Gabr et al.: Reproductive versus somatic growth during the life cycle of Sepia pharaonis 



803 



tion affects the growth and condition of somatic tis- 

 sue in squid (e.g. Rowe and Mangold-Wirz, 1975; 

 Hatfield et al., 1992), few studies have examined the 

 expenditure of nutrients and energy for the develop- 

 ment of reproductive organs in cuttlefish. Boucaud- 

 Camou (1971) found that the digestive gland of Se- 

 pia officinalis has high concentrations of lipids, sug- 

 gesting that it acts as a storage organ. Castro et al. 

 (1991) studied changes in the digestive gland of .S. 

 officinalis and S. elegans throughout their life cycles 

 and observed that the digestive gland weight of S. 

 officinalis decreased progi'essively with stai-vation. 

 However, there is no evidence that the digestive gland 

 provides stored energy for gonad development. 



Sepia pharaonis is widely distributed in the Indo- 

 Pacific from the Red Sea to Japan and Australia. 

 Together with S. dollfusi, it is the primary fishery in 

 the Suez Canal and the most valuable commercial 

 cephalopod in the northern Indian Ocean (Nesis, 

 1987). Little is known about the biology of this spe- 

 cies (Silas et al., 1985; Aoyama and Nguyen, 1989; 

 Gabr et al., 1998). In this paper, we examine aspects 

 of its life history to determine if a trade-off exists 

 between somatic and reproductive tissues. We mea- 

 sured food intake to determine whether nutritional 

 activity was involved in the development of repro- 

 ductive organs. 



Materials and methods 



Samples of Sepia phai-aonis were collected monthly 

 from Bitter Lake, the main fishing port in the Suez 

 Canal, from September 1994 to April 1996. In total, 

 1428 females and 1151 males were examined. The 

 dorsal mantle lengths ranged from 10 to 240 mm. 

 Specimens were dissected to determine sex and ma- 

 turity stage. Four maturity stages for each sex were 

 determined by using a modification of the scale pro- 

 posed by Mangold-Wirz ( 1963) and refined in detail 

 for Sepia pharaonis by Gabr et al. ( 1998 ): for females, 

 I=immature, II=maturing, III=pre-spawning, and 

 IV=spawning; for males, I=immature, II=maturing, 

 III=fully mature, and IV=spawning. 



Dorsal mantle length (ML) and nidamental gland 

 length (NGL) of females were measured to the near- 

 est mm. The following measurements of mass were 

 made to two decimal places (in grams): total body 

 mass (BM); total somatic mass (SM); mantle mass 

 (MM); head mass including arms and tentacles (HM); 

 digestive gland mass (DGM); viscera mass (VM) in- 

 cluding gills, stomach, caecum, pancreas, ink sac; 

 ovary mass ( OM ) including oviduct; testis mass ( TM ); 

 nidamental gland mass (NGM); and spermatophoric 

 complex mass (SCM). 



Relative assessment of somatic tissue growth 



Two statistical analyses were conducted to determine 

 if somatic tissue declined in relation to size during 

 maturation. All statistical analyses were carried out 

 by using the MINITAB statistical package and all 

 data were logjg-transformed. 



In the first analysis, a multiple regression was 

 applied to the female data only by using logjg mantle 

 length and logjQ nidamental gland length as inde- 

 pendent variables. These variables were selected 

 because they correlated with our maturity scale 

 (Gabr et al., 1998). Multiple regression equations 

 were obtained for log-transformed total somatic mass 

 and for masses of mantle, head, digestive gland, vis- 

 cera, and ovary (including oviduct and oviductal 

 gland, and nidamental gland). This analysis allowed 

 for the use of a continuous variable in the assess- 

 ment of maturity over the whole size range of the 

 sampled population. This multiple regression analy- 

 sis was carried out only on females because length 

 was used in the analysis. To carry out a similar exer- 

 cise with male cuttlefish, the relation of the log-trans- 

 formed mass of each organ with spermatophoric com- 

 plex mass (SCM) and body mass would have to be 

 used. These variables would lead to problems in scal- 

 ing because they would involve autocorrelation, 

 which would distort perceived relationships (LaBar- 

 bera, 1989). The ratio of SCM to ML has not been 

 tested before as an objective measure of maturity in 

 male cuttlefish and therefore no analysis was in- 

 cluded here. 



In the second analysis, a multivariate analysis of 

 covariance (ANCOVA) was applied to the total so- 

 matic mass, individual somatic organs, and repro- 

 ductive organs, whose relationships with the repro- 

 ductive cycle were considered dependent variables 

 (Garcia-Berthou and Moreno-Amich, 1993). Mantle 

 length was ti'eated as the covariate. A fundamental 

 assumption of standard ANCOVA (McCullagh and 

 Nelder, 1983) is the homogeneity of regression coef- 

 ficients (slopes) of dependent variable and covariate 

 relationships. This assumption can be tested with a 

 special design of ANCOVA, by analyzing the pooled 

 covariate by factor interaction. If the covariate by 

 factor interaction is significant, standard ANCOVA 

 should not be developed. Otherwise, if the covariant 

 by factor interaction is not significant, the standard 

 ANCOVA design is the preferred method. For the 

 cases with significant effect with factor (maturity 

 stage), the variation can be described by using the 

 predicted means for each cell, adjusted for the effect 

 of the covariate. We set the alpha level for statistical 

 significance at 0.001 to identify the strongest effects 

 of maturity on the dependent variables. 



