dissolves very slowly and incompletely ; the greater
part does not become transformed into syntonin, and
in this respect it resembles the flesh of calves, veal
being very refractory to the action of dilute hydro-
chloric acid.

I may have been dwelling rather a long time on
M. Gautier's experiments, but they seem to me very
interesting and suggestive. Of course we do not yet
know anything about the cause of variation, but it is a



II CAUSES OF COLOUR-VARIATION 67

great deal to have some information concerning the
phenomena which accompany it, and especially the
chemical and biological phenomena, and it is import-
ant also to be able to detect such phenomena even in
cases where no external or morphological differences
are obvious. Later on, I shall have to revert to these
matters, and for the present it is sufficient to show
that even in cases where the external variations seem
so weak as to be generally considered of no real
account by naturalists, very marked chemical varia-
tions underlie the slight morphological differences.

Upon the whole, then, if a steady and careful
investigation were made of the chemical differences
which accompany colour variation, doubtless we should
see that the latter is of great importance.

Concerning the causes of colour variation very
little is known. Mr. A. R. Wallace has given a
detailed discussion of the matter in his recent book on
Darwinism, and emphasizes the protective value of
colours and their consequent relation to natural
selection. This factor has doubtless been of much
importance, but on the other hand, as Wallace indeed
admits, colour variations have much to do also with
the general constitution, and some relationship between
vigour, or weakness, and colour is commonly recog-
nized. In 1823, Heusinger already contended that the
quantity of colour in animals is subject to two laws :

F 2



68 EXPERIMENTAL EVOLUTION LECT.

the first of which is that colour is more abundant
when the genital functions are most active ; and the
second, that colour is also in relation with the fatty
layer underlying the skin, being less abundant when
this layer is thicker, and vice versa. Recent observa-
tions, and the general experience of breeders, go to
show that there is much truth in this view, as the
breeders have remarked that the best breeds for flesh
and fat the Southdown for instance have less colour
than the other breeds of the same origin ; and the same
obtains among cattle, and among domestic birds,
where the white, or faintly coloured individuals are
known to be less vigorous than the others. It is also
commonly observed that the cause which induces
albinism generally affects reproductive functions in a
marked manner, and that partial or total sterility is
induced : in fact one may say that the absence of
pigment and the sterility are correlated expressions
of the same constitutional change. Such is the case,
for instance, with the Fredericksberg breed of horses.
These horses are very peculiar in their colour, which is
pure white, although they are not albinos. They are
sterile inter se, and in order to preserve the breed,
which is used by the royal family of Denmark, they
are mated with grey or black mares, and among the
progeny pure white horses are obtained. The sterility
between like and like is not complete, as in some cases



ii VIGOUR AND COLOUR 69

a progeny is obtained, but the latter is deficient and
lacks vitality. 1 As a last proof of the relation between
vigour and colour, we may notice the fact that accord-
ing to many authors, among them Settegast, Heusinger,
and Wyman, animals deeply coloured, animals of
black coats for instance, are well able to withstand
the influence of poisonous plants, which kill others
less coloured. There thus seems to exist a positive
relationship between vigour and colour, and this
cannot be wondered at, since colour is the result of
chemical processes which are carried on in the organ-
ism, and since this metabolism certainly varies in in-
tensity and rapidity in different individuals. Other
agencies, which may operate on vigour, exert an
influence on colour ; light is such an agency, but facts
seem somewhat contradictory, and we cannot wonder
at this when we consider that while light varies, other
influences, such as heat, may not vary, or may vary
in a different direction. The climate is also operative
although we can certainly not give a precise defini-
tion of this term and an interesting case is that of a
herd of Dishleys living in the vicinity of the sea, in
France. These Dishleys were always spotted on the
face and ears, and had large black spots which rather
discouraged buyers. The owner said these spots were

1 Tissernnd : fitudes economiqiies sur le D<nte'>wfc, le Hohtein, ft It

Slesivig.



70 EXPERIMENTAL EVOLUTION LECT.

the result of the climate of the sea-shore, and in fact,
, the descendants of these spotted animals bred inland,
at Lyons for instance, were always pure white. 1

So much, then, for colour variation. But animals
and plants vary in many other respects, and one of
the most familiar examples is that of dimensional
differences.

I am not aware that any naturalist has said
-unless in current English or French, happily not
in a Latin aphorism that dimensional differences
are of small importance ; but I presume many
think so. Variations of this sort are very frequent,
and although it is possible, by selection of extreme
variations, to create varieties of giant or of dwarfed
plants for instance, as any horticulturist may testify,
one might rightly consider this sort of variation as
more secondary than most others, if it could not be
shown that, while dimensions vary, other variations are
present at the same time, which may very well be of
high importance. These dimensional variations arc
in a large measure correlated with external influences.
Darwin has shown how, among horses, the dimensions
decrease in northern latitudes, in islands, and on moun-
tains. Man also is smaller in extreme northern climates,
and all the organisms of extreme northern parts
tend to be small. It may be, as Alcide d'Orbigny

1 Fact quoted by Cornevin : Traitt de Zootechnie, p. 278,



DIMENSIONAL VARIATIONS 71



believes ( Voyage dans ? Amerique meridionale, t. iv.),
that cold on one hand, and decrease of pressure on the
other, exert an unfavourable influence on growth.
But certainly food has a great deal to do with dimen-
sional variations. When food is abundant, and easy
to get, animals and man are prosperous and attain
large dimensions, while when it is scarce they remain
smaller. Japanese horticulturists rely in part on this
influence of the scarcity of food in their process for the
dwarfing of plants. Most persons have seen or at
least heard of these diminutive plants of theirs,
mostly conifers, such as Thuja > Juniperus, etc., which,
while aged 40, 60, 80, 100, or 150 years, are often
much less than a yard high, although their relative
proportions are well preserved, so that when you look
at them it is exactly as if you were looking at a normal
large tree through the wrong end of a glass. These
dwarfs are the result, in part, of mechanical processes
which prevent the spreading of branches, and in part,
of a starving process which consists in cutting most
roots, and in keeping the plant in poor soil.
Many of these Japanese dwarfs may be seen in
Europe, and they well illustrate the influence of
external conditions on growth and dimensions. Num-
erous instances show that plants or animals transferred
from unfavourable to favourable conditions, or vice
versa, acquire larger dimensions, or, on the contrary



72 EXPERIMENTAL EVOLUTION LECT.

become smaller. Such 'differences may be noticed
among all sorts of animals, from the highest to the
lowest. Gerard has noticed that bees transferred from
Burgundy to Bresse become larger in a generation ; A.
H. Curtiss has seen, in some places near the Potomac,
Bidens cernua acquire a height which is six times
the common average height of this plant, and he has
seen the same in Oxalis stricta ; C. Lemaire states in
D'Orbigny's Dictionary that, while cultivated hemp
grows no higher than a metre and a half in France, in
Piedmont it attains three and four metres ; and if
Italian stock is planted in France it rapidly reverts to
the small variety, in the course of two or three
years. Speaking of horses and of their dimensional
differences according to climate and environment, De
Quatrefages expresses himself as follows : " These
contrasts may be interpreted as due to the influence
which must be exerted on the first-named [Corsican
and Pyrenean stocks] by the stimulating and dry air
of the mountains, the frugal food with which they
must often be content, and, doubtless also, the hard
exercise which is rendered necessary by the roughness
of the soil. The others, on the contrary, [he refers to
the large heavy horses of the Bresse province,] always
immersed in a moist and heavy atmosphere, over-fed
with watery plants, and having none but easy work to
perform, must surely feel the effects of an environment



ii DIMENSIONAL VARIATIONS 73

whose influence exerts itself even on plants." That
such is the case, and that the influence of environ-
ment on dimensions is a very direct one, is amply
shown by the results of a change. Horses and oxen
become larger when transferred from Brittany to
Normandy, while the reverse happens in the reverse
case, for when some oxen were sent from Poitou to
Brittany, at the third generation the first named race
had acquired all the characters of the Breton stock.

Generally speaking, insular animals are smaller f
than their continental congeners. In the Canary
Islands the oxen of one of the smallest islands are
much smaller than those of the others, although all
belong to the same breed, and the horses are also
smaller, and the indigenous inhabitants are in the
same case, although, belonging to a tall race. It
would seem that in Malta elephants were very small
fossil elephants of course and that during the
Roman period the island was noted for a dwarf breed
of dogs, which was named after their birthplace, ac-
cording to Strabo. In Corsica also horses and oxen
are very small, and Cervus corsicanus> the indigenous
deer, is quite reduced in dimensions, although, accord-
ing to Polybius, this species was imported from
Europe 2,000 years ago, which makes it a descendant
of our Cervus elaphus ; and lastly the small dimensions
of the Falkland horses imported from Spain in 1764



74 EXPERIMENTAL EVOLUTION LECT.

are familiar to all. The dwarf rabbits of Porto
Santo described by Darwin may also be cited as a
case in point Dimensional variations in wild animals
are very numerous, and Locard (Etudes sur les Varia-
tions malacologiques a" apres la Faime vivante et fossile
de la partie centrale du Bassin du Rhone) notes among
a large number of similar cases, the fact that many
molluscs land or water common to France and
Algeria, are much larger in Africa, where their dimen-
sions are double those of their European con-
geners. 1 Isidore Geoffrey Saint Hilaire says that
Lymncea stagnalis is* much larger in ponds than
in rivers. Moquin-Tandon notes that in the same'
country the same species of molluscs exhibits im-
portant dimensional variations, and he has seen
Bulimus decollatns nineteen times larger in Africa
than in Europe.

Through careful selection these dimensional varia-
tions may become permanent, especially if no change

1 Such is the case particularly with Helix aspersa, vermiciilata,
lactea, melanostoma, Leticochroa candidissima, Fhysa contorta, and
many others. And when Leucochroa, for instance, is transferred from
Algeria to France it does not acquire a length of more than one centi-
metre, while in its African home it is two or three centimetres long.
Cf. Locard : L 'Influence des Milieux sur le Developpement des Mol-
lusques. Societe d 'Agriculture ', Histoire Naturelle, et Arts Utiles de
Lyon, 1891. It has also been issued in pamphlet form by J. B.
Bailliere, Paris, 1892. A large amount of facts of French origin are
quoted in this valuable contribution to the subject, and the author is
one of the leading malacologists.



II DWARFING AND STERILITY 75

occurs in the environment/ M. A. Roujon, of
Clermont-Ferrand, 1 by selecting the central, smaller
seeds of dwarfed plants of Helianthus annuus, Calen-
dula arvensis, and Zea rnais^ has been able to obtain
very small individuals of these three species. But the
most important fact, among those he has observed,
is that while the dimensions of the plants decrease,
their fertility is much impaired : the number of seeds
which are produced becomes smaller, and dwindles
down to 4, 3, 2, i, only, and finally no more seeds are-'
produced : a condition of absolute sterility is induced.
This concomitant sexual variation is of great import-
ance, of course, in showing that when dimensions vary
they are not alone variable ; there are other variations
which accompany the differences of dimensions.

One fact must be noticed concerning the point
which is now under consideration. It is the fact that
while we can easily, through a number of methods,!
induce unfavourable conditions, it is much more
difficult to induce favourable circumstances which,
lead to a better development. The advance of know-
ledge, however, may be expected to yield results
which shall prove more satisfactory, but we perceive
the difficulty of progress through the difficulty we
experience when we wish to maintain any natural or

1 De quelques Variations considerables observees chez les Vegetattx.
Journ, d'Ifist, Nattirelle de Bordeaux, t, iii., 1884, p. 156.



76 EXPERIMENTAL EVOLUTION LECT.

artificial race at its highest standard, and we all know
how readily degeneracy interferes with and ruins the
work of man or nature.

Dimensional variations, although very considerable,
cannot be regarded as unlimited. We cannot expect
to make any species of plant or animal become much
larger or much smaller than it is. Of course there are
natural or artificial conditions Under which all species
acquire a better development, and many facts display
this. But we cannot expect to be able to increase
the dimensions of any species beyond a certain point.
Such an increase would require numerous variations in
all the systems of the organism, stronger bones for
instance, a stronger heart, and so on. 1 And then, on
another side, giant forms would require so much food
that their number could never become very large, and
in fact, much goes to prove that such forms would
have much trouble to compete with others, while the
smaller forms could more easily live and maintain
themselves. So there certainly exists a limit to the
increase of dimensions a physiological limit which
cannot be passed without danger to the organism.
Conversely, there is also a limit to the decrease of
dimensions. Too small animals or plants are too

1 Paul Bert (Sur le Maximum de Tailk que puissent atteindre les
Animaux Vertebres : Soc. de Biologic, 1878) considers the maximal
dimensions of vertebrate animals as dependent upon the strength of the
cardiac muscle.



n EXPERIMENTS ON STARVING 77

weak to thrive unless considerable variations also occur
in their mode of life ; or their fertility may be very
much impaired, and thus the species is liable to go to
ruin. Thus it seems that, as things are, the condition
of every species including under this word condition
the state of all parts of the organism is exactly what
it should be to meet the present external circum-
stances, and departures from this condition are possible
only when necessary to the species itself, through a
change in circumstances. Concerning decrease in
dimensions, we may note that while a continuous
decrease must surely end in death, there are cases
where a large loss may be sustained without bringing
about this result. I made some experiments on this
point, a few years ago, and obtained the following
results. Wishing to ascertain the loss of weight
which animals are able to sustain without losing their
life, I weighed a number of Invertebrates, crabs and
medusae among others, and kept them in pure sea-
water without any chance to get anything to eat,
although I have reason to suspect some more enter-
prising individuals did eat some of their brethren.
But many mishaps befell this experiment, in one way
or another the course of true experiment seldom runs
smooth and at the end of a fortnight most of my
animals were gone, so that, in order to prevent a
complete disaster, I preferred stopping the process and



78 EXPERIMENTAL EVOLUTION LECT.

taking note of the results. Of all my animals only
two were left : two Aurelia aurita, a species of medusa
common on the Mediterranean coast. I had originally
put three in the aquarium, but one had died. The three
I had begun with weighed 98, 82, and 57 grammes at
first. At the end of the fortnight the two remaining
famished creatures weighed but 25 and 13 grammes.
Assuming for accuracy's sake, and in order to prevent
an over-estimation of the result, that these two were
those which weighed originally 82 and 57 grammes,
we see that the loss has been at least two-thirds in
one case, and three-quarters in the other. 1 This loss is
very considerable, for, as Chossat has shown in his
investigations on the effects of inanition, mammals die
before they have lost half of their original weight.
When the experiment was interrupted, my Aurelia
were in good condition, and seemed quite inclined to
live longer : in fact they did live. This experiment
should be repeated with Beroe ovata, or some other
species of this genus, for I have noticed that these
animals rapidly lose in dimensions when in captivity

1 Henry de Varigny, Bemerkung iiber den Gewichtsverlusl durch
Nahrungsmangel bei Aurelia aurita. Centralblatt fiir Physiologic,
12 November, 1887. Of course it must be said that in this case the
greater proportion of the loss of weight is due to loss of water, since
water is in such animals even more abundant than in higher terrestrial
organisms. But it must be noticed that even if the loss of weight is
especially due to loss of water, the latter is due to the loss of organic
tissues or substances with Which the water Was combined.



SEMPER'S EXPERIMENTS 79



with little to eat. Of course it is quite natural that
organisms which do not eat become smaller, and if
under-fed as a rule they must doubtless remain of
inferior dimensions. But there are cases where, not-
withstanding abundant food, animals are unable
to grow to their accustomed dimensions. Herbert
Spencer says that it is well known by all anglers that
trout and other fishes are small in small streams, and
large in larger rivers, and many naturalists are of the
same opinion. Is it that these animals remain small
because they get less to eat ? or is there some other
reason ?

I have also made some investigations on this
subject during the past two or three years, and may
be allowed to recall them. The starting-point of
these investigations was the fact announced by Karl
Semper some twenty years ago, in a special paper
on the matter, which he has since abstracted in his
Animal Life, that if the common pond snail is kept
in small volumes of water, of less than five or six
litres, the animal does not attain its regular develop-
ment, and remains more or less dwarfed. For instance,
if three young pond-snails (Lymncea stagnalis, or
L. auricular -ia), of the same brood and age, are
put respectively into aquaria containing 500, 1,000,
and 3,000 cubic centimetres of water, a difference
in their dimensions may be detected even after a




UNJVEESIIl



8o EXPERIMENTAL EVOLUTION LECT.



few days, and if the experiment is allowed to last
some months, we finally see that the inhabitant of the
largest volume of water is the largest in all ways,
that of the smallest being smallest, and that of the
intermediate aquarium being between the two as
concerns dimensions. Such is the general fact. But
many points are to be considered when we try to
explain it. The first explanation which suggests
itself is that in the larger space there is more to eat,
and that the pond-snails in small aquaria remain
small because they cannot secure food enough. This
objection and explanation are amply met by the fact
that in all my experiments care was taken to provide
superabundance of food in the form of aquatic plants,
and that the animals, whether in small or large
aquaria, had always at their disposal, a much larger
quantity of food than they could possibly eat, or than
they really did consume. So this explanation cannot
stand. Prof. Semper has thought of a curious
interpretation. He supposes that there exists in
common water some matter which, while not possessed
of nutritive properties, is conducive to growth and
development, and is a sort of incentive to both. If
the animal lives in a small body of water it has but
a small quantity of this matter at its disposal, and
does not grow as much as an animal in a larger
quantity of water. This interpretation is contradicted



ii SEMPER'S EXPERIMENTS 81

by a very simple experiment. Take two equal
volumes of water, 1,000 cubic centimetres for instance,
and put one of them into a broad and shallow basin,
so that it extends over a large surface, while the other
is poured into a spherical glass vessel, so that the
horizontal surface is very small. The two volumes
are equal, but their form is quite different. Into each
vessel, with an abundance of aquatic plants Myrio-
phyllum and Elodea especially : always submerged
sorts, so that they are not in need of a large surface,
and cannot interfere with it put one pond-snail
of the same brood, or cluster of eggs, recently
hatched. The difference after a few days is sur-
prising, and in the course of time it is seen that
the pond-snail of the large-surface vessel is much
larger than the other one. As the volume of water is
the same in both cases, we must conclude that in itself
the volume is not that which determines the variations
of growth, and also that Semper's interpretation
cannot be accepted, for, whether spherical or wide-
mouthed, the same quantity of the same water should
contain the same amount of Semper's hypothetical
matter.

If, then, we cannot admit Semper's explanation,
what is the cause of the observed facts ? This
question may be answered by new experiments, in
which various conditions may be made to vary at

G



82 EXPERIMENTAL EVOLUTION LECT.

will and in different known degrees. In my first
series of experiments I used equal volumes of water,
but with different surfaces. One of the volumes, for
instance, was poured into a large-surfaced vase of
fifteen inches diameter, while the other was poured
into a vase of only four or six inches diameter. In
such cases I always found that the animals living in
the large-surfaced vase became much larger than the
others. Why so ? Is it that the water has better
aeration in the large-surfaced vase ? But this is of
no account at all. In the first place I would call
attention to the fact which I have repeatedly ob-
served since I began this series of experiments,
(and of which I am at present a daily witness),
that the aquatic plants which I used in my experiments
(Myriophyllum and Elodea canadense] do positively
thrive and grow much better in narrow-surfaced
vases than in large-surfaced vessels, in spherical
glass balloons with a long neck, in which the water has
but a very meagre surface contact with the atmosphere
(two centimetres diameter for instance), than in twenty
centimetres diameter vessels. This shows certainly
that in spherical vessels, with small surface, aeration
must be very good. On the other hand there is no
reason to think that aeration is better in one case
than in the other, as the water contains a large
amount of plants which ensure good aeration and



II THE AUTHOR'S EXPERIMENTS 83

(in the last instance) the pond-snails care nothing
whatever about the aeration of water, since they
are not gill-bearers but pulmonated ; they breathe at
the surface, and do not breathe the air contained in
the water. So aeration has nothing to do with the
matter. It has so little to do that I have, in some
experiments purposely devised, been able to see that
pond-snails live exactly as well in two identical
vessels (identical in shape, surface, volume of water
and amount of aquatic plants) one of which remains
open, in contact with the atmosphere, while the other
is stopped by a paraffined cork, the amount of air