114 



nECOnp? OF TUT. ArsTrAT.TAX MTSFTM. 



broad, and not mncli broadei' rliaii rlie snpra-ocnlar ; Peters' orio-inal 

 description however, gives llie measurement as " frontal 4-2^," vrliicli 

 means rliat it is once and rliree-fiftlis as long as broad. In my specimen 

 of H. sfirliiigi, which species is considei*ed svnonymons with H.frenata by 

 Boulenger, the frontal is once and a thiid as long as broad. 



Examination of all the other specimens on hand shows tliat the frontal 

 varies in being from one and one-tifth to one-and-a-half times as long as 

 broad, and it is nearly always twice as broad as the snpra-ocalar. An 

 extraoi'dinarv range of variation in the shape of the frontal shield is 

 illiisMated in the accompanying tignres ; the difference between that of 

 the female as compared with one of it's young is especially notable. 



-t;l. 



Variations in the shape oi the frontal shield. 



Fig. J. The female from Willow Tree, N'.S.W. 



Fig. 4a. An embryo from the above. 



Fig 5. A variety intermediate between 4 and fi. 



Fig. 6. Normal variety possessed by the majority of specimens e.xamined. 



Fig. 7. An abnormal variety. 



The Ventral ami Suhcamlal Shirlil.-<. — In the description of P. snfa the 

 number of vential shields is stated to be 157-1(14, and the snb-caudals 

 25-30 ; D. fro)if all's is described as having ventrals 154, snb-candals oO. 

 In my series the specimens which agree with D. siitti have the ventrals 

 144-160 and the sub-caudals 27-35, a total of 175-193 ; while those of 

 J), frontalis form have the ventials 147-172 and the sub-candals 26-39, a 

 total of 174-207. 



Tahnlntiiiii uj Main Cliariirter.<. — The specimens leferred to in the 

 following table are selected as being re[»resentative of the thiity-six 

 examined by me ; tlie characters of the remainder vary slightly in inter- 

 mediate stages between these. For convenience of comparison I have 

 added the corresponding characters from the descriptions of the tvpes of 

 the various species. 



According to the several descriptions, the characters listed in the table 

 weie the main features used to separate the various species, but it will be 

 seen that none are sufficiently constant to maintain the species they have 

 beeji suppt)sed to characterise. 



