188 MEMOIRS OF THE CARNEGIE MUSEUM. 
to show that these species are subject to frequent injury. The percentage of regener- 
ating specimens may be a little high, as it is possible that the injured ones were 
more easily captured, and consequently occur in the collections in greater numbers 
than they do among equal numbers of the same species in a normal environment. 
Against this, it may be said, however, that in one locality, that in which collections 
No. 1 of E. virescens and No. 1 of S. macrurus were made, an entire trench was 
drained, and all of the fishes, both normal and injured, were secured. In making 
the other collections, it was also the object to take the entire fish-fauna at the 
point selected. Granting, however, that the percentages themselves are higher 
than normal, they nevertheless indicate that a very appreciable percentage of all 
the individuals are injured. This conclusion is substantiated by a review of the 
general data in the table on page 000, in which the percentages of mutilated individ- 
uals of these three species are 10, 12, and 15 respectively. 
The location and severity of all of these injuries are shown in Plate XX, 
figs. 1-4. These figures, as previously mentioned, are outline drawings of an average 
sized fish of each species, on which the line of each of the several injuries found 
in the collections of the particular species is indicated. 
If the line of injury crosses the body of the fish, the loss of the entire portion 
of the body behind the line is indicated. It is to be understood that these plates 
do not represent any one fish, but the entire series of injuries found among the 
specimens of a single species superimposed and drawn to scale on the outline of 
one fish. 
Considered in connection with the foregoing table Plate XX, figs. 1-4, shows 
one region, the dorsal quarter, to be uninjured. The majority of the injuries are 
in the caudal half, only about one-third of them occuring in the ventral quarter. 
The absence of any injury to the dorsal quarter is striking, and is made still more 
so by the following table, which locates all the injuries of all the specimens of these 
three species examined: 
Species. Regenerating Dorsal Quarter. Ventral Quarter. Caudal Half. 
| Specimens, | | 
E. virescens......... 72 | 0 20 4 52 
HEMOcrO DS 7 | 0 | 1 6 
ISERTILOCT UL Seater eee 20 | 0 | i) 15 
otal. Aes eee 99 0 | 26 | 73 
The ventral quarter and the caudal half share all of the injuries in a ratio of 
about 1 to 3. From a consideration of the gross anatomy and some experiments 
and observations made upon living specimens of these fishes in British Guiana, 
it was concluded that this distribution of injuries is the result of the joint action 
of two factors: (1) the liability of the several regions to injury; (2) the relative 
mortality resulting from injuries to the various parts of the body. 
The habits of these species explain the first factor. The several regions of 
