198 
PACIFIC SCIENCE, Vol. XVIII, April 1964 
Nemichthyidae. Bohlke and Cliff (1956), repre- 
senting the trend toward more extreme lumping, 
recognized only the Nemichthyidae (including 
Avocettina, Avocettinops, and Cyema ) and the 
Serrivomeridae (including Gavialiceps) . The 
larvae of Serrivomer and Cyema are quite well 
known, particularly from the studies by Beebe 
and Crane (1936) and Bauchot (1959) on the 
former, and by Bertin (1937) on the latter. 
These larvae clearly have differentiated along 
quite different evolutionary lines than have 
Nemichthys and Leptocephalus acuticeps, and 
are outside the scope of the immediate larval 
problems discussed in the present paper. Thus, 
regardless of which way one thinks it best to 
delimit the family Nemichthyidae (a very in- 
clusive, or a narrowly limited version), it is 
apparent that the adult stage of L. acuticeps 
can be sought in the more restricted group of 
forms that are thought to be phyletically the 
closest to Nemichthys. The forms in this limited 
group center around two principal genera, 
Nemichthys and Avocettina. Although the sev- 
eral pertinent genera are usually keyed out 
primarily on the basis of their lateral-line char- 
acters, a different grouping is more practical 
for the purposes of the present discussion, for 
it utilizes characters that can be determined on 
larvae as well as on adults. The widely dis- 
tributed Nemichthys and the less well-known 
genera Nematoprora and Cere omit us comprise 
the more extremely attenuated of the snipe eels, 
with excessively high vertebral counts that may 
continue to increase throughout life, and a thin- 
ner and more whiplike tail that has little or no 
trace of a defined caudal fin. Reported vertebral 
counts of Nematoprora exceed 259 (Trewavas, 
1932:649, for a specimen with an incomplete 
tail), and of Nemichthys range from 300 to as 
high as 660 (Beebe and Crane, 1937). The eels 
that may be grouped with Avocettina include 
Labichthys and, tentatively, the incompletely 
known Avocettinops. These eels are less ex- 
tremely attenuated, have a better differentiated 
caudal fin, have much lower vertebral counts, 
and apparently develop a fixed number of verte- 
brae rather than adding new units indefinitely. 
The known vertebral counts of the Avocettina 
group are only moderately high, compared with 
other eels in general. Beebe and Crane (1937: 
366, 375) reported a range of variation from 
about 170 to 198 in Avocettina, and from about 
175 to 180 in Labichthys. Bertin (1942:106) 
reported a count of about 192 vertebrae in 
Avocettinops. The known caudal-ray counts are 
low, compared with eels as a whole: five rays 
in Avocettina (Beebe and Crane, 1937:371), 
and four in Avocettinops (Bertin, 1942:107). 
The somite counts of 182 and 187 on the 
two eastern Pacific larvae of Leptocephalus 
acuticeps fall within the known range of varia- 
tion in vertebral counts of Avocettina. The count 
of 207 on the Atlantic larva of acuticeps is a 
little higher than the known maximum adult 
vertebral count. If the questionable Leptoceph- 
alus oxycephalus Pappenheim should prove to 
belong to this group, its somite counts of 220 
to 230 would indicate the existence of related 
populations that have vertebral counts substan- 
tially higher than the presently known maximum 
among the Avocettina- like eels. L. acuticeps re- 
sembles Avocettina also in its low caudal-ray 
count. 
Although metamorphosing specimens are still 
lacking, the available data warrant the tentative 
identification of Leptocephalus acuticeps as a 
larval stage of the Avocettina group. The os- 
tensibly extreme differences between the Ne- 
michthys larva and L. acuticeps seem inevitable 
consequences of the basic differences between 
Nemichthys and Avocettina. From the known 
vertebral and caudal-fin characters of the adults 
of these two genera, one can predict that their 
larvae must differ in somite and tail-tip charac- 
ters in precisely the way that L. acuticeps does 
differ from known larvae of Nemichthys. 
The available evidence suggests that acuticeps 
is probably significant above the species level. 
Similar problems on other leptocephali indicate 
that eel larvae in general tend to show strong 
group resemblances and relatively small or ob- 
scure species differences. For example, Anguilla 
larvae conform to a distinctive and easily recog- 
nized generic format wherever they occur, but 
in the Indo-Pacific Jespersen (1942) found that 
they are difficult to separate into species to 
match the approximate dozen named species 
that are currently recognized for adults in that 
region. Similarly, Bauchot (1959) found that 
two named larvae, Leptocephalus lance olat us 
