FISHERY BULLETIN: VOL. 81, NO. 3 



not know (and probably never will) precisely when 

 and how the species was introduced. Nevertheless, 

 some speculations based on available data and very 

 likely possibilities seem justified. First, it is almost 

 certain that//, quadrimaculatus was not inadvertent- 

 ly introduced along with the S. marquesensis in the 

 late 1950s. The former species is not known to occur 

 in the Marquesas Islands, and the extensive study of 

 Nakamura and Wilson (1970) would have likely 

 detected it if it were actually present. The absence of 

 any "sardines" generally and in our Kaneohe Bay 

 samples prior to 1975, as well as the very rapid in- 

 crease in abundance ofH. quadrimaculatus between 

 1975 and 1976, both indicate that this species had 

 probably not been in Hawaii very long before we 

 collected our first specimens. Finally, the number of 

 individuals released in Hawaii was almost certainly 

 less than the 144,000 fish released when the Mar- 

 quesan sardine was introduced (Hida and Morris 

 1963). 



If the above speculation is close to the truth, H. 

 quadrimaculatus appears to have increased from a 

 small group of individuals to a widespread and very 

 abundant species within a span of 2-3 yr. Even allow- 

 ing for multiple spawnings per lifetime in females, the 

 fecundity of H. quadrimaculatus is so low that sur- 

 vival from egg to adult during this period must have 

 been extremely high. Since the population appears to 

 have remained relatively stable since about 1976, 

 some mechanism must have acted to lower survival 

 rates. The most likely regulatory mechanism would 

 be the increasing abundance of the gold spot herring 

 themselves, i.e., intraspecific competition at some 

 point in the life cycle. It is also possible that local pre- 

 dators have responded to the herring as a new resource 

 and are presently regulating abundance. 



The known differences between Herklotsichthys 

 quadrimaculatus and Sardinella marquesensis seem 

 unlikely to account for the former's greater success in 

 Hawaii; to the contrary, the two species are very 

 similar in morphology and ecology. Berry and 

 Whitehead (1968) noted that in several respects S. 

 marquesensis is more similar to the Herklotsichthys 

 spp. than are other Sardinella spp. Fresh specimens 

 of H. quadrimaculatus can be distinguished from 5. 

 marquesensis by the gold spots on the operculae; 

 otherwise, close examination of the second su- 

 pramaxilla, the imbedded portions of the body scales 

 or the posterior anal rays (see Berry and Whitehead 

 1968) are required to separate the two species une- 

 quivocally. Sardinella marquesensis has higher gill 

 raker counts than H. quadrimaculatus, but Naka- 

 mura and Wilson's (1970) diet data from the Mar- 

 quesas indicate that, unlike most other Sardinella 



spp., S. marquesensis eats very little phytoplankton 

 and, in fact, eats zooplankton equivalent to those 

 recorded for H. quadrimaculatus in Hawaii. Other 

 features of the life history of S. marquesensis con- 

 sidered by Nakamura and Wilson are almost identi- 

 cal to those of H. quadrimaculatus. Investigation of 

 the physiology and larval ecology of the two species 

 and, if still possible, of the ecology of S. marquesensis 

 in Hawaii might better account for the greater suc- 

 cess oiH. quadrimaculatus in Hawaii and might also 

 be pertinent to the broader problem of factors limit- 

 ing natural distributions. 



ACKNOWLEDGMENTS 



We are indebted to W. J. Baldwin for correcting our 

 initial misidentification of the topic species and to P. 

 J. P. Whitehead for confirming the name used here. 

 We thank T. D. Cooney for his assistance in fitting 

 our age- length data to the growth model. This paper 

 is based in part on V. R. Williams' Master of Science 

 dissertation (Department of Oceanography, Univer- 

 sity of Hawaii). Partial support was derived from 

 NSFOCE 77-09202 and from the Hawaii Institute of 

 Marine Biology. 



LITERATURE CITED 



Ben-Tuvia, A. 



1960. Synopsis of the systematics and biology of Sardinella 

 maderensis (Lowe). In H. Rosa, Jr., and G. Murphy 

 (editors), Proceedings of the World Science Meeting on 

 Biology of Sardines and Related Species, Vol. 2, p. 497- 

 519. FAO, Rome. 



Berry, F. H., and P. J. P. Whitehead. 



1968. A new species of sardine (Sardinella, Clupeidae) from 

 the Marquesas Islands. Proc. Biol. Soc. Wash. 81:209- 

 221. 



Brothers, E. B., C. P. Mathews, and R. Laske-r. 



1976. Daily growth increments in otoliths from larval and 

 adult fishes. Fish. Bull., U.S. 74:1-8. 



Hida, T. S., and R. A. Morris. 



1963. Preliminary report on the Marquesan sardine, Haren- 

 gula uittata, in Hawaii. Pac. Sci. 17:431-437. 

 Hida, T. S., and J. H. Uchiyama. 



1977. Biology of the baitfishes Herklotsichthys punctatus and 

 Pranesus pmguis in Majuro, Marshall Islands. In R. S. 

 Shomura (editor), Collection of tuna baitfish papers, p. 

 63-68. U.S. Dep. Commer., NOAA Tech. Rep. NMFS 

 Circ. 408. 



Hunter, J. R„ and S. R. Goldberg. 



1980. Spawning incidence and batch fecundity in northern 

 anchovy, Engraulis mordax. Fish. Bull., U.S. 77:641- 

 652. 



Kwan-Ming, L. 



1960. Synopsis on the biology of Sardinella in the tropical 

 eastern Indo- Pacific area. In H. Rosa, Jr., and G. Murphy 

 (editors), Proceedings of the World Science Meeting on 

 Biology of Sardines and Related Species, Vol. 2, p. 175- 



596 



