1898.] BEEEDING OF THE DRAGOIOJT. 813 



appeared to exercise uo choice at all, but simply took the nearest 

 individual which offered the outward appearance of an able and 

 willing male. I should hesitate to believe that the enlarged 

 dorsals and brilliant colours of the male are other than a conspicu- 

 ous advertisement of his whereabouts. It is practically certain 

 that even in the small space of an aquarium tank the male cannot 

 see the female unless she is quite close to him, and it is difficult 

 to see why the converse should not hold good, were both sexes 

 equally inconspicuous. It is true that when the male has found 

 the female he continues to display his braveries, but in the absence 

 of any evidence of individual preference on her part the aesthetic 

 effect is at least doubtful. 



In a much less degree the males of the allied genus Gobius and 

 of Blennius and Clinus are distinguished by structure and colora- 

 tion from the females. The admirably careful observations of 

 Guitel ^ on the reproductive habits of these fishes give no indica- 

 tion of a sexual selection on the part of the female. 



Males of Clinus and Gobius minutv^ were observed to fight for 

 the possession of the female. Here the battles were of a serious 

 nature and were decided by force of arms, the females being left 

 to the victors. In G. ruthensparri the rivalry of the competing 

 males was not carried beyond the stage of menace, and the result 

 does not appear. In the other goby and in all the blennies 

 serious combats ensue if the possession of the nest is disputed. 

 Gobies and blennies appear to be polygamists, and if the females 

 are more numerous than the males, the selective proclivities of the 

 former are likely to be even less marked than in Callionymus, 

 where the males preponderate. 



It was observed by Savile Kent - that a male of Gobius niger, on 

 being disturbed, distended its gill-covers and branchiostegal mem- 

 branes, " with the evident intention of passing itself off as one of 

 those spiny-headed Cottidae which are not to be handled with 

 impunity." If the inference is correct, this observation is probably 

 important as bearing on the behaviour of Callionymus, since we 

 are led to suppose by the context that the male in question was 

 guarding its ova. In the forms studied by Guitel the same demon- 

 strations of form and colour were made by the males whether in 

 courting or quarrelling for possession of a mate, or in guarding the 

 nest. It is possible that G. niger does actually mimic Coitus. 

 It is perhaps equally possible that the unarmed gobies may be 

 descended from spiny-headed progenitors, and may have retained 

 the habit of protruding the once armed parts in courtship and de- 

 fence of the young, if not also for ordinary purposes of self-defence. 

 In Callionymus, as we have seen, certain demonstrations on the 

 approach of danger are to some extent common to all stages of 

 growth and to both sexes. I cannot find any important evidence 

 that these phenomena are primarily or finally mimetic of anything 



^ Arch. Zool. Exper. : G. niinutus, ser. 2, x. 1892 ; Climes and Blennius, ser. 3, 

 i. 1893 ; G. ruthensparri, ser. 3, iii. 1895. 

 2 Op. cit. p. 242. 



