493 



waste by the species of large possible sources of food available 

 if the species reacted more broadly. The iidvautage gained 

 by the narrower limitation is not obvious though perhaps the 

 species wastes no eggs and in naiure must rarely fail in readily 

 finding its host. 



The conipl(>x reflex of oviposition and the physiological 

 reactions must be, like the external physical characters of a 

 species, subject to variation whether Mendelian or Darwinian, 

 and it seems to me that these variations must have played a 

 considerable part in the evolutionary process. The elimination 

 of the olfactor}^ limitation of Allot i/pa to the smell of the ^V/r- 

 copliaga would obviously result in a wider range of food selec- 

 tion for the larva, perhaps to a wider extension of the range 

 of the species, and probably in an absolute increase in the num- 

 bers of individuals of the species prodnced so that it would 

 have greater opportunity for variation whether this might be 

 produced by internal factors, by diverse climatic conditions, by 

 change of food, by different nautral enemies, or whatever the 

 forces may be wdiich result in changes in the charactei's of 

 species. A new olfactory limitation might then arise and 

 serve as a factor in species limitation and segi'egation. 



It is interesting to revert to the Bruchidae, to speculate 

 on the few species which depart from the usual habits of the 

 family. Several species of Cari/ohonis are known to attack the 

 seeds of palms ; C. curvipes attacks several species of })alm 

 nuts inclnding the cocoanut, C. hactris and C. luteomarginatus 

 have been bred from the seeds of the carnanba palm (Coper- 

 nicia cerifcra) and an imdetermined species, like the others, 

 from South America, destroys the vegetable ivory nuts ( Plii/- 

 telephas macrocarpa), the ISTorth American C. ariliriticus feeds 

 in the larval stage in the seeds of palmetto. To account for 

 the development of such food habits and the breeding of Pseu- 

 dopachymerus pandani from ]\Iadagascar in the seeds of Fan- 

 danus, we need not assume that the parent species had any 

 greater variation in its oviposition reflexes than Caryohorus 

 gonagra has when it will lay its eggs on palm seeds and on 



