468 
niques of different design, blacktips were read- 
ily trained to avoid the negative end, especially 
with "easy” problems, e.g., no target vs. target, 
and small vs. large white targets. In the results 
reported here, only two sharks learned the 
avoidance response, and only in the apparently 
simple problems of no target vs. target and 
rectangle orientation. 
Difficulty in inducing avoidance may have 
been due partly to the lack of obvious visual 
cues marking the entrance to the punishment 
area, such as would be provided by a partition 
with an opening or by a barrier in a "shuttle 
box” such as that used by Wodinsky et al. 
(1962). Another factor may have been our 
technique of exhibiting the negative target for 
prolonged periods rather than single displays. 
Preliminary experiments indicated that the 
former method, although it complicated the 
learning process by first developing an associa- 
tion with the negative end of the tank rather 
than with the target itself, still resulted in faster 
learning of the required associations than the 
latter method. 
Another factor of considerable importance 
is the use of electric shock as an aversive stimu- 
lus. Church (1963) reporting upon the varied 
effects of punishment on behavior, points out 
that electric shock may elicit a variety of re- 
sponses, including avoidance and aggression. 
Since some subjects learned to avoid the 
shock in addition to discriminating the targets 
in simple problems, it seems likely that their 
failure in other problems may have been due to 
the fact that the problems were bordering on 
the threshold of the shark’s visual capabilities. 
This may also account for the heightened activ- 
ity and hypersensitivity which resulted after 
continued training to "difficult” problems. Such 
behavior occurred frequently, making it diffi- 
cult to assess visual capabilities and often forc- 
ing postponement or termination of training. 
There was also a suggestion that the sharks’ 
performance may have been influenced by prior 
training experience. After blacktips had been 
trained unsuccessfully with difficult or impossi- 
ble problems, they showed a relatively slow 
rate of learning when later presented with 
easier problems, e.g., orientation of rectangles. 
With continued training in attempts to de- 
PACIFIC SCIENCE, Vol. XX, October 1966 
velop the avoidance response, some blacktips 
exhibited an apparent attraction for the shock. 
Their behavior indicated that the punishment 
was anticipated, and, once shocked, they often 
persisted in the negative zone despite repeated 
shocks. Best (1963) notes a somewhat similar 
behavior exhibited by planaria subjected to in- 
strumental conditioning. After having demon- 
strated that they could make the required choice, 
their performance deteriorated as they chose 
the unrewarded alternative and became lethar- 
gic. He notes that higher animals, particularly 
cats, frequently exhibit such behavior, even 
choosing to lie on an electric grid and receive 
the shock rather than attempt to avoid it. He 
also states that "most workers agree that it may 
be due to overpunishment and . . . some kind 
of emotional response toward the entire test 
situation.” The behavior of blacktips, and, to 
a lesser extent, of grey sharks, can probably be 
attributed to an emotional response caused by 
extensive punishment in training them to diffi- 
cult or perhaps impossible discrimination prob- 
lems. 
V isual Ca p abilities 
It has generally been assumed that the shark 
eye is adapted for high sensitivity rather than 
acuity because of its rod-rich retina, high ratio 
of rods to ganglion cells, and the presence of a 
reflecting tapetum (Gilbert, 1963). Absence of 
cones in vertebrate eyes is usually correlated 
with poor retinal resolution and colorless vision, 
although Walls (1942) points out the possibil- 
ity that cones may not be the sole mediators of 
color vision. Cones have been reported to be 
absent in most shark retinas examined (Walls, 
1942). Recently, however, Gruber et al. (1963) 
found for the first time some cones in a car- 
charhinid shark, Negraprion brevirostris, and 
in two species of Carcharhinus as well. In a his- 
tological study of blacktip retinas, Kato (1962) 
found only a single type of visual cell, presuma- 
bly the rod, despite an intensive search for a 
second type. He also found a high ratio of 
visual cells to ganglion cells. A few grey shark 
retinas examined were similar (unpublished). 
Both retinas, then, are adapted for sensitivity 
rather than acuity. The behavior of captive 
sharks indicated that they could perceive small 
