460 TROPICAL NATURE IX 
botanists, farmers, gardeners, sporting-men, pigeon-fanciers, 
travellers, and any one who could possibly afford direct per- 
sonal information on any of the matters he was investigating. 
Then came his own observation and experiment, to fill up gaps, 
to settle doubtful points, or to determine questions the import- 
ance of inquiring into which no one had ever suspected ; and 
lastly, there was the power of arrangement and comparison, 
the originality and depth of thought, which drew out from this 
vast mass of heterogeneous materials conclusions of the highest 
value as bearing on the question of the possible change of 
species, and the means by which it had been brought about. 
In order to determine the nature and amount of the varia- 
bility of domestic productions, he prepared skeletons of all 
the more important breeds of rabbits, pigeons, fowls, and 
ducks, as well as of the wild races from which they are 
known to have been produced, and showed, both by measure- 
ments and by accurate drawings, that not only superficial 
characters, but almost every part of the bony structure 
varied to such an amount as usually characterises very dis- 
tinct species or even distinct genera of wild animals. Another 
set of experiments was made by crossing the different breeds 
of pigeons and fowls which were most completely unlike the 
wild race, with the result that in many cases the offspring 
were more like the wild ancestor than either of the parents. 
These experiments, supported by a mass of facts observed by 
other persons, served to establish the principle of the tend- 
ency of crosses to revert to the ancestral form; and this 
principle enabled him to explain the interesting fact of the 
frequent appearance of stripes on mules, and occasionally on 
dun-coloured horses, on the hypothesis, supported by a mass 
of collateral evidence, that the common ancestor of the horse, 
ass, and zebra tribe was a partially striped and dun-coloured 
animal. 
A number of very important conclusions were deduced 
from the facts presented by domesticated animals and plants, 
a few of which may be here referred to. For example, it 
was proved that the parts most selected or which had already 
most varied—as the tail in fan-tailed pigeons, which has more 
tail-feathers than any one of the 8000 different kinds of 
living birds—were most subject to further variation; and 
