e
manner, or of use and disuse. As every part of the organisation is highly
variable, and as variations are so easily selected, both consciously and
unconsciously, it is very difficult to distinguish between the effects of
the selection of indefinite variations, and the direct action of the
conditions of life. For instance, it is possible that the feet of our
water-dogs, and of the American dogs which have to travel much over the
snow, may have become partially webbed from the stimulus of widely
extending their toes; but it is far more probable that the webbing, like
the membrane between the toes of certain pigeons, spontaneously appeared
and was afterwards increased by the best swimmers and the best
snow-travellers being preserved during many generations. A fancier who
wished to decrease the size of his bantams or tumbler-pigeons would never
think of starving them, but would select the smallest individuals which
spontaneously appeared. Quadrupeds are sometimes born destitute of hair,
and hairless breeds have been formed, but there is no reason to believe
that this is caused by a hot climate. Within the tropics heat often causes
sheep to lose their fleeces, and on the other hand wet and cold act as a
direct stimulus to the growth of hair; it is, however, possible that these
changes may merely be an exaggeration of the regular yearly change of coat;
and who will pretend to decide how far this yearly change, or the thick fur
of arctic animals, or as I may add their white colour, is due to the direct
action of a severe climate, and how far to the preservation of the best
protected individuals during a long succession of generations?

Of all the laws governing variability, that of correlation is the most
important. In many cases of slight deviations of structure as well as of
grave monstrosities, we cannot even {420} conjecture what is the nature of
the bond of connexion. But between homologous parts--between the fore and
hind limbs--between the hair, hoofs, horns, and teeth--we can see that
parts which are closely similar during their early development, and which
are exposed to similar conditions, would be liable to be modified in the
same manner. Homologous parts, from having the same nature, are apt to
blend together and, when many exist, to vary in number.

Although every variation is either directly or indirectly caused by some
change in the surrounding conditions, we must never forget that the nature
of the organisati