90 E. T. NEWTON ON FISHES* TAILS. 



tails disguised. Prof. Alex. Agassiz, having worked out the de- 

 velopment of some of the Teleostean fishes, is led to the conclusion 

 that although MM. Agassiz and Vogt were mistaken as to 'their 

 facts, yet, in his opinion, their generalization was in the main correct. 

 He points out that, among the Devonian fishes, there are truly 

 diphycercal tails, and every intermediate stage between this and 

 the heterocercal form, and following on into the Secondary rocks, he 

 traces through several forms the gradual equalizing of the tail lobes, 

 and the gradual approach to an externally homocercal tail. 



While accepting the facts brought forward by Prof. A. Agassiz, I 

 cannot feel that this parallelism is altogether satisfactory. For 

 although the facts probably indicate a gradual advance in the 

 structure of the tail of Ganoid fishes, yet it must be borne in mind 

 that it is only in the Ganoid group, and besides this it must be re- 

 membered that, in the Devonian rocks, we find Ganoid fishes, not of 

 the lowest type only, but of every form, from the diphycercal to the 

 heterocercal, that is to say, of just those forms which are found among 

 Ganoids living at the present day. And it might be equally well 

 argued that, so far as we have any evidence, these types have been 

 persistent through all time. And still further, the oldest known 

 fishes' tail, that of Cephalaspis, from the Upper Ludlow rocks, is 

 heterocercal. On the other hand, it is true that some of the tails 

 of Ganoids in the Secondary rocks become more nearly homocercal 

 than any that are to be found in the Palaeozoic strata. It would be 

 very interesting to know whether these highest forms of Ganoids 

 have their internal skeletal structure more like the Teleostean homo- 

 cercal tails or not. In other words, we want to know whether these 

 Secondary Ganoids were really advancing towards the Teleostei, or 

 whether they were only more advanced on their own particular lines 

 of development. 



Some fresh points of interest are to be found by taking each group 

 of fishes and tracing the different forms of tails which they present 

 in the various geological formations. 



Representatives of the two lowest groups of fishes, the Pharyn- 

 gobranchii and the Marsipobranchii, we cannot expect to find 

 fossil, inasmuch as they have no hard parts which are likely to be pre- 

 served. The tooth-like fossils from Silurian rocks, which have been 

 called Conodonts, are probably not parts of fishes allied to the lam- 

 preys, as they were at one time thought to be, but remains of annelids. 



Dipnoi. — This group of fishes is known to us for the first time 



