256 BULLETIN: MUSEUM OF COMPARATIVE ZOOLOGY. 
the loss of the somatic musculature of this segment, a somatic (ventral) 
nerve is wanting, and as in the case of the 6th segment I assume that 
this is to-day represented by the abducens. Since the sensor fibres of 
the facialis innervate the skin of the hyomandibular (1st visceral) cleft, 
this cleft may be believed to have been primitively situated ventral to 
the constriction between the 3d and 4th somites. I find no evidence to 
support the view that the hyoid arch represents two splanchnic segments, 
d. RELATIONS or ENGEPHALOMERR IV. 
As has already been stated, there is no ganglionic nerve Anlage 
proliferated from encephalomere IV, and although the fibres of the major 
root of the trigeminus have their exit in early stages from its outer con- 
vexity, the probability is that such relation of nerve V is secondary, and 
therefore not of phylogenetic significance. I hold that this encepha- 
lomere, by virtue of its local thickening, affords evidence of a lost visceral 
arch, the loss of which would naturally be correlated with the loss of the 
dorsal nerve. Since, however, the disappearance of the splanchnic por- 
tion of this segment may have been due simply to a union with the 
corresponding portion of the anterior (mandibular) segment, it is also 
possible that the dorsal nerve has fused with the nerve of the anterior 
segment, the trigeminus. This conclusion seems indeed supported by 
the evidence that at least some of the fibres of the trigeminus roots have 
their nucleus in the lateral horn of this (4th) encephalomere, 
In a scheme of primitive segmental relations such as I am at present 
advocating, there is likewise difficulty in explaining the fact that the 
somite (van Wijhe’s 8d) which I assume to correspond with encephalo- 
mere IV is innervated by the abducens, whose fibres make their exit 
from encephalomere VII. The evidence which leads me to conclude 
that the abducens to-day represents the primitive ventral nerve of this 
encephalomere, as well as those of encephalomeres V, VI, and VII, has 
been partly given in connection with the study of its development ; it 
may be summarized as follows. (1) Its roots are many (4-6 in various 
Selachii) and more widely separated than those of any other nerve. 
(2) Not only do abducens fibres innervate pre-otic musculature (muse. 
rectus posterior), but fibres from this nerve may also be traced for a con- 
siderable distance in the mesoderm of the embryo posterior to encepha- 
lomere VIT (Fig. 20). (3) The variability as to the place where its fibres 
emerge, as shown by comparative embryological evidence, appears to 
indicate that its relations are not limited to any single encephalomere. 
(4) Its nucleus in the ventral horn of the neural tube is greatly elongated. 
