84 THE CHANGES FROM STAGE A TO F. 



mesoderm of the gut wall, and ventrally by a septum separating 

 it from the ventral division of the median pseudocoele (b. be). 



At the hind end of the same embryo the same relations are 

 visible but in a less developed state. Here (Plate VIII, fig. 41) 

 the somites are almost in contact in the middle dorsal line, 

 the heart space [b. h.) being very rudimentary. The dorsal 

 divisions of the somites {d. s.) themselves are still well-developed 

 structures with the generative nuclei in their floors. The two 

 new chambers {b.pc, b.bc'.) are present, but in a rudimentary 

 form. It is by looking through a series of sections of an 

 embryo of this age, such a series as that from which figs. 33 — 41 

 were selected, and comparing them with the previous stage, that 

 it is possible to settle conclusively the fact that in figs. 38 and 

 39 the space marked d. s. and *. 4 is the reduced somite — it can 

 be followed backwards, gradually enlarging in successive seg- 

 ments, until at the hind end of the body {in figs. 41, 42) it has 

 exactly the relations of the dorsal division of the somite of the 

 earlier stage — and that the spaces marked b. pc. and b. be', are 

 new formations in the walls of the dorsal divisions of the somites 

 and have nothing to do with the true enterocoele or cavities of 

 the somites. They, like the spaces b. h and b. be, are, from the 

 first, continuous structures. 



The determination of the relations of these cavities in suc- 

 cessive stages has been one of considerable difficulty, for this 

 reason, that the dorsal wall of the body often contracts at the 

 death of the animal and obliterates all traces of the complex 

 system of dorsal cavities. They then present merely the 

 appearance, which has been seen by Kennel and has led him into 

 error, of being an irregular system of spaces in the dorsal 

 mesoderm. 



Kennel's description of the origin of the heart and peri- 

 cardium from an irregular system of spaces in the dorsal 

 mesoderm, derived from the broken-up dorsal divisions of the 

 somites, is quite erroneous. 



It is easy to see the arrangements which I have described, 

 if embryos be used in which the dorsal ectoderm has not con- 

 tracted. 



