1898.] APPENDAGES IN THE ARTHROPODA. 949 



duced tarsi had started in the same instar, so it is evident that 

 their growth, always more rapid than that of the normal structures, 

 must have been subject to a special trophic control whereby those 

 which commenced their existence in the later instars grew more 

 rapidly than those of earlier origin. In a series of Stijlopyga 

 orienfalis there is, however, less evidence of such a special control, 

 for there is usually a well-marked difference in size between a 

 reproduced tarsus and its normal fellow in adult specimens as well 

 as in young. 



But this species affords an additional illustration of the special- 

 ized nature of the reproduced tarsus ; for, as has already been shown 

 in Table C, the several joints undergo changes in their relative 

 proportions at the attainment of the animal's maturity in the same 

 directions as do those of the normal. 



It seems probable that a statistical examination of the reproduced 

 legs of PhasmidaB would yield much the same results as above ; 

 but there is as yet not sufiScient evidence to justify a statement 

 that the reproduced antennae of Myriapoda, CoUembola, and 

 Hemiptera have so high a degi'ee of specialization. It is possible 

 that there are intrinsic differences in the two kinds of appendages 

 in their response to particular injuries. It has already been 

 pointed out that injury to the basal joints of an antenna in some 

 forms may result in numerical deficiency with its correlated 

 peculiarities, while the removal of distal portions alone is followed 

 by their regrowth with normal features. 



In this connection it is of interest that the more basal antennal 

 joints seem to be specially concerned in the formation of new 

 articulations in certain forms which progressively increase the 

 number of joints through the sequence of the ecdyses [Termitidse 

 (40), Ephemeridse (54), Phasmidas (73)]. And in the rare case 

 of a diminution in the number of joints with advancing age de- 

 scribed by Lubbock (54) as occurring in nymphs of Chloeon, it 

 is the 4th joint of a 20-jointed antenna which amalgamates with 

 itself the three joints distal to it. 



But admitting that there is greater variability in reproduced 

 antennae, it is clear that when they differ from the normal they do 

 so in the same kind of way as do the legs of Phasmidae and 

 Blattidae, and with them form a series of instances standing in 

 remarkable constrast mth those in which the reproduced appendages 

 are replicas of the normal. The causes which promote these differ- 

 ences of behaviour in allied groups are for the present quite 

 obscure. 



So also are we in the dark as to the factors which give constancy 

 to a growth which arises sporadically and has not been represented 

 in the normal ontogenetic development. It is surely therefore 

 somewhat meaningless to apply to such cases terms like " throwing 

 back "' and " reversion," as has been done by certain authors. It 

 is indeed true that reproduced parts in various animals display 

 characters similar to the normal characters of corresponding 

 parts in presumably allied genera. Perhaps the best known of such 



