NATURAL SELECTION 283 



tendency to use the selection theory to explain away the facts 

 on which it should be based. 



(b) Differences in structure. — We have already examined a 

 particular group (the Psammocharidae) and have shown the 

 difficulty of finding an adaptive meaning in the specific and 

 generic characters. We will now describe some instances in 

 which adaptive significance has been claimed for interspecific, 

 etc., characters. 



Robson (1928, pp. 191-4) reviewed a number of these, 

 which, with some additions, are reconsidered here. 



1 . Suckers offish living in currents of varying strength (Annandale 

 and Hora, 1922, p. 507). 



The differences between the oral structures in the species 

 of Glyptosternum are discussed under the general subject of 

 the adaptation of torrent-dwelling forms (p. 265). . 



2. Character of sculpture of sternites, etc., in the Scorpions Opis- 

 thophthalmus (Hewitt, igi8, p. 98). 



Hewitt states that the coarse granulation of the sternites 

 and of the lower surfaces of the anterior caudal segments in 

 various species of this genus ' perhaps serves the purpose of 

 securing a better grip on the substratum, and it is interesting 

 to note that such coarse granulation is completely absent in 

 the species characterised by weak and elongated hands in 

 the male, in which species apparently the characteristic 

 burrowing habit of Opisthophthalmus is lacking ; still it should be 

 added that certain species with smooth sternites are undoubtedly 

 burrowers.' He goes on to say that the granulation is re- 

 stricted to this genus, in which the burrowing habit is most 

 developed. He points out that a somewhat analogous adapta- 

 tion (?) is met with in Parabuthus brevimanus and Karasbergia 

 methueni, which have independently acquired a peculiar 

 modification of the crests of the anterior caudal segments, 

 ' which would seem to indicate an adaptation to the sandy 

 habitat in which they live.' 



On re-examining this case it seems to us that there is no 

 very exact correlation between the granulation and the sandy 

 habitat (' certain species with smooth sternites are undoubtedly 

 burrowers '). There does indeed seem a tendency for the 

 sand-living forms to develop some kind of roughness on various 



