1904.] Allen, The Tamandua Anteaters. 3^9 



outline, especially in the outline of the posterior border, 

 which is usually convex posteriorly, but may be square, or 

 even slightly emarginate. They are usually slightly narrower 

 near the middle than at the front border, and expand abruptly 

 at their extreme postero-lateral border, varying from this 

 general form only slightly in different individuals. The con- 

 stancy of the skull characters is thus in strong contrast with 

 the extreme variability of the coloration. 



Having passed in review the large Santa Marta series we 

 may take up the scanty material available for examination 

 from other localities, namely, Chapada, Brazil (6 skulls and i 

 skin), Chiriqui, Panama (4 skulls and 4 skins), and Passa 

 Nueva, State of Vera Cruz, Mexico (5 skulls and 6 skins), and 

 single specimens from elsewhere, 



(1) The Chapada series of skulls stands apart from all the 

 others in (a) the relative shortness of the nasals, due mainly 

 to the great anterior extension of the front als, which thus 

 gives a short rostrum ; and (6) in the greater posterior exten- 

 sion of the occipital condyles beyond the plane of the occiput, 

 as expressed in the tables of measurements r given beyond. 

 In this series the basal length of the skull slightly exceeds the 

 occipito-nasal length, while in all the specimens from else- 

 where it is considerably less, (c) The braincase is also natter 

 or much less convex, both antero -posteriorly and transversely, 

 than in the skulls from other localities. 



(2) The Santa Marta series differs from the Chapada series 

 in all the points above specified, and from this and all the 

 other series by much smaller size, the skulls averaging one 

 tenth less than those from any other locality, with correspond- 

 ingly smaller external dimensions, as will be shown beyond. 

 The average occipito-nasal length of the skull is only 1 18 mm. 

 against 131 in Chiriqui specimens, and the total length of the 



1 In this connection it should be explained that in all cases the basal length is taken 

 from the posterior border of the occipital condyles to the front border ot the maxillaries, 

 for the reason that in most museum skulls the premaxillaries have either been wholly 

 lost or are detached from the skull, with which in life their connection is merely liga- 

 mentary. They can be preserved attached to the skull only by the greatest care in its 

 preparation; and then they are liable to become separated later, even with careful 

 handling; and the same is true of the malar bone. 



