PAPILIONINAE : IPHICLIDP:S A.IAX. 



1277 



middle of Octulicr, wlicn tlic leaves were alni08t all fallen ; the parent 

 ought to know better if instinct is so infallible." Doubleday remarks 

 that on the Ohio the species was very numerous after the tenth of Septem- 

 ber ; if this is an indication of a new brood in Septenil)er, and if the second 

 brood of ajax appears early in July in this locality also, then the butterflies 

 seen by Doubleday must have been a fourth brood of ajax or the fifth of the 

 species. Tliere is, therefore, a continual flight of this species from early 

 spring to time of frosts ; and yet in Mr. Edwards's opinion an individual 

 life hardly exceeds two or three weeks. 



The results reached by Mr. Edwards in his study of this butterfly have 

 been the subject of some curious comments by Mr. llaphael Meldola of 

 England, in an interesting paper upon the "amount of substance waste 

 undergone by insects in the pupal state." (Ann. mag. nat. hist., (4) xii : 

 301.) This writer thinks he has shown that the comparative size of the 

 three forms of ajax is exactly opposite to what we should expect. In his 

 preliminary general remarks, it is presumed a jjriori that as there is in all 

 insects gain of matter in the larval state, and loss during the pupal, the 

 size of an individual imago of any given species "would be, cneter is paribus, 

 inversely proportional to the ratio of the pupal to the larval period, or 

 directly proportional to the ratio of the larval to the pupal period." 



He attempts to test this theory by tabulating the statements of Mr. Ed- 

 wards concei-ning the duration of the stages in the different forms of ajax, 

 and he finds that there is "a relationship, but exactly the reverse of that 

 which would be anticipated from the conclusions previously set forth." 



Marcellus, telamonides and ajax, as we have seen, succeed each other in 

 season ; they also inci'ease regularly in size in the same order. The fol- 

 lowing table represents the duration of the several stages and is taken by 

 Mr. Meldola fi-om j\Ir. Edwards's work. 



The next table is Mr. Meldola's attempted tabulation of the facts, by 

 which he comes to the conclusion stated. 



Name of variety.. 



Marcellus . 

 Telamonides 

 Ajax . . . 



Ratio of mean pupal to 

 mean larval period. 



-,fe = 0.549 



: 0.80G 



Ratio' of mean larval 

 to mean pupal period. 



2f;i= 1.821 



if :| = 1.320 

 H:4 = 1-240 



Mean ex- 

 panse $ . 



2.70 

 3.00 

 3.35 



"It is here seen," says Mr. Meldola, "that the size of the variety is 

 directly instead of inversely proportional to the ratio of the pupal to the 



