Table l. — Regression parameters for length-weight relations of seven species of laboratory-reared larval north- 

 west Atlantic fishes. 



IDala represent means for length and weigfit of samples of 10-25 larvae 



averaged and unaveraged data is not valid: al- 

 though he does not discredit the use of averaged 

 data by itself. Also, seven species, some of which 

 are closely related taxonomically, probably do not 

 constitute enough cases for drawing conclusions 

 about functional differences. Consequently, these 

 length-weight relations should properly be con- 

 sidered individually as empirically derived rela- 

 tions for each particular opecies. 



The length-weight relation of fishes usually ap- 

 proximates the cube law relationship in which the 

 weight is proportional to the cube of the length 

 (Beckman 1948; Rounsefe'l and Everhart 1953). 

 This is usually true for adult fishes; however, re- 

 sults of this research imply that it is not necessar- 

 ily .so for larvae. All the length exponents for the 

 species investigated in these studies were >3.6 

 with a mean value of 4.152. It would seem then 

 that the dry weight of larval fishes may be more 

 closely proportional to length to the fourth power 

 rather than cubed. Length-weight relations for 

 fish larvae are scarce in the literature. Examina- 

 tion of the data available (log,,, formulation) 

 seems to substantiate that the length exponent is 

 always greater than three and more closely ap- 

 proximates four. Marshall et al.( 1937) presented a 

 total length-dry weight equation for larval her- 

 ring, the only species with data available to com- 

 pare with this study, equivalent to log W = 

 -5.6990 + 4.52 log L. The length exponent is >4 

 and similar to the value of 4.295 for herring in this 

 research. Ehrlich et al. ( 1976) also presented a simi- 

 lar standard length-dry weight relation for Firth 

 ofClyde herring larvae ( log W = -5.7052 -(-4.5710 

 logL) as well as a relationship of log W = -4.3043 

 -i- 3.9155 log L for larvae of plaice, Pleuronectcs 

 plati'ssa. Stepien (1976) reported a standard 

 length-dry weight relationship for larval sea 

 bream, Anhosargus rhonihoidcilis, of log W = 

 -0.5144 + 4.2816 log L, and Lasker et al. ( 1970) 

 reported a standard length-dry weight relation- 



ship for northern anchovy larvae, Engrau lis nior- 

 dax. of log H' = -3.8205 + 3.3237 log L. 



It is acknowledged that variables such as tem- 

 perature and feeding conditions can influence 

 growth and complicate length-weight relations. 

 These factors may have contributed to some var- 

 iability in the present study. However, it is felt that 

 these influences were minimized by the experi- 

 mental feeding levels and temperatures which 

 were within ranges for adequate growth and sur- 

 vival, and any changes in length or weight were 

 most likely mitigated together causing little effect 

 on the form of the length-weight relation. This is 

 supported in studies of haddock larvae (Laurence 

 1974) where condition factors were similar and 

 randomly associated with prey concentrations 

 >0.5 organisms/ml. 



The use of larval length-weight relations for 

 extrapolation may result in some underestimation 

 or overestimation at the smallest and/or largest 

 sizes due to changes in growth rates for yolk-sac or 

 metamorphosing larvae. Farris (1959) suggested 

 that growth rates of larval marine fishes could be 

 separated into three different phases; the first two 

 prior to yolk absorption and the third following. 

 Zweifel and Lasker ( 1976) presented a mathemat- 

 ical interpretation of larval growth with age 

 defined by the Laird-Gompertz growth function. 

 They noticed two growth cycles; one extending 

 from hatching to yolk absorption and the other 

 following yolk absorption. This variability in the 

 small sizes is probably not inherent in the data of 

 this study because larvae were not included until 

 yolk was absorbed and active feeding had com- 

 menced. Some variability may be present in the 

 upper range of sizes in these length-weight rela- 

 tions. In some cases data for larger larvae are not 

 as extensive as for smaller larvae. Also, the major- 

 ity of the largest individuals for each species were 

 either undergoing or had completed metamor- 

 phosis where changes in growth rates of length or 



894 



