DISTRIBUTION AND GENERAL NOTES ON THE SPECIES 293 



length could be used to compute the ponderal index or condition factor K from each weight recorded. 



All the reliable data of this nature are given in Appendix lie. From them it has been found possible 



to trace the seasonal variation in K by using mean values from appropriate groupings of the results. 



Throughout this work the 'ponderal index', ' condition factor' or ' weight-length coefficient' K has 



been calculated from the formula 



„ w 

 K=™Xioo, 



where w = mean weight in grams and Z=mean length in cm. 



The results suggest important analogies with those Hickling (19306) obtained for the European 

 species, but there are divergences, some due to the different methods employed, others probably to 

 inherent differences between the two species. Here again it must be emphasized that Hickling, 

 dealing with a single species, was able to concentrate his efforts, and weighed individual gutted fish 

 and organs. Gunther and Rayner, on the other hand, were investigating virgin ground, so that it was 

 necessary for them to ' do something about' everything that came up in the trawl, including ' rubbish'. 

 The marvel is that they found time to make as many bulk weighings as they did, in addition to the 

 vast number of individual measurements. 



K values derived from heavily grouped data like these may still give us a good idea of the broad 

 outline of the seasonal cycle for the species, although they cannot be expected to prove so accurate as 

 more detailed results, especially if one were to attempt to apply them to individual fishes. They are 

 mean K values, derived from meati weights and mean lengths. Applied to whole samples of fishes they 

 permit of surprisingly accurate estimates of weight from known mean length (and the converse) as 

 will be shown later, and these may be of great practical value. If, however, they are applied to the 

 study of variations in condition of limited subsamples, they show features in contradiction to what 

 one would expect from Hickling's more detailed work on the European species. 



Apart from the seasonal variation in condition which this method is particularly adapted to show, 

 there is superimposed upon it a secondary variation related to the length (here the mean length) of the 

 fish. Older (longer) fishes tend to show a slightly lower level of condition throughout the seasonal 

 cycle, consequent upon the increased metabolic strain of spawning. The point of inflexion on a curve 

 showing this diminution of K with increasing length is thus a good approximate indication of the 

 length at which sexual maturity is attained. Analogous findings in several other species of fishes 

 could be quoted, but probably the best general exposition of the more important deductions to be 

 derived from the study of K values is that given by Sir D'Arcy Thompson (1942, p. 194 et seq.). 



In M. hubbsi the ' average 'i values of K in relation to length of either sex are shown in Fig. 23. 

 From this it can be seen that a majority of the males probably mature at a length of about 32 cm. 

 (some certainly do so when still smaller), while it is probable that most of the females below 42 cm. 

 length are immature. Considering this in conjunction with our previous findings of a difference of 

 over 10 cm. between the sexes (in trawl-caught samples), it seems possible either that the males 

 mature at least one year earUer than the females, or that the early growth rate of females is li times 

 as fast as that of the males. The former is more probable, but the great scarcity of males over 50 cm. 

 long while females diminish in numbers more gradually to twice that length, makes it fairly certain 

 that the growth rate of the older mature males is much lower than that of females of comparable size. 

 Condition factors for males of 50 cm. and over were too few to include in any 'average' curve, but 

 those we have (records of five individual fish) are all very low, suggesting complete analogy with 



1 I use this expression 'average' to indicate values based on means of means, not on the arithmetic mean that would be 

 derive^from full^integration of fll the constituent data. This device is necessary here m order to ehmmate the seasonal eflFect 

 when studying the relation between length and ponderal mdex. 



