MANIDAE. 197 



a specimen from Java measuriug about 25 c. m. with 25 



scales in the middle row of the tail. 

 a specimen from Java meas. about 39 c. m. with 24 scales. 

 » » » Banka » » 40 » » 29 



» » » Java » » 50 » » 25 



» » » ? » » 58 » » 24 



» » » Java » » 58 » » 24 



» » » » » » 70 » » 21 



» » » Salanga » » 80 » » 27 



» » » Borneo » » 81 » » 23 



» » » Sumatra » » 84 » » 25 



» » » Java » » 100 » » 22 



» » » » » » 100 » » 27 



» » ,> Borneo » » 116 » » 26 



The conclusion must necessarily be that the number of 

 scales of the tail in this species is entirely independent of 

 age or locality , is very inconstant and varies between 2 1 

 and 29, thus always exceeding 20, The specimens of 

 other species present analogous phenomena and in con- 

 clusion I may remark that neither the number of scales 

 of the head , nor those of the back , nor those of the un- 

 derparts of the tail, nor even the number of rows of the 

 back are constant in the same species , but that they vary 

 between certain limits in the same way as it was just now 

 demonstrated for the tail of Manis javanica. 



Without entering in osteological details I will summarize 

 what a comparison of our skeletons teaches as to the num- 

 ber of vertebrae, ribs, etc. Our collection contains only 

 a single skeleton of each of the following species, viz: 

 Manis javanica^ aurita, gigantea^ temminckii^ longicaudata and 

 tricuspis. As we possess no skeleton of Manis crassicaudata 

 I cite the number of vertebrae given by Cuvier: this will 

 facilitate comparison. 



Cerv. Cost. Lumb. Sacr. Caud. 



Manis javanica 7-|- 15-^ 5-4-4+30 = 61 

 » aurita 7-|- 16 -f- 6-[- 3-f-27 = 59 



» crassicaudata7-f 15(16)4-6(5)4- 3-1-26 = 57 apudCuv. 



Notes froHi the Leyden ÜVEuseuxn, Vol. IV. 



