GROWTH-CHANGES IN MAMMARY APPARATUS OF DASYURUS. 223 
the glands followed on the production of corpora lutea in non- 
pregnant rabbits. This growth went on for fourteen days and 
then regression set in. 
(5) Fourteen days is about the time taken in the formation 
and growth of the corpus luteum, and after this time it begins 
to disappear (Lane-Claypon and Starling). 
From this evidence it seems we are justified in drawing the 
following conclusion: The hormone causing the growth 
of the mammary glands during pregnancy, and at 
other times, is produced in the corpora lutea. 
The Cause of the Secretion of Milk. 
It is not intended here to discuss in detail the cause of the 
secretion of milk in the glands, a question by no means satis- 
factorily settled, and one that awaits further experimental 
inquiry, but at the same time a brief reference to it appears 
to be necessary. 
Hildebrandt (23) put forward the theory that the fertilised 
ovum during its development acted as a stimulus to the growth 
of the mammary glands, and at the same time protected the 
cells of the latter from autolytic disintegration. On the 
removal of this check therefore the cells underwent autolysis, 
and so produced milk. Reasons have been given above, under 
the section dealing with the changes after parturition, to 
show that this secretion is not to be regarded as the result of 
any necrobiotic changes, but is, on the other hand, a purely 
vital activity. The second part of the theory, i.e. that 
lactation follows upon the removal of an inhibitory stimulus, 
has been adopted by Halban (19), and also Lane-Claypon and 
Starling (loc. cit.). The former gives an account'of cases of 
extra-uterine fcetation in which the premature death of the 
embryo was not followed by its expulsion, and points out that 
there was a flow of milk from the breast a few days after 
the death may be supposed to have taken place. He records 
a similar phenomenon after the death of the foetus when this 
occurs in the uterus, and he ascribes the production of milk in 
