212 Pj'oceedings of the Royal Physical Society. 



many of these as are available, in order to reach a reliable 

 conclusion. 



Comparative Anatomy teaches us that of the two types 

 of tail, diphycercal and heterocercal, the former is un- 

 questionably the more archaic. As a piece of mechanism, 

 it is very much simpler and less perfect. In its muscular 

 and skeletal arrangements it shows far less divergence from 

 the ordinary arrangements seen in the regions lying anterior 

 to it, and of which it must be looked on merely as a modifica- 

 tion. Further, it is the type of tail which is alone found 

 amongst the Cephalochordata and the Cyclostomata. 



Embryology fails to afford a shred of evidence for the 

 former existence of a heterocercal tail. The tail is typically 

 protocercal throughout ontogenetic development. 



Palaeontology — or rather such knowledge of palaeontology 

 as we happen to possess at the present time — affords the 

 evidence upon which the view is based, which sees in the 

 diphycercal tail and continuous median fin of present-day 

 Dipnoans, characters which are not primitive, but developed 

 secondarily, the ancestral forms having possessed hetero- 

 cercal tails and divided-up median fins. The evidence upon 

 which this view, so opposed to morphological probabilities, 

 is based, consists of the fact that amongst palaeozoic Dipnoi 

 many forms are known with the last-mentioned set of char- 

 acters, while none are known with the first mentioned ; and 

 further, upon the fact that it is possible to select out from 

 such known palaeozoic forms a series, the members of which, 

 from successive geological formations, show a series of steps 

 from the heterocercal type, with segmented median fin, 

 towards, though not reaching, the diphycercal type, with 

 continuous median fin. Of these two pieces of evidence, I 

 personally, for the reasons already given — particularly on 

 account of the necessarily extraordinarily incomplete state 

 of our knowledge of extinct faunas — should attach 

 little weight to the last mentioned. The weight of the 

 first also is lessened fatally, so it seems to me, by the 

 following considerations. The palaeozoic times, from which 

 the evidence has been obtained, were times when fishes allied 

 to what we should to-day call Dipnoans, and Crossoptery- 



