Regeneration of the Tarsus in Phasinidie. 509 



with rust-coloured bands forming equidistant rings round the 

 body and legs. The latter have a mean length of 8 millim. 

 One would therefore at first be tempted to suppose that the 

 sniallost dirterencc in length between two limbs of the same 

 pair — the one normal, the other regenerated after autotomy — 

 could not be less than this iminber of millimetres. Under 

 these conditions how are wc to su|)po3C that a limb has been 

 regenerated after autotomy, because its length is less by 

 scarcely 3 or 4 millimetres than that of the corresponding 

 limb? I succeeded nevertheless in assuring myself by experi- 

 ment that the thing was possible. In order to do this I 

 provoked autotomy in a larva which had just hatched. The 

 amputated leg measured 8 millim. and was of precisely the 

 same length as the limb opposite to it. A priori one might 

 therefore suppose that the regenerated limb would always be 

 shorter by at least 8 millim. than the one which had remained 

 intact. However, immediately after the first moult, when the 

 larva had just quitted its skin, I perceived the regenerated 

 limb forming a little spiral, which unrolled itself after a few 

 days and assumed the appearance of the other legs. The 

 length of this regenerated limb was between 7 and 8 millim., 

 so that the limb opposite to it, having increased by only 

 3 millim. and then having a length of 11 millim., the difference 

 between the two legs was little more than 3 millim. There 

 is therefore a difference between the rate of growth of the 

 normal limb and the rate of growth of the limb in coarse of 

 regeneration, this latter growing more rapidly. This pheno- 

 menon certainly affords the interpretation of the doubtful 

 case that I mentioned at the beginning of this communication. 



The question may arise whether the variation in the 

 number of the joints of the tarsus is always a consequence of 

 amputation by autotomy or whether it is sometimes con- 

 genital. Although hitherto I have never seen tetramerous 

 tarsi in larva3 of Phasmidge examined immediately after birth, 

 tetramery might nevertheless very well be observed in the 

 young on emerging from the egg, and consequently without 

 having been preceded by the slightest mutilation. 



So far from regarding autotomy as a relatively recent 

 improvement, I should, on the contrary, be disposed to bslieve 

 that in primary ages certain groups of insects already benefited 

 by the advantages of this protective process. If, indeec*, vro. 

 examine carefully certain of the drawings given by M. Ch. 

 Brongniart in his handsome memoir on primary insects 

 (' itecherehes pour servir a I'histoire des Insectes fossiles des 

 temps primaires,' 1893), and especially the figures of pi. xlix. 

 and fig. 1 of pi. xxxvii., we notice in the limbs an arrange- 



