Length-Converted Catch Curves and the Seasonal Growth of Fishes*

A brief review of studies on the seasonal growth of fish is presented, followed by an equally brief review of length-converted catch curves. A new method for constructing catch curves from representative length-frequency data is presented. This new method explicitly accounts for seasonal growth and thus eliminates the bias in Z caused by such growth. Some practical and theoretical implications of the new m~thod are discussed. Fig. 1. Length-frequency data on the gobiid Chasmichthys dolichognathus, fitted with a seasonally osdllatinggrowth curve by means of ELEFAN I. The original length-frequency data, gathered from April to December 1974 (with the exception of the January-February sample, obtained in 1970), have been here plotted twice to show that the forward projection of the growth curvemeets the modal dass(es) of most samples. The curve has the parameters L_ =6 em, K= 1.0 yearl , C = 1.0 and WP = 0, the latter two values suggesting a period of no growth at the turn of the year. Adapted from Pauly and David (1981), based on data in Tamura and Honma (1977). Introduction That the growth of fishes displays seasonal growth oscillations was well known to the pioneers of fishery biology, notablytoT.W.Fulton (1901, 1904), who along with e.G.]. Petersen, invented length-frequency analysis. This awareness faded away, however, when fishery scientists gradually switched away from the analysis of length data and used "annuli" (on otoliths, scales and other bones) to estimate growth rate and draw growth curves (Went 1972). Thus, Beverton and Holt, in their classic of 1957, did not consider seasonal growth oscillations in more than a cursory manner, and particularly, saw no point in modifying the basic von Bertalanffy growth function (VBGF) to express such oscillations, although they occur in all the fishes they studied. Following a discussion of seasonal growth by von Bertalanffyand Muller (1943), the first version of the VBGF allowing for such oscillation was that of Ursin (1963a, 1963b). Other modifications of the VBGF were those of Pitcherand MacDonald (1973) and Daget and Ecoutin (1976). Improvements of these earlier models and various approaches for fitting them then fonowed in qUick successions (Cloero nnd Nichols 1978; Antoineetal1979; Pauly and Gaschiilz 1979; Hoenig and Chaudary ~ Hanumara 1982; Sager 1984a, 1984b, 1984c; ;J. & Appc1doom 1987; Moreau 1987; Somers 1988; Soriano and Jarre 1988; Soriano and Pauly 1989; Chaudary Hanumara and Hoenig 1990; Gaschiitz et aI. 1990). The appIica lion examples presented by these authors made it quite obvious that growth models which do not explicitly consider seasonal osciIIalions fail to capture an essential aspect of the growth process (Longhurstand Pauly 1987and see