198 BULLETIN: MUSEUM OF COMPARATIVE ZOOLOGY. 
section, that the greater part of the cells proliferated into the cavity of 
the myotome, the cells of which are at this stage already converted into 
elongated muscle cells, arise from its median wall, While the outer 
wall still maintains its primitive epithelial character, the inner wall has 
become many cells in thickness and some of these cells appear in the 
act of migrating into the now greatly diminished lumen of the cavity. 
Later, however, the cells of the outer wall also are converted into muscle 
cells, and thus both walls of the cavity participate in the formation of the 
musc. rectus posterior. We have therefore in the 3d somite of van Wijhe 
a pre-otic segment of the dorsal mesoderm, which becomes differentiated 
into myotome and sclerotome, and whose musculature is derived in greater 
part from its median wall. Furthermore, as is well known, its musculature 
is innervated by a nerve (abducens) which all the later morphologists, with, 
so far as I know, one exception (Kupffer, ’94, 95), regard as a ventral 
nerve comparable with spinal ventral nerves. Finding this to be the case 
with at least one pre-otic mesoderm segment, we are in a better position 
than we otherwise should be to understand the more modified, or at least 
more divergent, conditions presented by the remaining pre-otic segments, 
viz. the anterior, the lst, the 2d, and the 4th. That in these segments 
marked peculiarities appear is certain. In the 4th somite we have a 
segment of the dorsal mesoderm divided by constrictions from the 3d 
and 5th somites at a time when it presents essentially the same evidences 
of differentiation into myotome and sclerotome which appear in the 3d 
and 5th somites. That no muscle cells are formed in its inner wall, and 
that it soon breaks up into loose mesenchyma, are phenomena which are 
to be expected in a somite destined to become rudimentary. That it is 
more rudimentary than the 5th somite is due to the development of the 
otic capsule, under which it lies. ‘The 5th somite — in whose inner wall 
elongated cells appear, without however developing into muscle fibres 
(as stated by Sedgwick, ’92) — thus forms a natural transition to the con- 
ditions presented by the 4th. Tf the 3d and 5th are to rank as somites, 
it is in my opinion impossible to deny that the 4th, which lies between 
them, is serially homologous with them, even though it should lack some 
of the characteristics of a typical trunk somite. 
Passing forward in the embryo to the 2d (mandibular) somite, it 
seems to me indisputable that this is the anterior continuation of the 
dorsal mesoderm. In very early stages it grows ventrally to form the 
mesoderm (mesothelium) of the mandibular arch, a process which, accord- 
ing to Kupffer ('88), occurs in Petromyzon also. However, only the 
dorsal part of the “mandibular cavity,” which later becomes separated 
