NATURAL SELECTION 287 



the hyposternum is very powerfully armed, whereas in 0. 

 moubata, in which the nymphs are rapid feeders, the dentition 

 is reduced. The exact rapidity of feeding is not given. From 

 Nuttall's figures (I.e. p. 55) of the adults of savignyi, which he 

 groups with moubata as ' rapid feeders,' it seems that there is 

 a marked difference in degree of armature between the forms 

 grouped as rapid feeders. The situation is complicated by the 

 fact that Argus persicus, which appears to be intermediate in the 

 length of its attachment, seems (Nuttall's fig. 13) to be about 

 as heavily armed as moubata. The contrast between megnini 

 and moubata is sufficiently striking ; but its value is somewhat 

 minimised by the above-mentioned differences between savignyi 

 and moubata. It is to be regretted that more exact figures 

 as to the duration of fixation in the various forms were not 

 available. 



7. Number of gill-rakers in Salmo (Regan, 1926, p. 5). 



S. obtusirostris, which lives in the rivers of Dalmatia and 

 Albania, differs from the common Salmon parr in having more 

 numerous gill-rakers on the lower part of the first gill-arch. 

 According to Regan, the number of the gill-rakers in fishes 

 generally is related to the nature of the food, being numerous 

 in microphagous forms and few in piscivorous types. ' It has 

 been recorded that obtusirostris subsists mainly on the larva? 

 of Ephemeridae,' and it seems that the increased number of 

 gill-rakers, contrasted with that of the Salmon, which is a 

 piscivorous form, is due to this difference in diet. 



Like some of the preceding cases there is a good general 

 assumption that the difference in question is related to an 

 environmental difference, though it is open to question how 

 far the diet of the Salmon parr and of S. obtusirostris is exactly 

 known. 



8. Number of vertebra in Zoarces viviparus (Schmidt, igi8 ; 

 Regan, I.e. pp. 5-6). 



Schmidt showed that in the viviparous Blenny the number 

 of vertebras decreased the further they live up certain Danish 

 fjords. Regan suggested that this is due to the diminished 

 activity of the fish in the quieter conditions of the fjord water, 

 as there is a general relation between the number of vertebrae 

 and the degree of agitation of the water. 



As the number of vertebrae in fish has been in general 



