1898.] APPENDAGES IN 'XHE ABTHftOPODA. 939 



As iu the case o£ normal spurs, supernumerary ones were 

 always situated on the sides of the tarsal joints and never in the 

 mid ventral line. In all except three cases from S. orientalis the 

 supernumerary spurs were placed quite close to the normal spurs. 

 Hence the most common condition — that of one supernumerary 

 spur — was expressed as one spur on one side and two on the 

 other. In cases where the total number of spurs on a joint 

 exceeded three there were supeimumerary spurs on both sides : 

 for instance five spurs would be expressed as three on one side 

 and two on the other. " This tendency to uniformity in the 

 distribution of supernumerary spurs was observed in all cases 

 except in five from S. orientalis and one from P. americana. In 

 these latter there were either three or four spurs on one side of 

 the joint and only a single (the normal) one on the other. The 

 cases of abnormal armature showed no indication of any relation 

 between the number of supernumerary spurs on a single joint and 

 the extent to which a supernumerary armature occurred in the 

 tarsus as a whole. The joints seemed to vary quite independently 

 of each other in respect of their armature. But on the whole 

 there is obviously a greater frequency of abnormality of the 

 armature in reproduced than in normal tarsi. In this connection 

 it may be noted that Newport observed that the reproduced 

 appendages of Arthropods are particularly liable to disturbances 

 in their armature. Some or all of the normal spines may be 

 absent or there may be supernumerary spines. His observations 

 in this respect were chiefly on Myriapoda and Lepidoptera after 

 injury to the larval appendages. In respect of cases of abnormal 

 armature in the reproduced legs of Blattidae, it may perhaps be 

 suggested that the additional spines represent the normal terminal 

 armature of a joint whose normal articulation is absent, on the 

 supposition that iu the reproduced tarsus one of its four joints is 

 equivalent to two joints of a normal tarsus fused together. There 

 are, how"ever, objections to such a view. The supernumerary 

 spines in 4-jointed as in 5-jointed tarsi were, with only two or 

 three exceptions among all the tarsi examined, grouped closely 

 together in the proper position of the normal spines instead of 

 being situated near the middle of the joint. Moreover, the table 

 (see p. 938) shows the frequent occurrence of tarsi with a super- 

 numerary armature on more than one joint. 



(d) Malformed Tarsi. 



Among the several thousand tarsi examined there were found 

 10 in which one or more of the articulations were imperfectly 

 formed. The appearance of such cases is sufficiently explained by 

 Plate LVIII. figures 4-10, and they are probably much the same 

 as the " crippled " limbs often found in Coleoptera and other 

 groups. The present cases seem to be reproduced tarsi, though 

 those illustrated by figs. 4 and 10 are perhaps malformed 



62* 



& 



