1888.] MORPHOLOGY OF SUPERNUMERARY PHALANGES. 505 



and, except for a slight thickening of the distal one (most marked 

 in Proteus), there is nothing demanding general consideration beyond 

 that already given {ante). 



Spelerpes is particularly instructive in the fact that while its 

 digits terminate in well-marked and cup-shaped disks, its distal 

 syndesmoses are normal and comparatively thin, and in no way in 

 excess of the proximal ones («y.', fig. 15). 



There are one or two matters concerning this Order, in respect to 

 which our results are not in harmony with those of our predecessors. 

 Hyrtl, in dealing (8. p. 70) with the pes of Salamandra maculosa 

 writes : "Cartilago primae seriei, cum ilia secundae, fibrosis vinculis 

 cohaeret, quod etiam de metatarsi primi cum cartilagine secundae 

 seriei conjunctione valet. Omnes reliquae articulationes normaies." 

 The last statement we cannot confirm ; some of our sections show a 

 slit in the syndesmosis in question, but that is, almost to a certainty, 

 artificial. Leydig, in his short description of our supernumerary 

 phalanx (his " Zwischengelenkknorpel "), writes (14. p. 166): — 

 "Endlich sei an dieser Stelle bemerkt, dass auch bei Salamandra in 

 der bindegewebigen Substanz der Selinen des Zehenbeugers langges- 

 treckte Nester von Knorpelzellen vorhanden sind, wie solches von 

 ungeschwanzten Batrachiern salt langem bekannt ist." Our speci- 

 mens show nothing of the kind, and the syndesmoses are, in them, 

 throughout, uniform and simple. The figure which Leydig publishes 

 (/. c. pi. xi. fig. 26) in illustration of this statement greatly excited 

 our curiosity — for, did it hold good, it would follow that S. atra 

 would be, in respect to its joints, in advance of the Discoglossidce. 

 The fig. more nearly recalls the condition of the parts in a Hylid ; 

 and if it delineates that which it purports to do, it must be either a 

 bad drawing of a crushed or ill-preserved specimen or that of an 

 abnormal one. 



Peters undoubtedly regarded the supernumerary phalanx as a 

 correlation of the platydactyle condition. He did not actually 

 state this, but it is to be inferred from his classification. Boulenger 

 shows that his (Peters's) PolypedatiiKB was an unnatural group and 

 (2. p. 205) that "Cassina, though oxydactyle, and therefore placed by 

 Peters in his Ranince, has the additional phalanx." This investi2;ator's 

 demonstration that (/. c.) all the species of his genus Rana " have the 

 normal phalanges, irrespective of the presence or absence or size of 

 the digital expansions," goes far towards disproving a supposed con- 

 nection between the supernumerary phalanx and the expanded digit. 

 Our own researches reveal the presence of a fully developed super- 

 numerary phalanx in families other than the Ranidce, and they fidly 

 bear out Boulenger's deduction ; while the discovery that the syndes- 

 mosis does not become converted into a true piialanx in the platy- 

 dactyle Hylodes, Nototrema, and Dendrobates, amounts to a sub- 

 stantiation of the same. 



The condition of the parts in Spelerpes is especially interesting in 

 both its morphological and physiological aspects. The expansions 



