FROGS OF SOUTHEASTERN BRAZIL — COCHRAN" 117 



vermilion. Venter lemon yellow, underside of limbs orpiment orange. 

 Iris brassy, with many fine purplish brown spots. Pupil contracted 

 to a horizontal sht edged with gold. A pair of minute lemon spots 

 on elbow. 



A male, USNM 97354, taken with the female, was quite similar 

 in color, but the following differences were noted: Venter a little 

 paler. Central dorsal patch dark raw umber, surrounded by light 

 russet. Lateral stripe cinnamon-rufous. Upper femur ochi'e yellow. 

 Eye similar to that of female, but pupil less contracted, a more perfect 

 ellipse. 



Variations. — In this extremely widespread species, it is interesting 

 to note how the variable features coincide even in two series so 

 widely separated as one from Costa Rica and one from Rio de Janeiro. 

 In the series from Costa Rica five adults, 3 females and 2 males, were 

 taken for measurement. In these the heel may reach to center of eye, 

 to tip of snout, or beyond the snout. The same variability occurs 

 in the series of 18 adults and young from the State of Rio de Janeiro. 

 When reduced to measurement, the femur is 45 to 56 percent of the 

 total head and body length in the Costa Rican frogs, and 42 to 54 

 percent in the Rio de Janeiro specimens. The tibia measures from 

 50 to 60 percent of head and body length in the Costa Rican series, 

 and 48 to 57 percent in the Rio specimens. A single adult female 

 from Angostura, Ecuador, USNM 20601, has a short femur and tibia, 

 measuring 44 and 50 percent, respectively, of head and body length. 



It appeared at fu'st that the size and distinctness of the tympanum 

 might enable the Central and the South American forms to be 

 differentiated at least as subspecies, but this character also breaks 

 down upon the examination of even a half dozen specimens from either 

 place. The tympanum of the Costa Rican frogs is faint in most 

 examples, but very distinct in one or two, and when easily visible 

 is usually smaller than the largest finger disk. In most of the Rio 

 specimens it is rather distinct, and may be as large as the finger disk, 

 although occasionally smaller. Although the diameter of the 

 tympanum is difficult to gauge in those examples on which it is 

 indistinct, it appears to be about one-third to two-fifths the eye diam- 

 eter in most of the Costa Rican examples. In the Brazilian examples 

 it is usually about one-half the eye diameter, but occasionally not 

 over one-third the eye diameter. Thus, it does not seem possible to 

 separate the two forms subspecifically on the basis of this feature. 



The color pattern is apparently more definite in the Rio form, and 

 less subject to fluctuation, since the brown median dorsal rectangular 

 marking usually maintains its distinct form without any posterior 

 prolongation to form A-shaped markings on the sacral or lumbar 

 region. But in several Rio examples are invading areas of brown 



