proceed without saying a word of Weismann's posi-
tion in regard to these facts. In one of his essays,
On the Supposed Botanical Proofs of the Transmis-
sion of Acquired Characters, written a few years ago,
Weismann has discussed some cases which are similar
to some of those I have related. One of these cases,



222 EXPERIMENTAL EVOLUTION LECT.

brought forward by Detmer, is that of the shoots of
Thiija occidentalis. These shoots contain in their
upper side green palisade cells, while the under sides
possess green spheroidal cells. And when the branches
are turned upside down, and are fixed in this position,
the anatomical structure of the shoots which put in
their appearance later is reversed ; the side which was
destined to become the upper side, and which a change
in the branch's position has made the lower side,
assumes the structure of the lower side, and vice versa.
A similar case occurs with the climbing shoots of the
ivy, and in Tropaeolum leaves important structural
changes have been noticed by Detmer in response to
differences in external influences, or change of envi-
ronment. Now Weismann says, " Such differences [in
structure] do not by any means afford proof of the
direct production of structural changes by means of
external influences. How would such an explanation
be consistent with the fact that the leaves are, in all
these cases, changed in a highly purposeful manner ?
Or is it assumed that these organs were so constituted
from the beginning that they are compelled to respond
to external conditions by the production of useful
changes ? Any one who made such an assertion now-
adays, or who even thought of such a thing as a
possibility, would prove that he is entirely ignorant of
the facts of organic nature, and that he has no claim



iv WEISMANN'S CRITICISMS 223

to be heard upon the question of the transformation
of species." These are rather big words, and Professor
Weismann has perhaps written somewhat hastily. It
may be answered that all evolutionists, and more
especially " Natural Selectionists," to whom Professor
Weismann belongs, assume the production of " useful
changes " with or without change of external condi-
tions, since those only survive in the struggle for life
who offer beneficial modifications or adaptations.
And of course the production of such " useful
changes " is of much higher importance when the en-
vironment changes than when it remains unmodified.
Professor Weismann denies the importance and trans-
missibility of variations due to external modifica-
tions. But then how does he explain the fact now
repeatedly ascertained in all bacteriological labora-
tories that all micro-organisms, bacilli, bacteria, etc.,
undergo under cultivation in different external condi-
tionswhether of light, heat, or food, it matters little
such important modifications that they may be made
to lose their essential characters, and that these charac-
ters are lost as long as the external modification per-
sists ? Take Bacillus anthracis, for instance. Com-
pared with many other bacilli, it differs very slightly
in external characters ; the principal and all-important
difference is that it determines in many animals a
disease of a very precise character which can be mis-



224 EXPERIMENTAL EVOLUTION LECT.



taken for no other. But this bacillus, if you alter even
very slightly some of the external conditions it lives
in, gradually loses its most important character ; it is
deeply modified for the time being that is for the
time you compel it to live under particular conditions,
and perhaps even for some time the normal conditions
are restored. The usefulness or unusefulness of the
modification are not in the slightest degree apparent,
and to evolutionists it matters little, since modifications
may be useful, useless, or indifferent, or even injurious,
and natural selection destroys all that is not useful, or
at least indifferent ; but the main fact is that here is an
important modification which puts in its appearance
when some external conditions are changed, and dis-
appears when the normal condition reappears. Is envi-
ronment operative here, or must we assume that when
Bacillus anthracis is cultivated in some particular
manner, it loses its most important character without
any reference to the change of environment ? I dare
say no bacteriologist or physiologist could be met who
would venture to assert that the change of character
is not in direct relation with the change of external
conditions, whether the former is beneficial or not.

It seems, then, that Prof. Weismann goes certainly
too far when he asserts that we have no proof of
the direct production of transmissible changes by
means of external influences. It may be said that he



iv WEISMANN'S CRITICISMS 225

restricts his denial to the Metazoa ; but this is
assuming a physiological contrast to the unicellular
forms of life which it will not be easy to justify. On
the other hand, we possess, in the facts of domestica-
tion and cultivation, a large number of cases of
variation which occurs in every part due to environ-
ment, and transmitted by inheritance in various
degrees. Psychology affords similar instances : a
kitten which has never seen a dog is afraid from the
first moment it perceives one ; young birds of many
species instinctively fear the hawk and other birds of
prey, while remaining unaffected by the presence of
other birds. Are these not psychological " attitudes "
due to environment (acting on the mens of ancestors)
which have been transmitted by inheritance ; are
these not acquired characters / I would recommend,
in regard to this discussion, two recent papers : Mr.
J. A. Thomson's History and Theory of Heredity
(Proc. Roy. Soc. Edin., 1889), where the writer gives
his reasons for not accepting Weismann's extreme
views, and E. B. Poulton's Theories of Pleredity (Mid-
land Union of Natural History Societies, 1889). The
latter seems more favourably disposed to Weismann's
theory, which he has greatly contributed to spread
in England through his excellent translation of the
Essays on Heredity.



LECTURE V

Summary : Experimental Evolution based on the four preceding
Groups of Facts. These Facts illustrate at the same time its
Methods, which are : Change of Environment ; Use and Disuse ;
Natural Selection ; Sexual Selection ; and Physiological Selection.
These Factors of Evolution must all be subjected to Ex-
perimental Test in order to show what they can Effect. What is
wanted : A Direct Proof, which all may Perceive and Touch, of
one Species (or Form) giving Birth to another more or less
Different, and Permanent. Numerous Accessory Problems to be
Investigated at the same time. Scientific and Practical Import of
this Line of Investigation. Requirements : Farm and Laboratory ;
Animals and Plants ; Time ; Experiments must be able to last
20, 50, IOD Years or more. This Experimental Investigation
must and shall be performed. But who is to begin ?

WE have thus shown that a high degree of varia-
bility exists among animals and plants in the natural
state as well as under domestication, and that through
the modification of environment, in part or in whole,
we are able to determine some changes in organisms.
This variability is met with in all parts of the body,
even in those which seem to be the most permanent.

What does this demonstrate ? it may be asked.
And the objection arises : What does it matter that



LECT. v USE OF FOREGOING FACTS 227

variability occurs in all animals and plants, even to
a large extent ? Of course, if all parts of any animal
or plant were to vary to the extent which has been
shown, variability might explain the production of
new species ; but do they really vary to this extent
in any one individual or group of individuals ?
Certainly not, and we must admit that variability is
limited, and that it is not, so far as we know, sufficient
to create new species. The differences are not
numerous enough.

What, then, is the use of the series of facts which
have been examined, and what do they show ?

They provide a basis for the study of variation,
specific or otherwise, in showing that no species are so
very permanent in their structure or functions that no
departure from their type is possible. They form the
solid ground on which evolutionists stand, and
if this ground were missing the evolution hypo-
thesis would have no support at all, and would be
nothing more than an aerial and unfounded structure.
In the second place, they provide the basis and suggest
the methods for experimental transform ism.

Believing as we do that transmutation or evolution
must have taken place under the action of natural
causes and influences, we consider that these causes
have been natural selection and environment, in pro-
portions which we cannot determine. But these causes

Q 2



EXPERIMENTAL EVOLUTION LECT.



cannot have had any influence if the species do not
admit of some variability ; and this very variability
itself we consider as being in some degree the measure
of their operativeness and influence.

We are thus impelled to conclude that transformism
admits of experimental investigation, and this is the
main point we wish to establish. If the present species
have really originated from the more or less closely
allied species which have lived in the past, if the
present has really been evolved out of the past through
natural agencies, without any special intervention
of any force, we do not see why there might not be in
the future forms evolved out of the present, and why
we could not evolve them ourselves, in part, and help
towards their production, through the use of the
methods which we believe to have been used by Nature
herself. If we do not succeed, we are either mistaken
in our general idea, or mistaken in regard to the
methods through which evolution is supposed to have
taken place ; but we can really draw no conclusion at
all as concerns these methods, as long as we have not
subjected them to experimental test.

I do not propose to show all that can be done in
this line ; the matter would require more time than I
can devote to it, and on the other hand I am firmly
convinced that much is to be done of which we have
at present no idea at all. As Dareste rightly says :



FACTORS OF EVOLUTION 229



" We may rest assured that the execution of experi-
ments will cause a great number of questions to arise,
of which we can at present have no idea. This is one
of the great advantages of experiment. If we do not
always find what we are seeking after, while starting
from hypotheses which the facts do not support, we
often find what we were not looking for, and light is
thus cast on regions which till then seemed buried in
complete darkness."

This I consider as exactly true, and many unexpect-
ed questions will certainly turn up of which we have
no idea, while answers to others may also be found.
But, if it is impossible to state exactly what will be
done, we may at least gather some idea of the principal
methods and hints of experimental transformism.

The methods first.

What can the methods of experimental trans-
formism be ? The only answer to this question is
based on the consideration of what the Factors of
Evolution are or are supposed to be. At the present
moment five are usually recognised. 1 I quote from
Le Contc's able paper of recent date :

" First. Presence of a changing environment affect-
ing functions, and functions affecting structure, and the
changed structure and function inherited and integ-
rated through successive generations indefinitely.

1 Cf. Herbert Spencer's Factors of Organic Evolution, and Le Conte's
The Factors of Evolution in The Monist, April, 1891.



230 EXPERIMENTAL EVOLUTION LECT.

" Secondly. Use and disuse of organs reacting on
growth-force, and producing change in form, structure,
and relative size of parts, and such change inherited
and integrated through successive generations.

" Thirdly. Natural selection among individuals of
a varying progeny, of those most in accord with an
ever-changing environment, or, as it has been other-
wise called, survival of -the fittest in each successive
generation.

" Fourthly. Sexual selection. The selection of females
among varying male individuals, all competing for the
possession of the strongest or the most attractive.
Among mammals the selection is mainly of the
strongest as decided by battle ; among birds, of the
most attractive, as determined by splendour of colour
or beauty of song.

" Fifthly. Physiological selection, or selection of those
varieties, the individuals of which are fertile among
themselves, but sterile or less fertile with other varieties
and with the parent stock. This has also been called
segregate fecundity by Gulick, and homogamy by
Romanes."

These five factors are not all recognised by the same
group of evolutionists. The two first factors are
Lamarckian ; the third and fourth are Darwinian ; the
two first are supposed to operate during individual
life ; the third and fourth operate on the offspring,



v FACTORS OF EVOLUTION 231

which has more or less departed from the parental
type. For some time an important discussion has
been carried on, especially in England, Germany, and
the United States, concerning these factors between
Lamarckians and Darwinians, as to their efficiency
and frequence.

I cannot enter upon the discussion, which would
require much time, and in fact it might seem rather
early yet to discuss the quoinodo of a fact whose exist-
ence is not proved to the satisfaction of all whose
opinion is of any weight. Lamarckian views arc held
especially by American and some French evolutionists,
while in England and in Germany strict Darwinian
theory prevails. It must be said, however, as concerns
the Lamarckian theory, that, as Lc Conte has well
remarked, the Lamarckian factors of environment and
use and disuse, arc the most fundamental in importance,
and first in order of appearance. Selective factors are
conditioned by reproduction, and on sexual reproduc-
tion particularly in which the characters of two diverse
individuals are blended in different proportions in the
individuals of the same progeny; sexual generation
thus provides material for selection to operate upon.
But where no sexual generation occurs and this is the
case with the lower forms of life which were first evol-
ved, and out of which the higher forms are supposed
to have developed Weismann urges that selection is



232 EXPERIMENTAL EVOLUTION LECT.

impossible, since the progeny is nothing more than a
part of the parent form, an outgrowth thereof, the two
being so very similar that in most cases none can tell
which is the parent, and which the progeny. It then
results if really no selection can occur when sexual
reproduction is wanting, and this is a matter which is
not settled by Weismann's very sweeping assertion,
with which I cannot concur, as there is no reason
why some degree of variability should not exist in
unicellular organisms as well as among multicel-
lular plants or animals that Lamarckian factors
must have operated, and must operate even now, if
evolution has existed from the beginning, and has
been carried on through natural agencies. Weismann
strongly argues that at present, at all events, or as
concerns higher animals, the Lamarckian factors are
possessed of no influence, and his essays on heredity
are all against the heredity of acquired characters.

Such is the present state of the question. I have
no intention to discuss the matter, as I have already
said ; but I think it may be in the future discussed in
a much more profitable manner than has been done till
now. And this may be effected through experimental
transformism. If we are to subject the evolution
theory to the test of experiment, we can only do so by
the investigation of the efficiency of the factors of
evolution, and we must subject them to the said test.



v PROPOSED EXPERIMENTAL INVESTIGATION 233

This method cannot fail to be highly profitable to the
discussion ; I do not think of any other that, at present,
can settle the matter ; and if we are to know some-
thing some day about the general fact of evolution,
and the methods through which it has been going on,
and may in future go on, it is through experiment
that our knowledge will be acquired.

Whatever our opinions may be as to the real value
of the factors of evolution which have been suggested
on different sides, all must be subjected to the same
test, that of experiment : the results will allow us to
decide upon the theory itself, and upon the details of
the process.

Such is the general view. As to details, now, I
must confess that we are rather in the dark as yet. At
first we shall have to grope about somewhat, search-
ing for the ways in which the experiments may be
performed, and for suitable organisms upon which we
may experiment. Every one of the recognised factors
must be investigated.

Concerning environment, we may operate on many
sorts of animals and plants. A first and simple
method will consist in transferring animals and plants
from one country to another, or from mountains to
plains, or vice-versa, from dry to moist soil, from cold
to warm, from calcareous to siliceous soil, from one
pond in one sort of soil to another pond in another



234 EXPERIMENTAL EVOLUTION LECT.

soil, from light to semi-darkness, from land to water,
&c. The experiment may be performed in a thousand
ways, and all external differences, all changes in
environment which seem to operate on organisms
may be successively tested. Care must be taken,
however, to watch the experiment without interfering,
and the animal or plant must be left wild, without
domestication or cultivation, so that they are exactly
in the condition of a species transferred from one set
of conditions to another without being particularly
helped to support the change. Other experiments
may be performed in a different manner. Instead of
altering slightly all conditions of environment as in
the preceding case, we may alter only one : for
instance, while keeping the animal or plant in its
native climate, alter the proportion of chemical com-
ponents of the soil, alter the nature of food, add some
new compound to the one or the other, increase or
decrease motion around it, &c. All the facts men-
tioned as illustrating the influence of environment
show how numerous and varied are the experiments
which may be performed, and it is needless therefore
to repeat what has already been said on the matter.
If the variations of environment have really effected
the result noticed, such result must be again obtained
by direct experiments.

The fact is however, that in such experiments, the.



v PROPOSED EXPERIMENTAL INVESTIGATION 235



difficulty lies not in the experiment itself, but in
appreciating the results. Such results are not always
external and obvious ; many are internal and require
chemical and microscopical investigation in minute
details, and such differences in chemical constitu-
tion or in structure may have a great influence in the
struggle for life and operation of natural selection.

In reference to use and disuse, experiments may be
made to diminish or suppress the activity and use of
some organ, by keeping plants or animals in such con-
ditions as to render some character useless. For
instance, one might try to obtain unscented or plain
flowers through artificial fecundation of all the flowers
of the same plant which require insect-intervention,
as it is supposed to have developed scent or colour in
view of attracting insects. Or again, place any animal
in such conditions as to render any one function use-
less ; or also, such as to render the development of
some organ or function of great use and necessity.
Experiments on the inheritance of mutilations may
be repeated at the same time : those which have been
already performed have not been successful nor
sufficient, although a priori it seems most likely that
the result will be exactly what it has been. For
such and other experiments bearing on the question,
intending experimenters may be referred to the essays
of Weismann, and W. P. Ball's Effects of Use and



236 EXPERIMENTAL EVOLUTION LECT.

Disuse, where they will see what has been done and
what may be attempted. Care should be taken to
operate preferably on useless characters. But are
there any useless characters ?

It is a rather curious fact that while the operation of
selection is recognised by most evolutionists, even if
holding Lamarckian versus Darwinian views, but few
experiments, as such, have been yet performed,
although observations are plentiful. Among the best
which have yet been made, I must refer to those
of Vilmorin, 1 performed many years ago, and
published for the first time in the Transactions oj
the Horticultural Society in 1840 (2nd Series, vol. ii.,
p. 348). M. de Vilmorin, considering that most of
our food-vegetables are derived from species which
have been altered by man, and that the most interest-
ing point to investigate is the methods through which
the alteration has been obtained, notices the fact
familiar to all, that while species which have been
a long time under cultivation vary easily and in many
directions, those which have been less cultivated, or
have not been cultivated at all hardly exhibit an)'
variation. Such has been the case with this writer
in his experiments on Lactuca perennis, on Tetragonia,
on Solanum stoloniferum, on Brassica orientalis. But

1 See his Notice sur f Amelioration de la Carolte Salvage in Notices
sur t Amelioration des Plant es par la Cuttitre, Paris, 1886.



EXPERIMENTS ON SELECTION 237



in the case of the wild carrot circumstances have been
quite different, and through selection, artificial of
course, he has obtained very precise and interesting
results which it may be useful to quote here.

In 1832, M. de Vilmorin, wishing to obtain from the
wild carrot plants with thick and edible roots, planted
some seeds of the wild plant. All the plants thus
obtained grew quickly and yielded seed, while no
root was any better than that of the common wild
carrot. He began again in 1833, and among the seeds
planted, many were late in germinating and no seed
was produced, while some roots were somewhat larger
and thicker than usual. These roots he selected and
put apart so as to plant them in the following spring,
and they yielded seeds in 1834. The seeds were
again planted in 1835. Many gave plants with the
ordinary wild carrot root, but a rather large proportion
(-J-th) yielded plants with thicker roots. The seed of
these plants was selected, and planted in 1836. Selec-
tion again was performed, so that in 1837 many good
roots were obtained. In 1838 and 1839 the process
was continued, with the result of yielding a large pro-
portion of satisfactory carrots ( T %ths). While acquiring
different dimensions, the roots acquired also unusual
colour : yellow, lilac, and even red.

Here we have a good instance of the selective
process and of its influence and operation. Another



238 EXPERIMENTAL EVOLUTION LECT.

is yielded by experiments on the beetroot, performed
some years later by the same writer, with the view of
obtaining a variety of this plant containing more
sugar than is commonly the case. 1 It is worth
recording, as it shows that through selection it is
possible to influence physiological variability, and the
result has been to increase the proportion of sugar
from 10 on an average to 12, 14 and even 16 per cent.
M. de Vilmorin notices a fact which it is well to state
here, when he says that it is better to select seeds from
plants belonging to a group with high average than
from plants yielding high maxima but also low
minima.

A large number of facts from observation confirm
these results of experiment. Our domesticated
animals, our cultivated plants have been made to yield
so many varieties simply through selection. While
cultivation or domestication increases the tendency
towards variation as we all know, selection of
variations has led us to produce quite a number of
varieties of which we make use in very different
manners, because different variations have in turn
been selected according to the particular wish of
the selectors, or to the peculiarly interesting nature

1 Note sur un Projet if Experience ayant pour but tfaugmenter la
Richesse saccharine de la Betterave. Loc. cit. (1890), and Note sur la
Creation June nouvelle Race de Betterave a Sucre. Loc. cit. and Comptes
Rendus, Nov. 1856.