330 



Fishery Bulletin 90(2). 1992 



IRI = (%N + %V) X %F. 



The percentage IRI was calculated for the entire data 

 set of juveniles, males, and females, and by length 

 intervals (45 mm interval for C. arel, and 21mm for 

 C. lida). 



Stomach contents were sorted, identified to the 

 lowest possible taxa, and enumerated. Appendages and 

 remains of the unidentified crustaceans are classified 

 as "crustacean fragments." Volume of each taxonomic 

 group of prey was measured by water displacement. 

 To determine ontogenetic variations in feeding habits, 

 stomach contents of juveniles were analyzed separately 

 from those of adults. 



Age and growth 



Length measurement data were grouped into size- 

 classes at intervals of 15 mm for C. arel and 7 mm for 

 C. lida. Percentage frequencies were calculated by 

 month. Sexes were treated separately, to determine 

 whether there were differences in growth patterns be- 

 tween males and females. The Petersen method, prob- 

 ability plot method, and von Bertalanffy's equation 

 were used to determine age and growth. Attempts to 

 detect growth layers in hard parts (scales, otoliths, 

 opercular bones, and supraoccipital crests) were not 

 successful. 



Petersen method This method of growrth analysis is 

 based on the assumption that the lengths of individuals 

 of the same age in a population are distributed normal- 

 ly. When there are distinct intra-annual spawning 

 periods, the length-frequency distribution may be 

 multimodal, representing successive age-groups. The 

 rate of growth slows with age (Ford 1933), and as a 

 result the modes overlap, making interpretation dif- 

 ficult. In the case of fishes, such as tonguefishes, which 

 have a prolonged spawning period, various broods 

 entering the fishery overlap. In this case, it is necessary 

 to trace a size-group for as many months as possible 

 after it enters the commercial fishery and to find the 

 average monthly growth rate for different size-classes. 

 Approximate values of average size at different ages 

 may then be calculated. 



Probability plot method Plots of cumulative per- 

 centages of length distribution on probability paper 

 provide estimates of the length ranges of fish in each 

 age-group (Harding 1949, Cassie 1954). Hence fish 

 lengths were used to obtain an approximation of the 

 length-at-age structure. One difficulty in this method 

 is the uncertainty surrounding whether the deviations 

 represent virtual inflexion points of the lines. Another 

 difficulty is locating each inflexion point, since any 



bend in the line is considered an inflexion point. Follow- 

 ing this procedure, the line was divided into separate 

 parts and for each (Cassie 1954), partial straight lines 

 were drawn from which, a mean length was calculated 

 for each age-group. 



von Bertalanffy's equation The most widely ac- 

 cepted growrth model is that of von Bertalanffy (1938), 



U = L„(l-e-Mt-t„)) 



where Lt = length at age t, 



L^ = theoretical maximum attainable (asymp- 

 totic) length, 



k = a constant, expressing the rate of change 

 in length increments with respect to t, 



to = hypothetical age at zero length, and 



e = base of Naparian or natural logarithm. 



The value of to was calculated as follows: 



-to = 1/k [loge (L^) - loge (L^-U)] - t. 



Walford's (1946) procedure was used to substitute 

 Lf -H 1 for Lt . The equation now can be written as 



Lt + 1 = L^(l-e-'^) + L,e-K 



Length-weight relationship 



Length-weight curves were obtained by using the equa- 

 tion W = aL''. The least-squares regression of the log- 

 arithmic transformation, 



logic W = logioa + b logioL, 



where logioW = Y, logioa = a, logioL = X, b = n, was 

 used for estimating the values of a and b (Snedecor 

 1956). This linear equation was fitted separately for 

 males, females, and unsexed juveniles of C. arel and 

 C. lida from monthly data. 



To determine whether increased weight at a given 

 length was caused by increased gonad weight in mature 

 fish, the length-weight relationship was compared be- 

 tween different stages of maturity. Adults of both 

 sexes of C. arel and C. lida were classified into three 

 stages (Rajaguru 1987): 



Immature (Stage I for both sexes): n 56 male and 47 



female C. arel; 105 male and 54 female C. lida; 

 Maturing (Stage II for males. Stages II-III pooled 



for females): n 221 male and 224 female C. arel; 259 



male and 342 female C. lida; and 

 Mature (Stage III for males. Stages IV-VI pooled for 



females): n 359 male and 292 female C. arel; 363 



