Ovarian allometry and the use of the gonosomatic index: a case study in the Mediterranean sardine,<Emphasis Type="Italic"> Sardina pilchardus</Emphasis>

The gonosomatic or gonadosomatic index GSI=(ovary weight/fish weight)×100] has been widely used in fisheries science and experimental reproductive studies as a simple, low-cost measure of reproductive condition. However, its properties have not been fully evaluated, and several pitfalls, such as size-dependence and changes in ovarian allometry, may invalidate its use. In the present study, we examined ovarian allometry and the appropriateness of GSI for assessing ovarian activity in the Mediterranean sardine Sardina pilchardus. The analysis was based on a large sample of histologically scored females collected over an annual cycle in the eastern Mediterranean (Aegean and Ionian Seas). First, we examined GSI dependence on fish size by comparing ovary weight–on–fish weight relationships in different stages of oocyte development. The effects of recent spawning (incidence of postovulatory follicles) and intensity of follicular atresia were also addressed. In a subsequent step of the analysis, we applied generalized linear models (GLMs) to analyze the relationship between GSI and histological stage, taking into account the parallel effects of additional factors related to geographical region, month of capture, fish length and energetic reserves (fat stage, hepatosomatic index—HSI). Ovarian growth was isometric in all stages of oocyte development and states of atresia, but altered to positive allometric at the stage of oocyte hydration. Oocyte growth and intensity of atresia significantly affected GSI. Fish length, geographical region, month of capture, fat stage and HSI did not substantially affect GSI, further strengthening the appropriateness of the index for the Mediterranean sardine. Finally, we provide first evidence, based mainly on available (but yet limited) published information for other fish species, that: (1) the pattern of ovarian allometry may not be altered by fecundity variations and (2) the pattern of ovarian allometry may reflect the pattern of oocyte growth (i.e. isometric vs. allometric ovarian growth could reflect group-synchronous vs. -asynchronous oocyte development). Alternatively, the pattern of ovarian allometry could reflect the presence or absence of size dependency in the relative fecundity (eggs per gram of body weight) and/or egg size. The isometric ovarian growth in the Mediterranean sardine could be attributed to the lack of such size-dependent effects.Communicated by O. Kinne, Oldendorf/Luhe