82 Mr Creighton, On the order in ivhich the secreting [Feb. 26, 



series of intermediate steps to the cellular pancreas." The best 

 example, for the present purpose, is the mammary gland. In its 

 simplest form, in the Monotremata, it consists of 100 — 200 coeca 

 or follicles, which converge to a small spot of the skin on each 

 side of the abdomen, each follicle being attached to the skin by a 

 separate neck. This is an example of a gland that is follicular 

 at the beginning of the mammalian series, but is of a complex 

 acinous type in most mammals. The steps by which the follicular 

 type is changed into the acinous consist in certain acts of cen- 

 tralisation, by which single passages occupying the middle of the 

 follicles are united, near their termination in the skin, to form 

 one or more common outlets ; while smaller and still smaller groups 

 of follicles combine, at their outlying parts, to form their canals 

 into ducts of the secondary and tertiary order. The centralising 

 process may be represented diagram matically in more than one 

 way; the result is that the multitude of independent and co-ordi- 

 nate outlets are replaced by a main system of conduits and 

 subordinate branches. The starting point of the system of ducts 

 is the passage that forms in the middle of each follicle, which is 

 otherwise a mass of secreting cells. In the liver of Carcinus 

 moenas Goodsir describes the rudimentary duct as an irregular 

 passage existing in the midst of the cells near the attached end 

 of the follicle. Such passages, according to the same authority, 

 are formed by the breaking down of the ripest cells to yield the 

 secretion. The irregular passage thus formed is, in the further 

 adaptation, retained as a permanent outlet. Still further, the 

 central channels of the several follicles join to form a common 

 system of ducts, as already explained. At the same time the 

 follicles, which make up the primitive gland, lose their indi- 

 viduality, and the organ passes into the acinous type. 



The adaptation of the parenchyma, along certain central lines, 

 to become a permanent system of ducts, is an economical process. 

 The more permanent and convenient the outlets of the secretion 

 are, the smaller is the bulk of secreting parenchyma required. 

 The size of the mamma, during its active state, in Ornithorhynchus 

 and Echidna, is relatively very much greater than in other 

 mammals. 



An explanation has now to be given of the fact that the branch- 

 ing system of ducts, which is the latest acquisition of the secreting 

 gland, is the first part to appear in the embryo. The fact itself 

 may be readily observed in the developing mamma of the guinea- 

 pig, where a complete racemose system of ducts is found to extend 

 over the whole area of the gland before any of the secreting acini 

 have taken shape. The same fact may be readily made out in the 

 development of the salivary glands. That reversal of the 

 phylogenetic order may be accounted for by keeping in view the 



