Temperature and Vej'tebm 31 



specialized the vertebrae, increased their size and the com- 

 plexity of their appendages, while decreasing their number. 

 That with some exceptions and modifications this reduction 

 is characteristic of fishes in the tropics, and that it is so be- 

 cause in the tropics the processes of evolution are most active, 

 so far as the fishes are concerned. 



UNEXPECTED VARIABILITY IN THE NUMBER OF SEGMENTS. 



The most surprising feature in the present investigation is 

 that the number of segments in the adult animal should be 

 determined so late in the process of evolution and that it 

 should be so easily affected by the reaction from differences 

 in external conditions. There are several cases of species al- 

 most alike in external characters, differing one from the other 

 in the number of vertebrae, this difference being associated 

 with the distance of the range of the species from the tropics. 

 There are numerous cases in which such marked differences 

 distinguish species which no one would think of placing in 

 different genera (in Siphostoma, for example). 



In other cases {Sebastes, Sebasiodes and Sebastophis ; Lo- 

 phius and Lophiomus) genera commonly recognized are dis- 

 tinguishable only by their numbers of vertebrae. This fact 

 shows that the character in question is a recent one, arising 

 after all general matters of form, coloration and appearance 

 have become fixed. That the less number of vertebrae might 

 characterize tropical families as a whole as compared with 

 less specialized extra-tropical groups is not strange. That its 

 influence should be felt within the range of almost every 

 widely distributed family or even genus, and in some cases 

 even within the limits of a species, is certainly surprising. 



MATTERS FOR FURTHER INVESTIGATION. 



This matter has been thus far studied only in the skeletons 

 of adult fishes. It should be extended to their embryology, 

 that we may find out whether in fishes with 24 vertebrae a 

 larger number is present in the young. If so, we should 

 know by what process the segments disappear. 



