362 Georgina Sweet : 



while the left sacral process is carried by the 8th — also some- 

 what similarly in Bomhinator sp., by Howes, in which he found 

 the right sacral process on the head of the urostyle ( = " coccy- 

 geal sacrum "), the left being, presumably, on the 9th vertebra. 



Accepting the view that the positions of the vertebrae are 

 determined by the original myotomes of the embryo, the pelvis 

 being only secondarily affected, it is easy to understand that 

 not only may any vertebra become sacral in function, but, fur- 

 ther, if the iliac cartilages be disturbed or irregular in position, 

 the sacral processes will become correspondingly irregular in 

 their attachments to the relatively fixed vertebrae. 



The existence of the processes belonging to a potential 10th 

 vertebra, which is not infrequent in Hyla aurea, and is exempli- 

 fied in PI. XIX., Figs. 1, 4 and 5, has also been noted by Adolphi 

 in Bufo ci)iere'iis, etc. 



In view of our present unsatisfactory method of estimating 

 the homologies of similar structures (see Parker, '96, and Bate- 

 son, '92), I have simply recorded in detail the variations found, 

 nnuiberirg the parts concerned according to the conventional 

 manner, and g-iving: exact drawings to scale of the same. 



(ii). — Appendicular Skeleton. 



The curved epicoracoids characteristic of the Hylidae. as of 

 other families of the Arcifera, vary somewhat in their relation 

 to one another in HyUi rmrea. The more usual condition is 

 that in which, while not firmly attached to one another, the 

 right epicoracoid lies ventral to the left — but in approximately 

 4 per cent, of individuals the reverse is the case, the rightt 

 being dorsal to the left. 



In a few instances variations are found in the limbs. Thus 

 in one case, while one foot was quite normal, the phalanges of 

 the other foot showed a curious variation from the numbers 

 characteristic of each toe (see PI. xx.. Fig. 6). Thus, instead 

 of the 1st, 2nd, 3rd, 4th and 5th toes containing 2, 2, 3, 4, 3 

 plialanges respectively, as is usually the case, they here con- 

 tained 2, 3, 4. 3 and 3 respectively — i.e., the second and third 

 toes had each one more phalanx than usual, and the fourth 

 one less tlian usual, the third toe being the Utngost, instead of 



