ON THE VARIATIONS OF AGRION PUELLA. 329 



take on my specimens, having regard to the necessity that it 

 should be the same as nearly as was possible in each specimen. 

 From base of mandible to vertex : — 



In first instar 0*16 mm. In fourth instar 0*51 mm. 

 ,, second ,, 0*24 mm. ,, fifth ,, 0*81 mm. 



,, third ,, 0*36 mm. 

 (bybernating instar) 



If we assume that the increased size of the head at each moult 

 in the case of L. semiargus is as 3 to 2 (fifty per cent.) in linear 

 measure, we find we can construct a series thus : — 



First instar 0*16 mm. Fourth instar 0'54 mm. 

 Second ,, 0*24 mm. Fifth ,, 0'81 mm. 

 Third „ 0-36 mm. 



I certainly did not expect the actual measurements and the 

 calculated ones to agree so closely ; that, I think, is a matter of 

 (fortunate) accident. There might have been considerable dis- 

 crepancies, and yet the point to be shown would have remained 

 equally clear. It is obvious that no further instar could be 

 interpolated without entirely destroying the regular rate of 

 growth that obtains in all such cases. 



I don't know whether Mr. Frohawk's last paragraph is in- 

 tended to suggest that error underlies most observations as to 

 variations in the numbers of moults, notwithstanding his imme- 

 diately preceding appreciation of Mr. Merrifield's lecture ; if so, 

 it is obvious that Mr. Frohawk's work amongst the butterflies, 

 where, I believe, his view would be, so far as I know, quite 

 correct, should not lead him to generalize in the same sense ; 

 anyone working only with Arctiids might plausibly assert that 

 there are no fixed stages in lepidopterous larvae. 



ON THE VAEIATIONS OF AGRION PUELLA, Linn. 



(Odonata.) 



By F. W. & H. Campion. 



The colour characters of Agrion puella are, in both the sexes, 

 subject to a considerable amount of variation, and most, but not 

 all, of the aberrant markings which occur in the species imj)art 

 to the individuals exhibiting them a more or less close resem- 

 blance to the nearly related species A. puldielbim. The most 

 striking resemblance of this description is borne by the varia- 

 tions of the male and female which are numbered herein 3 and 

 2 respectively. There can be little doubt that these two forms 

 have been sometimes mistaken for the species which they 

 simulate, and it may be remarked that both sexes of imlchellum 



