498 PROCEEDINGS OF THE AMERICAN ACADEMY. 



the queen." (d) As Escherich (:06, p. 49) points out, if such great dif- 

 ferences as exist between the queen and worker of Carebara had arisen 

 suddenly (through mutation), unquestionably the destruction of the 

 species would have taken place, for the small worker is naturally gov- 

 erned by a wholly different sort of conduct from the queens, a wholly 

 different manner of feeding, of building, etc. If in a colony of Colo- 

 bopsis, according to Escherich, there had never been any transitional 

 forms between the workers proper and the soldiers, the latter might 

 not know how to work. 



There does not seem to be unanimity of opinion as to whether these 

 differences are due to qualitative or quantitative changes in feeding. 

 Emery is inclined to believe that qualitative feeding may play a con- 

 siderable part, since it is known that this is the case in bees ; but 

 Forel ('95, :04=^) thinks it is not safe to depend on this analogy, since it 

 has not been proved — but only conjectured — that different foi'ms of 

 ants can be created by different qualities of food ; ants have, moreover, 

 neither comb nor cells, such as bees have, but must make within their 

 own crops the differentiation of such food as they give to the larvae. 

 It seems to me that Forel's argument need not necessarily hold, in- 

 asmuch as there may be some specialization on the part of the ants 

 whereby the larvae are furnished by regurgitation with food of different 

 qualities. It is not impossible that a given ant may, at least for a cer- 

 tain time, busy herself in feeding larvae which are to be raised as 

 queens, while another is engaged in rearing a particular kind of worker, 

 etc., or certain larvae may (Wheeler, :08% p. 50; :10, p. 103) assimilate 

 only certain kinds of food. These are, it is true, only speculative sug- 

 gestions, for which there is at present no direct evidence, so far as I 

 know. However, I am inclined to think that quantitative feeding may 

 alone be sufficient to account for class distinctions, as. the organs which 

 tend to develop late may fail to receive sufficient nutriment before pu- 

 pating and therefore fail to develop. Though Wheeler (:00% pp. 27- 

 28 ; :02 ; :10, pp. 104-107) urges that there is much ground for believ- 

 ing it not only possible, but probable, that definite forms may arise 

 from differences in feeding, he found (:00'') that in Ponerine ants the 

 feeding habits were such as to make it improbable that a qualitative 

 difference of food exists. Later he (Wheeler, :07''; :08''; :10) gave 

 reasons for believing the differences of food to be quantitative rather 

 than qualitative, but adds, " a direct causal connection between under 

 feeding on the one hand and ontogenetic loss or development of char- 

 acters on the other hand, has not been satisfactorily established." 



Granting that some kind of food difference, probably quantitative, 

 exists, there are three principal views in regard to the amount of influ- 



