Ma?ichester Memoirs, Vel. I. (1906), No. 3. J J 



and destiny do not in any cell, or set of cells, invariably 

 and necessarily coincide ; homologies founded on the 

 former may be the diametrical opposite of comparisons 

 based upon the latter, and in this dilemma it is to the latter 

 only that we can continue to adhere without reducing 

 comparative anatomy to an absurdity. 



The facts of descriptive embryology might well be 

 deemed sufficient of themselves to warrant us in thus 

 finally relinquishing any hope of retaining the morpho- 

 logical or phylogenetic significance which for so long has 

 been attached to the germinal layers. Corroborative 

 evidence is, however, forthcoming from other quarters, 

 from the phenomena of budding and segmentation, from 

 pathology, and from the results of experimental embryo- 

 logy. Let us briefly consider these. 



We have seen that in the normal development of the 

 fertilized ovum there comes a moment, at any rate in 

 some forms, when it is possible to detect three layers or 

 sets of cells which, however much they may vary in their 

 origin, bear constant relations to certain definite parts of 

 the adult organization by reference to which they may be 

 defined. Even this constancy is, however, not absolute ; 

 muscles, for example those of the skin-glands in Amphibia, 

 may be derived from the ectoderm, and enamel is said 

 occasionally to arise from mesodermal tissues. In the 

 mode of reproduction termed budding this inconstancy 

 may be much more marked. 



The difficulties presented by the anomalous behaviour 

 of the germ-layers (or rather of their representatives) in 

 the budding of various Ascidians, is well known to 

 zoologists. The young bud is a two-layered sac, the outer 

 layer being formed in all cases of ectoderm ; the inner 

 layer, on the other hand, may be a diverticulum of the 



