64 bulletin: museum of comparative zoology. 



in which local swellings appear. The section represented by Figure 4 

 passes through such a swelling. It will be seen (even more clearly in 

 Figure 16) that, on the one hand, the mesomer is now entirely severed 

 from the sclerotome, and on the other retains only an indefinite connec- 

 tion with the somatic layer of the epimer. 



In the particular larva from which Figure 4 is taken there, is one 

 swelling in the blastema for each of the somites 8, 9, 10, and 11 (see 

 larva B, Diag. 2, p. 65). This would lead one to suppose that each 

 swelling corresponds to one of the original mesomers. Such a supposi- 

 tion, if true in this particular case, cannot be so generally, as there 

 are usually more swellings than somites. On plotting tlie swellings in 

 thirteen larvae {A-M. in Diagram 2) varying from 6.5 mm. to 9.5 

 mm. in length, it was found that 



somite 7 contains an average of 0.08 swellings 



8 " " 1.08 



9 " " 1.23 '' 



10 " " 1.6 



11 " « 1.8 



12 " " 0.4 " 



In other words, the seventh somite occasionally develops a swelling ; 

 somite 8 lias generally one, rai-ely two ; somite 9 often has two ; somites 10 

 and 11 liave more often two than one, and somite 12 shows an occasional 

 one. As these numbers agree fairly well with the number of primary 

 units in the fully developed mesonephros, the conclusion suggests itself 

 that the swellings are the fundaments of the blastulae. On that assump- 

 tion, it is true, there are not enough swellings in the posterior somites, 

 especially somites 11 and 12; but it is clear that in some cases they 

 have not all appeared, and it is also possible that some of the swellings 

 contain the fundaments of more than one blastula. As development 

 proceeds, their appearance confirms the idea that they are young blastulae. 

 They increase in diameter, and the deeply staining nuclei show a radial 

 arrangement (see Fig. 15, Plate 2). These blastulae are spindle-shaped 

 and more elongated antero-posteriorly than in Amblystoma, and are 

 usually, although not always, continuous with each other by their taper- 

 ing ends. There are thus formed cord-like connections, which are re- 

 tained in many cases to a late period ; these connecting parts may even 

 become massive, and seem to contribute a portion of the tissue from 

 which the secondary units are formed. 



It will be recalled that in Amblystoma the fundaments of the " units 



