376 
PACIFIC SCIENCE, VoL XVIII, October 1964 
Harengula konigsbergeri (Weber and de 
Beaufort 1912) 
Clupalosa lippa (Whitley 1931) 
Harengula maccullochi Whitley 1931 
Harengula macrolepis ( Steindachner 1879) 
Harengula ov alls (Bennett 1830) 
Harengula vittata (Valenciennes 1847) 
The last 1 1 species are Indo-Pacific and may all 
belong to Clupalosa. 
I have tabulated 110 nominal species that 
appear to belong to these four genera, of which 
only about 43-51 species may be valid. There 
are many conflicting opinions and uncertainties 
concerning the synonymies of these nominal 
species. In listing the species above I have used 
and attempted to reconcile the works of Chan 
(ms), Fowler (1941), Herre (1953), Regan 
(1917), Rivas (1950), and Whitley (1940, 
1941, 1948). 
The separation of Harengula and Sardinella 
was discussed by Chan (MS), who was the first 
to emphasize the differences in the larger and 
more posterior of the two supramaxillaries in 
these two genera. Chan also detailed differences 
in scale sculpture between these genera and 
commented upon the two enlarged terminal anal 
fin rays in Sardinella. Whitehead ( 1962 ) sug- 
gested that Harengula might have a greater 
number of parietal striae than Sardinella; but 
this difference, if valid, is complex, because the 
number of parietal striae in Harengula thrissina 
progressively increases from about 5 at 50 mm 
S. L. to about 13 at 130 mm. Chan and previous 
authors were unaware of the hypomaxillary 
bones in American species assigned to Haren- 
gula, however, and considered Harengula in its 
broad sense to include the Indo-Pacific species 
which lack hypomaxillaries. 
The phylogenetic significance of the hypo- 
maxillary and its importance in the classifica- 
tion of the Clupeidae are subject to various 
interpretations. A thorough knowledge of the 
morphology of the genera and species of Clupei- 
dae and of the origin and development of this 
bone will furnish a more definitive answer to 
these issues. 
The hypomaxillary is a specialized structure 
that must have developed independently in two 
phyletic lines of the Clupeidae — in the typical 
herring genus, Harengula, and in Pellona and 
Pliosteostoma of the group of clupeids with a 
high number of anal rays and greatly com- 
pressed bodies, sometimes referred to as "blood- 
less clupeids.” It probably arose as a permanent 
splitting off of a portion of the maxillary or 
premaxillary; it is less probable that its origin 
was the spontaneous development of a new site 
of ossification. 
In Aphredoderus ( Percopsif ormes ) and in 
certain species of Amblyopsidae ( Amblyopsi- 
f ormes) each premaxillary is divided distally 
into from 2 to about 7 distinct but closely as- 
sociated parts of progressively decreasing size 
(Rosen, 1962). These smaller terminal portions 
of the premaxillaries were termed segments, and 
Rosen ( 1962:23) suggested that these segments 
produced a flexibility to the upper jaw in full 
extension of the mouth. This might indicate 
that these segments were developed in response 
to a need for additional flexibility, or that, after 
they had developed, additional flexibility was 
possible. The clupeid hypomaxillary undoubtedly 
developed independently from the premaxillary 
segments of Aphredoderus and the amblyopsids, 
and, if any such functionalism were once a 
factor in the origin of the hypomaxillary in 
these clupeids, it has subsequently been lost or 
occluded. 
My studies have led me to believe that, within 
the two clupeid groups concerned, the hypo- 
maxillary must be significant in indicating a 
distinct phylogenetic (and taxonomic) differ- 
ence between the species which have it and 
those which lack it. Myers (1950) found the 
hypomaxillary present on one side and absent 
on the other in a large specimen of Pellona and 
questioned its significance; but in the hundreds 
of specimens of all species of Pellona, Pliosteo- 
stoma, and Harengula that I have examined, 
both sides of the upper jaw have hypomaxillaries 
of similar size. 
To stabilize the nomenclature until extant 
uncertainties are clarified, and because published 
generic names are available, I have proposed 
above that the presence of the hypomaxillary be 
regarded as a criterion of generic distinction, 
and restriction of the genus Harengula. 
ACKNOWLEDGMENTS 
I am grateful to Franklin G. Alverson, Iza- 
