198 bulletin: museum of comparative zoology. 



upper tooth row, 25; lower tooth row, 25; last molar to end of palate, 



S— 8 



19.4; palatals, 17; lower mandible, extreme length, 73.3; teeth, ^zg. 



Nomenclature. — According to Thomas (Proc. Zool. Soc. London, 

 1911, p. 141) Dasypus should replace the generic name Tatu, for this 

 armadillo. The type locality of Linne's Dasypus novemeinetus is 

 "America meridionali," and is generally taken to be the eastern coast 

 of Brazil. Linne also describes as Dasypus scptemcinetus an armadillo 

 which is characterized by "cingulis septenis," and lives "in Indiis." 

 Whether this name was actually based on specimens from the Antilles 

 or not seems impossible now to decide. Certainly, however, the 

 Antillean armadillo normally has nine bands; and it is more reason- 

 able to suppose an error in Linne's locality than that his specimen 

 was abnormal. The name scptemcinetus is doubtless best considered 

 as referring to the small armadillo of southeastern South America, 

 which does have seven bands. The name was so used by Schreber 

 (in his Saugetiere, vol. 2) and Gray. According to Thomas the 

 animal should be known as Dasypus scptemcinetus Linne. 



In his "Handlist of the Edentate, Thick-skinned, and Ruminant 

 Mammals of the British Museum" (1873) Dr. J. E. Gray recognized 

 no less than seven species of the large nine-banded armadillos from 

 the mainland of South and Central America, five of which he there 

 describes as new. These are: Tatusia granadiana, T. leptorhynchus, 

 T. brevirostris, T. leptocephala, T. bolivicnsis. He also recognizes 

 T. mexicana of Peters, but ignores the latter's T. fenestratus of Costa 

 Rica. His names are based mainly on minute and inconstant varia- 

 tions in the shape or arrangement of the head plates, and the outline 

 of the lachrymal bone. These differences disappear on the compari- 

 son of even small series, so that it is currently considered that Dasypus 

 novemeinetus of Brazil is the same as the Central American and Mexi- 

 can animal. Gray's names brevirostris and boliviensis are both based 

 on specimens from Bolivia, whence I have seen no material. 



A series of ten skins and skulls from Panama, Costa Rica, and Yuca- 

 tan, however, shows that the Central American animal, while essen- 

 tially of the same size as the large Brazilian armadillo, is very readily 

 distinguished by its absolutely much shorter palatal bones, which 

 do not usually reach the level of the posterior tooth, and by the very 

 marked inflation of the maxillary region of the skull directly in front 

 of the lachrymal bones, as is best seen from the ventral aspect. In the 

 Brazilian nine-banded armadillo, the palatal bones usually bow for- 

 ward at least to the level of the posteriormost tooth; but in the Central 

 American race there is commonly a space of from 1 to 3 mm. between 



