6i6 Subclass Dipneusti, or Lung-fishes 



generally to be of much use in locomotion on land. How- 

 ever, it is quite possible that it is occasionally compelled to 

 leave the water, although we cannot believe that it can exist 

 AMthout it in a lively condition for any length of time, 



" Of its propagation or development we know nothing except 

 that it deposits a great number of eggs of the size of those of 

 a newt, and enveloped in a gelatinous case. We may infer 

 that the young are provided with external gills, as in Pro- 

 toptcrus and Polyptcrus. 



"The discovery of Ccratodiis does not date farther back 

 than the year 1S70, and proved to be of 

 the greatest interest, not only on account 

 of the relation of this creature to the other 

 living Dipneusti and Ganoidei, but also 

 because it threw fresh light on those 

 singular fossil teeth which are found in 

 strata of Triassic and Jurassic formations 

 Yu!. 3S7 _Lower iaw of ii^ various parts of Europe, India, and 



,\,acer„tndn.'<for.^teri Gun- America. These teeth, of which there 

 tliLT. (After CTiintner.) . . 



is a great variety with regard to general 



shape and size, are sometimes two inches long, much longer 



than broad, depressed, Avith a flat or slightly undulated, always 



punctated, crown, with one margin convex, and with from 



three to seven prongs projecting on the opposite margin." 



Development of Neoceratodus. — From Dean's "Fishes, Recent 

 and Fossil," pp. 218-221, we condense the following account 

 (after the observations of Dr. F. Semon) of the larval history 

 of the Barramunda, Neoceratodus jorsteri: 



It offers characters of exceptional interest, uniting fea- 

 tures of Ganoids with those of Cyclostomes and Amphibians. 



The ncAvly hatched Neoceratodus does not strikingly re- 

 semble the early larva of shark. No yolk-sac occurs, and the 

 distributir)n of the yolk material in the ventral and especially 

 the hinder ventral region is suggestive rather of lamprey or 

 amphibian ; it is, in fact, as though the quantum of ^^olk mate- 

 rial had been so reduced that the l;)ody form had not been con- 

 stricted off from it. The caudal tip in this stage appears, how- 

 ever, to resemble that of the shark, and , as far as can be inferred 

 from surface views, a neurenteric canal persists. Like the 



