MANIS. 343 



of scales occur, these structures are arranged in perfect longitudinal series, and 

 increase in size from the central dorsal row to the line of scales occupying the 

 middle of the side, but as the Hnes are traced downwards to the belly they decrease 

 in size, and the last row is occasionally reduced to a Hne of three or four small, 

 almost aborted scales on each side, whereas in Avell-formed individuals this terminal 

 row is a reduction of the preceding row in the same ratio of diminution which 

 characterises the rows preceding it. In those instances in which the lowest row of 

 scales is more or less aborted, one row may be entirely lost on one side of the 

 animal, thus giving rise to 16 longitudinal rows, while in other examples both may 

 be suppressed, thus reducing the total of the longitudinal rows to 15. When the 

 number of longitudinal rows is in excess of 17, it appears to be due, not to the 

 development of a fresh series of independent structures, but, as in the case of the 

 multiplication of the normal number of scales in some reptiles, to be attributable 

 to the division of the scales of either one or more of the normal series of longitud- 

 inal Unes. When this occurs, the symmetry of longitudinal distribution is interfered 

 with, and difficulty is experienced in tracing the lines. It usually takes place in one 

 or other of the lateral rows, but a similar division of the head scales is not unfre- 

 quent, and in such instances the scales are smaller than in an animal with an un- 

 divided series. In all Asiatic species of Manis there are only five longitudinal rows 

 of scales on the upper surface of the tail, the outermost row by its lower half enter- 

 ing into the formation of the under surface of the tail, which has three longitudinal 

 rows of scales internal to these halves of the dorsal row. These ventral rows are 

 also frequently subject to division. 



With regard to some general characters : In the young state in all the 

 species, the three or four most inferior rows of scales on the sides of the body are 

 generally keeled, as are also the scales on the hind leg, but more so in M. 

 javanica than in the other species, and in 31. aurita than in M. pentadactyla. This 

 keeling, however, is of variable intensity in individuals of the same species, and 

 in the Indian and in the Himalo- Chinese forms it frequently disappears with age, as 

 the direct result of abrasion ; but in adults of M. javanica it is always more or less 

 present, so that it would appear that the habits of the animal are somewhat different 

 from those of the two previous ant-eaters. This supposition is also supported by the 

 circumstance that its scales, instead of having their edges regularly worn off, as in 

 M. pentadactyla and M. aurita, are generally chipped and broken. This, however, 

 may be due to the more loose manner in which the scales are set on the body compared 

 with the compact overlapping in the other two species ; but this very laxity of 

 arrangement of itself would suggest a difference of habit which the attenuated head 

 and much longer muzzle as compared with 3£. pentadactyla and Jf. aurita would 

 seem to render probable. The scales of the tail of aU the species, but more 

 especially of M. aurita, show a tendency to form an obtuse ridge or keel, and this 

 is soraetimes continued for a short way on to the back, but in other individuals of 

 the same species this character is obscure or altogether absent. The fine striae that 

 mark the scales, and which are most prominent in M. pentadactyla, are nearly all 



