449 



Abstract.— A gross measure of repro- 

 ductive condition (ovary weight ad- 

 justed for body size and oocyte volume) 

 is developed and evaluated as an alter- 

 native to commonly used gonad indices, 

 for classifying the maturity status of in- 

 dividual ehu {Etelis carbunculus) and 

 kalekale tPnstipomoides sieboldii), two 

 species of eteline snappers ( Lutjanidae ) 

 that contribute to the deep-slope hand- 

 line fishery in Hawaii. Discriminant 

 analysis and logistic regression, based 

 on body length, ovary weight, and oo- 

 cyte volume, were used to classify fish 

 as either immature or mature. Dis- 

 criminant analysis correctly classified 

 the maturity of about 98% of both ehu 

 and kalekale, with histological criteria 

 as the standard for comparison. Logis- 

 tic regression correctly classified ma- 

 turity of 97% of the ehu and 100% of 

 the kalekale. Misclassification errors 

 increased by 3.75-5% (discriminant 

 analysis) or 0-5% (logistic regression) 

 if oocyte volume was excluded and only 

 body length and ovary weight were 

 used as predictors of maturity. 



For kalekale, estimates of lengths at 

 which 50% are sexually mature 'L^yi 

 were identical (29.0 ±1.8 [SE] cm fork 

 length, FL; r-=0.92), when maturity 

 was classified histologically or by logis- 

 tic regression on body, ovary, and oo- 

 cyte metrics. For ehu, L^^,, estimates 

 were similar (r2=0.96l for maturity as- 

 signments based on histology (27.9 ±2.4 

 cm FL) and on logistic regression us- 

 ing gross metrics (27.8 ±2.3 cm FL). 



We conclude that gross morpho- 

 metries can provide adequate proxies 

 for histological evidence when catego- 

 rizing sexual maturity in asynchro- 

 nous, multiple-spawning fishes like 

 eteline lutjanids. The potential benefits 

 of using gross metrics for assessing 

 sexual maturity in other serial spawn- 

 ers are briefly discussed. 



Morphometric criteria for estimating sexual 

 maturity in two snappers, Etelis carbunculus 

 and Pnstipomoides sieboldii 



Edward E. DeMartini 



Boulderson B. Lau 



Honolulu Laboratory 



Southwest Fisheries Science Center 



National Marine Fisheries Service, NOAA 



2570 Dole Street 



Honolulu, Hawaii 96822-2396 



E-mail address (for E E DeMartini) Edward DeMartiniiginoaa gov 



Manuscript accepted 28 August 1998. 

 Fish. Bull. 97:449-458 ( 1999). 



Body size at sexual maturity is of 

 fundamental importance in fishery 

 stock assessment (e.g. when calcu- 

 lating a "spawning potential ratio" 

 [SPR, Goodyear, 1993; Somerton 

 and KobayashiM '• In fisheries ap- 

 plications, average size at maturity 

 is usually expressed as the body 

 length at which 50 percent of fe- 

 males are mature (Ljg), and is typi- 

 cally estimated by fitting a model 

 such as the logistic (Gunderson et 

 al., 1980; O'Brien et al., 1993) to 

 observed percentages of mature fish 

 within several length classes. Esti- 

 mating percentage mature by size 

 class assumes that individuals can 

 be accurately classified as either 

 immature or mature. 



Various methods have been used 

 to classify individual female fish as 

 sexually mature, including gross 

 inspection of ovaries, light micro- 

 scopic examination and measure- 

 ment of whole oocytes in fixed tis- 

 sue samples, histological stains of 

 intraovarian inclusions, and an ar- 

 ray of ratios, commonly referred to 

 as "gonad indices" or "gonadoso- 

 matic indices" (GSIs) that standard- 

 ize gonad mass to body size. These 

 methods differ in precision, accu- 

 racy, and processing time. Gross 

 inspection of ovaries is quick but 

 highly subjective and often cannot 

 be used to distinguish between im- 

 mature and spent (resting) mature 

 fish (West, 1990). Light microscopy 



of whole oocytes can effectively 

 identify the most advanced oocyte 

 mode only if the dynamics of oocyte 

 growth are known for the particu- 

 lar species and may be inaccurate 

 for ovaries that also contain oocytes 

 in intermediate stages of vitellogen- 

 esis (West, 1990). Histological stain- 

 ing of tissue samples for detection 

 of cytoplasmic inclusions (yolk, oil 

 globule presence and number) in 

 the most advanced oocytes provides 

 the best evidence of maturation and 

 is the most accurate method, but it 

 is also the most time consuming and 

 expensive (Delahunty and deVla- 

 ming, 1980). 



Usually only a single method has 

 been used in a given study, although 

 some studies (e.g. Cayre and Laloe, 

 1986; Hay et al., 1987; Kjesbu, 1991; 

 Olsen and Rulifson, 1992; Ramsay 

 and Witthames, 1996) have at least 

 attempted to confirm maturity clas- 

 sifications with more than one tech- 

 nique. Average size at maturity has 

 usually (e.g. Higham and Nicholson, 

 1964; DeMartini and Fountain, 

 1981), but not always (Erickson et 

 al., 1985a; Matsuyama et al., 1987; 

 McQuinn, 1989; Hunter etal., 1992; 



' Somerton, D. A., and D. R. Kobayashi. 

 1990. A measure of overfishing and its 

 application on Hawaiian bottomfishes. Ho- 

 nolulu Laboratory, Southwest Fish. Sci. 

 Cent., Natl. Mar Fish. Serv., NOAA, Ho- 

 nolulu, HI 96822-2396. Southwest Fish. 

 Sci. Cent. Admin. Rep. H-90-10, 18 p. 



