228 
character, such as in head and body form, 
prolongation of the lower jaw, dentition, gill- 
rakers, maxillary structure, lateral line, and 
lack of squamation, Macroparalepis group I 
is very close to Lestidium and can be differ- 
entiated from this genus only with difficulty. 
While there may be a consistent difference in 
the position of the intestine in the postlarvae 
of these two genera, I doubt that it is of great 
significance. 
Ege divided Macroparalepis into two re- 
markably distinct groups. The first division, 
which was just discussed, contains four spe- 
cies which have the anus situated behind the 
origin of the dorsal fin, 13-14 dorsal rays, and 
24-28 anal rays. Its high number of dorsal 
rays surpasses that known for any other form 
in the Paralepidinae. The dorsum pigment 
pattern is different from that of all other para- 
lepidids. This group most closely resembles 
the subgenus Lestidiops in general form, anal 
ray count, position of dorsal fin, and lateral- 
line form and is probably related to the an- 
cestral stock of this subgenus. It approaches 
the subgenus Lestidium in its nonossified pro- 
longation of the upper jaw. The genus Ma- 
croparalepis is hereby restricted to Ege’s group 
I (since no genotype has been previously 
designated). 
The second group of Ege’s Macroparalepis, 
comprising seven species, is very different 
from the first. The anus is situated in front of 
the dorsal fin and there are 9-11 dorsal rays 
and 37-50 anal rays. The species are very 
elongate and some even eel-like, the angle of 
the gape extends back almost to the tip of 
the maxillary, the tongue is unusually small 
and far back, the dentition is considerably 
reduced, and the lower jaw is very slender and 
pointed with only a slight elevation at its tip. 
Also, the shape of the lateral-line segments 
is very different from that found in group I 
and all other paralepidids, as the length of each 
segment is twice the height. Despite the fact 
that groups I and II have similar postlarval 
development, prolongation of the lower jaw, 
and pore pattern, their similarities end with 
PACIFIC SCIENCE, Vol. VII, April, 1953 
these characters. If these similarities to group 
I are put aside for the moment, group II 
clearly approaches the subgenus Lestrolepis 
with which it shares the prolongation of the 
lower jaw, the high number of anal rays, and 
the same position of the dorsal fin in relation 
to the pelvic fins. In fact, it may very well be 
that group II is more closely related to Les- 
trolepis than to group I, and that the similari- 
ties to the latter group are due to parallel 
development or convergence. In any event, 
group II is far more different from group I 
than group Lis from Lestidium, and it is here 
placed in the genus Stemonosudis. A provisional 
generic tree showing the degree of relation- 
ship of the genera as this now appears to me 
is given in Figure 4. 
Comparison of specific paralepidid anatom- 
ical structures throughout the family shows 
some interesting speculative results on the 
mode of barracudina evolution. In most char- 
acters I have examined there is evident a 
straight-line evolution from the primitive con- 
dition in the genus Magnisudis to the special- 
ized in Stemonosudis through the line of genera 
as shown in Figure 4. In almost every instance 
this evolutionary pattern involves a loss or 
reduction from the form as found in Magni- 
sudis. In general, specialization in this family 
means a loss or reduction in some pre-existing 
character. Such specialization is exemplified 
by the gillraker form, which is reduced in the 
Paralepididae to a flat bony base. This struc- 
ture is most elaborate in the genus Magni- 
sudis, being proportionately large and pecu- 
liarly notched. Each raker is armed with 4 long 
cartilaginous filaments with bony cores ar- 
ranged in a square. The genus most closely 
approaching this condition is Paralepis. In 
this group and all remaining genera the rakers 
are flat and simple in outline. Each raker in 
Paralepis is armed anteriorly with fairly long 
cartilaginous filaments grading posteriorly in- 
to short bony spines. The next step is found 
in the genus Notolepis where all the armature 
of each gillraker consists of short spines. 
Thus, we see a trend for the simplification of 
