194 BULLETIN: MUSEUM OF COMPARATIVE ZOOLOGY. 
The contention that the constrictions between van Wijhe’s somites are 
incomplete does not appear to me to militate greatly against the view 
that they have morphological value, inasmuch as their permanency has 
been repeatedly attested (van Wijhe, Hoffmann, Neal, and Sewertzoff). 
Nor does Rabl apparently consider this argument as of great weight, 
since he regards van Wijhe’s 5th (1st post-otic) somite — though the 
constrictions which are found in front and behind are incomplete — as 
a true somite. The reduction in the myotomic portion of the dorsal 
mesoderm accounts in great part for the incompleteness of the constric- 
tions. I believe that one who follows the development of the pre-otic 
and sub-otic mesoderm in Ammocetes, and observes the ontogenetic 
dissolution of the compact dorsal mesoderm into loose mesenchyma, 
which follows the great enlargement of the nerve ganglia and of the 
otic capsule, is in a position to understand the reduction of the dorsal 
mesoderm in this region in Vertebrates higher in the phylogenetic scale 
than Ammocoetes.! 
wick (’92), that this is not true of the mesoderm segments discovered by Dohrn 
(’90, 7908) in that form. Dohrn apparently did not endeavor to ascertain whether 
they were symmetrical or not. Iam unable to determine, even in carefully made 
reconstructions of well oriented frontal sections of embryos at the same stage of 
development as that described and figured by Dohrn (’90"), whether or not there 
is a correspondence of the mesodermal segments on the two sides of the head an- 
terior to the one which, in my opinion, corresponds with the 15th segment of Dohrn. 
While my own negative conclusions cannot be regarded as in any sense disproof of 
the segmental value of Dohrn’s somites, it is my opinion that the evidence of their 
variability shown by the conflicting results of Killian (’91) tends to throw consider- 
able doubt upon it. Since Killian (’91, p. 103) finds that of the anterior of these 
segments one is to be regarded as the sclerotome portion of a somite, while others 
are simply vesicular enlargements of the mesoderm of the mandibular arch, it is 
to be inferred that Dohrn subjected the head somites of Torpedo to little critical 
examination. To regard as evidence of somites all vacuolar spaces in the dorsal 
(and lateral!) mesoderm which appear between the somatopleure and splanchno- 
pleure at the time these layers separate, seems to be too uncritical. Similar phe- 
nomena appear in the mesoderm of Squalus in those early stages of development, 
when the coelom is in the process of formation, viz. in stages when the neural plate 
is widely expanded and the embryo possesses 4 or 5 somites, Recent studies by 
Sewertzoff (’98) render still more doubtful the results of Dohrn and Killian. 
1 A mechanical explanation of the constrictions between the head somites of 
van Wijhe, such as that offered, but without evidence, by Kastschenko (’88), 
seems hardly worthy of consideration. That such constrictions as those, for ex- 
ample, between somites 3 and 4, and 4 and 5, cannot result from the so called 
mechanical influence of visceral clefts, follows from the evidence already stated by 
Hoffmann ("94 and ’96) that in Squalus, the constrictions lie dorsal to the visceral 
arches. I cannot, however, agree with Hoffmann that we may conclude from this 
evidence that the visceral arches are intersomitic in position, as are the ribs in the 
