250 THE VARIATION, RACES AND MIGRATIONS OF THE MACKEREL. 



and only 4 females. During another season he counted 17 males to 

 8 females. Tt is true that on a third occa.sion Couch found the sexes 

 to be equally represented in a sample of the earliest spring fish ; but he 

 explains this exception by adopting the opinion of the fishermen that 

 the fish of this particular school belonged to a different class from the 

 ordinary spring fish — that they were old fish of the preceding season 

 which had not moved out into deep water, and that they did not 

 represent the usual spring immigrants from the Atlantic. 



The evidence of American observers on this point is conflicting 

 (Commissioner's Eeport for 1881, U.S. Fish Com., 1884, p. 114). 



My own observations on the proportion of the sexes in various 

 samples of mackerel at different seasons are recorded in table E (p. 293). 

 The only sample examined during March was one from Smerwick, 

 County Kerry, and it certainly supports Couch's statements that in 

 this particular sample the males were nearly twice as numerous as the 

 females, 63 out of a total of 99 fish being of the former sex. 



But it must be pointed out at the same time that an actual excess of 

 males over females in particular samples is also found at other seasons 

 of the year. Thus, out of 25 Lowestoft fish, captured October 12th, 

 1897, there were 15 males to 10 females; and out of 100 Eamsgate 

 fish caught a fortnight later 55 were males. In the American sample, 

 also caught during October, 55% were males. The preponderance of 

 males in these cases is certainly not so great as in the Smerwick 

 sample, but it is sufhcient to show the necessity of caution in accepting 

 conclusions of this kind based upon relatively small samples. Until 

 observations on the subject shall have been considerably multiplied, 

 I think it will be best to suspend judgment upon a matter which 

 may be, biologically, of considerable importance. 



In this connexion I would draw attention to the evidence submitted 

 in the succeeding section, which shows that the numerical proportion of 

 the sexes differs, as a whole, according to the size of fish under con- 

 sideration, owing to a difference in the rate and limits of growth of the 

 two sexes. Now it is an open question whether the so-called "schools" 

 of mackerel are formed by the chance association of fish which happen 

 to be near one another, independently of sex and size, or whether they 

 are not due to the selective association of fish having some common 

 characteristic. The evidence derived from samples of fish caught in 

 drift-nets is of little value, owing to the selective action of the mesh of 

 the nets. The evidence from samples caught in small-meshed seines 

 would be fairly conclusive, but the existing data are too few to be 

 of any use. Nevertheless, such evidence as we possess in the case 

 of other fishes tends to show that selective association plays a consider- 

 able part in the formation of shoals, and that similarity of size is one 



