THE GYMNOTID EELS OF TROPICAL AMERICA. 



193 



end of the caudal appendage thirty-seven miUimeters had been completely restored 

 as far as could be determined; (2) an irregular, semicircular piece fifty-seven milli- 

 meters long and twenty-two millimeters deep had been removed from the anal 

 fin and anal muscles quite well caudad on the anal fin. Here a strip twelve milli- 

 meters wide, bearing a narrow fringe of fin fused at both ends with the old anal, had 

 been regenerated; (3) a piece twenty millimeters long and thirty millimeters deep 

 was gone from the anal region about one hundred millimeters back of the head. 

 In its place was a regenerated mass twenty-two millimeters wide with a rather 

 complete fin on its ventral edge, this fin being fused with the old anal on both ends. 

 This fish showed no abnormalities to account for being thus mutilated. It an- 

 swered in every particular the specific measurements of the species. Plate XXI, 

 fig. 4 gives an outline drawing of this fish showing the regenerations first mentioned. 

 Source of Injury. — The source of these injuries was supposed to be predaceous 

 fishes. As many small alligators and snakes are found in the same habitat these 

 may be responsible for part of them. The wounding of one Gymnotid was ob- 

 served. A specimen of the "hooree," Hoplias malahariais, was seen to bite off the 

 caudal portion of an Eigenmannia virescens. It had been placed in a small trench 

 with several of the latter. The hoorees are abundant in all of the streams from 

 which either Eigenmannia or Sternopygus were collected in British Guiana. 



General Discussion. 



The power of regeneration in the Gymnotidce, as long as the injury is not fatal, 

 is quite general. All of the species of which more than a very few specimens were 

 examined showed some regenerated parts. As long as the mutilated specimens 

 amount to no large fraction of the whole number, they may be considered as 

 chance injuries. Out of a large number of individuals of any species of animal 

 some may be expected to have been injured in the natural course of events On 

 the other hand, when the number of injured in the collection of a given species 

 amounts to a considerable per cent, some other factor than chance alone has 

 probably been operative. Tabulating the total number examined in each sub- 

 family with the number injured, we have: 



This table and aU of the special data show the members of the first subfamily 

 to be subject to only chance injury, and those of the other two, to frequent injury. 

 The immunity from frequent injury of E. electricus, one of the species of the first 

 subfamil}^, is undoubtedly due to its remarkable electric power. The immunity 



