LENGTH-WEIGHT RELATIONS FOR 



FIVE EASTERN TROPICAL 



ATLANTIC SCOMBRIDS 



This paper presents an analysis of fork lengths 

 and body weights of five species of scombrids 

 measured from landings at several ports on the 

 west coast of Africa during 1967 and 1968: 

 yellowfin tuna, Thunnus albacares; skipjack tuna, 

 Katsuwonus pelamis; bigeye tuna, T. obesus; 

 little tunny, Euthynnus alletteratus; and frigate 

 mackerel, Auxis sp. Sampling of landings took 

 place between 26 September 1967 and 22 May 

 1968 at the ports of Dakar, Senegal; Freetown, 

 Sierra Leone; Abidjan, Ivory Coast; Tema, Ghana; 

 and Benguela, Angola. Samples were also taken 

 from fish stored at a cannery in Mocamedes, 

 Angola. Fish were captured by bait (pole-and- 

 line) boats, purse seiners, and combinations of 

 both. Only whole fish were used for this study, 

 landed in fresh, iced, frozen, and indeterminate 

 conditions. Fork lengths were usually measured 

 to the nearest centimeter. Weight was usually 

 measured to the nearest 0.1 kg. All nonmetric 

 data were converted to centimeters and kilograms. 

 The allometric length-weight equation is used 

 to describe the relation between length and 

 weight: 



W = aL^e (1) 



where W = weight in kilograms 



L = length in centimeters 

 a and 6 = estimated parameters. 

 e = error term 



Results 



Estimates of a and b were made for each sample. 

 A wide range in values of a and b occurred for 

 the same species and, in some cases, for the iden- 

 tical sample category (category is defined as port, 

 gear, and method of preservation offish), that was 

 at first alarming. However, examination of plots of 

 the estimated curves revealed only minor dif- 

 ferences among samples at sizes included in the 

 samples. It was also noted that estimates of a 

 are closely related to estimates of b (Figure 1), 

 again indicating the fish at the same length 

 weighed approximately the same. 



Analyses of covariance were used to test the 

 statistical significance of differences among 

 length-weight relations within a sample category. 

 F-tests for the significance of differences of the 



.001 



-J 1 I I 1 I I I I I 



2.3 2.4 2.5 2.6 2.7 2.8 2.9 3.0 3.1 3.2 3.3 3.4 3.5 3.6 



b 



Figure 1. — Relation between estimates of a and b of the allo- 

 metric length- weight relation from samples of Atlantic yellow- 

 fin tuna. 



estimates of both parameters a and b^ were made 

 instead of F-tests for each parameter as is usually 

 done, because I believe that the close relation 

 between estimates of a and b demonstrates that 

 no additional useful information would be 

 obtained by making the separate tests. F-values 

 for differences among samples within a category 

 were almost always significant for all species with 

 more than one sample. As mentioned previously, 

 plots of the fitted lines showed only minor dif- 

 ferences between samples for sizes found in both 

 samples. 



Analyses of covariance were also used to test 

 whether differences among sample categories 

 were present. Nested models were used because 

 the significant differences among lines within 

 sample categories indicated that samples rather 

 than individual fish should be used to estimate 

 the error term of the model. Only data for yellow- 

 fin and skipjack tunas were examined because 

 there were insufficient data for the other species. 

 Table 1 presents the analysis of covariance of 

 differences among all sample categories for 

 yellowfin tuna. The F-value for difference among 

 sample categories is statistically significant at the 



'Hr 



and 



of a from (th sample, 6j 



bj = bj where Cj = value 

 = value of b from ;th sample and 



848 



