314 ON THE BREEDING OF THE DEAGONET. [Apr. 19, 



in particular '. The preopercular spines being reserved as a second 

 line of defence, in case the animal be actually seized, it appears to 

 me that the object is either to simply disconcert the enemy by a 

 rapid change of form, or to convey an exaggerated impression of 

 size and strength. It is on some such line that one may suppose 

 that a finally mimetic condition (as instanced in the Puss-moth 

 caterpillar^) has been finally evolved, but in Callionymus it appears 

 to go no further in this direction. As it seems to me a primarily 

 aposematic feature has been seized upon and intensely developed, 

 by the aid of a coloration perhaps resulting from a primarily adven- 

 titious excretory process, by a sexual selection acting, as befits the 

 environment, rather in the direction of conspicuousness than of 

 aesthetic charm\ A term, " sematepigamic," must perhaps be 

 coined to suit the present condition. 



If Stolzmann is right in considering the dances of male birds, 

 not as a peaceful strife, but as a distraction to protect the female 

 againt the too constant attentions of the male, the same interpre- 

 tation can hardly be placed on the courting antics of the male 

 Dragonet. For here the only difficulty which appears to be felt 

 by the female is to get as much male society as she wants. 



Unless the female is degenerate, which I do not think we are 

 entitled to assume, the free use by the male of his special characters 

 for aposematic purposes appears to be of secondary origin; or, 

 perhaps more justly, the male simply continues the aposematic 

 demonstrations of his youth with apparatus that, fortunately for 

 himself, has been improved by a sematepigamic process of selection. 

 Beddard * has suggested that similar, but not sexually-diiferenti- 

 ated structures, the enormous pectoral of Dactyhpterus and the 

 dorsal fin of Thymallus, may be effective in diverting the attack 

 of an enemy to a non-vital part. This is possibly the case, since 

 we have seen that one male Dragonet struck at the first dorsal of 

 another. The tub-gurnard (Trigla Mrundo) furls its large and 

 beautiful pectorals when at rest ; but they are instantly expanded if 

 the fish is molested, and are kept expanded when the fish is driven 

 about or is simply swimming round the tank undisturbed. The John 

 Dory instantly erects the dorsal filaments when alarmed, and these 

 are supported by very powerful spines. On the whole it appears 

 most probable in the two cases last mentioned and in the Dragonet 



1 I have observed that Pike {Esox lucius), when quai-relling, menace each other 

 by inflating the whole gill-apparatus. In its natural environment it must be 

 long since the Pike was associated with any object of mimicry more formidable 

 than itself. 



» Cf. Poulton, 01). cit. p. 271. 



^ Though it is possible that the brilliant coloration was originally acquired 

 in shallow water, where details could be more readily appreciated. 



* Op. cit. p. 191. Messrs. W.L. Calderwood and G. P. Bidder have told 

 me that when a Dactyhpterus was placed in a tank at Naples containing some 

 small sharks, the latter bit pieces out of its pectorals, a liberty resented by 

 violent grunting. Gurnards and dories also grunt under circumstances of 

 discomfort, the sounds being of the nature which appears from experiment to 

 be perceptible by fishes. They may possibly subserve a function which is in 

 part protective. 



