into the cocoon loses about 50
per cent. of its weight; but the amount of loss differs in different
breeds, and this is of importance to the cultivator. The cocoon in the
different races presents characteristic differences; being large or
small;--nearly spherical with no constriction, as in the _Race de
Loriol_, or cylindrical with either a deep or slight constriction in
the middle;--with the two ends, or with one end alone, more or less
pointed. The silk varies in fineness and quality, and in being nearly
white, of two tints, or yellow. Generally the colour of {303} the silk
is not strictly inherited: but in the chapter on Selection I shall give
a curious account how, in the course of sixty-five generations, the
number of yellow cocoons in one breed has been reduced in France from
one hundred to thirty-five in the thousand. According to Robinet, the
white race, called Sina, by careful selection during the last
seventy-five years, "est arrivee a un tel etat de purete, qu'on ne voit
pas un seul cocon jaune dans des millions de cocons blancs."[508]
Cocoons are sometimes formed, as is well known, entirely destitute of
silk, which yet produce moths; unfortunately Mrs. Whitby was prevented
by an accident from ascertaining whether this character would prove
hereditary.

_Adult stage._--I can find no account of any constant difference in the
moths of the most distinct races. Mrs. Whitby assured me that there was
none in the several kinds bred by her; and I have received a similar
statement from the eminent naturalist M. de Quatrefages. Captain Hutton
also says[509] that the moths of all kinds vary much in colour, but in
nearly the same inconstant manner. Considering how much the cocoons in
the several races differ, this fact is of interest, and may probably be
accounted for on the same principle as the fluctuating variability of
colour in the caterpillar, namely, that there has been no motive for
selecting and perpetuating any particular variation.

The males of the wild Bombycidae "fly swiftly in the day-time and
evening, but the females are usually very sluggish and inactive."[510]
In several moths of this family the females have abortive wings, but no
instance is known of the males being incapable of flight, for in this
case the species could hardly have been perpetuated. In the silk-moth