1044 REPORT OF COMMISSIONER OP FISH AND FiSltERlKS. [64"J 



III Alosa the median fiu-fold is coutinuous and liigli, and at an early 

 period there is a slight vertical expansion of the caud;'.! portion of the 

 median fold, which is the prelude to a fan-shaped condition, which is 

 attained while the caudal is still eradiate and the chorda is straight at 

 its hinder extremity. In this condition the tail is outwardly homocer- 

 cal, but really lophocercal, because it is still without any other than 

 faintly developed embryonic rays. In this form there is no ontgrowth 

 of a distinct ventral lobe, but, on the contrary, the tail of the adult is 

 ileveloped directly from this symmetrical larval tail, the mesoblast of 

 the tail wandering outward directly into the caudal fold, where it is 

 transformed into the substance of the haemal pieces and medullary sub- 

 stance of the rays. This in like manner is an example of abbreviated 

 development, though it occurs in a Physostomous form, which is mani- 

 festly more primitive, because it possesses a ])erforate pneumatic duct, 

 than those Physoclistous forms which pass through that stage of devel- 

 opment of the tail which is permanent in the fishes of the Devonian 

 age. 



The causes of the failure to develop a stage in Siphostonia and Hip- 

 pocampus with a continuous median fold, such as is found in other forms, 

 are readily explicable on the ground that they are extreme modifications 

 of the ordinary Teleostean type, but when we consider the two cases ot 

 Alosa and Gamhusia amongst the Physostomes, with the modifications 

 of development which the one presents in respect to all of the median 

 fins, and the other in respect to that of the caudal, we are bound to 

 admit that they present singular and striking exceptions to the rules 

 governing the development of the tails of fishes first laid down by L. 

 Agassiz and C. Vogt. Fortunately these ('xcei)tions are very i'ew, and 

 they therefore do not affect the general principle upon winch these 

 authors insisted as much as might be supposed upon first thought, be- 

 cause after all, in the most important exception, Alosa, the tail finally 

 becomes heterocercal in structure just as in almost all other Teleosts, 

 the symmetry of the caudal of the larva) being simply an acceleration 

 of the process which ordinarily occurs, together with an apparent elis- 

 ion of some of the stages which usually accompany the tiansformation 

 of the lophocercal tail of the larva into the lioraocercal or heterocercal 

 of the adult, as the case may be. 



The writer has elsewhere expressed his views in relation to the con- 

 tinuity or discontinuity of the median and paired fins, and has sug- 

 gested the probable reasons for the existence of such differences be- 

 tween embryos of a similar age, in the following words : '■'Hippocampus 

 never develops a caudal fin, so that we would naturally not expect to 

 find the natatory fold prolonged over the end of the tail ; but the pos- 

 terior position of the early rudiments of the pectorals in Cyhium and 

 Parephippusj it appears to me, is a verj' strong argument against their 

 ■origin from a posterior branchial arch (a conclusion since reached by 

 Dohrn) ; indeed, it is the strongest yet offered against that doctrine by 



