THE GYMNOTID EELS OF TROPICAL AMERICA. 177 



as those collected by Dr. Eigenmann. In no case were the females at this time so 

 full of eggs as to be distended. 



Mutilations and Regeneration. 



The specimens of Gyinnotida' in different collections show an unusual amount 

 of mutilation and regeneration. This condition is undoubtedly in part due to 

 their peculiar anatomy and shape. It may be recalled that they are aU slender, 

 elongate fishes, with the visceral anatomy occupying a relatively small portion of 

 the fish just behind the head. The viscera, except the air-bladder, are so closely 

 packed that the mass occupies a space a little longer than the length of the head. 

 The compound air-bladder varies in size and shape in the several species. It lies 

 just below the spinal column on the posterior portion of the body cavity and 

 extends caudad to between the anterior third and the middle of the body. The 

 position of the anus is also noteworthy. The alimentary canal after several turns 

 bends down and runs forward along the floor of the body cavity and terminates 

 below the head or below the base of the pectorals. Back of the body cavity the 

 body tapers off gradually. The dorsal region bears no fins, hence the name, 

 Gymnolidce. Ventrals are also wanting. All species, however, have small fan- 

 shaped pectorals and a very long anal fin. The anal begins in the pectoral region 

 and extends caudad for the greater length of the fish. The number of rays in this 

 fin varies greatly among the different species and also among individuals of the 

 same species. In one species, at least, the number of anal rays may exceed five 

 hundred {Rhamphichthys rostratus). The species of one subfamily, the Stern- 

 archinoE, have a caudal fin. The tail of the other species tapers into a slender, 

 cylindrical, caudal appendage. 



Mutilation and the attendant regeneration are of general occurrence through- 

 out this family. Many of the accounts of the Gymnotidce note either in passing or 

 even in detail, cases of regeneration in the caudal and anal regions. Nineteen of 

 the twenty-seven species of this family are known to have regenerated lost portions. 

 The present account consists of three parts: the first deals with regeneration in 

 general in this group, and is based both on the collections examined and on the 

 cases recorded bj' others; the second relates to special collections of three species; 

 and the third gives an account of some experiments carried on during the Gimbel 

 ExiDedition. 



Part I. Regeneration !n General. 



The first table lists all of the species of the Gymnotidce and indicates those 

 known to have regenerated lost parts. Whenever specimens were at hand showing 

 injury and regeneration, the word "specimens" follows, while cases not examined 

 by myself are referred to the author mentioning them. Reference to several 

 interesting regenerations recorded for species of which mutilated specimens were 

 examined by myself is omitted from this table. It demonstrates the general 

 occurrence of regeneration throughout the family, regeneration being recorded for 



