618 MR W. E. AGAR ON DEVELOPMENT OF ANTERIOR MESODERM, AND 



probably also often contains the remains of anterior ccenogenetic head myotomes — for 

 the number of occipital myotomes which appear in development often varies in nearly 

 allied forms in which the occipital arch is certainly homologous. For example, the 

 occipital myotome w occurs in Ceratodus embryos, Semon — and apparently, v, also in 

 Sewertzoff's specimens — but not, or very seldom, in Lepidosiren or Protopterus. It 

 is well known also that variation in the number of occipital myotomes takes place in 

 individuals of the same species. It is very probable that in Petromyzon, where this 

 " somite " gives rise to muscles which form a continuation of the trunk muscles over the 

 head, it is really mainly a ccenogenetic segment. 



As regards the number of these somites whose myotomic parts have fused up and 

 ultimately disappear, which, if the gill slits were originally segmentally placed, corre- 

 sponds to the ancestral number of visceral arches, minus the first two, it is of interest 

 to notice that the number of vagus roots in Lepidosiren larvse is quite irregular. They 

 seldom correspond even on the two sides of the same individual. For instance, in an 

 embryo of stage 31+ the vagus fibres leave the medulla in twelve places on the left 

 side and eleven on the right. In one of stage 38 there are ten outlets on the left side 

 and seven on the right. It is significant, in connection with any attempt to estimate 

 the number of these fused segments by the number of vagus roots, that in the older 

 larva (1) the number of outlets is smaller than in the younger one, and (2) they are so 

 arranged that the ten outlets on the left side might be regarded as collected into six 

 "roots," and the seven on the right in five "roots," while it was impossible in the 

 younger larva to allocate the outlets into " roots " in this way. 



As will be seen immediately (p. 622), the region of the alimentary canal and its 

 musculature innervated by the vagus shows during ontogeny a backward movement 

 relative to the somatic trunk mesoderm (more correct morphologically, though less so 

 descriptively, a forward movement of this trunk mesoderm over the vagus innervation 

 region), similar to that which we suppose to have taken place phylogenetically in the 

 case of the visceral arches. As regards the branchial region itself in Lepidosiren and 

 Protopterus, no further backward movement of it relative to the trunk myotomes is 

 observable after the branchial outgrowths have made their first appearance. In fact, 

 in late larval life the posterior branchial arches actually move forwards, so that the 

 whole branchial region comes to lie in front of the middle of the auditory capsule. The 

 condensation of this region, thus brought about, obviously corresponds with its loss of 

 functional importance. If, however, we take the vertebrate in which the branchial 

 region occupies the greatest extent — i.e. Cyclostomes — we find the opposite. By the 

 time the eight gill pouches in Petromyzon are formed, the posterior one is ventral to 

 the seventh metotic myotome. This is the most anterior myotome to send a bud to 

 the sub-branchial musculature (Neal, Koltzoff). Later — in a larva of 8 mm. 

 (Koltzoff) — the branchial region extends to beneath metotic myotome 13, breaking 

 through the sub-branchial processes of myotomes 7-13 as it travels backwards. 



Now it is quite possible to give an interpretation of this process exactly the oppo- 



