[49] THE EVOLUTION OF THE FINS OF FISHES. 1029 



lations of the interspinous boDes of the dorsal and anal fius of Molacan- 

 thus and Molajuv. may be shown, without any doubt, to merely represent 

 stages of development of the median pieces which support those fins. 

 Tabulating- the relationship of the interspinous neural and haemal ele- 

 ments, as given by Mr. Putnam, this will become apparent. 



Dorsal interspinous bones connected with spines of fourth to seven- 

 teenth vertebrse in Molacanthus ; with spines of fifth to fifteenth verte- 

 brae in Mola. 



Haemal interspinous bones connected with spines of tenth to seven- 

 teenth vertebrse in Molacanthus ; with spines of tenth to sixteenth ver- 

 tebrae in Mola. 



The extension of the dorsal and anal fins posteriorly in Molacanthus^ 

 so as to be brought into relation with two more neural spines and one 

 more haemal spine than in Mola juv., is very readily explained by the 

 fact that these fins are, as a whole, apparently more posterior in posi- 

 tion in the specimens of Mola still younger than the one figured as Mola- 

 canthus by Putnam, and that there is an apparent but not real whole- 

 sale shifting of the bases of these fins in a fotward direction in order to 

 reach the condition found in the adult, due to the outgrowth posteriorly 

 of the caudal, in consequence of which the dorsal and anal are shoved 

 apart. This line of argument is shown to be in the very highest degree 

 the correct one, from the fact that in a very small pelagic fish, which 

 has been called Ostracion hoops, and now regarded by Liitken, Steen- 

 strui), and Giinther as a still younger stage of Mola than that repre- 

 sented by Molacanthus, the dorsal and anal fins are very closely approxi- 

 mated over the end of the body, so as to appear almost like the forked 

 tail of a normal Teleost. 



One apparent difficulty now presents itself with respect to the appar- 

 ent shifting backwards of the dorsal fin of Mola to the extent of one 

 neural spine farther to the rear in front, so that the first ray of the dor- 

 sal seems at first sight to be in relation with the spine of the fifth dorsal 

 instead of the fourth dorsal as in Molacanthus. This difficulty, how- 

 ever, disappears at once when we find that in the adult the interspinous 

 piece belonging to the fourth vertebra is really present, but is so closely 

 appressed to that of the fifth as to seem at first as if it were a part of 

 the latter. The dorsal spine of the fourth vertebra does not reach back 

 to the corresponding interspinous piece, but is greatly depressed and 

 lies partly in close contact with the upper side of the fifth vertebra in 

 almost a horizontal position. It is evident that the interval between the 

 posterior end of the spine of the fourth vertebra and its corresponding 

 interneural piece has arisen in the course of growth, and is primarily 

 due to the fact that the body grows rapidly in length anteriorly after 

 the Molacanthus stage has been passed, so that the intimate connection 

 of these elements would be sundered, so as to give rise to the interval 

 between them as observed in the adult. 



It is evident, therefore, that not a single one of these anatomical char- 



