192 THE POPULAR SCIENCE MONTHLY. 



the changes whicli occur in the development of an individual, the suc- 

 cessive forms of geological times, and the series of living forms, lower 

 and higher, or more generalized and more specialized. 



In the smallest gar here described, and presumably in still younger 

 examples, the axis of tl)e body, represented by the notochord or primi- 

 tive vertebral column, is nearly horizontal, about midway between the 

 upper and the lower borders of the tail. This is likewise the case with 

 the lowest known vertebrate, Amphioxus ; with the forms next above, 

 the hag-fishes [Myxine and Bdellostoma) and lamprey-eels {Petrotny- 

 Z07i) ; with the larvoe (tadpoles) of frogs and toads ; and with the 

 adults of the aquatic and tadpole-like salamanders, Menopoma and 

 Menobranchus. 



Finally, such a tail exists in the Dipnoans, or mud-fishes, of Africa, 

 South America, and Australia {Frotopterus^ Lepidosiren, and Cerato- 

 dus), which have some striking aflfinities with Batrachians, but are 

 usually regarded as fishes, and are, perhaps, the best illustration of 

 generalized forms. 



To this variety of tail, Cope has applied the name isocercal ; Hux- 

 ley calls it diphycercal, and gives as an example Polyjyterus, where, 

 however (as in Oalamoichthys), the "end of the notochord is hardly 

 at all bent up." Wyman, finding this kind of tail in the embryo of a 

 skate, called \t protocercal, and, on some accounts, this seems the more 

 suitable name. 



As the gars grow older, the relative length of the filament and the 

 infra-caudal lobe constantly changes. At first the former is the longer; 

 in a specimen 108 millimetres long, their tips coincide; in one 142 

 millimetres long, tlie lobe projects beyond the filament ; and in a 

 third, 300 millimetres long, the filament is much the shorter, is ragged 

 and attenuated, and during life was rarely employed. This second 

 stage, or rather series of stages, has several counterparts among liv- 

 ing Selachians and Ganoids. The most accurate resemblance is pre- 

 sented by the shovel-nosed sturgeon of the Mississippi River {Sca2:)hy- 

 rhynchus). The filament is excessively elongated in Chimera, and 

 exaggerated as to both length and breadth in the thrashing-shark 

 (Alopias). But, with many sliarks, the common sturgeons, and the 

 spoonbill {Polyod(m), the size of the infra-caudal lobe is so nearly 

 that of the filament as to give the whole tail a nearly symmetrical 

 outline, and lead zoologists to speak of the " uj^per lobe," whereas it 

 is really the bent-up end of the body. This kind of tail is called hete- 

 rocercal. 



The gars above mentioned are supposed to be the young of the 

 Zepidosteus osseus. Just at what size the filament wholly disappears 

 in that species is not known. But with the smaller and proportionally 

 shorter species, i. platystomus, tliere is no sign of the filament when 

 eighteen inches in length. The tail might then be thought, at first 

 sight, to be symmetrical. But the longest rays are a little above the 



