ARBOREAL RICE RATS — ^HERSHKOVITZ 557 



phase on upperparts and sides. Most of the other specimens are in 

 old worn pelage of one color phase with patches of new pelage of the 

 other phase. Underparts range from "dirty gray," sharply defined 

 from sides, to gray washed with buff on chest and belly and hardly or 

 not at all defined from sides. 



The arboreal rice rats of the rain forests of eastern Ecuador vary 

 like the preceding but average darker, especially on the underparts. 

 Twenty-two adults and subadults at hand collected in February, 

 March, June, August, October, and November reveal no seasonal 

 molt pattern. One specimen collected in each of the months of 

 February, August and October is in old pelage. Others taken at the 

 same times and during other months of the year are in good to prime 

 pelage. 



Specimens from higher altitudes in the valleys and slopes of the 

 Andes (Zamora and Chonta Urcu in Ecuador; Moyobamba, Amazonas, 

 and Hacienda Exito, Huanuco, in Peru) are smaller, have longer 

 pelage and are more richly colored, especially on underparts, than 

 their relatives from the base of the Cordillera. 



Remarks: Oryzomys concolor superans is the most distinctive race. 

 The transition in eastern Colombia from concolor of Orinoco drainage 

 to the larger darker superans of Amazonian drainage seems to be 

 abrupt. Likewise, the contrast between smaller white-bellied roberti 

 of northwestern Bolivia and superans of southeastern Peru is sharp. 



Oecomys palmeri Thomas is merely a small topotype of superans 

 collected the same day by the same collector. Four specimens from 

 Zamora, eastern Ecuador, first recorded by J. A. Allen as Oryzomys 

 auriventer, were described as Oecomys melleus by Anthony. They 

 average slightly smaller with underparts more generally ochraceous 

 orange. Individually, however, they are indistinguishable from other 

 eastern Ecuadorian representatives of superans. The type of Oecomys 

 osgoodi Thomas is another small individual of superans. Its author 

 compared it with 0. palmeri only. In 1927, Thomas (Ann. Mag. 

 Nat. Hist., ser. 10, vol. 2, p. 261) observed that palmeri was "not 

 really distinct from (E. superans" and that the Peruvian "bush rat" 

 was also the same. He concluded that "throughout the tropical re- 

 gions of S. America almost every district has two members of this 

 genus — a large one, the size of a rat [= concolor], and a much smaller 

 one, that of a large mouse [=bicolor]." 



Specimens examined: Sixty-three, from the following localities: 



Colombia: Florencia, Caquetd,, 1 (CNHM); Tres Troncos, Rfo Caquetd, Putu- 

 mayo, 7 (CNHM); Rio Mecaya, Rfo Caquetd,, Putumayo, 17 (CNHM). 



Ecuador: Rio Napo, Napo-Pastaza, 1 (UMMZ); Llunchi, Rfo Napo, Napo- 

 Pastaza, 1 (UMMZ); Chonta Urcu, Napo-Pastaza, 6 (MCZ, 4; CNHM, 2); 

 Canelos, Rfo Bobonaza, Napo-Pastaza, 3 (BM, types of superans and palmeri; 



