during the greater part of the period of lactation, which according to the
same authority is about four months. In the specimens of _Dasyurus_
described by O'Donoghue, in which the milk glands developed after
ovulation without ensuing pregnancy, normally developed corpora lutea were
present in the ovary. Of the five females which he mentions, the first
three, one with unfertilised ova in the uteri, two five and six days after
heat, could not have been pregnant, but the other two killed eighteen and
twenty-one days after heat might, since pregnancy lasts only eight days,
have been pregnant, the young having died at parturition or before. To
make certain on this point it would have been necessary to examine the
ovaries and milk glands of females which had been kept separate from a
male the whole time. There is no doubt, however, about the development of
the milk glands in the first three specimens, which were certainly not
pregnant.

It is difficult to reconcile entirely the evidence described by O'Donoghue
from _Dasyurus_, with that obtained from higher Mammals, although on the
whole there is reason to conclude that the corpora lutea have an important
influence on the development of the milk glands. According to Lane-Claypon
and Starling, if the ovaries and uteri are removed from a pregnant rabbit
before the fourteenth day the development of the mammary gland ceases,
retrogression takes place, and no milk appears in the gland. If, on the
other hand, the operation be performed after the fourteenth day, milk
appears within two days after the operation. It is to be concluded from
this that the cause of _secretion_ of milk is the withdrawal of a stimulus
proceeding from ovary or uterus. But O'Donoghue believes that milk is
secreted in _Dasyurus_ when no pregnancy has occurred. Ancel and Bouin
[Footnote: _C. R. Soc. de Biol._, t. lxvii., 1909.] have shown that the
growth of the mammary glands was produced in rabbits by the artificial
rupture of egg follicles and consequent production of corpora lutea: the