GGS ox rcSTEUioi! Diciio'idAiv FN A umajxc. [Apr. 1*.*. 



haiiils. It. was stateil that ilio autositc, in both instances, liad no 

 control over the supernumerary legs. 



Briefly to summaiise the result of the dissection, the following 

 features' were found to exist in the part dichotomised, wliich 

 occupied a median position on the ventral aspect, and, as usual 

 in such cases, was stunted generally in gi'owth. The pelvis was 

 reduplicated and malformed ; the femora were fused together, and 

 supported tltrcc, and not two stunted, tibiae. The lil)ula' and 

 iiatelhv were entirely absent. The two outside limbs were further 

 well-developed as described later, whereas the third lind) holding 

 a median position on the base of tlie confluent femur terminated 

 with a tarso-metatarsus, reduced to a mere spike-like projection, 

 covered, however, with scaly integument. 



The reduplicated pelvis did not appear to be attached in any 

 way to the body of the autosite, but was contained entirely within 

 the integument. Its form, as shown in the illustration (text- 

 fio-. 64), "would appear to suggest a tendency toward atavistic 

 structui-e, en\bodying as it does certain characters seen in the 

 iliac arch of l^atrachians. 



The writer desires merely to i)lace the fact of this aj:)parent 

 similarity before the reader, without throwing any particular 

 stress upon it, for as Gegenbaur has insisted, and his oj)inion has 

 been backed by subsequent investigators, atavism cannot be ac- 

 cepted as an explanation of rcsemlAances between foi'ms that are 

 systematically far distant. It is generally found that striking 

 instances of supposed atavism are the I'esult of misinterpretation. 



The total length of this presumed pelvis from its base to a 

 ix)int midway between the tips of the two furcations is 10*5 mm. 

 From paired slight processes occurring at the bases of the 

 furcations, there arise two unciform cartilaginous processes, 

 each extending upward in contact with the osseous parts, to a 

 point about 2 mm. from the apex of each furcation. The free 

 ends then deflect inward toward the body of the autosite, the 

 tips being obtuselj^ rounded and in no way connected with the 

 walls of the aVidon^en. 



The confluent femur occurs as a free joint covered with skiu, 

 and is attached by muscle, and lateral tendons, to the symphysis 

 pubis, a single caput femoris of minute size occurring directly 

 behind toward the ventral aspect of the autosite. The total 

 ien<»th of this bone is l(i mm. At its lower end it broadens out 

 to form three distinct series of condyles artieulatiug with the 

 stunted tibia'. TIk^ patella^ as already stated, appear to be absent, 

 and the fibuhv, if present, are fused with the tibia\ What may 

 i)erhaps be termed the outer pair of these bones are of even 

 lenizth, viz. 5 mm., whereas the single centre one is 7 mm. 



The two outer tarso-metatarsi are large in proportion to the 

 tibia\ each being 21 mm. in total length and covered with scaly 

 iutHiiument. The hind toe in ea.ch is lacking. The right foot 

 otherwise is normal in structure, but the left lacks the thii'd toe, 

 its place being taken by a broad (Mliplical tlap of web membrane. 



