number of species. One
can rarely infer from the external features of a plant, e.g. in a
herbarium specimen, or even in a living one, from what kind of
conditions it came. In the vast majority of cases, the most
minute morphological description will convey nothing as to the
habitat or the physiology, unless the plant happens to belong to
one of the great ecological groups like water-plants or climbers.
Can anyone read the characters in the most minutely descriptive
flora, and locate the probable types of habitat of the plants?

Taking genera with more than one species in the British flora,
the first, Thalictrum, the meadow-rue, has three. T. alpimim,

CH. XI] B. MORPHOLOGICAL 105

with a bi-ternate leaf, grows in alpine bogs, T. minus, with a
tri-pinnate leaf, in chalky pastures, and T. flavum, with a bi-
pinnate leaf, on river banks. In the next genus, Anemone, A.
Pulsatilla, with a bi-pinnate leaf, grows in chalky places, and A.
nemorosa, with a ternate leaf, in woods. Yet these two genera are
closelv related, and surelv. if the structural forms of the leaves
had anything to do with the conditions, the two with the bi-
pinnate leaves would occupy places not very dissimilar. The usual
reply of the selectionists to questions like this, that at some time
there must have been such conditional differences that a dif-
ference like that between these various types of leaf had a
physiological significance, is simply an appeal to ignorance, for
which there is not the slightest evidence.

If one takes the matter the other way round, one gets a good
argument against this contention of theirs. Why does one find
pinnate leaves, to take just a few examples from the British flora,
in Clematis, climbing in hedges, in Nasturtium in wet places, in
Cardamine in meadows, Anthyllis in dry pastures, Vicia climbing
in waste places, Sjnraea on downs, Potentilla by the roadside,
Rosa in hedges, Myriophyllum in water, and so on; and why in
Geum urbanum are the pinnate leaves only the lower, radical,
leaves of the plant? The argument of the selectionists is clearly
an admission of the point for which I am contending, that adapta-
tion is mainly an internal, physiological, or functional process,
without any necessary influence upon the outer, structural
features of the plants concerned.

The Englishman is successful enough in the conditions that
obtain in England, but if taken directly to India, and asked to
make good in the conditions to which the natives of that country
are subject, he would fail, primarily on account of the very
different climate. But he might succeed, if he were adapted by
nature's method of extremely slow change, say in a quarter or
half a million years. But by quick change he would be like the
potato and the dahlia, which have not yet become acclimatised
to Europe. Time is the needful thing in acclimatisation and
adaptation, and nature has plenty of it available. But it is of
course by no means unlikely that so great a change would be
beyond the limits of the Englishman's possible adaptation; there
are many cases in plants which seem to point to the existence of
such a limit. From what we know of man, it is not to be expected
that in the course of this adaptation the Englishman would suffer
great morphological changes, though he might acquire a darker

106 TEST CASES [ch. xi

skin, as apparently have other northern tribes that migrated into
India. The principal change that he would undergo would be a
gradual physiological adaptation to warmer climates.

Many, if not most or even all, of the characters of distinction
that mark families, sub-families, and even smaller groups, are
such that they can have no serious value upon the physiological
side, which is the only one that matters from the point of view of
natural selection or gradual adaptation. Only upon things with
functional value or disadvantage can natural selection operate,
and, as has frequently been pointed out by the writer and others,
its important work seems to be the killing out, probably rapidly,
of any variation definitely disadvantageous, though even here, as
the struggle for life is mainly among seedlings, disadvantageous
characters that only appear late in life may quite well survive.
There is no doubt that natural selection would encourage the
success of a new and improved form that had just arisen, but
there is no evidence that it can continue to call up small variations
or mutations always in the right direction, or that it can pass the
rough and ready line of distinction that exists between species,
that of mutual sterility, unless some mutation should happen
that will do so. But work of this kind will not ensure progress
such as seems to be the mark of evolution in general. Suppose a
whole family to possess a septicidal capsule, or diplostemonous
stamens. There is no evidence to show that there is any physio-
logical value attaching to this possession, which in any case only
appears in later life. One cannot imagine natural selection killing
out a member of the family that had adopted (or was varying —
if it could so vary — in the direction to adopt) a loculicidal capsule,
or obdiplostemonous stamens, or was even going so far as a septi-
fragal capsule. The family constancy of the capsule or the stamens
must be due to inheritance from a common ancestor. But how,
under selection, did the ancestor of one family obtain one kind of
capsule or stamens, of another family another? With the recent
revival of natural selection, there has been a recrudescence of the
idea that characters that are of no physiological value tend to be
very variable, but if so, why are family characters less variable
than generic and specific, though they are admittedly of less
physiological value?

Plants, animals, and man alike tend to produce so many off-
spring that, in a short time, but for various unfavourable condi-
tions, there would not be room for them upon the surface of the
earth. The illustration taken from the rapid multiplication of the

CH. XI] B. MORPHOLOGICAL 107

green-fly is well known (42, p. 188) and even in the Podoste-
maceae, annuals starting again every year, one plant might in
four years cover about 100,000 square miles. The fiercest struggle
for existence comes to a plant at birth, and any that is not suited
to the conditions as they are at that moynent will be killed out by
natural selection by reason of unsuitability, though of course mere
chance will have a large influence in the matter. But this is an
individual struggle, and we have no right therefore to assume that
species struggle as units. Nothing can come into permanent
existence without the permission of natural selection, but once
the newcomer has become established in a few places reasonably
far apart, the chance of its being completely killed out will
steadily diminish, and in course of time may be reduced to
vanishing point. Natural selection simply determines in each
individual case whether or not a given plant shall be allowed to
survive and reproduce.

Very few indeed of the morphological features that distinguish
one organism from another that is related to it have any physio-
logical significance at all, especially in those features that
separate the higher groups of plants from one another. Even the
bulk of the generic and specific characters come into the same
category. One cannot imagine any adaptational reason why
Ranunculus should have over 300 species, and world-wide distri-
bution, while its closest allies, like Myosurus or Oxygraphis, have
few species, are comparatively localised, and differ largely in the
fact that the wall of the fruit is not so hard. Still less can one
imagine adaptational reasons taking part in the separation of the
family Ranunculaceae into a group with achenes and another
with follicles, or one with alternate leaves and one with opposite.
Nor can one suggest adaptational reasons for the existence of
200 species of Clematis, and still less for that of a couple of
thousand Senecios or Astragali. If natural selection is to be held
responsible for the vast dispersal of, and numbers of species in
these genera, they must have some very great adaptational ad-
vantage over their close allies. And the adaptation which was so
successful must have been generic, for most of the species have
but small areas. There is no species in these very large and wide-
spread genera whose range covers that of the genus, though in
smaller and less widely dispersed genera this is very commonly the
case.

It has frequently been shown, e.g. by de Vries (66, p. 224) and
by J. T. Cunningham and others, that adaptation shows chiefly

108 TEST CASES [ch. xi

in generic and family groups, rather than in specific, so that any
theory that tries to explain it on the basis of a commencement
with the species, as does the Darwinian, must fail in its explana-
tion. What adaptation there is, is rather handed down by
heredity.

The things that are usually considered to be gradual adapta-
tions are steadily diminishing in number (cf. p. 116), and
though it may come as a shock to some, one must add such
things as climbers, parasites, saprophytes, lichens, fungi, herbs,
trees and so on, for in most of these cases no intermediates are
possible, or at any rate probable, and so much correlation (p. 129)
is also required, which could not be effected by gradual adapta-
tion. Trees, for example, are usually supposed to be older than
herbs, but can any one imagine them being gradually selected
down to herbs, especially when one remembers that both forms
may not infrequently appear in the same genus, so that it is
evidently, as in so many other cases, quite a simple matter to
pass from one to the other?

One might ask similar questions for the whole list of characters
of family rank (Appendix I). Is there any adaptational difference
between a superior and an inferior ovary, any between parietal
and axile placentation, trimerous and pentamerous flowers, a
dorsal raphe and a ventral, one cotyledon and two, or the various
kinds of zygomorphism? Incidentally, median zygomorphism is
looked upon as an adaptation to the visits of insects, but if so,
why do transverse and oblique zygomorphism exist also? Why
do the highly zygomorphic flowers of the Podostemaceae stand
stiffly erect, whilst they are wind-pollinated also?

Or, to go to generic characters, and taking a small family like
Styracaceae, is there any adaptational difference between a
flower with ten stamens and one with five? Between an ovary
3-locular below and unilocular above, and an ovary 3-locular
throughout? A flower with connate petals and one with free? Or,
in the Caryophyllaceae, between a glabrous and a hairy stigma,
a petal claw with and without wings, a capsule with teeth as
many as carpels and one with teeth twice as many? We may even
go on to species and still fail to find adaptational characters. It is
impossible to read into the distinguishing characters any adapta-
tional meaning which would be of any advantage in the struggle
for existence, especially when we remember that the great
struggle comes before the great bulk of these characters appear
at all. It is an axiom in taxonomy that the less that any character

CH. XI] B. MORPHOLOGICAL 109

has to do with the life of the plant, the more important is it from
a taxonomic point of view. The higher one goes from species to
family, the less connection have the characters with the life, and
if one tries to think out how a mechanism like natural selection,
depending upon improved adaptation, could thus have less and
less to do with adaptation, the more it separated into larger
groups the organisms with which it was concerned, one will
speedily arrive at a deadlock.

Two species usually differ in more than one character, even
when closely related, and among the supporters of natural selec-
tion it is, or has been, very often an implied assumption that a
species A shall change fully to B before it goes on to become C.
But one can see no reason whv this should be so : a variation in
the direction of C would probably be just as useful in a plant that
was only on the way to B. Natural selection can do nothing till
the right variation is offered to it. Let us suppose that A is
offered a variation in the direction of B and has started to adopt
it, and that then a new variation is offered in the direction of C,
obviously better, but in a different direction. What will happen
then? Will it go on towards B and ignore the later offer, will it
form C with a shade of B about it, or will it try to go back, and
get rid of the traces of B, with the risk that it may not get another
offer of C? There seems almost nothing for it but to demand that
variations shall not interfere with one another, but that the one
"in possession" shall be allowed to finish what it began, before
another one is allowed to start. But this will greatly slow down
the process of evolution, unless the variations are largely corre-
lated. But why under natural selection should there be so much
correlation? It is hard enough to find adaptational reasons for
one variation, let alone half a dozen correlated ones. It would
seem more probable and reasonable that in general the morpho-
logical characters have no necessary physiological value, and are
therefore not the result of any adaptational selection. If a new
structural character appears that has an adaptational value, it is
at once seized upon and perpetuated, unless in case of evil
chance. But to regard structural characters as necessarily
showing individual adaptational value — for example, that there
is some necessary value in a pinnate rather than a palmate leaf,
or vice versa — is to stretch the theory of gradual adaptation
too far.

It is a very remarkable thing that we do not find plants with
a superposition of variations, one complete, the other incomplete.

110 TEST CASES [ch. xi

The only reply the selectionist can make is to say that the consti-
tution of the plant does not allow of the mixture of characters,
or, in other words, that structural considerations override
adaptational. And as this reply comes in, in nearly all cases, it
does not leave much room, if any, for gradual adaptation.

The whole subject has suffered from the lack of proper thinking
out. Everyone can see the struggle for existence going on before
him at any moment. The individual who is in any way handi-
capped, be it by some physical disability, by poor health, by low
intelligence, by parental poverty (resulting in cheap schooling,
underfeeding, etc.), or by other difficulties, is on the whole the
one to be defeated. In the early days of the theories of Malthus
and of Darwin (which was based upon Malthus) the tendency was
to legislate (or rather not to legislate) in such a way as to leave
the struggle for existence uncurbed, the idea being that in this
way the best was brought to the top and the inferior left at the
bottom, if not killed out. The theory of "nature red in tooth and
claw" had, and still has, a great vogue. It was not realised that
the winners in the struggle for existence owed their success only
too often to some adventitious advantage which was not neces-
sarily part of their own equipment. Money, for example, pro-
viding the best food and education, was a great help. One has
only to examine the trend of modern social legislation to see how
we are drifting away from the old philosophy of the unrestricted
struggle for existence. Everything possible is now being done to
remove the handicaps that formerly were fatal to some of the
best men, and to give to everyone the best possible chance, and
there is reason to hope that in a few generations the results of this
work will show a great social advance.

Man is all of one species, and it is worthy of note that in his
struggle for existence against members of other species, he has
owed his success not to the slight morphological differences that
distinguish his different varieties, but to internal adaptation of
brain, etc., leading to greater skill in handling the difficulties
that beset him.

TEST-CASE IX. DIFFERENCES IN GENERIC RANK

This test will be rendered more intelligible by aid of the figure (8),
in which A represents a family of two genera only, B a family of
intermediate size, and C a large family, both B and C being
imagined a good deal larger than here shown. All are supposed
accepted by the same systematists, to make their rank fairly

CH. XI] B. MORPHOLOGICAL 111

equal. The diagram will serve for the growth of the families under
differentiation, in which progress is supposed to work downwards
from the original species and genus that began the family, A, B,
or C, both species and genus of course being the same plant. As
the family grows, it will form new species and genera, and all will
on the average survive, so that the now existing family is in each
case represented by all the dots under A, B, or C Whether the
whole family, if seriously old, survive like this, will depend upon

Level 1

C. Cc. Cbb Cb

CbChbCc C

Fig. 8. Diagrammatic origin of small, medium and large families under
differentiation, to show relative rank of genera in each, which goes more
or less with the line 1, 2, 3 etc. upon which they happen to stand.

what geological or other catastrophes it has met with, and whether
any general change may occur in a genus, causing its death, or
transforming it, or more probably one of its species, into another
genus. The well-known fact that the smell was lost at the same
time by all known examples of the common musk, once universal
in cottage windows by reason of its sweet scent, shows that
though a species may be represented by innumerable individuals,
something may happen simultaneously in the internal make-up
of all of them. And there is nothing to show that larger mutations
than this are not possible. The way in which the successive fossil
species of Stratiotes appear in different geological horizons, each
specifically different from the preceding one, shows the kind of

112 TEST CASES [ch. xi

thing that may possibly happen, and again we have no reason
why the change should not sometimes be generic as well as
specific (3).

Under the theory of natural selection, the existing plants will
only be the lowest row in each family, for evolution, as we have
already explained, is at present supposed to work by the forma-
tion of slight varieties, which gradually increase to larger, to
species, and so on, by the killing out of the less well adapted
ancestors, while the ultimate survivors in the way of genera tend
to be those that show the greater divergences. And the smaller
the family, the greater are the divergences between its genera.

To return to the diagram, under the supposition of evolution
by natural selection, the various genera that occupy the bottom
row in each family, whether in line 3, 4 or 5, will simply be genera,
or generic stages in the evolution that is continually going on.
Under this supposition, there is no reason why the genera in the
lowest line, 3, of family A should in any particular way be any
different from those in line 4, of family B, or these from those in
line 5, of family C. Nor does natural selection offer any test by
whose application we may gain any idea as to the relative degree
of divergence that there may be between the genera of A, B,
and C But upon the theory of differentiation or divergent muta-
tion we shall expect that the divergence between the two genera in
family A will he about equal to the first divergence in the families B
and C, i.e. equal to the divergence between their tribes or subfamilies.
At the same level in the diagram, in other words, there will be
more or less equal divergences. It has long been known as an
axiom in taxonomy that genera in a small family are much better
separated than genera in a large one, and here is a simple expla-
nation of this. As the family grows in size, new mutations will
come in at more and more frequent intervals, but within, or close
to, the original divergent mutation. In other words, all the
family, sprung from the original genus and its first mutation,
will show some at least of the characters shown by these first two
genera, which by the hypothesis of divergent mutation will tend
to be very divergent. The original family characters will show
best in the largest genera, which will be the oldest in the families,
and carry the most of the earliest characters. The genera sprung
from later mutations will not have, so well marked, many of the
characters of the earlier mutations. The generic characters will
necessarily become on the whole closer and closer together as the
mutations to which they are due are less and less far back in their

CH. XI] B. MORPHOLOGICAL 113

ancestry. It would also seem as if there were a tendency in each
family for mutation to become less pronounced as time goes on,
so that the appearance of what we usually call family characters
(list in Appendix I) becomes less frequent in proportion to the
total of characters that appear. On the whole, there is more room
for wider divergences the nearer one is to the starting-point of
the family, i.e. to the original genus which gave rise to it, upon
the theory of differentiation. Upon that of natural selection, it
has always been a great difficulty to explain why the divergences
became greater the higher one went in the key from species up-
wards. Why should natural selection cause the disappearance of
just those forms necessary to make the divergence increase? This
is inexplicable by natural selection, working upwards from small
differences, but simple to differentiation, working the other way.

This being the expectation, we have only to look at Appen-
dix III which gives the distinguishing characters of the two genera
in those families that contain two only, to see that the facts
agree with what was expected. The divergences are obviously of
the same rank as those given in Appendix I as being "family"
characters. If, on the other hand, one compare the generic
characters in larger and larger families, one finds that as one goes
up the scale, the genera, as one will expect under the theory of
divergent mutation, get closer and closer together as new ones are
"squeezed in" among the old. In a really big family, like the
Umbelliferae, Compositae, or Gramineae, it is a familiar ex-
perience that it is as difficult to make out the genus, as in a small
family to make out the species.

It is clear that we have not properly taken into consideration
the relative rank of genera and other groups. In a very large
family, where the genera have become closer and closer together
by the continual appearance of new ones, the generic rank is
evidently lower than it is in a normal small family. The ranks of
all divisions in the classification, whether tribes, families, or
genera, depend to a very great extent upon their relative sizes in
their circles of relationship. This conception of relative rank has
gone neglected during the reign of natural selection, to which a
genus is simply a generic stage upon the upward road.

If, on the other hand, the small family is to be regarded as a
relic, as is done by the supporters of selection, it becomes neces-
sary for them to explain why the divergences of the two or three
genera that are left is so great, and equal to the divergence of the
sub-families in a large family. Often one hears people say that

WED 8

114 TEST CASES [ch. xi

the sharpness of definition in a small family is due to the fact
that the family is small, with few genera. But this does not
explain the fact that those genera have ahiiost without exception
the rank of sub-families in a large family.

The result of this test case is thus very strongly indeed in favour
of the theory of differentiation as against that of natural selection
with gradual adaptation.

TEST CASE X. THE PERFECTION
OF CHARACTERS

The fact, which seems to have been completely ignored, that
structural characters are practically alwaj's shown both by
animals and by plants in their perfect condition, is one which is
simply incapable of explanation upon the ground of gradual
acquirement, but simple if it be the result of a sudden mutation.
The astonishing thing in the latter case would be to see an im-
perfect acquisition. The perfect condition is best shown by the
very widely divergent characters, like opposite or alternate leaves
and many others that have no intermediates, in fact most of the
characters shown in Appendix I. How can the divergence, under
natural selection, have become not only larger but more perfectly
marked? Supposing for the moment that an intermediate were
possible between alternate and opposite leaves, and that there
was such an adaptational urge that a plant began to progress in
the direction of the latter. It is clear that once the leaves began
to be nearly opposite, the urge would rapidly fall off, till at say
95 per cent of perfection it would be quite small, and almost
infinitesimal at 99 per cent. How comes it then that opposite
leaves are exactly opposite? How comes it that a drupe or a berry,
a capsule or a schizocarp, is always the same in structure (inci-
dentally, why has evolution made no apparent attempt at im-
proving them?), and always complete? In the same way, a
disadvantageous character would be unlikely to be completely
■got rid of.

It is, I think, safe to say that natural selection could not dis-
tinguish between 96 and 100 per cent of perfection, and that
there must be some other principle that is responsible for the
perfection that is always shown. By far the simplest explanation,
and the only satisfying one at present, is that the perfection is
due to a direct mutation. One can multiply examples to an
almost unlimited extent.

CH. XI] B. MORPHOLOGICAL 115

For that matter, how is it that all the leaves upon a plant
match one another so closely as they do, or all the flowers? The
only explanation that the supporters of natural selection can
give is that morphological considerations are more important in
evolution than is natural selection (cf. pp. 120, 121). But how did
natural selection begin to develop different types of leaf, and
to make them so constant in size and form, and to put different
types upon closely allied species (cf. the Thalictrums on p. 104)?
There is not the faintest reason to suppose that evolution worked
by different rules at different stages in its history, but the selec-
tionists seem to think that if by aid of assumptions and supple-
mentary hypotheses they can produce some kind of explanation
of the phenomena seen at the present day, the past can take care
of itself. What we are contending for is that morphological and
anatomical considerations are more important than natural selec-
tion, and that the latter has not been, unless to some small extent
or in some recondite way, responsible for the appearance of
important structural characters. It acts upon what is given to it
by the process of evolution, which goes on regardless of whether
its products are acceptable or not. If they are killed out by
natural selection, that is the end of that line, but others will
appear. The simple and easy explanation of the phenomena of
morphology is that they are due to mutations, which as a general
rule probably produce a new form at one operation. To some
extent at any rate, there is probably some definite factor in the
parent, perhaps some arrangement or structure of the chromo-
somes, that determines what will appear in the offspring (and
here again perhaps only in certain conditions, as for example
possibly under the influence of cosmic rays). But w^e are as yet
too completely ignorant of the whole subject to be able to hazard
any definite opinion.

This test evidently gives very strong evidence in favour of the
large mutations that are required by the theory of differentiation.

TEST CASE XI. THE EARLY STAGES
OF CHARACTERS

One of the great difficulties that have always dogged the path of
the supporter of natural selection as a cause of evolution, is to
explain the beginnings of the various structural characters. This
is a problem with which he has had little or no success. We have
instanced many of the characters that divide species, genera, and

8-2

« .

116 TEST CASES [ch. xi

families, and have shown that even when fully fledged it is im-
possible to find any functional reason for their existence, and
equally impossible to show why one should be preferred to the
other, or to any (not commonly possible) intermediate, for any
adaptational reason whatever. The adaptation to their surround-
ings that is possessed by all living beings is primarily an internal
affair. Descending from ancestors not too far away in distance,
they presumably in most cases possessed an adaptation that was
not very difl'erent from that of their parents — at any rate those
that did not possess it would soon be destroyed by natural selec-
tion. The adaptation might cover a greater or slightly different
range of temperature or moisture, etc., that would enable them to
reach places unattainable by the parents, thus ensuring ultimately
a different distribution.

All evidence goes to show that adaptation is rarely shown
in structural characters, and it will be of interest to draw up a
short list of some of those things that were considered as adapta-
tions in the writer's early days; hundreds more might be added:

Phyllodes in Acacia
Thorny roots in Acanthorhiza
Reversed leaves in Alstroeyneria
Adventitious embryos
Self-burying fruit in Arachis
Hooked bracts in Arctium
Clasping hooks in Artabotrys
Cauliflory in Artocarpus
Pollinia in Asclepiadaceae
Thorns in Astragalus

and many more in genera beginning with A.

Red seeds in Paeonia

Gutta-percha in Palaquium

Horizontal fruit wing in Paliurus

Scaly fleshy fruit in Palms

Distribution by animals of stem joints in Panicum

Distribution of Papaver seeds by pores in capsule

Protogynous flowers in Paris

Neuter flowers in Parkia

Extrafloral nectaries in Passiflora

Biennial life in Pastinaca

and many more in genera beginning with P.

It only requires that one should quote such cases as these,
which are not selected, but simply taken in alphabetical order
from my Dictionary, to show how the idea of universal adaptation,

CH. XI] B. MORPHOLOGICAL 117

at one time held by almost everyone, has passed away, though
natural selection, which is looked upon as depending upon
structural adaptation, survives.

But the great difficulty which has always hindered the selec-
tionist is to explain how natural selection got a grip upon the
early stages of any of these characters. If they were produced in
one operation, as differentiation demands, everything is simple,
but in that case it is clear that natural selection can have little
or nothing to do with their appearance. One must drop out
natural selection as a guiding cause in evolution ; it could get no
grip upon the evolution of these structural features by gradual
adaptation, and it could have nothing to do with it if they ap-
peared fully-fledged. This test is in full favour of differentiation
and what would seem the most probable order of things is that
evolution, strictly so-called — the appearance of continually new
structural forms — had little or nothing to do with adaptation of
those forms to the conditions bv which thev were surrounded.
They would inherit from the parents a reasonable probability of
not being too unsuitable to survive at all, and it would then be
"up to" natural selection gradually to fit them in minute detail
for some particular combination of the conditions of life that
existed near to the spot where they began, or to destroy them if this
could not be done. Natural selection, in other words, strenuous
though its action may be, has apparently nothing to do with the
evolution of plants, though it has everything to do with the way
in which they finally become best suited to some detail of com-
bination of the conditions by which they are surrounded. Evolu-
tion and natural selection, in other words, may be represented as
working more or less closely at right angles to one another, and
the evolution goes on by large steps, as required by the theory of
differentiation.

The theory of gradual formation of the structural features of
plants seems to be left with little or no support, and a much
simpler explanation of everything is provided by that of sudden
appearance. People say that we have no evidence of such an
occurrence, but we have no evidence of gradual acquirement, and
a mere glance at the table of family characters in Appendix I will
show that a great number of them are so divergent that they
allow of no intermediate, and if one therefore cannot derive them
by stages, they must have come in one step. And this is especially
true when one finds that gradual adaptation will not do as a cause
for change.

118 TEST CASES [ch. xi

TEST CASE XII. ALTERNATE OR OPPOSITE LEAVES

Here is a familiar pair of contrasting characters, occurring in so
many different places in the flowering plants that it is clear that
they must be very easily acquired, while sometimes one of the
two may be shown by a whole family, as are alternate leaves in
the grasses. We have already shown (74) that many or most
large families show, somewhere in their make up, exceptions to
most of the characters that usually mark the family. Thus in
Rubiaceae,^ a very large family, one can find alternate, whorled,
anisophyllous, pinnate, and gland-dotted leaves, leafy, and intra-
petiolar stipules, dioecious, zygomorphic, and solitary axillary
flowers, different male and female inflorescences, male and female
flowers so different that at one time they were regarded as
separate genera, flowers united in pairs, male flower 4-5-merous
with female 8-merous, calyx convolute, imbricate, opening
irregularly, with calyculus, with one large sepal, 5-merous in
male and 2-merous in female; corolla aestivation descending;
stamens united, unequal, 8-12, two only with a 5-merous
corolla; anthers opening by pores, or by valves, multilocular,
heterostyled, with poflinia; ovary superior, united in pairs,
1- 3-5- 4- 6-10- or oo-locular; stigma 10 -lobed; capsule both
septi- and loculicidal or circumscissile, berry, schizocarp; endo-
sperm none, ruminate; embryo with curved radicle, or with no
cotyledons.

This is a very extensive list of exceptions, but most large
families show something of the same kind, whilst even in the
small ones divergence, usually just as pronounced as the diver-
gences just given, is the common phenomenon, usually showing
in them between the first two genera, or in the division into
species if there be only one genus.

It is clear that if one were to combine in a group of plants a
number of the "abnormal" characters that have just been given
for the Rubiaceae, say alternate leaves with intrapetiolar
stipules, dioecism, zygomorphic flowers in male and female
inflorescences different from one another, the male 5-merous and
the female 8-merous; calyx imbricate with one large sepal,
corolla with descending aestivation, united, unequal stamens,

1 Usual characters decussate entire leaves with interpetiolar stipules;
regular flowers in cymes or heads, 5-4-merous ; K usually open ; C valvate or
convolute ; A 4-5, epipetalous ; G (2), 2-loc., each with 1- co ovules, style 1 ; fruit
various; usually endosperm.

CH. XI] B. MORPHOLOGICAL 119

anthers opening by valves, ovary superior with oo locuU ; fruit a
schizocarp; embryo with no endosperm, no cotyledons, and
curved radicle, a family would be produced that no one at any
rate would imagine to have any relationship whatever to the
Rubiaceae, and yet half-a-dozen to a dozen mutations might
produce it.

Divergence such as that shown by alternate and opposite
leaves, or any of the divergences shown in the list of "abnormal"
characters of the Rubiaceae is a matter of extraordinary difficulty
to explain by aid of the hypothesis of natural selection.
Neither of the divergent characters has any functional value to
the plant that anyone has ever been able to prove, or even to
suggest; nor as a rule is there any possible intermediate, nor
could it have any value or the reverse. Yet the divergences show
in so many different places among the flowering plants that they
must be very easily acquired; they are even found quite com-
monly between one genus and the next, or between some species
and the next. But for such differences to be quickly acquired by
natural selection, there would have to be some very pronounced
advantage to be gained by their acquisition, and that is just
what no one has ever been able to indicate. There is nothing to
show that either opposite or alternate leaves have any advantage
the one over the other, whilst an intermediate would still have
alternate leaves, with a particular phyllotaxy. A point which is
usually lost sight of, but is of great importance, is the difficulty
of passing by aid of natural selection from say 95 to 100 per cent
of perfection, already dealt with in Test case no. x.

This question of the relative value or disadvantage of a
character is another thing that has been completely ignored
during the long reign of natural selection. The great struggle for
existence is among the seedlings, and a character that is of im-
portance one way or the other to a seedling has a far greater
relative value than for example a character of the flower or fruit
which only appears in later life, when the plant is more esta-
blished and has greater reserves of food and vitality. Leaves, for
example, are much more important, individually and even collec-
tively, when the plant is young. Even if a character were defi-
nitely disadvantageous it might still survive if it only appeared
when the plant was old, whilst a disadvantageous character of
any kind would probably be fatal to a seedling.

The only reasonable explanation of alternate and opposite
leaves would seem to be to suppose that they are determined by

120 TEST CASES [ch. xi

single mutations. The supporters of natural selection can only
explain the exact nature of the oppositeness in the one case, or of
the phyllotaxy in the other, by supposing that anatomical neces-
sities are more potent than selection. Differentiation is much the
most simple explanation, when one sees the well and exactly
marked divergences that show so well, not only in leaves, but
throughout the whole list of the characters that mark the dif-
ferences in relationship of plants, and show that evolution has
gone on.

TEST CASE XIII. STAMINAL CHARACTERS

One may work through the whole list of family, or even of generic
characters, and find similar phenomena in all, inexplicable by
the theory of natural selection or of gradual adaptation, though
simply explained by differentiating mutation. Why in so many
families and other groups should a great and important dif-
ference be that one has one whorl of stamens, while the other has
two, or more? This dropping (or addition) of whole whorls of
stamens cannot easily be exjDlained upon adaptational grounds.
Fewer stamens are usually regarded as a mark of progress in
evolution. But why, for example, in a family mostly provided
with ten, like the Caryophyllaceae, should the "advanced"
members (which in actual fact look less advanced) only have five,
with no indication, fossil or other, that they have ever had ten?
Why does one find no trace of plants with nine, eight, seven, or
six? If it be of any advantage to reduce the number of stamens,
surely nine would be an improvement upon ten, and so on. Why
should the whole whorl be got rid of with no trace of intermediate
stages? The supporters of selection, when confronted with a
morphological problem like this, are obliged to defend themselves
by bringing in another supplementary hypothesis, this time a
"tendency", supposed to exist in plants, to vary the number of
the stamens by whole whorls at a time, which of course is more
in keeping with the general morphology of the flower, though it
is a very remarkable thing that this tendency is so widespread in
flowering plants, there being extremely few cases, so far as the
writer can remember at the moment, of intermediate stages in
regular flowers. In other words, the supporters of selection admit
that morphological facts weigh more in evolution than does selection,
and they also admit that large mutations can take place. And
whence did this tendency come, if it was not handed down from

CH. XI] B. MORPHOLOGICAL 121

the first ancestor of the whole family? Many of the families that
now exist go back unchanged through the fossil records to more
and more ancient times, or rather some of the larger and more
widely distributed ones (the older, by age and area) do. There is
no record of any preliminary stages in the development of a
family, so that to imagine its characters as having been handed
down from a first (single) ancestral form requires no stretch of
the imagination, though it is not quite in keeping with the views
derived from Darwinism. And as such large changes as the loss
(or gain) of whole whorls of stamens are admitted, there seems
no reason left why it should not be admitted that the family
ancestor can appear by a single mutation.

One more example must suffice — the opening of the anther by
slits, by valves, or by pores. Here again, one of these characters
may be found in a whole family, in part of one, in a few genera, in
one, or even in some species only in one genus. But where does
natural selection get any leverage upon the character? In what
way can it possibly matter to a mature plant which of the methods
of anther dehiscence is employed, or to a young plant how its
anthers are going to open at a later period? And how can the
differences arise except by direct mutations? Gradual stages are
almost inconceivable. The only adaptational value ever suggested
is that the valve or pore might localise the pollen better upon a
visiting insect, but unless the stigma is also arranged so as to
touch the part bearing the pollen, there will be no gain, but rather
loss. And this brings up the question of correlated characters,
about which something must presently be said.

It is a matter of very great difficulty to account for morpho-
logical uniformity unless it arise by direct mutation, and unless
it be handed down from above, as differentiation demands. How
did the widely distributed tap root come into existence in so
many flowering plants ? How did the pore of an anther come to be
like that of a fruit? How did leaves appear? Why have such a
vast number of them much the same dorsiventral anatomy?
Why are so many exactly opposite? Why are they in definite
phyllotaxies ? Why are some simple and some compound, why
are they entire or toothed, palmate or pinnate, and so on? How
could all the Cruciferae, and they only, get tetradynamous
stamens, which have incidentally no adaptational value, and
have these together with the other well-marked characters of
this family? Once more it must be admitted that morphology, or
what the selectionists call tendencies, can override natural

122 TEST CASES [ch. xi

selection, and that natural selection can do nothing to explain
staminal morphology. The phenomena shown can only, at
present, be explained by the supposition of sudden mutation,
causing, for example, the loss (or gain) of five stamens, or the
formation of a new method of dehiscence, etc.

TEST CASE XIV. THE BERRY FRUIT

The berry, as seen in the gooseberry or grape, is a well-marked
and distinct type of fruit, the only hard part being the seeds,
though there is a skin upon the outside. In the drupe, as seen in
the cherry or the plum, the innermost layer of the fruit wall is
hard, and the seed (kernel) inside this is usually soft.

There are berries in about a third of existing families, and as
these contain more than half the total of genera, they are upon
the large (old) side. But only a portion of their genera have
berries. Berries occur all through the flowering plants, including,
for example, the Araceae, Bromeliaceae, Rafflesiaceae, Annona-
ceae, Vitaceae, Myrtaceae, Ericaceae, Solanaceae, and Cam-
panulaceae.

The fleshy fruits have always been a standby of the supporters
of selection, who of course had to find adaptational reasons for
phenomena, and supposed these fruits to be adaptations for dis-
persal of the seed. But if the seed be carried far, it will likely be
dropped into another association of plants, where the competi-
tion will be equally severe, and the conditions probably different,
so that it will be, if anything, at a disadvantage. One rarely
finds another plant growing in an association to which it is
foreign.

The berry and the capsule go together very much in related
groups, but the capsule is much commoner, though it shows no
adaptational advantage; the seed may be shaken out in a wind,
but that is all. Some berried families, like Annonaceae, are com-
mon and widespread, but so are capsular families like the
Caryophyllaceae. There is no evidence to prove any adaptational
value in a berr}^ An instance which was sometimes brought up
was the family Taccaceae, where Tacca, with a berry, is wide-
spread through the tropics, and Schizocapsa, with a capsule, the
only other genus, is confined to Siam and South China. But in
Dioscoreaceae, one berried genus, Tamus, is confined to Europe
and the Mediterranean, and has only two species; the other,
Peter mannia, with one species, occurs in New South Wales

CH. XI] B. MORPHOLOGICAL 123

(incidentally, therefore, the berry fruit must have been acquired
independently of that of Tamus) ; while Dioscorea, with 600 species
and a capsular fruit, is in all warm countries. Such cases are
common in many different instances of various fruits. There is no
evidence to prove that advantage is gained by the possession of a
berry. In fact, as was pointed out in Age and Area, p. 21, nothing
in the distribution of plants would lead any one to suppose that
the " mechanisms for dispersal " have produced for the plants that
possess them any wider dispersal than usual. Tithonia, with no
pappus, and with mainly vegetative reproduction, spread as
widely as, and not much more slowly than, the bird-carried
Lantana in Cevlon, to which both were introductions.

The berry may occur in the whole or part of a family, in a few
genera, in one, or in part of one. Considering the wide taxonomic
separation of many of the berry families, it is clear that it is very
improbable that all derived it from the same ancestor, unless the
character could remain dormant for immense periods of time and
change. It would rather seem to be one that is easily acquired,
perhaps through some kind of kaleidoscopic change in the assort-
ment of genes.

To explain why the berry is more common than the drupe,
which is equally well adapted to transport by birds or animals,
the selectionists have to bring up one of their many supple-
mentary hypotheses, this time a "tendency" to vary rather in
the direction of berry than of drupe, or again an admission that
morphological facts weigh more in evolution than does selection.
Presumablv there is a still greater tendencv to varv in the
direction of the capsule, the least "efficient" fruit of the three.
And whence did the tendencv come, unless it were handed down
from a common ancestor in each group, for whole families like
Epacridaceae show a tendency towards the drupe, while their
close relatives Ericaceae show a tendency chiefly towards the
berry, but sometimes towards the drupe? Rhamnaceae have a
dry fruit or a drupe, their close relatives the Vitaceae a berry. In
the genus Chironia (Gentianaceae), mainly African, a small
group of species in South Africa and Madagascar have a berry,
the rest capsules. Why are there no berries in most of the generic
area? This geographical localisation of structural features is
common ; e.g. in Styrax, the first genus to come to hand, most of
the species have sixteen to twenty-four ovules, but there are
some with three to five in Cuba and in Peru. It is a matter of
great difficulty, if not impossibility, to explain such cases by

124 TEST CASES [ch. xi

means of selection, but quite simple by mutational differentia-
tion.

In the Caryophyllaceae, all otherwise dry-fruited (usually
capsules), Cucuhalus alone, with one species in Europe and Asia,
has a berry. In Ranunculaceae, admittedly a very old and widely
dispersed family, Actaea has a berry, while in Annonaceae all have
berries but Anaxagorea.

In Myrtaceae, half the family, the Leptospermoideae, have a
dry fruit, the other half, the Myrtoideae, a berry. How did
natural selection, working upon the ancestors, ensure that all
those with the berry should be more closely related to each other
than to those with the dry fruit? Again "tendencies" have to be
called in, but the differentiation answer is simple; an early
mutation split off a genus with a berry from one with a dry fruit,
and the descendants have inherited one or the other. An excep-
tion like Cucuhalus is explained by a later mutation which in-
volved a change of the chief fruit character of the family.

A great difficulty is to explain why the berry is always the
same in its general structure, though it must have been picked
out upon so many separate occasions. Why is it usually associated
with the capsule, while the drupe is usually associated with the
achene or the nut? In some families both may be found, but each
keeps strictly to its own morphology, though under selection one
would have expected more variety. How did capsules and other
kinds of dry fruits that occur in close relatives all manage to
change to berries of the same morphological construction? No
intermediate forms occur, with few and slight exceptions. It is
clear that the phenomena of berries are better explained by
differentiation.

TEST CASE XV. ACHENES AND FOLLICLES

Here again are types of fruit found all through the classification
of the flowering plants. Alismaceae have achenes, while their
near relatives the Butomaceae have follicles. Half of the Ranun-
culaceae have one, half the other. Two of the three groups of
Spiraeoideae have one, one the other. How were all these groups
produced by natural selection with gradual adaptation? The
question has hardly been properly thought out. How did the one
fruit obtain, by this method, a completely closed wall, the other
(when ripe) a completely open one? The value of selection would
become less and less marked as the fruit approached perfection

CH. XI] B. MORPHOLOGICAL 125

in either of these respects. Yet both the follicle and the achene
show perfection — the one in its complete closure, the other in
opening from one extreme to the other of the wall, and only on
one side. Why, again, did selection cause only one side of the
follicle to open, and that exactly, while the pod opens with equal
accuracy upon both sides? No ada^^tational difference between
them can be shown to exist.

Ovules, again, cannot be developed in stages, from nothing to a
complete ovule, though the reverse process is possible, but
usually leaves some rudiments, which are not found in an achene.
Nor can one imagine any transition — direct or through some
intermediate form — from a multi-ovulate dehiscent fruit to a
one-ovulate indehiscent. Nothing but mutation, and that con-
siderable, could effect such a change, and as there is no adapta-
tional reason behind it that one can conceive, a single mutation is
more probable than a series of mutations. And again the morpho-
logical question comes up — why are all follicles structurally
alike, and why were they produced in preference to pods or to
achenes? In the author's opinion, nothing but a complete muta-
tion of considerable size can have produced the difference, and
nothing but inheritance from a common parent can have caused
it to be shown by whole groups of species, genera, families, etc.
In other words, differentiation is the most probable explanation,
and natural selection in any direct form is out of the question.

Other types of fruit lend themselves to similar explanations,
in which adaptation has but little if any part. It is, when one
comes to think about it, a matter of extraordinary difficulty to
show that the different fruits have any real adaptational value.
What is the value to a tree like a Dipterocarp, which grows in
dense practically windless forest, and often in a forest of one
species only (pure stand), of its characteristic winged fruit? How
could it, under the circumstances, have been developed by
natural selection? Under gradual variation, all the sepals would
vary alike, so that it must have begun with a mutation. And
why should this be small, and not complete? In any case the
calyx does not appear till the tree is perhaps thirty years old, and
can anyone pretend that the struggle for existence between trees
of this age and size is so severe that natural selection can get a
leverage upon so slight a difference as the fact that two or three
of the sepals are slightly longer? If anything, as the elongation
will use more material, the longer sepals are more likely to be
disadvantageous. If one say that the winged fruit gives the

126 TEST CASES [ch. xi

advantage of dispersal to some little distance — an almost certain
advantage if not pressed too far — how were the non-winged
parents killed out, unless the winged offspring were also superior
to them in some functional character that enabled them to kill
out the parents upon ground which they already occupied, and
where they had the great advantage given by the fact that they
were already established there, and that transport of seeds in
windless forest was a very difficult thing?

TEST CASE XVI. THE ORIGIN
OF LARGE GENERA

Another troublesome problem for the selectionist is to explain
how the method of selection gave rise to large genera. Upon what
grounds of adaptation did Senecio come to have about 3000
species, and other genera also have enormous numbers, com-
bining with the numbers a vast distribution over the earth's
surface? If they owe their wide dispersal ("success") to adapta-
tion, that adaptation can only have been generic. There are no
characters in the individual species that one can point to as
adaptive, and how could an adaptive and generic feature be
produced in a genus formed from below upwards by the dying
out of intermediates between it and its near relatives? If one of
the species that were going to form Senecio had a really fine
adaptation, one would expect it to go ahead and rather form a
genus of its owti than simply join the rest. The bulk of the species
in these big genera are local in distribution, and it is far simpler
to explain the whole matter by differentiation and by age, which
simply says that on the average the wider-dispersed species are
the older.

Other remarks on "generic" adaptation will be found on
pp. 18, 59.

TEST CASE XVII. SOME MORPHOLOGICAL

PUZZLES

Even in the comparatively few cases where a plant shows some
structural feature that may be looked upon as a definite physio-
logical advantage, like the tentacles of the Droseraceae, natural
selection is hard put to it to explain how they could be formed by
gradual adaptation. How, for example, did it produce the mar-
vellously sensitive tentacles of Drosera itself, when the first steps

CH. XI] B. MORPHOLOGICAL 127

in their formation would be absolutely useless, and when their
movement would be of no value until it was perfected? Why did
it also evolve Drosophyllum, with no movement, and with two
kinds of tentacles ? And how did it place the tentacles in straight
rows, and make them all alike? Again the reply has to be that
morphological considerations inherent in the plant override the
effects of natural selection. And why so, when they must them-
selves have been derived in the same way? Further, how did
natural selection evolve, in the same small family, Aldrovanda
and Dionaea, with leaves that close up like a book? One does not
expect to find, in so small a family, such marked differences; it
reminds one of the Podostemaceae. The differences are much more
marked than in a whole large family like the Compositae or the
grasses.

There is almost no end to the inexplicable difficulties in
structure that can be brought up for the selectionist to try to
explain. Here, for example, are a few picked out in hastily
running through the list of family distinctions given at the end of
my Dictionary:

The windows in the leaves of Aponogetonaceae.
The complex inflorescence of Zostera.
The three-ranked leaves in Cyperaceae.
The spiral or disc-like flowers of Cyclanthus.
The pitcher of leaves in many Bromeliaceae.
The resupinated flowers of Orchidaceae.
The Equisetum-like stems of Casuarina.
Chalazogamy.

The explosive stamens of Urticaceae, etc.
The integumentless ovule of Opiliaceae.
The tetradynamous stamens of Cruciferae.
The pod of Leguminosae (why not a follicle?).
The obdiplostemonous stamens of Oxalidaceae.
The cyathium of Euphorbia.
The explosive capsule of Impatiens.
The stinging hairs of Loasaceae.
The asymmetrical leaf of Begonia, etc.
The one-sided flowers of Lecythidaceae.
The vivipary of Rhizophora.
The free-central placenta of Primulaceae.
The corona of Asclepiadaceae.
The scorpioid cyme of Boraginaceae.
The didynamous stamens of Labiatae, etc. (why
do they match in several different families?).
The four nutlet fruit of Labiatae.
The pappus of Compositae.

128 TEST CASES [ch. xi

Nothing but common descent will explain most of these, and,
if so, the family must have been very ancient, and why are there
no fossil traces of any family formation, which must have gone on
for an immense period of time if they were made by the method of
dropping intermediates involved in the explanation by natural
selection? Not only so, but the bulk of the characters described
in the list above, to which hundreds more might be added, are
such that they must have arisen at one step; either no inter-
mediates are possible or they would have been completely useless,
and therefore incapable of being chosen by selection.

TEST CASE XVIII. THE SMALL GENERA

If the small genera are to be regarded as failures and relics, it is
somewhat remarkable the way in which they are closely grouped
round the large ones, usually regarded as the successes. If one
take the two largest genera in a family^ — the two which upon the
theory of differentiation represent, upon the average, the result of
the first throwing of a new genus by the original genus which was
the first parent of the family — one commonly finds them marked
by a large divergence. But this same divergence is shown
(cf. p. 84) by the groups of "satellite" genera round them, and
these include the bulk of those which are classed as relics. Their
characters are the chief characters of Ra^iunculus, for example.
Upon the theory that Ranunculus owes its success to some of its
visible characters, we should expect these to be the characters.
Why then are the satellite genera so "unsuccessful"? Very few
small genera are known which are not classed in the sub-families
which are usually marked each by a fairly important genus at the
head. And why should this be, unless the satellites were derived
from the large genera? If this happened in the earlier days of the
big genera, it is somewhat remarkable that one so rarely finds any
fossil traces of the little ones, and if in the later days, w^hy should
the big genera throw off, at such a late period, genera that were
only to be relics or failures? It seems much more probable that
the small genera were thrown off at a late period in the life of the
large ones, by some larger change than would give rise merely to
new species, but a change that could not have been the result of
the work of natural selection. The test favours differentiation
much more than it favours natural selection.

<^"-^i] B. MORPHOLOGICAL 129

TEST CASE XIX. CORRELATED CHARACTERS

The difficulty of imagining that evolution worked in the direction
Irom species towards genus is vastly increased when we come to
deal with the correlations that exist in the characters of the
various flowering plants. Though there is usually no conceivable
adaptational reason behind them, the characters of whole families,
for example, usually go together in groups, for whose connection
we can see no reason at all, unless it be simplv that the common
ancestor happened to possess this combination. In the grasses
there go together alternate leaves, in two ranks, a split sheath, a
ligule, jointed stems, a spikelet inflorescence with glumes and
paleae, and so on. How did natural selection pick out all these
characters to go together, even if by any stretch of the imagina-
tion one could imagine it picking out a split sheath in a grass,
and a closed one in the allied sedges, or in fact any of the other
characters? They must have been derived from a common an-
cestor, and if so, where did selection and adaptation come in? If
all the structural characters of a family, those characters in fact
that mark it out as a family, are hereditarv characters, there is
comparatively little room left for any adaptive characters at all
and once again it is clear that morphological characters override
selection. Even if there be no specially adaptive characters in
the grasses or the sedges, there must have been some disadvan-
tageous ones in the plants that were suppressed in the struggle for
existence which made the wide gap that now separates these two
allied families. But is it conceivable that a series of intermediate
forms with, for example, a sheath partiallv split should have been
so inferior that they were killed out? Still more difficult is it to
imagine intermediates which showed intermediate characters
m all the characters of diff^erence, if one suppose for an instant
that such a thing were possible; there can be no intermediates
between 2-ranked and 3-ranked leaves, or between the two types
of inflorescence, etc. Direct mutation must have occurred in
many cases; and gradual adaptation is hardly conceivable,
especially when so many characters have to go together, and each
has to be brought to the point of perfection (cf. p. 114).

The larger the family, the greater on the average is the variety
of conditions that it occupies, as may be seen in the grasses. Yet
natural selection is supposed to form a family by gradual adapta-
tion, and it is therefore clear that, as indeed the^ geological record
shows, families must have come verij early in evolution, and the

WED

130 TEST CASES [ch. xi

great variety of conditions in which the large families now live
must have been due to subsequent adaptation. But this leads to
the somewhat surprising conclusion that adaptation must have
been very strongly in evidence in early days, with a corresponding
amount of destruction to separate the families, for which we have
no evidence.

Or why were one-fifth of the flowering plants picked out to
have only one seed-leaf instead of two, to have the parts of the
flower in threes instead of in fives, to have leaves with parallel
veins instead of netted, and to have so different an internal
anatomy, with no simple process of growth in thickness? And
still more difficult is it to explain, on the theory of gradual
selection, why all these characters should go together, when they
have no adaptational meaning, either singly or in combination.
One can conceive that the anatomy of the Monocotyledons was
definitelv disadvantao^eous, which mav explain whv there are
comparatively few trees among them; yet the palms seem
successful enough, or the bamboos. But the important fact
remains unexplained, and not to be explained upon the theory of
gradual selection, that, as already pointed out, the Monocoty-
ledons maintain their proportion of one in five in all important
parts of the world.

An interesting case of correlation incidentally showing the
totally useless nature of many, or nearly all, of the generic and
specific characters may be seen in the genus Pyrenacantha in
Icacinaceae, which has a drupe with the inner side of the shell
thorny; correlated with this are definite holes right through the
endosperm to leave room for the spines. Here is a case that it
would puzzle the selectionist to explain, and there are many more.
And it is somewhat difficult to imagine intermediate stages.

To try to explain these correlations in terms of gradual adapta-
tion is a practical impossibility, and if they were formed at one
step, how does adaptation come in? Take, for example, the case
of climbing plants, already considered (p. 57). Or take parasites,
which must also have been a later development than non-
parasitic plants. Until the sucker has actually penetrated the
host, the habit will be of no value, so how did it begin under the
operations of natural selection with gradual adaptation? And
incidentally, such parasites as the fungi live almost entirely
within the host, where the conditions must be more or less the
same for all, so how did they come to develop such numbers of
species with definite structural diff'erences? How did the ordinary

CHxi] B. MORPHOLOGICAL 131

leaf come to develop stomata, intercellular spaces, palisade and
spongy tissue, and the fine network of veins, and how did it
develop these last in so many patterns of netting, parallelism, etc.?

Correlation, if large, implies that most characters have no
bearing upon natural selection, and do not interfere with the
results gained by the first character. And as differences in one
character only do not usually cause mutual sterility, one wonders
how that comes to be so common a mark of specific difference.

One must look with great suspicion upon such easy interpreta-
tions of things as calling them direct adaptations. If they were
formed as such, the work was too complicated for natural selec-
tion. It is more probable that they were formed at one step, and
not being harmful, were allowed by natural selection to survive.

9-2

CHAPTER XII

SOME TEST CASES BETWEEN THE
RIVAL THEORIES

C. TAXONOMIC

JL HESE cases might equally well go under morphology, for
taxonomy or systematic relationship is founded upon that sub-
ject. The separation is simply used to prevent the morphological
chapter from growing too large.

It is of interest to note how easily the axioms of taxonomy
that are given by Darwin in the Origin of Species are explained
by the theory of differentiation. The first one, for example,

Wide-ranging, much diffused and common species vary most

fits in admirablv with much that has been said above, and with
what the writer hopes to bring out in another book. It should
also be compared with Guppy's remarks about the wide-ranging
species that so often accompany endemics, and with what is to be
said about the wide-ranging species that so often do the same
thing in India (p. 158). One may also refer to what will be said
about contour maps (p. 149).

The current view is that the large and widely distributed
genera and species are the "successful" ones, and that they are
breaking up into new species by the formation of what as yet are
only small varieties. On the view taken by the adherents of
Darwinism, the Linnean species of the taxonomist is an abstrac-
tion, consisting of an agglomeration of smaller forms that really
breed true, and that may be more or less well assembled into a
Linnean species which can be reasonably well marked off from
others that are closely related to it. But upon the theory of dif-
ferentiation the case is turned the other way round. There is little
or no doubt that many of the very local endemic species, which
are often supposed to be relics, but which upon the theory of age
and area are regarded as young beginners, are well and clearly
marked Linnean species. Take, for example, the Coleus elongatus
(p. 24), or the Indian local species described on p. 159. The whole
species, in cases like the Coleus, is made up of so few individuals
that it is impossible that there should be a great range of varia-
tion, for mere lack of numbers. It is after the formation of the

I

CH. XII] C. TAXONOMIC 133

species, when it begins to move into a greater range of conditions
and climates, that it begins to show a range of smaller forms,
which to the writer represent later stages in the continually
diminishing mutation that began with the formation of the
family, the genus, and this particular species. It is to be noted
also that the great range of form only shows as a rule in species of
the larger (or older) genera. When it does occur in species of
small genera, the genus usually has a wide range, showing that
it is probabl}^ old in its own circle of affinity.

It is verv difficult to see whv on the Darwinian scheme the
only genera of a very small family ("relics ") should be separated
by as large distinctions as those that separate the sub-genera of a
large family (p. 112), and why those distinctions should be so
often such as are incapable of having intermediates, like many of
those given in Appendix I. And it is equally difficult to see why
the species of a single genus making up a family by itself should
be grouped by such wide divisions as are instanced upon p. 79,
again distinctions that do not often admit of intermediates.

From the differentiation standpoint, the puzzle presented by
these little "Jordanian" species, such as were described in
Draba, for example (22), and which no stretch of imagination can
show to be the commencement of new species derived by gradual
adaptation upon the Darwinian plan, becomes quite simple. They
are simply the last wavelets of the great disturbance that was
made when the parent of the Cruciferae was formed from some-
thing else by a "large" mutation that gave it tetradynamous
stamens and the rest of the outfit of the Cruciferae.

The second axiom is

2. Sjjecies of the larger genera in each country vary more frequently
than the species of smaller genera.

Here again the variation w^as put down to the "success" of the
larger genera, which were going on to develop new species, but,
as explained above, it is much simpler to put it down simply to
the age of the genera and size or area of the species, the larger
being older, and having had more time to develop smaller muta-
tions than that which gave the ordinary species.

3. Many of the sjjecies included within the larger genera resemble
varieties in being very closely but unequally related to each other,
and in having restricted ranges.

This is exactly what shows in the hollow curves. In the large
genera there is a great proportion of species of small area, far

134 TEST CASES [ch. xii

more than of medium or large (cf. p. 98). In places where there
are many of them close together, as with the big genera Eugenia
or Memecylon in Ceylon, they are all more or less closely related,
though many of them are quite good Linnean species. Attention
may also be drawn to the "Jordanian" species in big genera like
Draba or Hieracium.

4. The varying species are relatively most numerous in those
classes, orders, and genera which are the simplest in structure.

5. As with species, so with genera and families. . .upon the
whole those are the best limited which consist of plants of complex
floral structure.

6. Those classes and families which are the least complex in
organisation are the most widely distributed, that is to say that they
contain a larger proportion of widely dispersed sjjecies.

7. This tendency of the least complex species to be most widely
diffused is most marked in Acotyledons (Cryptogams) and least so
in Dicotyledons.

8. The ynost widely distributed and commonest species are the
least modified.

All these latter axioms go together, and obviously fit exactly
with what would be expected under the law of age and area,
which makes the older (and therefore on the whole the simpler)
forms to occupy more area than the younger and more complex.
The fact that all these dicta are axiomatic does not say much for
the supposed continual improvement in adaptation under the
operations of natural selection, especially as this theory also tries
to explain greater range by the same improved adaptation. The
whole of the axioms are rather against the Darwinian theory of
progress, and are in much better accord with that of differen-
tiation.

TEST CASE XX. THE POSITION OF THE LARGEST

GENERA IN A FAMILY

On the theory of natural selection, one can make no prediction
whatever as to the position in the classification of a family of
the largest genera in it. There seems no reason whatever in any-
thing that we know about them which should show that they
should be near together, or that they should be far apart. But
upon the theories of diff'erentiation and age and area, the largest
genera should on the whole be the most widely separated, in-

CH. XII] C. TAXONOMIC 135

asmuch as they will have inherited their characters from the
point that is the farthest back that is possible — the earliest
mutational divisions that took place in the family concerned.

In deciding this point we must, of course, work with the keys
with which the taxonomists have provided us, but the latter
have, of course, taken the greatest possible pains to find the most
widely different characters that mark the different groups. In
their keys they usually begin with very divergent characters,
inasmuch as they have learnt by experience that these mark the
largest divisions in the great majority of cases, separating the
genera first of all into tw^o large groups. These groups again are
separated by the most different characters that can be found, but
which do not mark the whole, but only a part of the first group.
And so on, breaking up the family into allied groups within
allied groups — the general principle of all classification.

One will, therefore, expect that the first one, two, or at most
perhaps three separations that are given in any ordinary good
key will separate not only the chief sub-families or tribes, but
also the largest genera, and one will expect these to be separated
by such distinct and divergent characters that there will be little
or no difficulty in picking them out from one another. When such
difficulty occurs, it should be in genera that have become so large
that their outlying species, which will have been liable to more
change than the earlier and more "genus-like" ones, have in one
or two cases reached almost to the overlapping point. We should
expect, but have not had sufficient time to test the matter, that
these difficult species would prove in general to be comparatively
local, that is to say, on the whole, the youngest species in their
genera, which will have gone through the greatest number of
mutations since the first throwing of the genus.

As a test of this case, we may take the family Ranunculaceae,
which is already described from this point of view in the chapter
upon Differentiation. The first division of the family in most keys
throws the largest genera on both sides. Here, for example,
Aconitu77i, Aquilegia and Delphinium have follicles, and Anemone,
Clematis, Ranunculus and Thalictrum have achenes. But as the
two actually largest genera are Clematis and Ranunculus,
separated by the very divergent character of opposite or alter-
nate leaves, it is possible that this was the very first mutation,
and that Clematis mutated off no other genera with opposite
leaves. Or yet again, we must always bear in mind the possi-
bilities of such complex mutations as are indicated in Hayata's

136 TEST CASES [ch. xii

work (16). All that we are at present contending for is that
species and genera were formed each at one mutation, and that
the change went downwards to the species, not upwards from it,
as required by natural selection (of course a genus cannot exist
without one species.

In the same chapter (ix) we have also described the sub-family
Silenoideae of Caryophyllaceae, and shown that the first split in
the key throws Silene with 400 species to one side, Dianthus with
300 to the other.

This phenomenon, which is so common that it must have a
reason behind it, occurs in a great number of cases. Picking up a
few copies of the Pflanzenreich as they come, the first is Maran-
taceae, where Calathea and Maranta, the two largest genera, are
separated by the first split. In Myrsinaceae, Ardisia goes one
side and Rapanea the other. In Amarantaceae, Alternanthera
goes one side, Ptilotus the other. In Cyperaceae Cyperus and
Carexdo the same; in Eriocaulaceae£'Wocat//on and Paepala?7thus.
In Hydrophyllaceae, Phacelia and Nama go one side, and Hy-
drolea, with only nineteen species but very wide distribution, the
other. In Monimiaceae (p. 33) Siparuna goes one side and
Mollinedia the other. And so on indefinitely.

It is thus clear that as the position of the largest genera, and
their sharp distinction in the great majority of cases, agrees with
what is required by the diff'erentiation theory, while that of
natural selection can give no idea where they will be found
in a family, the evidence of this test is in favour of the former.
Inasmuch as the classification of animals is equally possible, with
equally good results, when conducted upon the same lines as that
of plants, and as it shows the same hollow curves, it would seem
highly probable that the same general principles have guided the
evolution that has gone on in them also.

TEST CASE XXI. THE POSITION OF
THE LARGE FAMILIES

We may even carr}^ the supposition outlined in the last test a
stage farther, and apply it to families, saying that the very large
ones will be very widely separated. We are still so undecided
about the proper classification of the larger groups of plants that
it will not do to push this very far, but one may note that the
three largest families of all, from the latest figures in my posses-
sion, are Compositae (18,039 species), Leguminosae (12,754) and

CH. XII] C. TAXONOMIC 137

Orchidaceae (10,088). Here, incidentally, is a case for the provisos
with which I hedged age and area, that one must never com-
pare anything but close relatives as regards age. To say that the
Compositae are older than the Leguminosae is obviously a state-
ment with nothing to back it. But the fact is there, that one
could not easily find greater divergence than is shown by these
three families, which incidentally contain nine out of twenty-nine
of the genera containing over 500 species each. If we go over the
first ten families in point of size, we find the fourth, Rubiaceae,
to have one of these genera, the fifth, Gramineae, one, the sixth,
Euphorbiaceae, three, the seventh, Melastomaceae, two, the
eighth, Labiatae, one genus, while there are none in the other
two. But these eight families contain seventeen out of the
twenty-nine of these big genera, and most of the rest are in large
families, though there are a few small ones which contain very
large genera, like Begoniaceae.

The twenty-seven large families with over 1500 species are
Acanthaceae, Apocynaceae, Araceae, Asclepiadaceae, Bora-
ginaceae, Bromeliaceae, Caryophyllaceae, Compositae, Cruci-
ferae, Cyperaceae, Ericaceae, Euphorbiaceae, Gesneraceae,
Gramineae, Labiatae, Leguminosae, Liliaceae, Melastomaceae,
Myrtaceae, Orchidaceae, Palmaceae, Rosaceae, Rubiaceae,
Rutaceae, Scrophulariaceae, Solanaceae, Umbelliferae. It will
be seen at once how wide a range they cover in the classification,
in fact touching all important parts of it.

The evidence of both these test cases is strongly in favour of
divergent mutation, forming the whole family or genus at one
step.

TEST CASE XXII. DIVERGENCE OF VARIATION.

SYSTEMATIC KEYS

In making keys to families or genera, by whose aid one may
determine the relationships and position of the plants with which
one is dealing, the taxonomist is concerned with providing the
easiest and most certain method of so doing. And it is a very
remarkable fact, that has hardly been sufficiently recognised,
that it is usually possible, without any very great difficulty, to
make a dichotomous key (sometimes trichotomous at certain
points), beginning at the top with characters that will separate
one sub-family from another, and working right down through
tribe, genus, and species, to variety, in the same way. This fact,
which upon the theory of differentiation must occur, does not

138 TEST CASES [ch. xii

agree very well with the theory of natural selection, nor with
that of gradual adaptation. There is no doubt that as one proceeds
up the scale from variety, the divergence of the characters be-
comes greater and greater, and upon these latter theories it is a
matter of extraordinary difficulty to explain why the destruction
of the intermediate forms should proceed in such a way as to
leave groups that present divergences that are more and more
marked the higher that one goes in the scale, while at the same
time they are quite simple divergences, such as ovary uni- or
multi-locular, anther opening by slits or by pores, leaves opposite
or alternate, and the rest. Nothing but differentiation can at
present explain such phenomena*

TEST CASE XXIII. DIVERGENCE FROM USUAL
FAMILY CHARACTERS

It is a very noteworthy thing, which the selectionists have found
so difficult of explanation, that they have had to fly to their
usual refuges, that plants that show great divergences from the
characters usual in their families occur, not in the small families
(relics or failures) but almost only in the large ("successful")
ones. We have given an instance from the Rubiaceae on p. 118
and the matter is discussed in detail in (74, p. 621). In the large
families one would be inclined to expect constancy, for if it were
settled by the early ancestors of Papaveraceae, for example, that
a hypogynous flower was the best, why did Eschscholtzia adopt
a perigynous one? Solanum, by far the largest genus in its family,
opens its anthers by pores, while most of the rest open by slits.
Is this the generic adaptation that caused Solanum to become so
"successful"? One could go on bringing up cases like these, and
there is no escape from the conclusion, so far as our present
knowledge goes, that characters of all kinds, however important
in classification, may be acquired by single genera at any stage,
so that their acquisition is evidently easy, and must almost
certainlv be due to direct mutation. Like all the other tests this
speaks in favour of differentiation.

TEST CASE XXIV. PARALLEL VARIATION

A puzzling case, which the natural selection theory can in no way
explain, except by the favourite suggestion of "tendencies", is
the parallel variation that so often may be seen. A good instance
is afforded by the related families of Eriocaulaceae, Centro-

CH. XII] C. TAXONOMIC 139

lepidaceae, and Restionaceae. Each family makes its first divi-
sion (in some classifications) into Haplanthereae and Diplan-
thereae, or groups with monothecous and with dithecous anthers,
a well divergent and clearly marked division. In the Erio-
caulaceae the Diplanthereae contain half the genera of the
family, including Eriocaulon and PaejJalanthns, which are by far
the largest genera, while the Haplanthereae include only very
small genera, whose species make only about 2 J per cent of those
in the other group. And whilst the Diplanthereae cover the
warmer parts of the world, the Haplanthereae are found only in
warm America. In the Centrolepidaceae and Restionaceae, on
the other hand, the larger group is the Haplanthereae. In the
former, they include five genera and thirty-five species, against
one and two in the Diplanthereae ; and in each case the distri-
bution is much more extensive.

There is no conceivable reason why dithecous anthers should
suit America better, and monothecous the Old World, and yet
the former are more common in the one, the latter in the other.
It is clear that we must be dealing here with a divergent muta-
tion, and that one family began with dithecous anthers, the other
two with monothecous, and that probably each one subsequently
split off the other division. The Eriocaulaceae, for example,
beginning dithecous, spread over the world, but split ofP the
monothecous group in America. Perhaps the splitting off was too
late for the plants to cross to the Old World in any case, or it may
have been that as we have elsewhere explained the early growth
and dispersal of the new forms was too slow for them to be in time
to cross.

Cases of the same kind, showing exact parallelism, are very
numerous indeed. To take a few examples, the Marantaceae
divide into a group with 3-locular ovary, and a group with
1-locular, and each of these divides into a group with two lateral
staminodes, and a group with one. In Amaryllidaceae both the
groups Amaryllideae and Narcisseae divide into groups with
many ovules and with few, whilst this is the first division in the
related Haemodoraceae. In Araceae, most of the principal groups
divide into groups with endosperm and without. In the Palma-
ceae, several widely separated groups have fan leaves, others
feathers. And so on, in hundreds of cases.

This phenomenon has always been a great difficulty to explain
upon the theory of selection, for it makes it obvious that none
of these characters — for example, those of climbing plants, else-

140 TEST CASES [ch. xii

where described (p. 57) — can be difficult of acquisition. In
many cases differences of this kind can be seen between closely
related species. The only reasonable explanation is that their
appearance has nothing directly to do with adaptation, and is the
result of simple mutation, which is so very commonly divergent.
In other words, this phenomenon, which is so very common
throughout the vegetable kingdom, and which is not unknown in
the animal, is an expression of the operations of differentiation,
not of those of natural selection, while at the same time it
suggests complications in evolution, perhaps like those suggested
by Hayata (16).

TEST CASE XXV. GREATER LOCALISATION
OF HIGHER TYPES

That the higher groups of organisms, for example the flowering
plants, are more localised in distribution than the lower groups,
such as the ferns, has long been an accepted axiom, and has often
been put down, as for example by Darwin and by the author,
largely to the greater antiquity of the lower groups. But if we
carry this principle into greater detail, it is clear that if in any
family or group of families some forms are more widely distri-
buted than others, those forms should on the whole be the older
— the principle for which the author contended in the hypothesis
of age and area. But the explanation of geographical distribu-
tion that is given by natural selection, or gradual structural
adaptation, involves the assumption that the forms that have
spread the most widely will be those that are the best adapted,
though to what they are adapted is left vague. Upon this view of
evolution, one cannot regard genera like Carex, Draba, Eryngium,
Eugenia, Eujjhorbia or Senecio as being poorly adapted when
compared with the vastly more numerous smaller and more
localised genera. But when one asks why such families as Cepha-
lotaceae, Hydnoraceae, Nepenthaceae, Orobanchaceae, or Sar-
raceniaceae have not spread widely, with such "adaptations"
as they show, one is told that their adaptation is too special to
have allowed them to do so. But why should Nepenthes, for
example, be well suited to the variety of conditions with which it
meets in Malaya, Ceylon and Madagascar, and yet not capable of
withstanding those of tropical Africa, America, Polynesia or
Australia? The Sarraceniaceae, with not dissimilar adaptations,
can do so, and do not occur in the Old World. It is not even as if
there were only one species in each of the genera ; there are scores

CH. XII] C. TAXONOMIC 141

of Nepenthes^ for example, so that the adaptation which enabled
the genus to spread must have been generic, perhaps principally
the pitcher. But if so, why could not some species have been able
to live in America, or some Sarracenias in Europe? No feature
can be pointed out in the pitcher or any other character of
Nepenthes, which should limit it to its present distribution.
Sarracenia, as pointed out on p. 56, is naturalised in a bog near
Montreux. Nothing but an explanation based upon age and
area will answer the innumerable questions like this w^hich come
up in a study of distribution.

This feature, that the enormous distribution of large genera
like Carex or Senecio can only be explained by generic adapta-
tion, if one is to accept the "explanation" given by natural
selection, is a very fatal objection to the theory. The six genera
above mentioned average a thousand species each, and it is a
very astonishing thing that the original adaptations should have
been such that they remain in their progeny after all this degree
of change.

As in general we are not alwavs verv sure of what we mean
when we say that one genus is more complex than another, and
as opposite views are frequently expressed in any particular
case, it is fortunate that in the Podostemaceae and Tristicha-
ceae we have families where it is almost impossible to be in
doubt, for the obvious change that has gone on is from a slight
to a great dorsiventrality. The comparatively primitive forms
are widely dispersed, the more modified are local.

It is fortunate that we have this evidence, for usually it is not
easy to draw conclusions from the morphology. It is often said,
for example, that reduction in number of stamens and carpels is
evidence of progress, yet we can find the widely dispersed species
in some families showing the one thing, the narrowly dispersed
in others. For example, with leaves alternate/opposite, the
Erythroxylaceae go one way, the Caryocaraceae the other; with
flow^ers regular/irregular, Aristolochiaceae and Commelinaceae
go one way, Dichapetalaceae the other. With corolla valvate/con-
volute, Quiinaceae go one way, Cistaceae the other, with stamens
cX)/few we have Loasaceae and Papaveraceae/Quiinaceae and
Velloziaceae. With carpels oo/few, Papaveraceae/Portulacaceae,
and so on indefinitelv.

CHAPTER XIII

SOME TEST CASES BETWEEN THE
RIVAL THEORIES

D. GEOGRAPHICAL DISTRIBUTION

X HIS group of test cases is placed last, as the author is at
present writing a book upon geographical distribution, and many
tests that could be given would require such long quotations
from that work that they are not suitable to the present one.

Geographical distribution, properly so called, unlike ecology,
is so bound up with the question of the origin of the species with
which it deals, that it must be based upon some theory of that
origin, and this theory must be able to explain all or most of the
well-known facts of distribution without serious difficulty. To
take one case only, special creation could not explain the
relationship of species in one country, say Britain, to those in
another far removed, like New Zealand. It was succeeded by
natural selection, which, however, did its great work rather in
establishing evolution, and thus opening out a great field for
research, than in explaining geographical distribution. Not only
did it show that resemblances were mainly due to relationship,
but it also seemed to show that wide dispersal, or successful
spread, as it now began to be called, must be due to unusually
good "adaptation". This latter, however, has never been proved.
The struggle for existence was undoubtedly in full operation
among individuals, but even there, chance had probably a much
greater effect, for the great struggle was amongst the young, and
better water supply, better light, better soil, earlier arrival or
germination, etc., etc., would have greater effect than any slight
advantage that the young plant could carry in itself.

Natural selection had to explain geographical distribution, and
there seemed no other way to explain it than by transferring the
hypothesis from individual to species; but as yet we have no
evidence in favour of this great assumption. We do not know that
species or varieties can come into direct competition with one
another as units in a war a Voutraiice, especially as in general
they will occupy more or less different areas, and one would
hardly expect that species B would follow its defeated rival A
into all its habitats, and kill it out there. If this was the way in

CH. xiii] D. CxEOGRAPHICAL DISTRIBUTION 143

which one species won at the expense of another in the struggle
for existence, one ought to find many cases of this internecine
struggle going on in many places, but one does not. One only
finds a struggle between individuals, in one place a member of
species A being successful, in another a member of B.

The supporters of selection say that the intermediates, which
also came into the competition, have been killed out, and that the
two survivors are now adapted to slightly different conditions.
This is of course possible, but it is a very remarkable thing,
when one thinks of all these processes going on gradually, as must
be the case under the old theory, that one does not find inter-
mediates in the fossil deposits. What are sometimes called inter-
mediates are really a very different thing, usually plants with
some of the characters of one, some of another, really a very good
argument for differentiation. And further, why does one not
find intermediates at the present date? Is the competition now
finished? One would expect to find some cases in which it was
still going on. We have already seen that in a great number of
cases, especially in those high in the scheme of classification, inter-
mediates between the characters are actually impossible, and
how mutation, crossing the whole gap between the two at one
operation, is the only probable explanation. It is no argument in
favour of this supposition, that species can act as units, to say
that masses of men of (to some extent) the same race, like the
Fijians or the Hawaiians, can act together as units. Man has
sufficient intelligence to be able to combine to some slight extent,
though it is a somewhat ironical commentary upon that intelli-
gence that his chief and most efficient combination is for the
purpose of making war, whose results are more against natural
selection than for it.

The new and better adapted form was supposed to kill out the
less well-adapted parent. But as they would usually meet only at
the edges of their respective territories (p. 13), where they
would tend to cross, and to lose their identity, it would require
a vast amount of time for the new one to invade the territory of
the unimproved parent, and to kill it out entirely. Almost cer-
tainly examples of the old species would be left in many different
spots, where they had been overlooked, a feature which in actual
fact is very rarely seen.

Incidentally, the new species would have to kill out all the
hybrids at the meeting place of the new and the old, and if it
had not crossed the "sterility line" it would continue to make

144 TEST CASES [ch. xiii

more hybrids, so that the only result of an incipient species
trying to gain territory at the expense of its parent would be the
continual formation of hybrids. Only when the sterility line had
been crossed would the new species really be able to conquer the
old, and to supplant it. But it is very hard indeed to see how this
line can be crossed in any case without a large mutation that will
create a new species at one step; one cannot easily imagine a
species gradually crossing the line of sterility, nor even a series of
small mutations doing it.

There is evidence to show that on the whole the parent will
continue to gain in dispersal upon the offspring (66, p. 34), and
if this be so, it could not be altogether killed out, unless the
assumption that the offspring, by becoming better adapted to
place A, became thereby better adapted to B, the home of the
parent, were correct. There is little or no evidence that a species,
and still less a variety, fights as a whole, and an organisation that
is based upon such a contention, as so much political organisation
is at present based (the operation of the dead hand, so well
described in Woolf's Aftei' the Deluge, chap, i), has no strong
scientific backing.

To carry out evolution by natural selection involves a vast
amount of destruction, for which we have no evidence in fossil
botany or elsewhere, whilst such destruction is not involved in
the theory of differentiation. To try to explain the phenomena of
geographical distribution upon the supposition that one species
has conquered and destroyed another is to build upon a somewhat
insecure foundation. It has hitherto been assumed that a widelv
dispersed species owes its dispersal to the fact of its superior
adaptation. But to ivhat is it adapted, and how in country A did
it become adapted to the conditions of country B ? If its range be
large, it must come into greater variety of conditions than if its
range be small, and that must mean that as it moved about it
became functionally adapted to all these conditions in turn, but
that is no proof that in becoming adapted to B it retained the
adaptation to A. But in any case much time must be allowed,
i.e. that wide-ranging species are usually old, a supposition that
agrees with age and area. The more local species, which do not
occur in such variety of conditions, are the younger. It would,
therefore, form a much more probable explanation to say that
the widely dispersed species were the old ones, dispersed before
the land was broken up into its present divisions, and before the
climates showed so much differentiation as they do at the present

CH. xiii] D. GEOGRAPHICAL DISTRIBUTION 145

time. These old forms, being simpler, would show less adaptation
to any particular conditions, but would probably show greater
adaptability. This conception agrees much better with the facts,
which go to show, as was pointed out by Darwin, that the organi-
sation of the widely dispersed species is definitely simple rather
than complex, when allies only are considered, as must always
be the case in general comparisons with regard to age (cf. p. 29) or
dispersal.

All the facts that are known go to show that in the majority of
cases an individual plant arises in a place at no great distance
from that where its parent is to be found. If it survive, and grow
to the reproductive stage, one may conclude not only that chance
has favoured it, but also that it has probably passed through the
sieve of natural selection, and may be said to be more or less
suited to that locality. If the seed, however, be carried to a
greater distance than usual, say to more than 250 or 500 metres,
whether it prove so suited to its new locality as to survive and
reproduce there will depend upon a number of things. It may
find a good deal of difference in the soil, though not perhaps in
the climate, and if it has been carried beyond the range of the
particular association of plants in which it has been growing,
there may be considerable biological differences, which again may
be accompanied by soil changes and the like. It will then be a
matter of chance whether it prove suited to the new locality — to
talk of adaptation in a seed only newly arrived, though it may
prove suited to the place, would be going too far. If it survive to
the reproductive stage, it will probably have begun by that time
to adapt itself to its new surroundings. In each successive genera-
tion this adaptation will continue, until, after a time which is
probably different in each case, it has again become fully adapted
to local conditions. This process may continue until, after a very
long period, the species may cover, as does Hydrocotyle asiatica
(p. 58), a very large area of the surface of the globe. If we
abandon the notion that adaptation is shown by the structural
characters of plants, but that it is much more the physiological or
functional adaptation that must go on in any plant that moves
about and comes continually into new conditions, the supposi-
tion that we have just given explains, with the aid of age and
area, why species are arranged over the world in "wheels within
wheels", why the largest numbers are found upon the smallest
areas, and those that occupy larger areas decrease in a "hollow

curve".

WED

146 TEST CASES [ch. xiii

If the conditions begin to change in any place, the new ones
may encourage some plants, and discourage others, so that
natural selection may in time effect a change of the local flora,
some plants coming in from other near-by regions where condi-
tions are more or less like those which now obtain in the locality
under consideration, and some of the local ones perhaps dying
out in that region. Possibty, even, under the stimulus of changed
conditions, new endemics may appear. But while plants that are
really very local may be completely killed out by a serious change
of climate or other conditions, it is very unlikely that this will
happen with plants that are already widely dispersed into a
considerable variety of conditions. To imagine that a species that
has become well adapted to certain conditions that occur in one
country has become thereby adapted to those that may occur in
some country widely separated from the first, is to press the idea
of adaptation altogether beyond possibility.

TEST CASE XXVI. AGE AND AREA

There is no need to add much to the description already given
in chap. iii. One of its striking features is the proof that it gives
that the distribution of a plant within a country, such as Ceylon
or New Zealand, goes on the average with its total distribution
outside that country. When one considers the differences in con-
ditions that must exist, this goes to show that natural selection,
in the sense of gradual structural adaptation, can have had little
or nothing to do with the distribution. What kind of an "adapta-
tion" can a species have acquired that enables it to go so far
afield, into so great a variety of conditions? And still more diffi-
cult is it to explain why the species that are endemic in any given
country are usually closely related to these species of large and
widely ranging genera.

In Ceylon, for example, and the same can be said of other
places, the species that are most widely dispersed locally, on the
average, are those that range beyond the South Indian peninsula,
i.e. beyond a line drawn from Bombay to Calcutta. The next most
widely dispersed occur in Ceylon and in the peninsula only,
while the least dispersed are the local or endemic species that do
not occur outside Ceylon. All, of course, as pointed out in Age
and Area, must be taken in averages, as an endemic in an old
genus (in Ceylon) might be much older, and occupy more ground,
than a newly arrived "wide", even if the latter also ranged to

CH. xiii] D. GEOGRAPHICAL DISTRIBUTION 147

tropical Africa or America. But on averages there are very great
differences between the species of the three groups, and the
statement above made as to relative distribution is fully borne
out in all cases that have been investigated. Between the widely
distributed species and the local endemics in New Zealand, there
is a great difference in range (average length for wides 742 miles,
for endemics 414).

On the theory of natural selection, it is quite impossible to
make any prediction about what is likely to be found in studying
the distribution of plants in such a place as Ceylon. The sup-
porters of that theory tried to answer the author's attack by
calling in two supplementary hypotheses, which as already
shown (p. 24) are mutually contradictory. The Ceylon local
species were supposed in the first to be local adaptations to the
Ceylon conditions. But this did not get over the difficulty of the
intermediate distribution of the species that also occurred in
South India. Were they suited to the conditions that occurred in
both countries, and if so what were those conditions, and how did
natural selection adapt plants in such a way that some Ceylon
things were confined to Ceylon, some reached as far as say
Cochin in South India, while some got as far as Goa and some to
Bombay? This overlapping of areas, which shows in all parts of
the world, is a fatal objection to the theory of local adaptation as
a general rule for the explanation of endemics, without something
else to explain the varying distribution that they show. But in
any case, it was a very remarkable thing that if they were really
local adaptations to local conditions, they should be the rarest
plants in those very conditions. Their general distribution was
simply a reproduction on a smaller scale of the kind of distribu-
tion that might be seen in any big genus or family, or in the
flora of any big country — all gave the same "hollow" curves.
There was nothing peculiar about local endemism to distinguish
it from any other type of distribution.

The rival supplementary hypothesis, which contradicts the
first, and is the popular explanation at the present time, is that
the endemics of a country are the relics of a previous vegetation.
The tenacity with which this opinion is held, in spite of all
evidence to the contrary, is really noteworthy, though a weakening
is to be seen in the tendency to expand the idea of a relic. Such
things as Ceanothus in North America may perhaps be brought
into this category, though the genus has about forty species,
which puts it very definitely into the large genera, but it does.

10-2

148 TEST CASES [ch. xiii

however, seem to belong to the vegetation that was largely
destroyed there by the ice. But things like Artocarpus, with
over sixty species, common in warm Asia, are now being called
relics, because they have fossils in places not now occupied by
them. But if these plants are to be counted relics, one might as
well say that all widely distributed things, but probably not the
local or endemic, are relics, for there are few widely distributed
things that have not the possibility of fossils somewhere, for
example the whole British flora that anywhere reaches the coast.
There are rarely any fossils of the small and local genera that are
usually called relics.

But the hypothesis of relicdom is no better than that of local
adaptation in explaining the intermediate position of the
Ceylon-South Indian things in the distribution. Are they half
relics? No hypothesis other than that which we have termed age
and area can explain the "hollow curve" into which all kinds of
distribution fit. No theory involving natural selection or gradual
adaptation can explain why 38 per cent of the genera of the
world have only one species, 13 per cent two, and only 7 per cent
three, and why the proportions are very much the same wherever
one may go. There is no escape from these facts, and to say that
they are accidental is simply to admit that the distribution of
plants is largely accidental, and to ignore the rule under which
they have probably come into being, the simple doubling of
every species at intervals as time has gone on (cf. Yule, 75). The
author has lately shown that the distribution of family sur-
names in the mountainous regions of Switzerland follows exactly
the same rules as does the distribution of plants. No invocation
of natural selection can explain why Rochat, which is a common
name in its place of origin (the valley of Joux), should have spread
more widely in the canton of Vaud than Capt, which is less
common, or why the surnames should be arranged in "wheels
within wheels" just like the species of the Ceylon or other floras.
Nor can one invoke gradual adaptation to explain why in the far
north-east of its range, Rochat is replaced by Rojard, which is
much more easily explained by the general illiteracy of former
days, and largely matches the way in which plant varieties occur.
It wfll perhaps be well to quote part of the original note, by kind
permission of the Linnean Society (and cf. fig. 6, p. 40).

CH. xiii] D. GEOGRAPHICAL DISTRIBUTION 149

Surname-distribution of farmers
in Canton Vaud {Switzerland)

As a sequel to Guppy's study of surname-distribution of
farmers (who move about less than usual) in Britain, which
showed a good "hollow curve" by counties, the author has
studied Canton Vaud, which is about as large as Gloucestershire,
but much broken up into more or less isolated valleys by moun-
tains larger and smaller. Its nineteen "districts" average 64
square miles each, and they show as good a curve as, or even
better than, that of the English counties. A. very great number
of the villages, especially in the more rugged districts, contain
endemic names found nowhere else in the Canton. Not infre-
quently these occur on more than one farm, and then they
usually show a curve just like that of plants, with the greatest
number upon the smallest area (here one farm). The spread of
a name may be due to various causes that can hardly be regarded
as other than chance, as for example the chance that a farm may
fall into the possession of a woman of family X. If she marry
a man of family A, that family will rise in status by one farm,
and X may even be extinguished. The same process happens
with plants, and the plant (or the surname) that increases its
numbers increases its chance of spreading. The bulk of the
villages in the Canton have one or more names exceeding the
rest in number, and in general these names show greater dis-
persal in the Canton (just as the commoner plants in Ceylon,
for example, show greater dispersal outside the island). Spread
is alike in the two cases, so that it becomes very difficult to call
in adaptation or natural selection as the chief causal agent in
distribution. Rochat, for example, is the commonest name in the
valley of Joux, and has spread the most widely of the Joux names
in the Canton. But there is no adaptation, nor any handle for
natural selection, in the possession of Rochat as a name. No
shigle plant, and no single owner of a name, of course, can
become established anywhere without passing through the sieve
of natural selection, but that is its chief action. The effect of
selection upon a name, or upon a species, will be the sum of
its effects upon the individuals, and one must remember the
failures.

Age and Area is very strongly indeed in favour of differentia-
tion.

TEST CASE XXVII. CONTOUR MAPS

It will commonly be found, in studying the distribution of the
species of a genus, especially if it be of small or moderate size,
that they are more densely congregated towards the centre of the
distribution of the genus, and fall off gradually towards the

150 TEST CASES [ch. xiii

edges, so that when one draws a Hne round the outermost locaUties
of each species one obtains a picture not unlike that which is
called a contour map by the geographers, such as may be seen in
any good guide-book to hilly country. If the genus be small,
there will probably be only one generic centre, whilst the larger
that it becomes, the more broken will the central part be, with,
more and more regions in which there is a concentration of
species, like regions of the higher peaks in a geographical contour
map. So long as a genus is of small or moderate size, the outer-
most or boundary species seems usually to be one species only,
but as it grows larger it becomes rarer for there to be one species
occupying the whole generic area, and one begins to find local
concentrations of species in widely separated parts of the world,
like that which is shown here in the map of New Zealand, with
the species of Ranunculus there found. Here one finds three
"wides" (as I have called the species which have a dispersal
outside the country in question) occupying the whole area of the
islands of New Zealand, and also reaching eastwards to the
Chatham Islands, 375 miles away. Their distribution is thus by
far larger than that of any other buttercups in New Zealand
(fig. 9). The fourth wide has a distribution not very much less
than that of the most widely dispersed endemic. The total length
of the islands is 1080 miles and the breadth does not vary very
much from 100 miles, so that the longitudinal range may be taken
as a reasonable measure of the dispersal of a species. The en-
demics are evidently crowded together rather south of the middle
of the South Island, whilst they fade out completely before the
north end of the North Island is reached. Of the twenty-eight
endemics, ten have a range not exceeding 60 miles of the length
of New Zealand. If one take the ranges in differences of 200 miles
— 200, 400, etc. — one finds that fourteen, seven, five, one, one
species have these ranges, or, in other words, the figures form the
usual hollow curve of distribution, and this is shown by any New
Zealand concentration of the larger genera. The general impres-
sion that one gains from a map like this is that the genus
Ranunculus entered New Zealand probably from the south, and
at some place in the southern half of the South Island, where the
incoming species began giving rise to endemics, and on the average
each species, wide or endemic, spread to the distance allowed by
its age, and suitability to the conditions with which it met.

The same type of contour distribution is shown by the genera
of a family, as fig. 10 shows. Incidentally, these contour maps

CH. xiii] D. GEOGRAPHICAL DISTRIBUTION 151

¥\g. 9. Diagram showing the areas occupied by species of Ranunculus in
New Zealand. Wides dotted; extension East includes Chathams.

(By courtesy of the Editor, Annals of Botany.)

152

TEST CASES

[CH. XIII

a
O

o.

a

/-^
^^

o

II

o

o

o

c

c

rt

o

• rH

;_

O

f?^

Qi

•

M

<D

O

c3

tC

a^

r^

a

o

o

C-i

f-i

^^ ^

CTi

o

S

o

^5

<x>

Oh

CO
•r-l

S

o
zn
:^

o

d

K^

t:)

pH

'i^-|

cS

+J

bfi

O

^.

o

o

•1— '

N

-t-3

c

:3

-4^

-M

• l-H

<1

o

-4^

o

Z/i

• *-H

>.

t!

Q

K

£-1

'"»— ^

D

r™ ^

a

ri2

o

{- bXj
o C<J

o

-1^

-1-3

CH. xiii] D. GEOGRAPHICAL DISTRIBUTION 153

show the absurdity of trying to draw a definite line of distinction
between endemic and non-endemic.

Working upon the theories of Age and Area and of Differentia-
tion, this distribution is exactly what one would expect to find,
but it is extremely difficult to account for upon the theory of
natural selection or of gradual adaptation. On that theory the
widely dispersed things are supposed to be the best adapted. But
to what? It is clear that if the distribution is very wide, each
individual or group of individuals found in any small region can
only be adapted to that region. Suitability to other regions that
differed to some extent from the first could not be such an advan-
tage to a species that it would help it to settle in the first region.
Natural selection, picking out species suitable to A, would not at
the same time pick out qualities that would suit the species
to B ; it could not even know, to put it in a kind of personal way,
that B existed, and that A would gain in area of distribution by
being able to settle there without further adaptation. A species
must become adapted in turn to every change of conditions with
which it may meet, whether differing soil, temperature, moisture,
or biological conditions, and so on, and when at last it meets with
conditions that go beyond its possible range of adaptation, then
it will have met one of the boundaries that limit distribution,
already fully enough described in Age and Area. Probably there
is some kind of limit to adaptation (or it may be only to speed of
adaptation) in most or all species. Sooner or later they will come
up against a barrier, most often probably climatic, which they
cannot pass. But at the meeting place of such barriers, e.g. in
Ceylon at the junction of the dry with the wet zones, one not
infrequently finds different species of the same genus, some on
one side, some on the other. This is apt to suggest that at some
time and place, one or the other species was becoming adapted to
one or the other zone, and that some kind of turn of the kaleido-
scope took place which resulted in the formation of the second
species, better adapted to the new conditions, though its morpho-
logical differences probably had nothing to do with physiological
problems, but were perhaps in some way a correlation.

The general evidence of contour maps, of which a very good
example {Beta) may be found in Nat. Pflanzenfamilien, 2nd ed.
vol. XVI c, 1934, p. 461, is entirely in favour of differentiation and
age and area. It is sometimes suggested that at the centre of a
contour map the conditions are more varied, but very little
thought is required to show the absurdity of this contention. The

154 TEST CASES [ch. xiii

conditions in Britain are perhaps more varied than in any part of
Europe, yet no genus has the centre of its map there, and several
hundred genera have the one marginal species in Britain, and that
only. If this conception were correct, the variety of conditions
would tend to increase away from the sea. If one take family
contours such as those shown in the map of Menispermaceae on
p. 152, the case is even better marked. Such families as Umbelli-
ferae or Cruciferae have their centres of aggregation well marked
in the Eastern Mediterranean and Central Asia. But not only do
they show there the maximum of species in general, but also the
maximum number of monotypic genera with one species only.
These are usually set down as relics, and why should relics be
most numerous at headquarters? In their anxiety to prove the
validity of natural selection people have worked upon more or
less independent lines, which often clash badly with one another.
Workers with floras of islands or of mountain chains have urged
the conception of endemic species and monotypic genera as relics,
regardless of the fact that other workers have shown that these
relics are most abundant at the "headquarters" of the family,
and are regarded as showing the great suitability of the family to
that particular region.

A very difficult problem for supporters of the idea that condi-
tions and their variety have anything to do with the contours is
provided by their behaviour in New Zealand. The northern inva-
sion of plants shows contours beginning in the north, with the
last species of the genus somewhere in the south. The southern
invasion begins in the south, and its contours fade away to the
north, but each invasion passes over the centre of the other
(where the conditions are supposed to be so varied) without
taking the least notice of it. There cannot be conditions that only
affect northern plants, or only southern, as the case may be.

TEST CASE XXVIII. TAXONOMIC RESEMBLANCES OF
(GEOGRAPHICALLY) WIDELY SEPARATED PLANTS

This case has already been published, and the following descrip-
tion is largely quoted, by kind permission of the Linnean
Society, from a paper in their Proceedings of 21 April 1938:

Whilst when first published both these conceptions — Differen-
tiation and Young Beginners^ — were much opposed to current
beliefs, there is no doubt that the latter, at any rate, is gaining

1 I.e. the conception that the bulk of the very local endemic species,
especially in warmer countries, are young species starting life.

CH. xiii] D. GEOGRAPHICAL DISTRIBUTION 155

ground, as may easily be seen by looking at various recent
publications in systematic botany, where a great part of the
endemic species are now admitted to be new. The first conception
is also beginning to receive support. Systematists have in recent
years made important additions to the evidence for mutational
origin of species and genera, though themselves only trying to
place these species and genera nearest to those which appear
to be most closely related to them.

If one take as illustrations some of the more recent mono-
graphs in Engler's "Pflanzenreich", one notices at once the great
geographical separations of closely allied species, genera, sub-
families, families. In Cardamine, for example, species no. 70
is in New Zealand and Polynesia, no. 71 in the Azores, no. 72
in Chile. In Euphorbia one finds allied species in Venezuela and
Cape Colony, in Persia and in Africa, in central Asia and in N.
America, and so on. If in the Drabeae (of Cruciferae) one join
the consecutive related genera by a line, one crosses the Atlantic
five times and the Pacific once, and usually goes well into the
continent also. In the Arabideae the crossings are seven and
six respectively, and in the Lepideae the whole map is covered
with a web of lines.

Now with relationship like this, which is so much complicated
by the great separations over the surface of the globe, to get
an explanation by the method of accumulation of small differences
is an extraordinarily difficult matter, and it is much simpler to
call in the Unking genera that cover the enormous gaps than to
suppose that the related genera, say in Chile and Siberia for
example, once overlapped or nearly overlapped each other, and
that then destruction took place upon an enormous scale, and
through all varieties of conditions and climates. All three of
these subfamilies have various genera that cover the whole or
much of the range, and it is much simpler to regard these as
connecting links — as in fact the ancestors, directly or at times
indirectly through intermediate genera, of the small scattered
(though so often closely related) genera. One may take any view
one pleases as to how they were derived from these large and
widely distributed linking genera, though personally I hold to
the views expressed in 1907, when pointing out how all the
existing Dilleniaceae might have been derived, directly or in-
directly, from Tetracera, the most widespread and about the
simplest of the family. The only necessary thing is to get rid of
the idea that small genera and species of restricted area are
necessarily relics, and we have seen that this conception is now
definitely losing ground.

If one suppose a genus to give off new species more or less
in proportion to the area that it covers^ (which again will be
more or less in proportion to its age among its peers), it is clear

^ For the mathematical consideration of the question, cf. Yule in Phil.
Trans. B, 213, 1924, p. 21.

156 TEST CASES [ch. xiii

that all the offspring will carry a large proportion of the cha-
racters of the parent, and that therefore while offspring arising
near together will be most likely closely to resemble one another,
there is no reason why a close resemblance should not arise
with a wide geographical separation.

It is rare to find a genus going far outside the limits of the
genus that may be looked upon as the linking genus (e.g. Draba
in Drabeae). When it does, one may imagine that in its " make-up "
there was included a greater suitability to conditions that may
be a barrier to the parent — it may be capable of growing in
warmer (or colder), wetter (or drier), or otherwise different
localities.

The author is not attempting to set up this " parent and child "
theory as a universal rule, nor at present attempting to apply
it to zoology; but there is no doubt that it will apply very well
to most of the small families of plants, to a great number of the
larger families, to a great number of the subdivisions of families,
and to a great number of genera whose species behave as do those
genera that we have been dealing with.

The theory of accumulation of small differences makes many
of these and other phenomena very difficult to understand. To
get two closely related genera or species so widely separated
geographically by aid of the selection of small differences would
be very difficult, for one would have to assume — if the differences
be regarded as adaptational — that the conditions in the two
places were very similar, though there is little evidence to that
effect, or that the genera once touched one another in their
distribution, and that there has been a vast amount of destruc-
tion. Not only so, but this destruction must have gone on through
every variety of conditions to which the genera must have been
adapted. There is some change and variety to be passed through
between Greece and California, for example, or between Persia
and Cape Colony, to take a couple of examples from the Lepideae.

It is probable that cytological study will throw some light
upon this difficult problem and it is clear that what has been
said here is fully in favour of the theory of differentiation,
affording no support to that of natural selection.

TEST CASE XXIX. VARIETY OF CHARACTER
WITH UNIFORM CONDITIONS

We have seen (p. 18) that the Podostemaceae and Tristichaceae,
growing in the most uniform conditions that it is possible to
imagine, yet show a very great variety of character and of
structure. And not only so, but the characters are at times very
definitely divergent, such things showing as bilocular and unilo-
cular ovary, one stamen or two, many seeds or two to four, and

CH. xiii] D. GEOGRAPHICAL DISTRIBUTION 157

so on. There are about forty well-separated genera, with well-
marked characters of flower and fruit, as well as strongly marked
structural characters of the vegetative body. It is impossible to
suppose that such structural characters as a bi- or uni-locular
ovary can matter in the struggle for existence to a family like
this, whose life is passed under water, the flowers only appearing
a few days before their final death. The flowers in Asia, and to a
considerable extent elsewhere, are fertilised by wind, so that
their structural features are even less important to them than
usual, though they mostly show the extreme of zygomorphism
and stand rigidly vertical. The fruits produce a vast mass of seed,
among which perhaps one in ten thousand may produce a new
plant. The seeds have no adaptation for clinging to the rock, so
that the survivors must be determined by chance.

A great many other families also show great variety in form
though living in conditions that are comparatively uniform.
Larger families are in general found to be living in a greater
variety of conditions than small, but there are no general rules.
But to co-ordinate the number and variety of the genera and
species with the variety of the conditions was always an in-
soluble problem until it was shown that mere age had a great
deal to do with it. With few exceptions, the older a family was,
the greater the variety of conditions that it occupied, but there
was no arithmetical relation between the two.

We may take a few examples of families and genera that show
a considerable variety in themselves, without occupying a cor-
responding variety of conditions. Pandanus, which is found
almost entirely in the uniform conditions of seashores or marshes,
has 180 species. The Naiadaceae (1 genus with 35 species) and the
Aponogetonaceae (1/25) are water-plants of very uniform condi-
tions. The Cyperaceae, mostly in swamps or in sandy places, both
of which must be very uniform, show 85/2600. The Bromeliaceae,
epiphytic or on rocks, and therefore in very uniform conditions,
are 65/850, the Juncaceae, in damp and cold places, 8/300. The
orchids, largely epiphytic, where the conditions must be very
uniform, are 450/7500. The Salicaceae, mostly mesophytic trees,
are 2/180, the Loranthaceae, woody semi-parasites, are 30/520.
The Balanophoraceae, internal parasites, whose conditions must
be very uniform, show 15/40, and Orobanche, a semi-parasite, has
90 species. The halophytic Chenopodiaceae have 75/500, the
xerophytic Aizoaceae 20/650, the water-inhabiting Nymphaea-
ceae 8/50, the insectivorous marsh-loving Drosera has 90 species.

158 TEST CASES [ch. xiii

Nepenthes has 60, all living in very much the same conditions.
Begonia, mostly in the undergrowth of damp forests, has 750, the
xerophytic Crassulaceae 25/1500. Impatiens, mostly in the moun-
tain flora of India and Ceylon, has 350 species. There are six
species of Sonneratia, all mangroves, whose conditions of life
must be the saine. And so on. From these one can work down-
wards through smaller and smaller families, showing less and less
variety, down to single species like Hipjyuris vulgaris of family
rank. The smaller the family, on the average, the smaller is the
area that it occupies (size and space, p. 113 of Age and Area).

Perhaps the most striking example of a great number of species
all occupying practically identical conditions is the existence of
the great group of the Fungi, more especially those that are
internal parasites, where the conditions must be exactly the
same, except for the chemical differences in the sap of one host
and of another, differences which must be discontinuous, as the
different chemical substances that occur are discontinuous. The
eight genera Clavaria, Fonies, Marasmius, Miicor, Penicillium,
Peronospora, Puccinia and Saccharomyces, all living in extremely
uniform conditions, have 2500 species among them.

As in related forms the number of species goes up with the age
and distribution of the genus or family, it is much simpler to
look upon it as going simply with the age — the larger genus or
family, with the larger distribution, is the older. If the conditions
also become more varied with increasing age of the family, as they
almost always do, this probably helps to increase the number of
species by the stimulus that it gives. There is nothing to be
extracted from the figures that will go to show that natural
selection, or variety of conditions, is responsible for the numbers
of forms that exist. Probably as time goes on, and at any rate if
there is any stimulus, evolution has to go on.

TEST CASE XXX. A COMMON TYPE OF DISTRIBUTION

IN INDIA AND ELSEWHERE

A proposition very difficult of explanation is put before the sup-
porters of natural selection by what is a very common type of
distribution, long ago pointed out by Dr Guppy in the islands of
Polynesia, and by the writer in India, Ceylon, and elsewhere.
This is the polymorphous widely ranging species, accompanied
by few or many species confined each to one part only of its
range, and endemic to the regions that they occupy. Guppy noted
three stages in the development of local endemism. First, the

CH. xiii] D. GEOGRAPHICAL DISTRIBUTION 159

island was occupied, so far as a given genus was concerned, by
one or more widely ranging species, usually very variable, such for
example as Metrosideros j^olymorpha. Then the wide-ranger was
accompanied by one or more local endemics, allied to it, and
finallv there were onlv the endemics. He thousrht that the wide-
ranger had given rise to the endemics, and might, or even did,
ultimately disappear (swamped, cf. 66, p. 95) (cf. 74, pp. 611-
13).

It is the general experience of systematists that it is only in
numerous and widely ranging forms that this variability occurs
(cf. p. 132 for axioms). Linnaeus (12th ed., ii, 324) gives a list of
thirty such polymorphous genera, including willow and saxifrage
in Europe, oak and Aster in North America, Cactus in South
America, heather and everlastings at the Cape.

Another way to bring out this point is to look at the synonyms
in generic indices like the Index Kewensis. The first forty-five
generic synonyms at the beginning of C are referred to genera of
an average size of 94, the mean for all genera being 14-15. The
first seventy in de Dalla Torre's Indea; are merged in genera with
an average of 70, or in both cases definitely large genera.

Let us begin with the Indian Anemones, which show

A. rivularis All mountains of India and Ceylon

rupicola Kashmir to Sikkim

vitifolia Himalaya, Mishmi Hills

Griffithii Sikkim, B ho tan, Mishmi

Falconeri W. temperate Himalaya

ohtusiloha Temp, and alpine, Kashmir to Sikkim

rupestris Alpine, Kashmir to Sikkim

trulUfolia Sikkim to Bhotan

demissa Alpine, Sikkim

polyanthes Kashmir to Sikkim

tetrasepala Western Himalaya

elongata Garhwal, Nepal, Khasias

Or take Clematis, § Cheiropsis; C. montana is common all along
the Himalaya, while C. napaulensis, C. barbellafa and C. acut-
angula are confined to particular sections. These two genera are
simply the first that occur in the flora, and almost any Himalayan
genus will show the same thing, whilst it is also shown by the
genera of lower levels, e.g. Portulaca:

P. oleracea All India and Ceylon, and all warm countries

quadrifida All India and Ceylon, and palaeotropical

WighUafia Carnatic to Ceylon. Endemic

tuherosa Behar to Ceylon. Endemic

siijfruticosa W. Peninsula, Ceylon. Endemic

160 TEST CASES [ch. xiii

Other genera, e.g. Amoora, Celastrus, Hippocratea, Leea, Limacia,
OIcLV, Salacia, Tinospora, Zizyphus, from the first volume of
Hooker's Flora, show the same thing. Clarke (4) says that in
the Himalaya closely allied species of Didymocarpus are confined
to single districts, though there appears no reason in soil or
climate why they should not spread to adjoining valleys.

Now to explain such phenomena as these by aid of natural
selection is very difficult. The range of the wide-ranging Anemone,
for example, is put down to its "adaptation", though to what
exactly it is adapted is not explained. If it suit (as it does, place
by place) the very varied range of conditions in which it is found
that must be due to functional or physiological adaptation as it
moved from one region to another. We have no evidence that a
seed from say Ceylon would at once suit a station in the north-
west Himalaya, without first acquiring the necessary local adap-
tation which it would have received as a matter of course had it
been slowly transported from place to place by nature's method.
But it would then, in all probability, cease to be fully suited to the
Ceylon habitat. But why should it be accompanied by eleven
local species? All these are endemic to their own regions. In
their anxiety to disprove my contention that such local endemics
are young species as compared with the wide-rangers, my
opponents have gradually pinned their faith to relicdom. But
why should A. rivularis leave eleven defeated relics in its range
of distribution? It looks as if selection had been very strenuous,
and was greatly diminishing the number of species, not increasing
it (p. 90). There is absolutely nothing to prove that any of them
are relics, and no feature in A. rivularis that gives even a faint
suggestion that it may be adaptationally superior. It covers all
the mountains of India and Cevlon, and whv are there no local
relics in any of the southern mountains? None occur south of the
Khasias. But there can be little doubt that Anemone advanced
from north to south in the Indian region, reaching Ceylon last of
all, so that it would be younger in the south than in the north.
By the theory of age and area, its peculiarities are at once ex-
plained. A. rivularis arrived first somewhere in the Himalaya,
where only the local endemics are to be found, and it has not been
long enough in the southern mountains to mutate off new en-
demics there. The relic explanation is altogether too fanciful to be
accepted, as is also that of local adaptation, which also will not
explain the crowding together of the endemics in the north.
Nothing hitherto proposed with the exception of age and area

CH. xiii] D. GEOGRAPHICAL DISTRIBUTION 161

is capable of explaining problems like these, which occur in
hundreds and all closely similar. Natural selection is completely
incompetent to do so. What Anemone shows in the Indian region
is a contour map, which we have already shown on p. 149 to be
completely favourable to differentiation and to age and area.

TEST CASE XXXI. LARGE GENERA THE
MOST "SUCCESSFUL"

One cannot accept the large genera as the most successful in the
light-hearted way in which this has been done under natural
selection. They are not usually composed of numbers of widely
distributed species — their successes are limited to comparatively
very few. We have seen above how few of the numerous Siparunas
or MolUnedias are widely distributed, and yet these are by far
the largest genera in their family. And the same phenomenon is
almost universal. If we take as another example the Styraceae,
of which the monograph is lying beside me, we find a family of
six genera, four with three species each, one with two, and Styrax
itself with 100. Here surely is a family with one conspicuously
successful genus. But when we look at the whole distribution,
there are only four widely dispersed species in the whole family,
one in Pterostyrax and three in Styrax. The distinction between
these genera is mainly that one has a superior, the other an in-
ferior, ovary. Upon the hypothesis of natural selection, therefore,
the family consists of about four successful species and 110 relics.
And not only so, but the Styrax that is by far the most widely
dispersed has a very discontinuous distribution (W^. As., Eur.;
W.N.Am.), a thing that does not occur with the small genera,
usually looked upon as relics. It is much simpler to regard the
widely distributed species as older, the local as younger, as differ-
entiation requires.

TEST CASE XXXIL CHARACTERS MORE CONSTANT

THE MORE USEFUL

This is sometimes advanced as a corollary of the theory of
natural selection, and indeed it seems almost necessarilv to
follow. How much substance there is in the argument, however,
may be judged from the fact that the most constant characters
in plants are notoriously those that are the most important in the
classification (for the obvious reason that they are the most
constant). But the higher one goes in the classificatory characters,
from those of species to those of families, the more constant do

WED II

162 TEST CASES [ch. xiii

the characters become, and the less functional value do they
have, as is universally admitted. This test is entirely against
natural selection, even if it do not specially favour differentiation.

TEST CASE XXXIII. RELATION OF MONOCOTYLEDONS

TO DICOTYLEDONS

A feature in geographical distribution to which Hooker called
attention in 1888, and Avhich was never explained until Age and
Area gave the key to it, is described in the following quotation
(18): "The conditions which have resulted in Monocotyledons
retaining their numerical position of one to four or thereabouts of
Dicotyledons in the globe and in all large areas thereof are, in the
present state of science, inscrutable." The exactness of the rela-
tion is remarkable. The latest figures in the writer's possession
add up to 36,639 species of Monocotyledons and 145,718 of
Dicotyledons, or almost exactly one to four.

So long as one keep to large areas, and to the centre of the land
masses, the relation keeps wonderfully steady, but when one
comes to the edges of vegetation, especially to the north or to the
south, one finds fluctuation beginning, as also in the tropical belt
from Malaya (which has 26 per cent of Monocotyledons) through
Ceylon (27 per cent). While the average proportion is just 20 per
cent, and in the Kermadec Islands north of New Zealand is 21 per
cent, it is 31 per cent in the Chatham Islands to the east of New
Zealand, and 45 per cent in the Aucklands to the south, and again
26 per cent in Juan Fernandez and 30 per cent in Tasmania, all
these figures suggesting that the old southern continent was a
great home of Monocotyledons. In Europe there is a belt of high
proportion of Monocotyledons from Sardinia through France and
Britain to Iceland. In the Canaries the proportion is only 15 per
cent.

Now there is no " monocotyledonous " mode of life to which
this group can have been adapted. Every kind of life is repre-
sented, and there is nothing in common in mode of life between
such things as orchids, grasses, lilies, aloes, bulrushes, water-
soldiers, palms, aroids, duckweeds, rushes, Bromeliads, yams,
bananas, gingers, etc. The steadiness of the proportion of Mono-
cotyledons to Dicotyledons goes to show that in their dispersal
adaptation played but a small part, and that it was primarily
governed by the laws of age and area, as is demanded by the
theory of differentiation.

CH. XIII] D. GEOGRAPHICAL DISTRIBUTION 163

TEST CASE XXXIV. OVERLAP OF LARGEST
GENERA IN A FAMILY

If the differentiation explanation of the origin of a family be the
correct one, the first two genera of a family, the largest upon the
whole, should overlap in their distribution, as one of them sprang
from the other, but there is no reason why this should be so under
natural selection. Geological or other changes may, of course, at
times have rendered this impossible. Upon examination, we find
that in the majority of families this overlap does occur, though
there are a number of families like the Apocynaceae with a large
genus in each of the continents, or in the Old and New Worlds.
Exceptions are frequent among the families of the southern
hemisphere, with their broken areas of distribution, but in the
greater number of families the rule holds. Among the small
genera in a large family, this is rarely the case, but in a small
family it usually occurs. It is not impossible for a grouping like
this to have been produced by natural selection, but there must
have been something upon which it could get a grip, and one can
scarcely ever find anything of this kind.

II-2

CHAPTER XIV

GENERAL DISCUSSION

X HE results to which this work leads being somewhat sub-
versive of current opinions, it wiU be well, perhaps, briefly to
restate parts of the argument in other words. Some of it appeared
eighteen years ago in Age and Area, but the propositions there
put forward were not accepted, though the arguments brought
up against them appeared to the writer to be lacking in logical
force, and he has remained faithful to his published opinions.
Bateson alone among reviewers realised that the discovery of the
"hollow curves" was one of importance, and the only thing that
opponents have been able to bring up against them is that they
are "accidental", just as the curve of names from the telephone
book, or of the names in Canton Vaud (p. 35) is "accidental",
which is exactly what the writer was out to prove.

Until some eighty years ago, the appearance of the vast
numbers of forms of life that people the world, and that are
usually known as the species of animals or plants, was put down
to a somewhat crude intervention of the Supreme Power, which
was supposed to have created all the hundreds of thousands of
them, each species in its existing form, and to have placed each
in a more or less definite region, where it is still commonly to be
found (p. 2). When studied in more detail, however, many
difficulties cropped up, difficulties that became ever more insis-
tent, and that at last resulted in the sweeping away of the old
theory of special creation, then the background of biological
work. One great difficulty, for example, was to explain the
evident likenesses that one may see in the tiger, the leopard, and
the cat, or the daisy and the sunflower, resemblances so great
that they seemed to point to definite relationship, as indeed had
been suspected since the days of Aristotle.

In 1859, with the appearance of the Origin of Species, there
began the long reign of " Darwinism ", lasting to the present time.
Darwin's immortal service to science was to establish the theory
of evolution — that every living species has been derived from
some other by direct descent, accompanied by such modification
that for example the tiger, the leopard, and the cat might all be
derived from a common parent sufficiently far back. Unfortu-
nately the name of Darwinism was popularly given rather to the

CH. XIV] GENERAL DISCUSSION 165

mechanism by which these changes were to be effected, i.e. to the
struggle for existence that was a famiUar everyday experience,
allowing those gifted by nature or by parents, or by chance, to
succeed, while the non-gifted usually failed. As every living
being tends to produce more offspring than there is room for,
some must obviously be picked out, and this selection, or
"survival of the fittest", Darwin called natural selection. Being
so familiar, it had a great psychological appeal, and was soon
taken up in all directions. It was evidently an almost complete
reversal of special creation ; instead of being created, beings were
evolved, and instead of being discontinuous, the process was
continuous.

Picking out only variations that gave some advantage, natural
selection worked by what we may call gradual adaptation
(p. 4), which was an essential feature of the theory. But it is
clear that a small improvement in adaptation would not be
enough to create a new species, which is usually more or less
sterile with its near relatives (a functional difference), and shows
various structural differences as well. It had to be assumed,
therefore, that the process would go on until the line of mutual
sterility had been passed, and the differences had become great
enough to mark it as a new species. It was the structural dif-
ferences that showed that there had been any evolution at all,
and so it had to be assumed also that they were adaptational,
marking the adaptational advantages that had accrued to the
organism. Functional adaptation was ignored, though the mor-
phologists had long insisted that structure had little or nothing
necessarily to do with function.

The freedom of the position of natural selection was really lost
very early in its history, when Darwin had to give way to the
criticism of a well-known professor of engineering, Fleeming
Jenkin, who pointed out that unless a great many individuals
varied in the same direction over the whole of a considerable
area, the improvement would promptly be lost by crossing.
Darwin therefore stipulated for such a beginning, which seems
only likely to happen under the action of some external force, and
which practically excludes the action of biological factors, which
are usually local. Improvement seemed unlikely in the fluctuating
variation upon which Darwin usually relied, for some might go
up when others went down, and crossing would level them. This
criticism took much of the spring out of the action of natural
selection, for instead of remaining a simple affair of individuals,

166 GENERAL DISCUSSION [ch. xiv

as it was in daily life, it was assumed to be a competition of
groups. Whatever may be the case with animals, there seems
little or no reason to imagine that plants compete as groups. It
is this assumption which has become so marked a feature in
social and political life — that the best, and incidentally the most
satisfactory, solution of a difficulty or of a competition lies in
the conquest and dominance, or even in the extermination, of the
opponent. Species, to begin with, are not structural units with
all individuals just alike, any more than are language groups of
mankind. The nearest approach to this condition is in such cases
as Coleus elongatus (p. 24), a well-marked "Linnean" species
where there are so few individuals — perhaps a dozen in this case
— that they do not allow of a great range of variation. There is
also less range in the small "Jordanian" species, but these, on
the theory of differentiation, are later phases in evolution than
are the Linnean species. As one of the latter increases in number,
and in occupied area, from its first beginning, and thus probably
comes into greater variety of conditions, and into more crossing
with other individuals, the more variation does it show, on the
whole (p. 159).

The following quotation shows the point of view that is being
taken up as the result of the work of agricultural geneticists;
*' Studies of crop populations have shown that natural selection
does not result in the survival of the fittest type, but of the fittest
'population, and the fittest population is almost always a mixture
of many types" (78). This agrees with the ordinary observation
of everyday life, that natural selection is individual in its action.

The plants (or group, occupying the whole of the locality) that
did not show the useful improvement (or another as good) were
killed out in the struggle for existence, that also killed out the
parent, which was assumed not to become adapted.

It is clear that there are many weak points in the Darwinian
position, and to support them all kinds of assumptions and sup-
plementary hypotheses have been brought up. But there has
never been any good proof (1) that evolution proceeded essen-
tially by improvement in adaptation, (2) that it was gradual and
closely continuous, (3) that the phenomena of the structure of
plants reflect the adaptation that has gone on in them, or (4) that
groups of plants can compete as units.

When one comes to look into the matter, one soon realises that
the theory of natural selection rests upon a great many assump-
tions, sometimes backed by more or less proof, sometimes not.

CH. xiv] GENERAL DISCUSSION 167

Yet I have been assured by one of its most eminent supporters
that it contains none, but rests upon proved facts. The following
list gives the most important assumptions :

1. That a small structural variation may be advantageous
enough to call in the action of natural selection (pp. 4, 13).

2. That a small advantageous variation may be inherited
(p. 4).

3. That it mav be added to, and become more and more
marked in succeeding generations (pp. 4, 54, 106, 165).

4. That the whole number of individuals upon a considerable
area will show the same advantageous variation, i.e. probably.

5. That the variations are controlled by external conditions
(pp. 5, 165).

6. That the whole number of individuals carrying a useful
variation can, and does, fight as a unit (pp. 107, 142, 144, 166).

7. That the parent form does not also become adapted
(pp. 4, 13, 54).

8. That adaptation is structural rather than functional

(p. 4).

9. That structural characters are the means of expression of
adaptation (p. 14).

10. That differences in structure mean differences in adapta-
tion (pp. 52, 109).

11. That the variety with the advantageous variation, slight
though it would be at first, will defeat the parent in the struggle
for existence (p. 4).

12. That the new form produced by natural selection, and
adapted to area B, became thereby also better adapted to A, the
area occupied by the less well adapted parent species (p. 144).

13. That all variations that survive must be useful, or must be
correlated with variations that are useful or at least that are not
harmful enough to be of serious disadvantage (pp. 57, 58).

14. That the new form will invade the territory of the old, and
kill it out there, without being lost in hybrids (p. 143).

15. That the defeated species w411 gradually become relics, and
ultimately disappear (pp. 4, 17, 91, 97-8).

16. That fluctuating variation is irreversible.

17. That fluctuating variation is qualitative as well as quanti-
tative.

18. That fluctuating, or even small, variations can be added
up so that they pass the sterility line that usually divides one
species from the next.

168 GENERAL DISCUSSION [ch. xiv

19. That the needful variations will appear at all (p. 55).

20. That natural selection can act continuously upon them
(p. 54).

21. That most or all of the individuals that do not show the
favourable variation will be killed out (p. 54).

22. That conditions will continue to vary in the same direc-
tion long enough to enable the sterility line to be passed (pp. 54, 55).

23. That natural selection is so strenuous in its action that the
sterility line will be passed (p. 55).

24. That when a species has become well started upon a
variation in one direction, there will not be offered to it one in
another direction, obviously better (pp. 55, 109).

25. That the adoption of one variation does not interfere with
the adoption of another (pp. 55, 109).

26. That when one variation has done its work, it shall be
followed by another of those that mark the species (p. 55).

27. That morphological and anatomical necessities override
the effects of natural selection (p. 110).

28. That economic botany is of no importance from the point
of view of natural selection (p. 8).

29. That advantageous structural variations are so desirable
that they will commonly be followed up to a result of 100 per cent
(p. 114).

30. That natural selection will produce uniformity in structure
of a morphological feature (pp. 55, 114, 124).

31. That there was some reason why transitions were dropped
out more and more as evolution went up towards families (p. 113).

32. That varieties are incipient species, species incipient genera.

33. That numbers would increase greatly under selection
(p. 90).

This is a very formidable list, and a mere glance will show that
many or even most of the assumptions still remain such, though
by the adoption of mutation in place of gradual variation several
of them have been removed. It is, therefore, clear that the
theory of evolution by the agency of natural selection, picking
out gradual improvements in adaptation, chiefly structural, is
still a very long way from being established, and as no evidence
has been found in seventy-five years to prove many or most of
the assumptions, one may be permitted to feel somewhat sceptical
of its discovery. Evolution is now thoroughly well established,
and whether natural selection carried it on or not is a matter of
indifference to it.

CH. XIV] GENERAL DISCUSSION 169

The theory of natural selection, holding as it did that every-
thing was gradually acquired, went to show that evolution must
be gradual and continuous from one structure to its successor of
different form, and this soon led to difficulty. The facts of
economic botany (pp. 8, 89) among others, though dismissed as
unimportant since they did not favour natural selection, showed
that there was much discontinuity in evolution, and Bateson's
work (1) showed the same thing.

The continuous small fluctuating (infinitesimal) variations upon
which Darwin chiefly relied were not fully hereditary (p. 10);
they were not differentiating, but simply up and down in the
same character, nor were they irreversible; and they could
not be accumulated beyond a certain point (p. 10). They could
all but never be found to show adaptation, whilst the differences
became more and more marked, and less and less adapta-
tional the higher that one went from species to family, this
illustrating the principle that we have termed the divergence of
variation (p. 74). Species, again, usually showed several points
of difference which were unconnected with one another so far as
anyone could see, and it was very hard to see how selection could
deal with so many. Species also proved to be mostly local in the
big or "successful" genera, so that their adaptation must have
been generic, and it was very hard to understand how this could
have been the case. If it were so, natural selection, working
upwards from the species, could hardly explain it. If all specific
characters were correlated, then the greater portion of evolution
did not show the effects of natural selection (p. 11). It was
almost impossible to see how gradual selection could pass the
rough and ready line of distinction between species, the fact that
they are almost always more or less mutually sterile. No transi-
tion stages, again, were to be found among the fossils, though one
would have expected to find such upon a theory that was based
upon the separation of genera and families, to say nothing of
species, by the continual destruction of transitional forms on
account of their inferiority to the more perfect. Nor could one
find among the fossils any indication of the gradual formation of
existing families, etc. These seem to appear already fully de-
veloped, and in widely separated sections of the classification of
flowering plants.

Evolution could only go on if the right variations were to
appear; natural selection would kill out any that were harmful,
and would be indifferent to any that showed neither value nor

170 GENERAL DISCUSSION [ch. xiv

the reverse. Also evolution could only go on provided that
natural selection could act as desired, another assumption. Then
the freedom of action of selection was destroyed by Fleeming
Jenkin's criticism, though the fact was hardly realised. Finally,
selection proved itself incapable of explaining many of the facts
of geographical distribution, a subject which is completely
bound up with evolution.

Immense effort was put into the study of adaptation fifty to
sixty years ago (p. 52), but with little or no result other than to
show that no one in his wildest dreams could attach adaptational
value to the bulk of the structural characters that distinguish
one plant from another, and show that evolution has really gone
on. There was also no doubt that what little adaptation did show
decreased rapidly as one went up the scale above the rank of
genus ; but the higher divisions were supposed to be made by the
killing out of transitions, which would imply that selection came
more and more into play to make larger and larger divisions. The
facts, when judged in the light of the theory of natural selection,
are evidently somewhat incompatible.

Early in this century de Vries brought in the theory of muta-
tion or sudden change, which in many respects got over the worst
difficulties of Darwinism, and would have surmounted more had
not people taken up a somewhat illogical attitude with regard to it.
It was admitted that small mutations could take place, but people
were averse to admitting large ones, for that would probably
remove any effect of natural selection in guiding evolution. It
would be almost absurd to suppose that it showed its work by the
production of large and sudden differences, though it is not im-
possible, for one may imagine it perhaps selecting slight genie
changes, and these being added up till the strain in the nucleus
produced some kind of kaleidoscopic effect by a readjustment.
The writer suggested in 1907 that "a group of allied species
represents so many more or less stable positions of equilibrium in
cell division" (70). But though by the adoption of small muta-
tions the power as a determinant of evolution was taken away
from natural selection, so that it could no longer start the im-
proved adaptations, it was expected to carry them on and to
increase them, gradually or by further mutation. Of course it
would only carry on those which had, so to speak, passed through
its sieve, and had proved to be of definite value. We were still,
however, without any indication that the characters produced in
the small mutation had any adaptational value, so that their
survival would usually be due simply to the fact that natural

CH. xiv] GENERAL DISCUSSION 171

selection was indifferent to them. But if this were so, it could not
carry them further. This being so, why not go the whole course at
one effort, and admit that selection had little, or perhaps even
nothing, to do with the evolution of the organisms that now exist
in the world, however much it may have improved them after
their evolution, or fitted them to the local conditions in which
they were trying to live? Everything, before it can become
established, must pass through the sieve of natural selection, and
each new individual, in any place, must do the same, but the
characters of that new species or individual were not selected by it.
If selection does not begin a species or an individual, it has no
responsibility for its arrival, but it will kill it out if it be un-
suitable to the conditions of the place in which it appears, at the
time at which it appears. Why then should natural selection be
needed at all for structural change, if it does not begin it, and
when one can generally find no adaptational value in it?

This is very much the position that the author took up in 1907,
basing his change of view more or less upon this line of reasoning,
and admitting that no mutation that may be needed for the
purpose in view — the formation of species, genera, or families —
is too large for possibility. There is little or no evidence that
structural differences in root, stem, leaf, flower, fruit, seed, have
any adaptational value, yet it is these things that make up the
characteristic differences that separate one species, genus, or
family from another (cf. Thalictrum, etc., on p. 104, and lists in
Appendices). There is no evidence that climbing plants (p. 57)
have gained by the fact that they can climb. The same genus
sometimes contains both climbers and non-climbers, and the
former must have erect plants upon which to climb, with few
exceptions. Supposing that they smothered all the erect plants
by their success, as they might very easily do if really "suc-
cessful", both they and the latter would be in a bad way, yet
there is nothing to prevent it.

The view that mutations are necessarily small rests to some
extent upon the opinion that a Linnean species is composed of a
great assemblage of micro-species which breed true. But it can
only be so if it consist of a great number of individuals and
occupy a large area. Upon the theory of age and area, as well as
upon that of differentiation, this means that it is older than the
small and local (allied) species, which is so often Linnean in the
sense of marked difference, but cannot show great variety through
lack of numbers (p. 132).

We imagine, then (under the theory of differentiation or diver-

172 GENERAL DISCUSSION [ch. xiv

gent mutation), that families, genera, and species may any of
them be the result of a single 7nutation, more divergent in genera
than in species, in families than in genera. These ideas receive
great confirmation in the actual structural differences that
separate plants. As one goes up the scale from species to family,
the divergence of the characters of separation increases upon the
whole, as is at once shown by any good dichotomous key. A
feature of special interest is that the divergences become more
and more such as allow of no intermediates or transitions at all,
as for example, a berry and a drupe, an achene and a follicle, an
anther with slits and one with pores. But if this be the case, the
character, one or the other of a divergent pair, must have
appeared at one step, so that, so far as one can see, natural
selection can have had no hand in its appearance. The higher
that one goes in the direction of the family, the less adaptational
value can one find in the characters, so that the less is the handle
that is offered to natural selection. Competition is greatest
among individuals, less among species, still less among genera,
and so on upwards. Yet the distinctions become greater along the
same route, and the puzzling question is put as to how the
diminished competition can bring about the larger and more
permanent distinctions. Why also are these characters of so
slight (if any) functional use? If natural selection be the active
agent in evolution, it must have been working at its highest
pressure among the highest groups to separate them as they are
separated, and also must have been working all the time to pick
out characters with greater adaptational value; whereas in fact
one finds the characters to be of less and less value as one goes
upwards in the supposed track of natural selection. In seventy-
five years no one has been able to prove any functional value for
them. The uniformity of the statistics of the various continents
and other large areas (66, p. 180) in the proportions of genera
of various sizes, in their distribution, in the relative sizes of
families and genera, etc., shows that one area is just like another,
and that evolution must be going on in the same orderly way
in all.

Though going to Ceylon an enthusiastic supporter of natural
selection, the author found it needful to change his views after
some years of tropical experience, both in the forest, and as the
result of a minute study of the Podostemaceae (p. 18). Though at
first glance looking as if they showed great adaptational dif-
ferences, these plants all live under the most amazingly uniform

CH. XIV] GENERAL DISCUSSION 173

conditions, with nothing special to which to be specifically
adapted. The life in moving water, and the loss of proper polarity
of the plant (p. 20), are common to all. Yet there are about forty
genera with many species. The most reasonable explanation is
that evolution must go on, with or without adaptational reason,
and is not necessarily a matter of local adaptation. All are com-
pelled to "adapt" themselves more or less to the action of the
permanent force that acts upon them, which cannot be escaped in
any way.

The author's work with endemic plants, which has occupied
many years, also showed that the old (and still more or less
current) view, that they are relics of previous vegetation, had no
sound basis. Under the Darwinian theory, there had to be found
somewhere some at least of the species that had been defeated in
the struggle for existence and were dying out. The harmless
endemics just came in conveniently to fill this position, but while
there are undoubtedly many relics in regions that were cooled in
the glacial periods, one cannot suppose this of most of the local
species of warmer climates.

As one result of this work, the writer discovered the "hollow
curve" of distribution (chap, iv) that shows in plants and
animals, and in many other cases, such as surnames, and even in
inanimate objects (cf. p. 35). It shows well in areal distribution,
many species of any genus occurring on small areas, few on large.
It shows still better in the distribution of the genera in a family
by the numbers of species that they contain. On the average,
from which there is but small variation, with reasonable numbers,
about 38 per cent of genera have only one species, 12 per cent
two, and 7 per cent three. The curve turns the corner between
three and five, and tapers away in a long tail, and the larger the
family the more accurately does it follow the curve. When plotted,
as 38 + 7 make more than twice twelve, the curve has the dip in
the middle that gives it its name (fig. on p. 36). When plotted in
logarithms, the curves form close approximations to straight
lines, showing that all have the same mathematical form, and
have appeared as the result of the same law, working upon all.
The mathematical treatment of the subject will be found in Yule's
paper (75), the introduction to which should be read by all in-
terested in evolution. The general law, as he showed, that imme-
diately governed it was that at the end of certain intervals,
probably very variable in length, a genus became two, and both,
as a rule, survived. The parent genus of the two was not neces-

174 GENERAL DISCUSSION [ch. xiv

sarily killed out, as was the rule under natural selection. In fact,
as the curves must be the result of uniform pressure, they could
not result from, or under, natural selection.

The logarithmic curve undoubtedly shows some marked devia-
tions from the straight line at the further end (cf. p. 37), and
Longley (25) says "the hollow curve, we may therefore reasonably
assume, results from some sort of compounding of a series of
geometric series of different common ratio, but all lying between
the limits of J and 1 ".

But if genera and species are formed like this, it must almost
certainly have been by single steps, and if the old ones were not
killed out, natural selection can have had little or no influence in
the matter. They cannot have been formed by structural adap-
tation. Gradual mutational change is no more satisfactory than
the gradual changes that were supposed to have been effected by
natural selection, for in the vast majority of cases where such
small changes have been seen, there has been no possibility of
imputing to them any functional value whatever. On the Dar-
winian theory, where the parent is killed out, it is very hard
indeed to see how there can have been any increase in numbers
(Test case i, p. 90).

One great advantage of the large mutations for the formation
of species, and still more of genera and families, that are de-
manded by the theory of differentiation, is that at one step they
will cross the "sterility line", the rough and ready distinction
that separates a species from its nearest relative. The new form
will at once become isolated (chap, vii), and there will be little
likelihood of its being lost by crossing. As we have seen, isolation
may be of very great importance in the establishment of new
species, if not also in their evolution.

When one considers the fact (pp. 132-4) that the more
primitive things are more widely distributed, that a genus (unless
monotypic) occupies a wider area than at any rate all but one of
its species, and again that there is no evidence to show that there
is any adaptational reason why a small variety should become a
larger one, or the latter a species, a species lead to a genus, and
so on, it would seem, as it seemed to (the late) Dr Guppy and to
the writer over thirty years ago, that we have been to a large
extent trying to make evolution work backwards. It was in-
finitely simpler to work forwards throughout evolution, beginning
always with the family, deriving the genus from that, the species
from the genus, and the variety from the species. In fact, with

CH. XIV] GENERAL DISCUSSION 175

the absence of adaptational reasons for progress, and with the
frequent impossibiUty of transitions (especially in the characters
of the higher groups) it seemed to be almost the only way.

My friend Dr H. B. Guppy was perhaps the first to call proper
attention to the fact that the Darwinian theory was trying to
work evolution backwards. He says (12): "It follows from the
foregoing remarks that no plant groups, in the sense of the great
orders, could have been produced on the evolutionary lines im-
plied in the Darwinian theory" (i.e. beginning with small
varieties and going through species to genera and families), and
continues "to lay down, as the Darwinian evolutionist does, that
the order of development begins with the variety . . . species . . .
genera. . .families, is to reverse the method followed in nature,
since it implies that the simpler, least mutable, and less adaptive
characters that distinguish the great families are the last de-
veloped. This could never have been. Nature has ever worked
from the simple to the complex, from the general to the particular.
Had she followed the lines laid down by the Darwinian school
of evolutionists, there would be no systematic botany. All would
be confusion. There would be no distribution in the sense in which
the term is generally understood, and the plant world would be
a world of oddities and monstrosities."

It is upon such propositions and facts that the pre-Darwinian
theory of differentiation or divergent mutation is now founded.
Natural selection is no longer to be regarded as the mechanism of
evolution; it does not choose what shall be evolved, but it decides
in each case, individually , what shall be allowed to live. Probably
the bulk of the structural characters make little or no difference
one way or the other, and so are indifferent to natural selection.
Evolution ceases to be a mere matter of chance, and comes into
that scheme of things of which Jeans has said that all the
pictures which science draws of it are mathematical pictures.
What causes it to go on we have yet to discover, but we can make
one important step by finding out in which direction evolution
moved, for that involved in the theory of differentiation is the
exact opposite of that involved by natural selection. One goes
from the family downwards, the other from the variety up, and
as there is as yet no evidence to show that it moved in one
particular direction, we are free to take that for which there is
the better evidence.

After fifty years of work, the author has come to the conclusion
that evolution and natural selection work at right angles to one

176 GENERAL DISCUSSION [ch. xiv

another, with but slight mutual interference, the latter being
quite possibly greater in animals. The evolution provides the
structurally different forms of life, while natural selection works
upon the functional side, and adapts them in detail for their
places in the local biological economy. There is no obvious reason
why selection should not develop small structural variations,
though one will not expect specific changes, unless rarely. In
general, selection will simply kill out those individuals, whether
new species or not, that commence anywhere with functional
characters that are unsuited to the conditions of the moment, or
that simply have ill-luck. Each new species, by mere heredity,
will probably have functional characters more or less closely
suited to the place in which it arises, but as time goes on, and the
number of species increases, chiefly by arrivals from elsewhere,
more and more careful adjustment will be needed to fit in each
newcomer. It is in this work that natural selection is of the first
importance, doing work that nothing else could do with the same
efficiency.

Whether evolution must go on in all circumstances, we do not
know, for there is evidence like that of the widespread Hippuris
that seems to show that it is not perhaps absolutely necessary.
The evidence of the Podostemaceae seems to show that it may go
on without change of conditions, though perhaps only under the
action of a permanent force. If a plant suddenly arise with a
suitability to any particular mode of life, like a climber or a
parasite, natural selection will not kill it out, and it may go on
living, and perhaps do very well.

As things show more and more definite adaptation to some
peculiarity of the conditions, they come up sooner and sooner in
their distribution against actual barriers to further spread, so
that they tend to occupy lesser areas than older and less adapted
species, perhaps closely related to them. As we have seen, a
species may become adapted to many regions, one by one (p. 145)
as it travels through them, but it need not show this adaptation
in external characters, nor have we any reason to suppose that
when it has become suited to B it remains necessarily suited to A.
It is possible that this functional adaptation, with or without
isolation, may result in genie changes that may be added up until
they cause a structural mutation. Admittedly we have not yet
solved the whole problem of adaptation as one may see it in so
many characters, but there is no evidence for the gradual adapta-
tion in structural characters that is demanded by natural selection.

CH. XIV] GENERAL DISCUSSION 177

Longley (26) thinks that it will be found that adaptation comes
automatically.

Though natural selection comes more and more into play as
the species in a given region become more numerous, or as a
human society becomes more complex, its action is always
primarily individual. There is little or no competition between
entire species, varieties, or races (cf. pp. 107, 142, 144).

The conclusion to be drawn, therefore, is that natural selection
has not been the driving force under whose influence evolution
has been carried on, though it has been the selecting force by
whose action the individuals best suited to the conditions of any
time and place have been continually picked out. In this way a
continual adaptation has gone on, and except in casual and im-
permanent cases, has ensured that the plants that occur under
natural conditions are very closely indeed adapted to those con-
ditions. This adaptation is not structural, but functional, as is
illustrated, to take one example only, by the structural resem-
blance of the members of a large family growing in a great variety
of conditions, and the great structural differences of a large
ecological "association " of plants growing in very uniform condi-
tions (p. 53). The work of Hutchinson and other agricultural
geneticists shows that natural selection picks out a mixture of the
most suitable individuals, not a type, as indeed may be seen every
day in ordinary life by any observant person, and as is shown by
the composition of any of the larger nations at the present time.

There are a great many difficulties which to a logical mind are
fatal to the supposition that natural selection was responsible for
the great evolution of living forms that has gone on. Take for
example the facts of economic botany, always dismissed as
unimportant since they do not agree with the theory of natural
selection. The definite similarities and relationships that exist
among the various products belonging to the same family show
that whatever was responsible for the production of the family
must also be responsible for the economic products, while at the
same time the discontinuity in structure of the latter, and the
impossibility of gradual transitions between them, shows that
their evolution must have been by large mutations. The difference
between the distribution of a family and that of an association as
given in the last paragraph is another very strong argument in
the same direction.

In the second place, natural selection would make the whole
great process of evolution, including man, the result of chance

WED 12

178 GENERAL DISCUSSION [ch. xiv

selection of favourable variations, whereas the recent progress of
the physical sciences goes to show that in their case the whole
evolution is proceeding upon a well-marked "mathematical"
programme. The theory that is beginning to be indicated in the
work that has been described above, goes to show that evolution
also, one of the greatest recent facts of the physical universe, has
proceeded upon a course underlying which there is some physical
law, probably electrical, which also can be expressed in mathe-
matical terms. This has already been shown to be the case with
the law of age and area, which is evidently only a corollary of
the larger law thus indicated.

To go on to some of the minor objections to natural selection,
of which there are a great number, it is impossible to explain by
its aid the characters that divide species, and the difficulty be-
comes greater and greater as we go up the scale through genus to
family and beyond, while at the same time the distinctions
become also greater and greater, and any functional value to be
attached to them becomes less and less, whilst possible transi-
tions become rarer and rarer.

It is almost impossible to explain the perfection in which the
characters show themselves, a clean-cut perfection which again
becomes more and more marked the higher we go (p. 76). If
natural selection cannot perfect either of such divergent charac-
ters as opposite leaves and alternate (showing a definite phyllo-
taxy), their perfection must be due to heredity, or to direct
mutation, for there cannot be a gradual passage from one to the
other. In the latter case natural selection is excluded, while in
the former one has to remember that the way in which the
ancestor obtained the perfect character must be explained by
natural selection. There is the further difficulty that so often the
two characters occur side by side in species of the same genus.
As a special case we may take the family Rubiaceae (p. 118).
Members of the family can be found showing alternate leaves (as
against opposite, the "family" character), pinnate (entire),
intrapetiolar stipules (interpetiolar), male and female flowers (^),
zygomorphic flowers (regular), solitary axillary flowers (cymes
or heads), 8-merous flowers (5-4), convolute calyx (open),
descending aestivation of corolla (convolute or valvate), anthers
by pores (slits), ovary 10-locular (2), endosperm none (present),
ruminate (not), whilst the whole family shows an amazing variety
in the fruit. All the characters that distinguish the family are
found at times to be replaced by something quite different,

CH. XIV] GENERAL DISCUSSION 179

whilst at the same time no transitions are possible, a fact which
would indicate that all the characters were due to direct muta-
tion. How, then, was the family Rubiaceae, whose actual general
characters are those shown in the brackets, evolved by aid of
natural selection? If all these vagaries are to be explained on the
supposition that morphological necessities override selection,
there is nothing at all of structural nature left for selection to act
upon. The selectionist is content, and seems to think that his
case, that evolution is due to natural selection, is proved, if he
can explain a single, and probably very minor character, upon
that supposition. He forgets that it also has to explain the corre-
lated characters of a whole family or other systematic group, to
say nothing of the great differences that characterise the great
divisions of the vegetable kingdom like ferns, mosses, and liver-
worts, as well as the flowering plants. One cannot employ one
machinery to explain one feature or one portion of the vegetable
kingdom, another for another.

There is no evidence to show that natural selection is collective
in its action rather than individual. It is obviously the latter in
daily life, and the work quoted on p. 166 shows that it is probably
the same among plants. It seems, therefore, that once the idea
that adaptation — ultimately reducible to the chance appearance
of favourable variations — is mainly responsible for the distribu-
tion of plants and animals has given way to a more scientific
conception, the study of plant distribution and of its dynamics
will become associated to that of human populations, each giving
valuable aid and assistance to the other.

Plants seem to behave like a mixed and more or less casual
population expanding in a country where there are barriers of
many kinds to interfere with the regularity, and where the dis-
tribution is determined in detail by natural selection, working
upon the individuals. The same kind of thing has marked the
distribution of races in Europe, etc. We have seen that it seems
to pick out a mixture, not a type, and we may add to this the
curious fact that has lately been exciting some interest, that one
kind of cotton may do best when mixed with another (79).
This fact may have important bearings upon racial intermingling.
One thing at least seems fairly certain, that a whole group A will
not conquer and destroy a whole group B, but that the result
will be an intermingling of the individuals of both that are best
suited to the conditions at the time and place.

Nothing but sudden mutation, usually large, will explain why

12-2

180 GENERAL DISCUSSION [ch. xiv

the same character, and that in perfection, should so often be
found at widely separated places in the same family, and not
only so, but also at numerous places in other families and in
other classes. There is no morphological difference in a berry,
whether it be found in the Dicotyledons or in the Monocotyledons.

Only the conception, which is so largely borne out by the facts,
that mutations on the whole were larger the further back into
the past that one goes, from species through genus to family and
class, can easily explain the remarkable fact that this is definitely
the case, as both differences, and impossibility of transitions,
increase together. Neither in life nor in the fossils do we find any
evidence of serious transitional stages, and it is therefore evident
that the further back we go from the individual the greater are
the differences, whereas natural selection cannot be shown to be
more and more efficient in destroying transitions upon the same
route.

We have now to consider the actual differences seen between
organisms. There is no doubt that specific differences are usually,
but 7iot necessarily, small (p. 79), while generic are on the whole
larger, though there are large differences between different kinds
of genera, as for example between those of small families and those
of large (p. 110). Family differences are on the whole the largest
of the three. Looking at the list of family characters in Appen-
dix I, one notices their divergence when taken in pairs — alternate
or opposite leaves, cymose or racemose inflorescence, and so on.
Many of these pairs do not allow of intermediates or transitions,
but this shows less in generic or specific characters. Practically
all of the family characters, however, may at times appear as
generic or specific ; there is nothing about a character to place it
only in one of these classes. In fact, as we have seen on p. 110,
the rank of genera or of species differs with the size of the groups
to which they belong.

One may almost say that a family has a combination of most of
these "family" characters, though sometimes the one, sometimes
the other, of any particular divergent pair. The important points
are this divergence, and the fact that the character is shown in
full perfection (p. 114), a feature that one would certainly not
expect under the operation of natural selection, for the adapta-
tional value of the character would diminish as it approached
perfection, and probably 95 per cent or less would be as good as
100 per cent. It is all but inconceivable that selection should
produce perfection in a character, especially one like most

CH. xiv] GENERAL DISCUSSION 181

structural characters, in which one can neither find nor imagine
any adaptational value whatever.

If differentiation be accepted, the process of evolution may be
quickened up considerably, for a single mutation may effect in
one step a change which might take an immense time under the
action of natural selection, especially when one fully realises that
the vast bulk of structural differences have no adaptational value.
And if, as upon this view would seem highly probable, mutations
were, on the whole, larger as one went further back into past
times, the difficulties of explaining the origin of great groups like
the ferns will be greatly lessened. It must not be forgotten that
these also must be explained by natural selection, which as yet
has shown itself quite incompetent in this respect. So long as we
try to explain these by adaptational changes, or by dying out of
transitional stages, so long shall we be in great difficulty. The
theory of divergent mutation requires nothing of this kind, and its
capacity of explanation is far greater than is that of the theory of
gradual adaptation. It seems to the writer that the theory of
natural selection leads to too many untenable positions to be any
longer acceptable, and that differentiation, working downwards
towards the species, and by large mutations, diminishing as one
comes downwards (on the whole) should take its place.

The evidence is clearly in favour of differentiation, or diver-
gent mutation, rather than natural selection. The largest and
most divergent mutation gives rise on the whole to the family,
while the later and usually less divergent ones give rise to the
later genera and species, which come as a rule within the limits
marked out by the first. This agrees completely with the familiar
fact that the key to a family can be so easily made upon these
lines, with the largest differences coming first, followed by
smaller and smaller ones down to the specific and varietal dif-
ferences at the bottom. But this feature is a matter of extra-
ordinary difficulty to explain upon the Darwinian theory, under
which two species form by progress in gradual adaptation in
slightly different directions, the unmodified and the transitional
forms being killed out, until at last the difference is so great that
they have become new species. But if a slight variation in a
favourable direction is enough to give an advantage over the
forms that have not varied, what is to be gained by going on with
the variation until it becomes specific, and how is this to be done?
What adaptational need made one species adopt an alternate
leaf with a phyllotaxy of 5/8, its nearest relative opposite leaves?

182 GENERAL DISCUSSION [ch. xiv

Nothing but sudden mutation can account for such differences,
and natural selection has probably nothing to do with it. It
sorts the products of evolution into their most suitable places. It
is as if the evolutionary train dropped a passenger or two at
every station, who has then to make good in the particular
conditions that there obtain, in the society or community in
which he happens to find himself, and with the equipment for
the task that he happens to carry with him.

But if differentiation or divergent mutation be the more correct
explanation, it is clear that evolution moved in a direction the
opposite of that in which it would move under natural selection.
The latter works upward /rom the small variety, which is assumed
to be an incipient species, whilst the former works downward to
the variety. The latter, under differentiation, represents the last
ripple of the disturbance which gave rise to the family, not the
first ripple which is to give rise to a disturbance becoming ever
greater and greater. Natural selection kills out the ancestor and
the transitional forms, differentiation does not kill the ancestor,
nor expect transitions.

The question as to which explanation is nearer to the truth,
therefore, may be settled by an answer to the question as to
which was the direction in which evolution moved. To obtain
this answer, the author has devised some thirty-four test cases,
given in Chaps, x-xiii, and as all of them give good, and a number
give very strong, if not convincing, evidence in favour of the
direction required by divergent mutation, it becomes in a high
degree probable that this is the more correct explanation, and
that natural selection had little or nothing to do with the fact
that evolution went on.

There is some other law behind the latter, which at present we
do not understand, though probably when we learn what is the
driving force in cell-division, we shall be nearer to the goal. My
friend Dr Charles Balfour Stewart suggests that the law is
probably electrical, and that perhaps the development of a new
form may have some relation to the transfer of energy in some
way. The divergence of mutation may perhaps become a little
less unintelligible by some explanation of this kind.

To commence with the Numerical test cases (chap, x), it is
shown in case i (p. 90) that selection would have great difficulty,
as its very name suggests, in causing the evolution of vast and
increasing numbers of plants, whilst under differentiation this is
automatic, and follows the rule of the hollow curve. In case ii

CH. xiv] GENERAL DISCUSSION 183

(p. 94) it is shown that while natural selection can make no
predictions, under differentiation it is clear that on the average
the size of the largest genus in the family must go with the size
of the family itself, which proves to be the case, whilst in case iii
(p. 95) the gap in size between first and second genera, second
and third, and so on, is predicted as, and proves to be, a rapidly
diminishing one. In case iv (p. 97) it is predicted, and proved,
that the proportions of very small genera, considereti as relics
under natural selection, must on the average be larger the larger
the family, while it would be expected to be the reverse under
selection. In case v (p. 99) it is shown that the hollow curve is
entirely in favour of differentiation, and in case vi (p. 100) that
"Size and Space", a corollary of Age and Area, is equally so.
Case, VII (p. 100) refers to a paper by Yule and Willis (76),
showing that "the manner in which evolution has unfolded itself
has been relatively little affected by the various vital and other
factors, these only causing deviations this way and that from the
dominant plan", a conclusion which obviously does not harmo-
nise with the action of natural selection. Case viii (p. 101) shows
that while on the average the parent genus in small families has
as many species as all the rest, more and more genera are required
to halve the family when it grows larger. This could be predicted,
and is against natural selection. The numerical tests are all clearly
in favour of differentiation.

Morphological tests are described in chap. xi. In the important
case IX (p. 110), differences in generic rank are dealt with.
Natural selection can make no predictions, and simply regards
all genera as generic stages in evolution, and of rank as nearly
the same as the systematist can compass. Differentiation, how-
ever, says that the rank of a genus of a very small family will be
approximately equal, on the principle of divergent mutation, to
that of the sub-family of a large family. This proves to be the
case, giving very strong evidence indeed for evolution by diver-
gent mutation, and showing that the rank of a genus varies with
its position, and the size of it and of its family. In case x (p. 114)
the fact, hitherto almost totally ignored, is considered, that the
characters of plants are generally shown in their perfect condi-
tion, and especially so those of the higher groups. This could not
happen under selection, to which 95 per cent or less of perfection
would be as good as 100 per cent. This is a simple, but destructive
argument against gradual acquisition of characters. In case xi
(p. 115), the difficulty as to how natural selection got a grip upon

184 GENERAL DISCUSSION [ch. xiv

the early stages of non-adaptive characters is considered, and it
is pointed out that by differentiation there need not be such
stages, nor is adaptation called in. In case xii (p. 118) is con-
sidered the case of alternate and opposite leaves, a very common
case of divergent characters with no transitions, and where it
is almost impossible to suggest any adaptational value in the
difference between them. Selectionists have to admit that
anatomical needs are more potent than adaptational. This ques-
tion of the relative value of characters has been somewhat
ignored. In case xiii (p. 120) staminal characters are dealt with
in the same way, and give similar evidence. In case xiv (p. 122),
the berry is dealt with, and incidentally it is shown that there is
little evidence of adaptation in this phenomenon, so often quoted
as an illustration of it. Case xv (p. 124) deals primarily with
achenes and follicles. Natural selection could not produce these
in their perfect form, nor could it produce the perfect pod, and
distinguish between this and the follicle in the marked way that
one always finds. In cases xvi, xvii and xviii various other
structural puzzles are considered, all much more easily explained
by differentiation. In case xix (p. 129) the puzzle of correlated
characters in so many of what are usually called adaptations is
discussed, and it is shown that while it is quite inexplicable by
natural selection, it is somewhat more easy with differentiation,
which does not demand an adaptational value in everything.
These (adaptation) correlations are useful, while most are not,
but some day, perhaps, cytology will bring us the explanation of
correlation phenomena.

In chap. XII some further tests are considered under the head
of Taxonomy, though largely a continuation of the last. The
position of the largest genera of a family is dealt with in case xx
(p. 134). On the theory of natural selection one can make no
prediction about them, but on that of divergent mutation it is
clear that in general they will be widely separated, inasmuch as
they will have inherited their characters from the earliest muta-
tions that took place in the family. This is just what proves in
general to be the case, as is illustrated by the cases of the Ranun-
culaceae and the sub-family Silenoideae, etc. In some cases, e.g.
Clematis, the second largest genus in its family, the genus does
not seem to have given rise to a sub-famih' inheriting its most
obvious divergent character, the opposite leaves, but more usually
this is the case, and one finds the large genera heading sub-
families or other divisions. This is as predicted, whilst natural

CH. XIV] GENERAL DISCUSSION 185

selection is quite helpless to explain it. Good evidence is thus
given for differentiation. In case xxi (p. 136) the three largest
families are shown to occur one in each of the three great divi-
sions of the flowering plants ; this seems to indicate the proba-
bility of very large mutations in very early divergences. In
case XXII (p. 137) the mere fact that one can usually construct
dichotomous keys goes to prove differentiation. In case xxiii
the fact that divergence from the usual family characters is more
pronounced the larger the family, goes the same way. In case xxiv
(p. 138) the puzzling but frequent case of parallel variation in
one, or in two or three related, families, very difficult to explain
by selection, is simple to differentiation, whilst in case xxv
(p. 140) widespread organisms are shown to be the simpler,
though it does not say much for the advance in organisation
supposed to be the result of selection. Darwin himself puts it
down to their greater age, as does the writer.

In chap. XIII a few tests based on Geographical distribution
are given, but the full development of this attack upon the
current theory of evolution must be left for the publication of a
book which the writer has in preparation. In case xxvi (p. 146)
the difficulties brought up by age and area are considered,
especially the fact that on the average the distribution within a
country goes with the distribution outside. As the conditions
must vary, this goes to show that gradual adaptation other than
physiological can have had little or nothing to do with the distri-
bution. Natural selection has had to call in two supplementary
hypotheses to explain the facts brought up about endemics and
their distribution in Ceylon, and these hypotheses are mutually
contradictory. The fact that the distribution of family surnames
in Canton Vaud (p. 149) also matches that of plants, including
endemics, goes to show that natural selection had very little to
do with the latter other than purely locally. In case xxvii
(p. 149) it is shown how contour maps may be constructed for
most genera, especially when they are small, and have in the vast
majority of cases only one centre. As no one genus takes any
notice of the contours of any other (p. 154), the contours can
hardly be determined by any local conditions. Great Britain,
with its great variety of conditions, has nothing but margins of
contours. In case xxviii (p. 154), already published, it is shown
how the relationships of the smaller genera and of the species in
a genus, in any given family, often show such great geographical
divergence, with its near relatives separated by distances which

186 GENERAL DISCUSSION [ch. xiv

may even be enormous, crossing the oceans, or even the equator.
Natural selection could not explain this by any destruction of
transitions, for the separations are of all sizes and in all directions.
The only simple explanation so far proposed is that the local
genera or species are due to direct mutations from the linking
large and widely distributed genera or species that cover the
places in which they occur. This of course involves the acceptance
of differentiation. In case xxix (p. 156) it is showTi how variety
in structure shows no necessary relation to variety in conditions,
as one would expect under natural selection. In case xxx (p. 158)
the difficulty is pointed out, of explaining, under natural selec-
tion, a very common type of distribution. Many genera show one
or more widely distributed species, usually very polymorphous,
accompanied by local endemic species of the same genus in
various parts of their range. The only simple explanation is that
put forward by Guppy and by the writer that these endemics
have been derived from the widely distributed species by one or
more mutations. The incomprehensibility of selection is further
developed in case xxxi (p. 161). Incasexxxii (p. 161) the incon-
sistency of the contention that characters are less constant the
less useful they are, is pointed out, and in case xxxiii (p. 162) the
bearings of Hooker's discovery of the constancy of the numerical
relation between Mono- and Di-cotyledons are pointed out, with
the fact that there is no monocotyledonous mode of life. Case
XXXIV treats of overlap of genera.

It is clear that the tests give very strong evidence indeed in
favour of the theory of differentiation or divergent mutation,
according to which the course of evolution is in the opposite
direction to what has hitherto been supposed, and by mutations
which tend to diminish as time goes on, but go in the direction
family — genus — species. The organism that first represents the
family is, of course, at the same time its first genus and species,
but these are of different rank from genera and species in a
larger family. By further mutations this will then give rise to
further genera and species. The first new genus formed will
usually be widely divergent from the parent genus of the family,
even if the family be quite small, e.g. of two genera only. Later
formations will be less and less divergent on the whole, but will
show some of the characters of divergence of their first parents.
The main lines of divergence are therefore given by the latter,
and later genera fill them in, as shown by a good dichotomous
key. As time goes on, new genera will necessarily be evolved at a
continually increasing rate, and each, given time enough, may

CH. XIV] GENERAL DISCUSSION 187

ultimately become the parent, not only of many species, but of a
group of generic offspring forming a sub-tribe or larger division.
The whole family will at last end in the tail of genera containing
one species each, as shown in the hollow curve. The oldest genera
will have the most species, and the number will diminish as does
the age, till we come to the tail of monospecific genera.

There is thus very strong evidence to the effect that evolution
has gone on without any direct reference to natural selection so
far as we can at present see. The new form will appear, whether
it be desirable, or suitable, or not, and whether it then survive
will depend upon the action of natural selection, with reference
to the conditions at the moment. The business of natural selec-
tion is (1) to kill out everything in any way unsuitable to the
conditions that surround it at the time, either at its first ap-
pearance upon the scene, or when a change of conditions occurs ;
and (2) to adjust to its surroundings, if possible, every new form
that comes into the place, whether a new species just born, or a
species newly arrived from somewhere else. There is thus plenty
of occupation left for natural selection, and in a field where its
usefulness and value have never been questioned. The early
pioneer species will, of course, get the best chance, and as each
newcomer arrives, increasingly close and careful adjustment will
be needed, adjustment which natural selection will apply without
fear or favour.

Lastly, the evidence is equally strong that in the process of
evolution, at any rate as a general rule, the new species formed
(which might also be a new genus or even new family if the muta-
tion were a little larger) would appear at one step by sudden
mutation. Evolution goes on, but we can see no reason at present
that will determine that it shall go in any particular direction,
especially in one that shows greater adaptation. The mere fact of
the survival of the "lower" forms in such numbers, like mosses,
ferns, and liverworts, is against the idea, of any rapid progress in
adaptation, but probably when an "adaptation" appears, such
for example as climbing habit, it will be allowed or encouraged to
survive, though why it should appear is at present a mystery.

It is an inspiring thought that so great and complex a process
as evolution must have been has not been a mere matter of
chance, but has behind it what one may look upon as a great
thought or principle that has resulted in its moving as an ordered
whole, and working itself out upon a definite plan, as other
branches of science have already been shown to do. Darwinism
made the biological world a matter of chance. Differentiation,

188 GENERAL DISCUSSION [ch. xiv

backed by the universal occurrence of the hollow curves, shows
that there is a general law, probably electrical, at the back of it.
And if evolution goes on without reference to adaptation values,
each genus giving rise to another, and both surviving (as a rule),
then the hollow curve becomes an integral part of it. Further
refinements must be left to the mathematicians, and will doubt-
less provide interesting results.

Differentiation is not based upon adaptation at all, the
latter remaining a primarily physiological phenomenon. The
ordinary type of adaptation, that is familiar to agriculturists,
is described in the following extract from a paper by Cockerell.
" In California certain scale insects, subjected to poisonous fumes
by the horticulturists, have, by a process of the survival of the
fittest, developed resistant races, not distinguishable by any
morphological characters." The writer was very troubled by
cockroaches in his (tropical) house, and used a certain much-
advertised poison, but though at first the death-rate was very
high, presently there appeared a race of cockroaches that was
immune, but looked exactly the same as their predecessors. "The
chance of introducing from outside an all-round superior strain
diminishes as the adaptation of the local strain to its environ-
ment increases" (77, p. 283). Many similar extracts from agri-
cultural papers might be quoted.

There is good reason to suppose that in some way the genes
and chromosomes are immediately responsible for the evolution
that is going on. Their divisions and fusions strongly suggest some
electrical process, with which the suggested action of cosmic rays
may have something to do. Or again, something of the nature of
genie changes may be going on, and occasionally result in the
taking up of "more or less stable positions of equilibrium in cell
division" as suggested by the writer in 1907. The apparently
purposeless way in which distinguishing characters go together is
verv like the similar behaviour seen in anv mutation involving
more than one character. There is no evident reason, nor sugges-
tion of reason, why a Monocotyledon should have at the same
time one cotyledon, a trimerous flower, a parallel-veined leaf, and
a peculiar anatomy. Nor why Cruciferae should have tetra-
d}aiamous stamens, Dryas eight petals, and so on. Nothing but
the direct effect of the genie composition, with heredity, will
explain why the characters are shown in perfection.

The continual appearance of characters with complex correla-
tion that could not be due to selection, such as the characters of a

CH. XIV] GENERAL DISCUSSION 189

family, genus, or other group, or such as climbing stems accom-
panied by the means of climbing, goes to show that the family
characters, or the climbing habit, must have been produced by
some sudden chromosomic change, but by what, and how, deter-
mined, we are as yet completely ignorant. Many other "adapta-
tions" come into the same categforv.

A very probable large mutation, giving the ancestor of what is
now a large genus, is that which perhaps gave rise to the colum-
bine (Aquilegia), which can easily be imagined as arising from the
larkspur (Deljjhinium) by a mutation like that which often gives
a symmetrical sport in the toad-flax flower.

It would seem probable that the early future development of the
study of evolution will be largely based upon the study of cytology,
for it would seem that the conception of gradual adaptation, at
any rate in its present form, must be abandoned. The larger
groups seem to have appeared before the smaller, upon the whole,
the force or size (if one may use such a term) of the mutations
that went on diminishing as time went on, the number of smaller
mutations on the whole increasing in proportion to the larger.
What actual part the external conditions took in the matter
is at present inexplicable, but there is nothing in the structural
characters, as a general rule, to show that the part was a large one.

One must not lose sight of the hybridisation that is so easily
possible, and of which Lotsy (27) made so important a feature in
evolution. At the same time, if mutation can take place, as
seems highly probable, in such a way as to cross the "sterility
line" between species, and so to isolate them, it does not seem
very likely that fertile species -hybrids will be produced in such
numbers as to have an important influence upon evolution
generally, though one must not forget the possible influence
of the cosmic rays or other factors in causing the doubling of the
chromosome numbers. Hybridisation seems very unlikely among
the widely separated genera that seem to be the firstcomers, in
most, if not in all, families, but as one goes down the scale, one
seems to come among genera that are closer and closer together
in their taxonomic characters, and with these hvbridisation
would seem to become more and more possible, and more likely
to occur. Still more would this be the case among the species,
and here again rather in the species of large genera. It seems to
the writer that this question of hybridisation, with its increasing
possibilities in the genera and species of later formation, may be
one of some importance, though one must, of course, not forget

190 GENERAL DISCUSSION [ch. xiv

that these later genera and species will be of much less wide
distribution than the earlier.

The conceptions thus put forward have several possibly even
unexpected bearings. If new species and genera can thus arise in
widely separated places, though related, there seems no reason
why the same character, produced of course by some particular
arrangement of genes or chromosomes, should not at times arise
from ancestors in which it did not itself occur, i.e. should arise
polyphyletically, or from different ancestors. One may even
imagine more than one character arising in this way, so as to
form, though probably only with great rarity, a polyphyletic
genus. In some such way as this one may imagine the case of one
genus coming through another, as suggested by Bower in the
ferns (2). One must remember, too, that what look like species
of the same genus and closely allied, need not necessarily be such,
and one must compare their chromosome numbers. It is even
possible that originally separate types may converge until they
may be able to become cross-fertilised.

The sudden appearance of similar mutations at widely separated
places may be easily accounted for by a similar construction in
the chromosomes of their ancestors, which might give rise to
similar mutations. There is no definite reason that one can see —
though, of course, this is unfamiliar ground to the writer — why
the same genie distribution should not appear in two new species
formed from one genus, thus giving rise to a new genus of two
species, and possibly even discontinuous in distribution.

Finally, a very strong argument in favour of differentiation,
just as with Age and Area, is that by its aid one may make a
great many predictions as to what will be found to occur, and
find that these predictions are borne out by the facts. A number
of such are to be found in many of the test cases given in
chaps, x-xiii, and others may be found elsewhere. Now upon
the theory of natural selection it is as a rule impossible to make
any predictions at all, and when, as for example in several test
cases, one may venture a prediction, this is found to be opposed
to that made upon the theory of differentiation, and is not borne
out by the facts, which always favour the latter theory. This seems
to the writer to be a very strong argument in favour of differentia-
tion or divergent mutation. At first, owing to the fact that one
has to think, so to speak, in the reverse direction from that to
which one has been accustomed (i.e. from family to variety, not
from variety to family), it is not always easy, but one soon gets
into the new direction of thought.

CHAPTER XV

FINAL SUMMARY OF CONCLUSIONS

1. The world has undoubtedly been peopled by an evolution of
forms one from another, giving rise, as time has passed, to beings
of increasing complexity.

2. The process of evolution appears not to be a matter of
natural selection of chance variations of adaptational value.
Rather it is working upon some definite law that we do not yet
comprehend. The law probably began its operations with the
commencement of life, and it is carrying this on according to
some definite plan.

3. Evolution and natural selection are probably to a great
extent independent, and they work at right angles to one another,
with (in plants at any rate) little mutual interference.

4. Evolution most probably goes on by definite single muta-
tions, which cause structural alterations, which mav, but bv no
means necessarilv must, have some functional advantao^e at-
tached. If such an advantage appear in the mutation, natural
selection will likely allow it to survive. There is no necessary
reason why the immediate ancestor should die out.

5. Evolution goes on in what one may call the downward
direction from family to variety, not in the upward, required by
the theory of natural selection.

6. Evolution thus moved in the opposite direction to that
required by natural selection, and thirty-four test cases are
given, all giving evidence to that effect.

7. Evolution is no longer a matter of chance, but of law. It
has no need of any support from natural selection.

8. It thus comes into line with other sciences which have a
mathematical basis.

9. The theory of natural selection has been trying to work it
backwards.

10. Mutation tends to be divergent, especially in the early
stages of a family. The family, consisting probably of one genus
and one species, is probably first created by a single mutation,
whilst later ones are usually less marked than the first, and give
rise to further genera and species. The earliest mutations ulti-
mately give rise to the chief divisions of the family.

11. The Linnean species is not necessarily a conglomeration of
forms made from below upwards, but is rather a stage on the way
downwards to the Jordanian species.

192 SUMMARY OF CONCLUSIONS [ch. xv

12. Varieties are the last stages in the mutation, and are not,
as a rule, incipient species.

13. Chromosome alterations are probably largely responsible
for the mutations that go on.

14. The theory of natural selection is no longer getting us any-
where, except in politics (influence of the dead hand).

15. It comes in principally as an agent to fit into their places
in the local economy of the place where they are trying to grow,
the forms there furnished to it, whether newly evolved, or only
newly arrived, killing out those in any way unsuitable.

16. It has, therefore, not been responsible for the progress that
has been made by the actual evolution of new forms, but it has
been all-important in fitting them into their places in the economy,
which is always increasing in complexity.

17. The theory of natural selection makes evolution a con-
tinuous and gradual process, diff'erentiation a discontinuous one.

18. Natural selection (the struggle for existence) works rather
upon individuals than on groups. It causes the survival of the
fittest population, rather than the fittest type in the mixture.

19. It can make few or no predictions, while diff'erentiation,
like age and area, can make many, which are usually successful.

20. Adaptation has been mainly internal or functional, rather
than external or structural.

21. Differences in structure do not necessarily mean difiPerences
in adaptation.

22. The mutations supposed in diff'erentiation would at one
step cross the "sterility line" between species, which has always
been a great stumbling block to natural selection; and thus at
once isolate the new form, preventing its loss by crossing.

23. Diff'erentiation makes it possible for evolution to go on
more rapidly than under natural selection.

24. It explains the great discontinuity seen in the facts of
economic botany.

25. It explains the difficulty, almost insuperable to the theory
of natural selection, of the increasing divergences of characters
as one goes up the scale from species to family.

26. It gets over the difficulty of early stages, and of the fre-
quent correlation of characters, and the need of calling in "mor-
phological necessity"; it does not need to call in adaptation, as
the theory of natural selection has to do ; and it explains why the
large genera are the most variable.

27. It explains the fact that adaptation is so often generic.

CH. XV] SUMMARY OF CONCLUSIONS 193

28. With its probably genetical basis, it explains the difficulty
of the perfect form in which characters, and especially those of
the higher divisions, are exhibited, which was almost impossible
to the theory of natural selection.

29. It gets over the difficulty caused by the fact that few tran-
sition stages are found, either in living or in fossil plants.

30. It explains the universal hollow curve, as well as age and
area and size and space, all impossible to the theory of natural
selection.

And one may add :

The 34 test cases given often bring out new and sometimes
unexpected relations, e.g. the grouping of a family (or sub-families
if large) into large, medium, and small genera.

Adaptation, isolation, and other phenomena are discussed from
somewhat new points of view.

Upon pp. 76, 139, and elsewhere, indications have been given,
more or less unintentionally, about things that will only appear
in a forthcoming book upon geographical distribution. Therein
the writer hopes to show that the adoption of age and area and
of differential or divergent mutation, for both of which good
proof has now been given, reduces the problems of distribution
to a simpler form. By abandoning the supposition, necessarily
inherent in natural selection, that plants may be divided into
successes and failures, the one expanding and the other contract-
ing the area occupied, all may be regarded as behaving in much
the same way as their near relatives. One thus obtains a more
satisfactory picture of how evolution and geographical distribu-
tion went on, and how thev fitted into one another.

WED 13

APPENDIX I

THE COMMON CHARACTERS THAT
DISTINGUISH FAMILIES

The list is made up from the key at the end of my Dictionary, and
includes the necessary characters to distinguish one family from
another. They are arranged as far as possible in divergent pairs,
and it will at once be noticed that most of them do not lend
themselves to possessing intermediates or transitions.

Herbs, shrubs, trees; parasites, saprophytes, epiphytes, thalloid.

Roots from tap-root, or adventitious.

Stem, rhizome, bulb, etc.; creeping, climbing, or not; herbaceous
or woody; jointed or not; mono- or sympodial; angled or not;
with latex or resin, or not.

Leaves radical or cauline ; alternate, opposite, or whorled ; in two
ranks, or in three or more; sheathing or not; ligulate or not.

Leaves simple or compound ; palmate or pinnate, etc. ; entire or
lobed or toothed ; fleshy or hairy or not ; pitchers or not ; with
oil cavities, glandular dots, with chalk glands, or not.

Leaves stipulate or exstipulate; parallel- or net-veined; dorsi-
ventral or isobilateral ; asymmetrical or not.

Inflorescence racemose, cymose, or mixed; ^ or unisexual; a
raceme, corymb, catkin, mono- or dichasial cyme, etc. etc.;
with bracts or not; with spathe or not; with bracteoles or not.

Receptacle convex, flat, or hollow; with or without effigurations.

Involucre or none; epicalyx or none; disc or not.

Flower spiral or cyclic; ^ or ^ $; mon- or dioecious; with
perianth or not; homo- or hetero-chlamydeous ; iso- or hetero-
merous; with parts in twos, threes, fours, etc.

Flower regular or zygomorphic; zygomorphism vertical, trans-
verse, or oblique; with ray florets or not; heterostyled or not;
resupinate or not.

Perianth petaloid or sepaloid, or none.

Calyx whorled or spiral; convolute, imbricate, or valvate; poly-
or gamosepalous ; odd sepal anterior or posterior.

APPENDIX I 195

Corolla of free or united petals; regular or two-lipped; convolute,
valvate or imbricate; alternate with sepals, or superposed;
corona present or not.

Androphore, gynophore, column, etc., or not.

Stamens in one, two, or more whorls, or spiral; staminodes or not;
epipetalous or not; on disc or not; changed to nectaries, etc., or
not.

Stamens in one, two, or more whorls ; in one whorl all present, or
not ; spiral and oo or not ; free, or united in tube or in bundles ;
diplostemonous or obdiplostemonous ; antepetalous (if one
whorl) or not; epipetalous or not; on the disc or not.

Stamens branched or not; tetra- or di-dynamous, or not; odd
stamen anterior or posterior; staminodes or not; changed to
nectaries or not; exploding or not; bent inward in bud, or not.

Anthers dorsi- or basi-fixed, or versatile; extrorse or introrse;
mono- or di-thecous; opening by splits, valves, pores, teeth,
etc. ; connective with or without appendages.

Pollen spherical, polyhedral, etc.; smooth, prickly, warty, etc.;
in tetrads, pollinia, etc.

Ovary superior or inferior, etc.; 1-2-3-4-5-more carpels; 1-2-3-4-5-
more loculi.

Carpels spiral or in whorls; apo- or syn-carpous; united only by
style; 1-2-3-more; transverse or anteroposterior to flower;
some abortive, or not ; in superposed whorls or not.

Placenta parietal, axile, basal, apical, free-central, etc. ; bilobed.

Ovules 1-2-few-many per loculus; in one, two, or more rows as
seen in transverse section; stalked or sessile; erect, horizontal,
or pendulous; orthotropous, anatropous, campylotropous ; on
surface.

Raphe ventral or dorsal; micropyle up or down.

Style basal or terminal ; present or not, one, or as many or twice
as many as carpels; entire or divided; with pollen-cup.

Stigma capitate, lobed, divided, etc.; petaloid or not; sessile or
not.

Fruit fleshy or dry; achene, follicle, siliqua, schizocarp, capsule
(loculi-septi-cidal, septifragal, etc.), drupe, berry, etc., etc.;
dehiscent or indehiscent, etc.; simple or compound; winged or
not; with pappus, or hooks, etc.

13-2

196 APPENDIX I

Replum or not; individual carpels divided by horizontal or
longitudinal walls, or not.

Seeds per flower, or per carpel 1-2 -few-many; albuminous or
exalbuminous ; with endo- or perisperm; with aril or not,
winged or not; hairy or not.

Embryo with one cotyledon or with two ; large or small in propor-
tion to endosperm; straight, curved, twisted, folded, etc.

Endosperm oily, starchy, fleshy, cartilaginous, etc. ; ruminate or
not.

APPENDIX II

CHARACTERS OF FIRST DIVISION
INTO SUB-FAMILIES OR TRIBES

Vegetative organs

Land plants — waterplants

Roots — none

Green plants — parasites

Climbing — not

Shrubby — annual

Shrubby — undershrub or herb

Leaves cauline — radical

Leaves 2-ranked — not

Leaves opposite — usually alternate

Leaves opposite, stipulate — alternate,

exstipulate
Leaves palmate — simple or pinnate

Leaf-thorns in axil — not

Cystoliths — none

Glandular or stinging hairs — none

Hairs simple or none — usually branched

Pedahaceae

Lemnaceae

Convolvulaceae

Lardizabalaceae, Polemoniaceae

Capparidaceae

Juncaceae

Flagellariaceae

Iridaceae, Zingiberaceae

Gentianaceae, Myrtaceae

Rhizophoraceae

Bombacaceae, Datiscaceae, Lar-
dizabalaceae
Salvadoraceae
Hernandiaceae
Loganiaceae, Urticaceae
Cruciferae

Corolla

Free or slightly united — long tube
Broadly campanulate — salver-shaped
Valvate or other aestivation

Spurred or not
Labellum — none

Petals with appendages — without
Petals outside disc — on margin
Honey-leaves — none
Lateral teeth of corolla overlap —
underlap

Dichapetalaceae, Tamaricaceae
Nolanaceae

Ebenaceae, Elaeocarpaceae, Gen-
tianaceae
Papaveraceae, Violaceae
Stylidiaceae
Sapotaceae
Burseraceae
Berberidaceae
Scrophulariaceae

Fruit and Seed

Berry or drupe — dry fruit

Other varieties of, or variations in,
fruit

Bromeliaceae, Commelinaceae, Fla-
gellariaceae, Myrtaceae, Oxali-
daceae, Pittosporaceae, Rhizo-
phoraceae, Ulmaceae, Zygophyl-
laceae

Connaraceae, Epacridaceae, Faga-
ceae, Geraniaceae, Labiatae,
Malpighiaceae, Oleaceae, Pole-
moniaceae, Proteaceae, Ranun-
culaceae, Valerianaceae, Vochy-
siaceae

198

APPENDIX II

Embryo straight — curved, or otherwise
differently shaped

Endosperm — none

Endosperm ruminate — not
Seeds basal— not
Seeds embedded in placenta — not
Seeds in one plane — in more than one
Capsule many-seeded — one-seeded in-
dehiscent

Basellaceae, Butomaceae, Cheno-
podiaceae, Convolvnlaceae, Eu-
phorbiaceae, Droseraceae, Her-
nandiaceae, Melastomaceae, Sol-
anaceae, Ulmaceae

Goodeniaceae, Myrsinaceae, Nym-
phaeaceae, Ochnaceae, Rhizo-
phoraceae

Polygonaceae

Connaraceae

Sonneratiaceae

Nolanaceae

Caryophyllaceae

These will serve as examples of the degree of divergence of the
characters that separate the sub-families or tribes of the various
families, without going into too great detail.

APPENDIX III

CHARACTERS OF GENERA IN
BI-GENERIC FAMILIES

Wing on fruit unilateral
P 5, bracteoles

Aceraceae
Achatocarpaceae

Balanopsidaceae
Balsaminaceae

G (2), style slender, 2-
armed

Capsule elastic. Ovules
pendulous one above
another

Embryo spiral with nar-
row cots.

K and C alternate. A free Caricaceae

Cannabinaceae

Leaves opposite. Plumule
thick and straight

Sepals with distinct midrib

Leaves opposite

K in tube. Disc normal

P of ^5-8. 0\nile basal. ?
flowers in groups of 4

Caryocaraceae

Elatinaceae
Erythroxylaceae
Hippocastanaceae
Julianiaceae

Sta. in one whorl. Ovules Hydnoraceae

free
5-merous
Ovary 1-locular
C valvate. A 15-30

Disc of few teeth. Sta. few.
Bracts entire

Pets. 6-7. Sta. unequally
long, anthers by longi-
tudinal slits

Bracteoles. (C). Pollen
spiny. Perennial

Berry

L. opp. C 5, very unequal,
one spurred. A to 10

Partial fruit not winged

Stamens not more than 6
C. 3 stds. G 1-loc. with
basal or parietal placenta

Limnanthaceae

Nyssaceae

Quiinaceae

Salicaceae

Scytopetalaceae

Stackhousiaceae

Taccaceae
Trigoniaceae

Tropaeolaceae

Velloziaceae
Xyridaceae

Wing all round the fruit
P 4, exc. term, flower. No

bracteoles
G (3), style thick, 3-armed,

arms bifid
Berry. Ovules pendulous

side by side

Embryo curved with broad

cots.
K and C superposed. A

united
Leaves alternate. Plumule

long and spiral
Sepals with no midrib
Leaves alternate
K free. Disc excentric
P of cJ usually 4. Ovule

lateral. $ flowers in

groups of 3
Two whorls. 0\ailes sunk

in placenta
3-merous

Ovary 6-10-locular
C convolute. A oo

Disc hollow. Sta. oo.

Bracts divided
Pets. 3. Sta. equally long,

anthers by apical slits

No bracteoles. C. Pollen

smooth. Annual
Capsule
L. alt. C 3, alike. A 3-5

Partial fruit with 3 broad

wings
Stamens more than 6
(C). No stds. G3-10C. with

axile placenta

LIST OF LITERATURE

1. Bateson, W. Materials for the Study of Variation. London, 1894.

2. Bower, F. O. Phylogeny of the Filicales. Ann. Bot. xxxi, 1917,

p. 1.

3. Chandler, M. The arctic flora of the Cam valley at Barnwell.

Q.J.G.S. Lxxvii, 1921.

4. Clarke, C. B. In DC. Monogr. Phanerog. V. Cyrtandreae. Paris,

1888.

5. Darlington, C. D. The Evolution of Genetic Systems. Cambridge,

1939.

6. Darwin, C. The Origin of Species, 6th ed. London, 1872 seq.

7. Life (one vol. ed.), by F. Darwin. London, 1888.

8. Darwin, C. and Wallace, A. R. In Journ. of Proc, Linn. Soc.

p. 45, 1 July 1858.

9. DoBZHANSKY. Genctics and the Origin of Species. New York,

1937.

10. Evolution. Essays presented to E. S. Goodrich. Oxford, 1938.

11. GuppY, H. B. Observations of a Naturalist in the Pacific. London,

1906.

12. Plants... in the West Indies and Azores. London, 1917.

13. Plant Distribution from an old Standpoint. Trans. Vict.

Inst. April 1907.

14. Plant Distribution from the Standpoint of an Idealist.

Journ. Linn. Soc. xliv, 1919, p. 439.

15. Harland, S. C. The genetical Conception of Species. Mem.

Acad. Sci. U.R.S.S. 1933, No. 4.

16. Hayata. Ueber das dynamische System der Pflanzen. Ber. D.

Bot. Ges. XLix, 30 July 1931, p. 328.

17. Hooker, Sir J. D. Botany of the Voyage. ..'"'' Erebus'''' and " Terror''\

(a) II. Flora Nov. Zeal. 1853; (b) III. Flora Tasmanieae,
1860.

18. In Biologia Centr.-Amer. p. Ixii.

19. Life and Letters. London, 1918.

19a. On Insular Floras. Journ. Bot. 1867, p. 23.

20. Huxley, T. H. The Gentians.... Journ. Linn. Soc. xxiv, 1888,

p. 101.

21. Jenkin, Fleeming. In North Brit. Review, June 1867.

22. Jordan, A. Diagnoses d''especes nouvelles ou meconnues. Paris,

1864.

23. Kerner, a. Flowers and their unbidden guests. London, 1878.

24. Kerner, A. and Oliver, F. W. The Natural History of Plants.

London, 1894.

25. Longley, W. H. In Science, lxix, 3 May 1929, p. 462.

26. In Science, lxxiii, 26 June 1931, p. 700.

LIST OF LITERATURE 201

27. LoTSY, J. p. Evolution by means of Hybridisation. The Hague,

1916.

28. IVLvcBRiDE, E. W. The present position of the Darwinian theory.

Sci. Progr. xviii, July 1923, p. 76.

29. ]VL\CLAREN, J. J. On some chemical difficulties of Evolution.

London, 1877.

30. Malthus, T. R. An Essay on the Principle of Population, 6th ed.

2 vols. London, 1826.

31. MiVART, G. S. The Genesis of Species. London, 1871.

32. MoRGAX, T. H. Evolution and Adaptation. New York, 1903.

33. OwEX, Sir Richard. Anato7ny of Vertebrates, iii, chap, xl, p. 786.

London, 1868.

34. Palmgrex, a. Chance as an element in Plant Geography.

Proc. Internat. Congr. Plant Sciences, i, 1929, p. 591.

35. Pax, F. and Hoffmax^x, K. Caryophyllaceae. Engl. Nat. Pfl.,

ed. 2, 16c, 1934.

36. Perkixs, J. Styraceae. Pflanzenreich, 1907.

37. Monimiaceae. Pflanzenreich, 1907.

38. Ridley, H. X. The Dispersal of Plants throughout the World.

Ashford, 1930.

39. DU RiETZ. The fundamental Units of Biological Taxonomy.

Svensk Bot. Tidskr. xxiv, 1930, p. 333.

40. RuHLAXD, W. Eriocaulaceae. Pflanzenreich, 1903.

41. St Hilaire, G. Lectures. Resume in Rev. et Mag. de Zool.

1 January 1837.

42. Shipley, A. E. Life. Cambridge, 1923.

43. Small, J. Quantitative Evolution in Compositae. Proc. R. S.

Edin. LVii, i, p. 26; iii, p. 215; lviii, i, p. 14.

44. Taxsley, a. G. Types of British Vegetation. Cambridge, 1911.

45. Trimex, H. Note on the Botany of Ritigala. Journ. R. A. S.

Ceylon, 28 December 1888.

46. Trimex, H. and Hooker, J. D. Flora of Ceylon. 5 vols. London,

1893 to 1900.

47. TuRRiLL, W. B. The Plant Life of the Balkan Peninsula.

48. DE Vries, H. The Mutation Theory (Engl. ed.). London, 1910.

49. Waxgerix, W. Cornaceae. Pflanzenreich, 1910.

50. Wext, F. a. F. C. Ueber Zwecklosigkeit in der lebenden Natur.

Biol. Centr. xxvii, 1907, p. 257.

51. Willis, J. C. A re\ision of the Podostemaceae of India and Ceylon.

Ann. Perad. i, 1902.

52. Studies in the Morphology and Ecology of the Podoste-
maceae.... Ann. Perad. i, 1902, p. 267.

53. The lack of adaptation in... Podostemaceae. Proc. R.S. B,

Lxxxviii, 1914, p. 532.

54. The Origin of the... Podostemaceae. Ann. Bot. xxix, 1915,

p. 299.

55. The Evolution of the... Podostemaceae. Ann. Bot. xl,

1926, p. 349.

202 LIST OF LITERATURE

56. Willis, J. C. The Flora of Ritigala, a study in Endemism. Ann.

Perad. iii, 1906, p. 271.

57. Hill-top Floras of Ceylon. Ann. Perad. iv, 1907, p. 131.

58. The Flora of Naminakuli-Kanda. Ann. Perad. v, 1911, p. 217.

59. The Endemic Flora of Ceylon.... Phil. Trans. B, ccvi,

1915, p. 307, and correction in Proc. R. S. B, lxxxix, 1916.

60. The Evolution of Species in Ceylon. Ann. Bot.xx^x, 1916, p. 1.

61. Distribution of Species in New Zealand. Ann. Bot. xxx,

1916, p. 437.

62. Further Evidence for Age and Area. Ann. Bot. xxxi, 1917,

p. 327.

63. Relative Age of Endemic Species. Ann. Bot. xxxi, 1917,

p. 189.

64. The sources and distribution of the New Zealand flora.

Ann. Bot. xxxii, 1918, p. 339.

65. Endemic Genera in relation to others. Ann. Bot. xxxv,

1921, p. 493.

66. Age and Area. Cambridge, 1922.

67. Age and Area; a reply to criticism. Ann. Bot. xxxvii,

1923.

68. Some further Studies in Endemism. Proc. Linn. Soc.

cxLviii, 1935-6, Pt. 2, 27 March 1936, p. 86.

69. Some Conceptions about Geographical Distribution and

Origin of Species. Proc. Linn. Soc. cl, 1937-8, Pt. 3, 24 June
1938, p. 162.

70. Some evidence against the theory of Natural Selection of

Infinitesimal Variations. Ann. Perad. iv, 1907, p. 1.

71. Further evidence. Ann. Perad. iv, 1907, p. 17.

72. The Geographical Distribution of the Dilleniaceac.as

illustrating... Mutation. Ann. Perad. iv, 1907, p. 69.

73. Is the Theory of Natural Selection adequate? Nineteenth

Century, xcii, October 1922, p. 615.

74. The Origin of Species by large... mutations, and by Guppy's

method of Differentiation. Ann. Bot. xxxvii, 1923, p. 605.

75. Yule, G. Udny. In Phil. Trans. B, ccxiii, 1924, p. 21.

76. Yule, G. Udny and Willis, J. C. Some statistics of Evolution

and Geographical Distribution in Plants and Animals, and
their Significance. Nature, cix, 9 February 1922, p. 177.

77. Hutchinson, J. B. Note on a Policy of Introduction of new

varieties of Cotton in Africa. Empire Cotton Growing Review,
XV, 1938, p. 283.

78. Some problems in Genetics, whose Solution would help the

Plant Breeders. Empire Cotton Growing Review, xv, 1938,
p. 286.

79. The third Conference of workers on Cotton-growing

problems. Empire Cotton Growing Review, xvi, 1939, p. 1.
Cf. p. 4 for "mixing". And see various papers in the same
journal for the two preceding years.

INDEX

(r.C. = Test Case)

Achenes, T.C. xv, 124

Acrotrema, 50

Adam's Peak, endemics, 61

Adaptability, 60, 145

Adaptation, 4, 13, 14, 15, 23, 39, chap.
VI, 52, 131, 146, 170, 176, 188;
generic, 18, 60, 107 ; geographical
distribution based on, by selec-
tionists, 56; gradual, 108, 117,
1 65 ; importance exaggerated,68 :
internal, 15, 44, 57, 58, 116; in
Podostemaceae, 18 ; in structural
characters, 4, 44, 52, 116; phy-
siological, 15 ; structural charac-
ters of adaptational value, 52-3 ;
to movable conditions, 56

Age and Area, 27, 29, 30, 31, 69, 86,
140, 146, T.C. XXVI, 146; com-
parison with allies, 29; essence
of theory, 69, 95 ; gradual devel-
opment, 27, 29; opposition, 30,
31

Alternate and opposite leaves, T.C.
XII, 118

Anaphalis, 26

Anemone, 105; in India, 159

Animals, 136

Aquilegia, 49, 189

Artocarpus, 31, 148

Assumptions of Natural Selection,
see Natural Selection

Author's change of ^^ews, 18-20, 24,
26, 39, 44, 46, 67, 171-2

Axioms of taxonomy, 132

Balkan endemics, 63

Barriers to spread, 59, 69, 153,

176
Bateson, W., 164
Berry fruit, T.C. xiv, 122
Beta, contour map of, 153
Bog plants, 53
Breeding, 10, 56

Cardamine, distribution, 155
Carludovica, 76
Caryophyllaceae, 96, 108, 124
Ceanothus, 31, 147
Centrolepidaceae, 49, 139

Ceylon, 18, 27, 33, 39, 50, 59, 61, 146,

147
Chandler, Miss, 64, 72
Change of conditions, 55, 146
Change of views, see Author
Characters, differences in, 16; early
stages, T.C. xi, 115: easy of
acquisition, 139, 140; famUy,
16, 74, 76, 106, 108, 180; more
constant the more useful (?),
T.C. XXXI, 161 ; of distinction of
families, genera and species,
106 ; value of, 71
Chemistry of plants, 8
Chromosomes and evolution, 188
Classification, 1, 2
Clematis, 70, 135; in India, 159
Climatic conditions, differentiation,

59
Climbing plants, 57
Colem, 24, 132, 166
Common names, 2
Compositae, 48, 98, 102
Conditions of life, 3; equal, with
great structural difference, 44;
variation in, 91
Continuous variation, 10
Contour maps, T.C. xxvii, 149
Convergence of evolution, 135, 190
Correlation, 11, 48, 57, 58, T.C. xix,

129, 188
Cosmic rays, 26, 62, 63, 115
Cruciferae, 154; distribution, 155
Cvclanthaceae, 76
Cytology, 89, 189

Darwin, Charles, 3, 65, 74, 91, 110,

132
Darwin, Sir Francis, 9, 18
Darwinism and its difficulties, 6, 10,

13, 21, 30, 39, 48, 115, 117, 139,

164, 166, 167, 187
Dead hand, influence of, 6, 144
Death of species, 72
Delphinium, 49, 189
Destruction under natural selection,

144, 155
Dianthus, 86
Dicotyledons, 15, 43, 47

204

INDEX

Differences in character, 16

Differences in generic rank, T.C. ix,
110

Differentiation, 6, 17, 39, 46, 50, 88,
96, 112, 181, 186, 187, and chap.
VIII, 65 ; author's pubhcation of,
68; course of evolution under,
68, 69; diagram of process, 69,
70, 111; direction the oppo-
site to selection, 68; generic
rank under, T.C. ix, 110;
growth of family under, 71 ;
Guppy's publication of, 66;
Hooker on, 74; in Ranuncula-
ceae, 81; in Silenoideae, 86; of
climate, 59; survival of parent
under, 66; and see Test Cases,
chaps, x-xiii

Dilleniaceae, distribution, 44, 45

Dipterocarpaceae, 125

Discontinuity in evolution, 8, 169

Dispersal mechanisms, 123

Distribution, 21 ; and age, 29, and
cf. Age and Area; barriers to,
59, 153; discontinuity, 5; not
determined by selection, 39

Divergence of character, 74, 119,138,
180

Divergence of variation, 16, 82, 85,
112, 137, chap, ix, 74 ; increasing
upwards, 76, 113; more or less
equal at corresponding levels in
large, medium, and small fami-
lies, 112

Droseraceae, 126

Dry and wet zones, 59

Early stages of characters, T.C. xi,
115

Ecology, 9

Economic botany, 8, 89, 169, 177

Endemics, as young beginners. 30,
154, 160; in large genera, 26; in
North America, 30; local adap-
tation, 30, 147 ; may prove very
useful, 89; not usually mori-
bund, 28; of Balkans, 63; of
Ceylon mountains, 26, 61 ; of
New Zealand, 29 ; often Linnean
species, 49; relics, 30

Endemism, 24, 26, 27, 30; on
mountains, 26, 61

Englishman, acclimatisation of, 105

Eriocaulaceae, 49, 139

Eugenia, 26

Euphorbia, distribution, 155
Evolution, 3, 21, 22, 41, 46, 50, 51,
65, 66-9, 89, 95, 100, 117, 175,
187, etc. ; at right angles to
natural selection, 117, 175, 187;
backwards, 22, 32, 65, 66, 68,
88, 98, 175; course under Dif-
ferentiation, 68, 69; de luxe, 21;
direction of movement, 67, 182 ;
downwards, 46, 65 ; in structural
characters, 4; mechanism of, 41,
89; must go on, 21 ; no longer a
direct expression of improving
adaptation, 95 ; not a matter of
chance, 187; nothing to do with
adaptation, 41 ; on mathemati-
cal lines, 50, 175, 178; plan, 51 ;
statistics, 50, 100; survival of
parent, 66; two theories dia-
metrically opposed, 89 ; and see
Test Cases, chaps, x-xiii
Extermination under natural selec-
tion, 69, 155; of parent, 4, 13, 46,
167, 173

Families, ditype, 78; monotype, 79
Family characters, 16, 106, 108, 143,

180
Follicles, T.C. xv, 124
Fossils, 12, 72
Frequency distribution, 10
Fungi, 21, 158

Gaps between larger genera in a

family, 97
Gene change with separation, 62
Genera, relative sizes, T.C. iii, 95
Generic adaptation, 18, 59, 107, 126,

141
Geographical distribution, 9, 24, 26,
39, Test Cases, 142; based on
adaptation by selection, 56, 68,
142
Geographical localisation of struc-
tural features, 123
Geological catastrophes, 73
Guppy, H. B., 16, 39, 66, 68, 74, 89,
132, 174, 175, 186

Halving of species in a family, T.C.

VIII, 101
Harland, S. C, 62
Hedyotis, 26
Hollow curve, chap, iv, 33; T.C. v,

99; 164, 173

INDEX

205

Hooker, Sir J. D., 9, 17, 47, 74
Huxley, T. H., 17, 74
Hybrid formation, 143, 189
Hydrocotyle, 58, 104, 145
Hydrophytes, 52

Increase in number with evolution,

T.C. I, 90
India, distribution in, T.C. xxx, 158
Infinitesimal variation, 10
Intermediates, 12, 16, 44
Island floras, 62
Isolation, 25, 26, 27, chap, vii, 61

Jeans, Sir James, 90, 175

Jenkin's criticism of Darwin, 5, 13,

25, 165
Jordanian species, 133

Keys to families, 77, 85, 137

Large families, position of, 136;
larger the family, the larger the
variety of conditions, 129

Large genera, 17, 26, 80, 93, 94, 97,
126, 134, 136, 163; among en-
demics, 26 ; divergent, 136 ; gaps
between, 97; origin, T.C. xvi,
126; position of, T.C. ii, 94;
successes, by selection, 93, 161 ;
supposed to be best adapted, 17

Likenesses of organisms, 1, 2

Limit to life of species, 72

Linking genera, 19, 155

Linnean species, 61, 132, 166, 171

Local adaptation of endemics, 147;
conditions have small effect, 55

Localisation of higher types, T.C.
XXV, 140

Lofgren, A., 62

Logarithmic curves, 35, 174, fig. on 37

Malthus, T. R., 3, 110

Mechanism of evolution, 41, 89

Menispermaceae, 152, and fig. 10

Mesophyt€s, 52

Mivart,'St G., 7

MolUnedia, 33

Monimiaceae, 33, 92, 136

Monocotyledons, 15, 43, 47, 130;
monocotyledonous mode of life,
15, 45, 47, 162 ; relation to Dico-
tyledons, 162

Moors, plants of, 44

Morphological test cases, 103

Musk, loss of smell by, 66, 72, 111
Mutation, 13, 20, 25, 41, chap, v, 43,
48, 59, 63, 89, 115, 129, 170, 172,
187, 189; differentiating, irre-
versible, hereditary, 43; large,
not seen, 48, 89; single, 59, 67,
172; smaU, 59

Natural Selection, 4, 5, 6, 13, 14, 15,
21, 22, 24, 25, 41, 45, 54, 57, 58,
69, 77, 78, 90, 97, 103, 106, 107,
109, 110, 114, 115, 117, 139, 153,
155, 166, 175, 177, 181, 187;
adaptation by, 54 ; assumptions
of, 4, 13, 54, 55, 107, 109, and
especially 167; destruction
under, 69, 155; difficulties, 10,
21, 30, 39, 48, 58, 77, 78, 115,
117, 139, 166; fascination, 103;
individual, 15, 110, 177, 179; no
room for operation, 25 ; not the
driving force of evolution, 177;
results in survival of fittest
population, not type, 166; work
too complex, 57

Nature red in tooth and claw, 6, 110

Nepenthes, 140

New Zealand, contours and condi-
tions, 154; distribution of en-
demics and wides, 29; propor-
tion Monocotyledons to Di-
cotyledons, 162

Opening of anther, 121
Origin of plants and animals, 2
Origin of Species, 3, 8
Overhead force acting, 20
Owen, Sir Richard, 7

Paeonia, 84

Parallel variation, 138

Parasites, 130

Parent and child, 50, 156

Perfection of characters, 45, T.C. x,

114, 178, 179-80
Pinnate leaf, distribution of, 105
Podostemaceae, 18, 19, 63, 141, 156
Polyalthia, 50
Polyphyly, 190
Portulaca in India, 159
Prediction, 89, 94, 97, 147, 190
Pure stand, 91, 125
Pyrenacantha, 130

Rafflesiaceae, 21

206

INDEX

Range large, structural difference

sometimes small, 44
Rank and Range, 100
Ranunculaceae, analysis of, 81, 135
Ranunculus^ 50, 70, 135, 150, 151,

and fig. 9
Regression, 10
Relative rank of genus and species,

68, 113; sizes of genera, T.C.

Ill, 95; value of characters, 119
Relics, 4, 17, 26, 30, 31, 61, 79, 81,

91, 93, 95, 113, 128, 132, 147,

160, 173
Restionaceae, 49, 139
Rhipsalis (endemic), 62
Ritigala, 24
Rubiaceae, unusual characters, 118,

178

St Hilaire, G., 65

Sedum, 54

Senecio, 126

Silene, 86

Silenoideae, 86, 136

Single steps to genus or species, 59,
117, 172, 187

Siparuna, 33, fig. 1, 34, 63

Size and Space, T.C. vi, 100

SmaU, J., 72

SmaD genera, failures (by natural se-
lection), 93 ; proportion in famil-
ies, T.C. IV, 97; satellites round
large, 81, 128, T.C. xviii, 128

Social legislation, 110

Solanum, 138

Special creation, 2, 164

Species assumed to fight as units,
107, 142, 144, 166, 179; formed
at one step, 20

Specific characters, 11, 109, 180

Staminal characters, T.C. xiii, 120

Statistics of continents, etc., very
uniform, 172

Sterility line, 12, 46, 143, 174, 189

Stewart, C. Balfour, 47, 182

Stratiotes, 64, 111

Strobilanthes, 26

Structural characters, 4, 8, 15, 44,
52, 103; no necessary adapta-
tional value, 109, 115; nothing
to do with Ufe, 54

Structural considerations override
adaptational, 110, 115, 120,
121, 129; difficulties for natural
selection, 127

Struggle for existence, 1, 3, 5, 22, 39,

49, 54, 106, 110
Styracaceae, 108, 161
Successful species, 24; genera and

species, 132
Super-plants 90

Surname distribution, 35, 39, 40, 99
Survival of the fittest, 165

Taxonomic axioms, 132; resem-
blances of geographically widely
separated plants, T.C. xxviii,
154; tests, 132

Tendencies, 120, 124, 138

Test Cases between the rival theories
(reviewed on p. 182):

I. Increase in Number with Evo-
lution, 90

II. Size of the Largest Genus in a
Family, 94

III. Relative Sizes of Genera, 95

IV. Proportions of Small Genera
in Families, 97

V. The HoUow Curve, 99

VI. Size and Space, 100

VII. Some Statistics of Evolution
and Distribution, 100

VIII. The Halving of the Species
in a Family, 101

IX. Differences in Generic Rank,l 10

X. ThePerfectionofCharacters,114

XI. The Early Stages of Cha-
racters, 115

XII. Alternate or Opposite
Leaves, 118

XIII. Staminal Characters, 120

XIV. The Berrv Fruit, 122

XV. Achenes and Follicles, 124

XVI. The Origin of Large Genera,
126

XVII. Some ^Morphological
Puzzles, 126

XVIII. The SmaU Genera, 128

XIX. Correlated Characters, 129

XX. The Position of the Largest
Genera in a Family, 134

XXI. The Position of the Large
Families, 136

XXII. Divergence of Variation.
Systematic Keys, 137

XXIII. Divergence from usual
Family Characters, 138

XXIV. Parallel Variation, 138

XXV. Greater Localisation of
Higher Types, 140

INDEX

207

Test Cases (cont.)

XXVI. Age and Area, 146

XXVII. Contour Maps, 149

XXVIII. Taxonomic Resem-
blances of (Geographically)
widely Separated Plants, 154

XXIX. Variety of Character with
Uniform Conditions, 156

XXX. A common Type of Distri-
bution in India and Elsewhere,
158

XXXI. Large (ienera the Most
"Successful", 161

XXXII. Characters the More
Constant the More Useful, 161

XXXIII. Relation of Monoco-
tyledons to Dicotyledons, 162

XXXIV. Overlap of Largest Ge-
nera in a Family, 163

Tetracera, 45
Thalictrum, 104
Transitions, 12, 16, 44,

143
Trimen, H., 18, 24
Tristichaceae, 21

Tropical Forest, 22
TurrUl, W. B., 63

Umbelliferae, 154
Unsuccessful genera, 128

Variation, differentiating, 11, 14;
inherited, 11; in sudden steps,
10; irreversible, 11, 14; of
character with uniform condi-
tions, T.C. XXIX, 156

Vaud, names in, 35, 40 (fig. 6), 148,
149

Vries, H. de, 14, 170

Wallace, A. R., 3
Weather effects, 55
Went, F. A. F. C, 104
Willisia, 19
Woolf, L. S., 6, 144

74, 125, Xerophytes, 52

Yule, G. Udny, 50, 90, 93, 100, 155,
173

CAMBRIDGE: PRINTED BY W. LEWIS, M.A., AT THE UNIVERSITY PRESS