PAIRED FINS WITH THEIR NERVES, IN LEPIDOSIREN AND PROTOPTERUS. 613 



leaving the lateral plate as the only remaining epithelial mesoderm in the head. This 

 lies entirely ventral to the branchial ridge, and so is uninterrupted by the gill out- 

 orowths. The mesothelium included in the first three visceral arches has also become 

 mesenchymatous. The lateral plates now meet in the extreme front end of the head 

 in the middle line, below the " stomodaeum," which has now been folded off from the 

 underlying yolk sac. 



In the trunk, the pronephrocceles have appeared opposite the pronephrostomes, but 

 the mesoderm ventral to this is unsplit. In the head the paired rudiments of the 

 pericardial cavities have met in the middle line in front. The cavity extends backwards 

 for some distance on each side into the diverging lateral plates. 



From this point it will be convenient to consider the development of the different 

 parts separately. 



1a. Segmentation and Musculature of Head. 



As already mentioned, there is no segmentation of the somatic portion of the head 

 mesoderm. The development of the muscles derived from the three anterior head 

 somites of Elasmobranchs is as follows in Lepidosiren and Protopterus. 



At about stage 30 we find the future eye muscles and the temporal muscle repre- 

 sented by dense condensations of heavily yoked mesenchyme (text figures 2, 3, 4). 

 The condensation round the optic vesicle (text figure 2) is continued backwards as a 

 well-defined narrow strand, which passes into the larger mass forming the rudiment of 

 the temporal muscle (text figure 3), joining it on its medio-dorsal border. This strand 

 of mesenchyme can be traced further back still (text figure 4), curving in somewhat 

 towards the middle line, and breaking up posteriorly into the general mesenchyme. 

 This description applies to both genera. 



This may be interpreted as showing that the mesenchyme concentrated round the 

 eyeball derives its source partly from a concentration in the orbital region (correspond- 

 ing to the premandibular somite), partly from the same source as the temporal muscle 

 (mandibular somite), and partly from further back still. To be brought completely 

 into line with the development in Elasmobranchs, etc., the posterior end of the 

 mesenchyme strand should come from the same " somite " as the hyoidean muscles. 

 No connection could be traced, however, between the mesenchyme in the hyoid arch 

 and this strand, which in fact does not reach so far back. 



The anterior portion of the eye muscle rudiment (premandibular and mandibular 

 portions) can be traced back to a portion of the somatic head mesoderm lying between 

 the first gill outgrowth and the optic vesicle, which does not completely break up into 

 loose mesenchyme. A similar compact mass of mesoderm is apparently not left above 

 the second visceral arch when the mesoderm in this becomes mesenchymatous. 



It is impossible to say from which source the different eye muscles are derived, as 



