FROGS OF SOUTHEASTERN BRAZIL — COCHRAN 239 



the microscope, minutely glandular on upper and lower surfaces, 

 except on throat and concealed limb surfaces, which are fairly smooth; 

 no ridge above the tympanic area but a considerable swelling of the 

 head behind it; a heavy bilobed projecting fold of skin above the anus, 

 extending well over it, and a slighter one below the anus which is 

 thus enclosed in a deep pocket of skin; no definite parotoid glands; 

 two very hea vy skinfolds across the throat, indicating a median external 

 vocal sac. 



Dimensions. — Head and body 17 mm.; head length 5.5 mm., 

 width 7 mm.; femur 7 mm.; tibia 7 mm.; foot 6 mm.; hand 4 mm. 



Color in alcohol. — Above pale immaculate wood brown, lighter below; 

 disks of toes and fingers nearly white, spines black. 



Variations. — In the six additional examples from the same locality, 

 very little variation occurs. The peculiar anal pocket is more devel- 

 oped in some than in others, but is prominent in all. The smallest 

 one, measuring 15 mm., does not have the posttympanic swellings, the 

 full development seeming to occur only in adults. This individual is 

 quite immature, having only one spine on the first finger, the number 

 of spines in this species evidently increasing with age, as in the genus 

 Crossodactylus . In one adult, spines are totally lacking. The others 

 have 7-8, 7-9, 8-?, 9-9, respectively, while the type has 12-12, as 

 stated. No true vomerine teeth are to be perceived in any of the 

 specunens, but in all except the very young individual a slight rough- 

 ness of the vomerine ridge can be felt with a sharp instrument. It 

 would not be surprising if in a larger series a few individuals with a 

 small patch of true teeth might be found, as this happens occasionally 

 in the genus Crossodactylus, in which the tooth development seems to 

 be similar. 



Remarks. — The shape of the head and body of Crossodactylodes 

 suggests that of Zachaenus parvulus much more than that of a 

 Leptodactylus, but its true relationships are apparently not with the 

 former because of its digital disks and its lack of definite, heavy 

 vomerine teeth. 



Its closest relationships seems to be with the genus Crossodactylus, 

 since both genera have a very similar tooth development, and the 

 patches of black spines on the inner surface of the first finger are nearly 

 alike. Crossodactylodes, however, has all the digital disks round and 

 without pads or furrow in the well-preserved specimens, while Cross- 

 odactylus has the pads and furrow always distinct on the upper surface 

 of the toe disks, and usually on the finger disks as well. But when a 

 toe of Crossodactylodes was dried out, a distinct difference in the tissue 

 down the middle of the disk was observed ; the edge shrank in so that 

 the dried toe gave a bilobed appearance, indicating that the tissues 



