252 BULLETIN: MUSEUM OF COMPARATIVE ZOOLOGY, 
A study of neural segments anterior and posterior to the medulla has 
led me to the conclusion that the local thickening is a more essential 
characteristic of a hindbrain neuromere than the commonly accepted 
criteria, viz. the radial arrangement of cells in the neuromere, and the 
crowding of them in the regions of constriction between neuromeres, 
both of which may be the result of mechanical influences. 
The shifting of the point of exit of the roots primitively related to 
encephalomeres VI and VII may easily be explained as the result of the 
crowding caused by the ear capsule. Since four hindbrain neuromeres 
are clearly related to four visceral arches, we should expect the remain- 
ing one, encephalomere IV, to have been primitively related to a vis- 
ceral arch, That such an arch has been present in the region of this 
neuromere during phylogeny, has been made probable by the studies of 
an Wijhe (82), Miss Platt (91), and Hoffmann (94). The evidences 
from the study of neuromerism and mesomerism are mutually confirma- 
tory, and to the effect that a visceral arch has been lost in the region of 
encephalomere IV and van Wijhe's third somite. Having established an 
exact numerical correspondence between encephalomeres and somites 
(head cavities), and a probable primitive correspondence of hindbrain 
encephalomeres with visceral arches, I conclude that in the head region 
there existed primitively a correspondence between neuromerism, mesom- 
rism, and branchiomerism. Since this correspondence is not to-day 
exact in Squalus or in any other known Vertebrate, it seems necessary 
to discuss somewhat in detail the constituent parts of the anterior or 
more highly modified metameres, and to inquire what may be inferred as 
to their previous conditions. The table on the opposite page, although 
in part theoretical, will help to make the discussion clearer. 
I have in this table included neuromeros as far posteriorly as the 
eleventh. Accepting Hoffmann’s (’94) conclusion that vertebral arches 
as far back as that which corresponds with van Wijhe’s tenth somite 
fuse into the cranium of the adult Squalus,* it would follow that neuro- 
meres I to XI would be included in the cranium. The variability 
in the number of segments added to the occipital region of the cra- 
nium in different Selachii and Ichthyopsida (Fúrbringer, Sewertzofl) 
makes the exact number in Squalus a matter of no great morphological 
importance. 
We see that the cephalic segments are highly modified segments 
altered by reduction or enlargement (possibly even by substitution and 
change of relation, as, for example, in the case of the vagus segments) of 
1 Recently confirmed by Sewertzoff (’98). 
