4/8 Johnson, A Four-winged Wild Duck. [bet. 



or allantois, play any part in the formation of polydactylism, and 

 believes that these membranes cannot be adduced as causative 

 factors in the production of such anomalies. Painstaking search 

 in embryonic stages showing incipient polydactylism — one series 

 in particular having the embryonic membranes faultlessly preserved 

 — failed to suggest the possibility of amniotic influence. Further- 

 more the early appearance of the anlage of the supernumerary 

 digits, at a time when the foot-plate possesses no indentations what- 

 ever, speaks against such external agency and justifies the view 

 that if in any other amniote, in much later stages, amniotic bands 

 or folds are found in the clefts between supernumerary digits, they 

 have invaded the depressions secondarily. Kaukmann-Wolf holds 

 the view that polydactylism is due to internal influences which 

 in our present state of knowledge cannot in detail be satisfactorily 

 analyzed. 



In certain amphibians which possess notably marked capacity 

 for regeneration, such as Siredon and Triton, Barfurth, and Tornier 

 ('97) produced with regularity supernumerary limbs by means of 

 more or less complex amputations and other forms of injury, the 

 accessory parts being here produced by regeneration at the wound 

 surfaces. From the results of his experiments Tornier concluded 

 that embryonically initiated extra digits or limbs in Anamnia are 

 due to influences analogous to those produced by the embryonic 

 membranes of Amniota; that is, to some stress producing a warp- 

 ing, twisting or splitting of the developing part, thereby inducing 

 regenerative processes or complete division. In both vertebrate 

 groups Tornier thus believes that the underlying causes are of 

 external nature. 



From the opposing views here briefly outlined it will be seen that 

 the problem of causes is far from a satisfactory solution. 



In regard to the case recorded in this paper it would seem that 

 the embryonic membranes must be excluded as causative factors. 

 The fact that the radial branch of the N. brachialis longus inferior 

 lies between the primary and accessory upper arm bones, indicates 

 that the latter of these bones is not the result of a splitting off from 

 an originally normal embryonic humerus by the ingrowth of an 

 amniotic band, or other mechanical agency, for in that case we 

 should expect to find the nerve which precedes the skeletal parts in 



