504 



Fishery Bulletin 89(3), 1991 



values covering the full range of sizes, with energy rate 

 terms expressed in fish tissue gram equivalents. Age 

 and standard length were computed from the von Ber- 

 talanffy model, and energy components were calculated 

 as described in the Methods section. Entries within a 

 row in Columns 1-7 are all raw, unweighted values for 

 an individual of a particular weight-class. Thus, Col- 

 umn 7 contains the estimated ration of a weight-class. 

 In Column 9, this value is weighted by the fraction 

 (Column 8) which that weight-class contributes to the 

 total population. Thus, for C. melampygus the sum of 

 all 38 "proportional ration" values such as the exam- 

 ples in Column 9 constitutes the ration of a hypothet- 

 ical individual representative of the entire population, 

 i.e., 47.82 kg/year of prey. Similarly, for C. ignobilis 

 the sum of all 71 "proportional ration" values provides 

 the annual ration of such a representative individual, 

 i.e., 150.69kg/year. 



The best estimates of species populations for FFS 

 that our census data permit are about 230,000 C. me- 

 lampygus and 130,000 C. ignobilis (Table 6). Combined 

 with the estimate of annual ration of the representative 

 individual of each species, these values produce esti- 

 mates of total annual consumption of about 11,000 

 metric tons (t)/year by C. melampygus (about 22 t/year 

 per km 2 of prime habitat), and about 19,600 t/year by 

 C. ignobilis (about 39.2 t/year per km 2 ). Table 6 shows 

 the annual amount eaten of each major prey category, 

 based on results in Tables 3 and 4. 



Discussion 



Age and growth 



Length-weight data for these species have been previ- 

 ously reported. Seki (1986a) reported a power function 

 regression based on 124 C. ignobilis specimens caught 

 in the Northwestern Hawaiian Islands, ranging in FL 

 from 207 to 1330 mm. Applying the FL-SL relation- 

 ship of Table 1 to his result produced the expression, 



W = 3.44 x 10" 5 (SL - 6.0) 2 



913 



Over the range of sizes collected in both studies, predic- 

 tions of W from this relationship and that of Table 1 

 (based on 118 NWHI specimens) agree within about 7% 

 at worst, and for a large majority of the specimens they 

 agree within 2-5%. Their predictions are most diver- 

 gent (over 8.5%) as L^ is approached. 



Seki (1986b) also reported an expression for FL 

 versus weight based on only 24 specimens of C. melam- 

 pygus, 93-710 mm FL. Applying the FL-SL relation 

 of Table 1 to this result produced the expression, 



W = 3.0053 x 10 - 5 (SL - 1.5) 



2.941 



The agreement between predictions of this expression 

 and those of Table 1 (based on 149 NWHI specimens) 



