Leptocephalus acuticeps — Orton 
197 
vertical artery that precedes the main renal 
artery (Fig. 4). The location of the third spot 
averages about 30 somites farther back in 
Nemichthys, but in both kinds of larvae it is 
situated in the same region about 24-31 somites 
behind the posterior end of the kidney. The 
internal morphological affinities of the third 
spot are not clear from the present data, but may 
be determinable from histological examination 
or from study of the anatomical changes that 
occur in this region during metamorphosis. 
From Bertin’s data one can diagram the posi- 
tions of the spots and of the major blood vessels 
in the holotype larva of L. acuticeps, and the 
result is essentially the same as in the Scripps 
larvae of acuticeps and in Nemichthys : the first 
spot coincides with the position of a median 
vertical artery in the pyloric region, and the 
second coincides with a similar vessel about 10 
somites anterior to the main renal artery. Bertin’s 
illustration (his fig. 4) includes this spot and 
artery. 
It might seem reasonable to suppose that 
color -pattern elements that are directly beside 
the somites are associated primarily with these 
immediately neighboring somites rather than 
with some other structure. The contrary expla- 
nation in the present example traces to the fact 
that the spots in question are internal rather 
than external to the somite surfaces. In eel 
larvae, the visceral complex is displaced far 
downward, usually completely below the lower 
edges of the somites. The essential links between 
the body axis and the viscera (blood vessels, 
supporting connective tissue, etc.) are greatly 
attenuated and occupy a thin median vertical 
plane sandwiched between the laterally com- 
pressed right and left halves of the somites. 
The available evidence indicates that position 
of the diagnostic internal three-spot pattern in 
this group of larvae is primarily a function of 
the visceral and arterial positions and therefore 
only indirectly dependent on somite number. 
Evolutionary changes that have shifted the 
critical internal landmarks farther forward or 
backward along the body have also shifted these 
characteristic spots correspondingly, and thus 
the markings have maintained their predictably 
constant relationship to the specific blood vessels 
and visceral organs. The more conspicuous posi- 
tional association is with the pylorus and the 
end of the kidney, but experimental studies (if 
such work is ever feasible on deep-sea lepto- 
cephali) might show that the specific vertical 
blood vessels are the more direct determinants. 
This problem demonstrates graphically the 
importance of noting precisely whether a color - 
pattern element that is "on the somites” actually 
occurs outside or Inside of the transparent 
muscle layer. The morphological affiliations, and 
hence the evolutionary potentialities,' of these 
two locations are quite different. 
Once it is established that L. acuticeps and 
the Nemichthys larva share a uniquely inte- 
grated structural and color-pattern character in. 
the predictable detail that hints genetic rela- 
tionship, then the probable significance of the 
similarities and differences that they show in 
other characters becomes clear. The Nemichthys 
larva is, essentially, an exaggerated acuticeps that 
has specialized in extremely attenuated shape,, 
very high somite count, and the probably con- 
tinuous addition of new somites in its filament- 
like tail tip. These chief larval differences in- 
volve the same characters as do the differences 
that separate the adults of Nemichthys from 
those of certain related genera, and these char- 
acters suggest the probable correct generic 
placement for Leptocephalus acuticeps . 
ASSIGNMENT OF Leptocephalus acuticeps 
TO Avocettina 
Opinions on the generic composition and 
best nomenclatural treatment of the Nemichthys 
group differ greatly. Some authors divide the 
presumed relatives of Nemichthys into several 
families, and set the entire group apart from 
all other eels at the subordinal level. Others 
reduce the number of families, chiefly by lump- 
ing rather than by deletion from the group, 
and either accept or omit the subordinal cate- 
gory. Roule and Bertin (1929) proposed subordi- 
nal rank for these eels and, on the basis of small 
differences, divided them into six families: Ne- 
michthyidae, Avocettinidae, Avocettinopsidae, 
Gavialicipidae, Cyematidae, and Serrivomeridae. 
Trewavas ( 1932) reduced this assemblage to 
three families: Serrivomeridae ( including Ga- 
vialiceps in part; she placed G. toeniola in the 
nettastomid genus Saurenchelys) , Cyemidae, and 
