SENSE ORGANS OF 
theoretical maximum, the absence of 
obstruction is in itself worth while, and 
is possibly a sufficient reason for the 
geometrical arrangement which _ is 
found. In birds the situation is differ- 
ent. There are no retinal blood ves- 
sels and no apparent advantage to be 
gained by displacing cells of the neural 
layers whose transparency in life is very 
nearly equal to that of the vitreous 
humour. The extreme regularity of 
the bird fovea (in contrast to the pri- 
mate) suggests that the shape is of func- 
tional importance. We owe to Walls 
(1937, 1940) both the suggestion that 
the fovea has a lenslike action and the 
resuscitation of the measurements by 
Valentin of the refractive index of 
vitreous humour and neural retina— 
measurements which support the sug- 
gestion. Walls infers that the effect of 
the central fovea of hawks is to magni- 
fy the image falling on the fovea by 
30 percent in area, and concludes that 
acuity is increased in proportion (fig. 
4). Later he states that, ‘‘foveally, the 
visual acuity of some hawks and eagles 
reaches a value at least eight times 
that of man” (1943, p. 662). It must 
be said at once that this statement 
could not possibly be justified on the 
most extreme assumptions. And, in 
fact, Walls’ analysis is oversimplified 
in its neglect of aberration, which is so 
great that the static acuity of the 
central fovea is in all probability sub- 
stantially less than it would be if there 
was no refraction at the vitreo-retinal 
boundary (Pumphrey, 1948). It is 
significant in this connection that 
the profile of the temporal fovea 
of Aquila is very shallow and the 
effect of refraction negligible; for 
the temporal fovea is used binocu- 
larly and must, presumably, obey 
the same conditions as the similarly 
shallow fovea of man. It may be 
inferred that where binocular fusion 
of images is a prerequisite of accurate 
judgment of distance and speed, the 
aberration and loss of definition asso- 
ciated with a steep foveal profile 
cannot be tolerated (fig. 5, c, d). 
The central fovea is not associated 
8173693—49——_-25 
BIRDS—PUMPHREY 314 
with binocular vision, and it must be 
remembered that a central fovea with 
a steep profile is found in fish and 
reptiles as well as birds. Moreover, 
the high static acuity associated with 
the anthropoid fovea is probably a 
late acquisition. Is there a visual 
function which (in contrast to acuity) 
will be assisted rather than hindered 
by foveal refraction? I suggest that 
there is. Consider the effect of the 
fovea on the image of an object which 
is passing across the visual field with a 
constant angular velocity. While the 
image is traveling over the fovea, the 
effect of refraction is to change mo- 
mentarily its apparent course, its ap- 
parent size and its apparent shape. It 
is reasonable to suppose that the 
saliency of the object’s movement is 
thereby increased, just as to human 
eyes the saliency of movements viewed 
through a defective piece of window 
glass is increased. Conversely, foveal 
fixation of moving objects against a 
featureless background should be facili- 
tated. Asaircraft watchers well know, 
aircraft vanish into the distance not 
because they have really become 
invisible to foveal vision, but simply 
because the fovea loses them owing to 
inadequate fixation, and, once lost, 
the chance of picking them up again 
is remote. I suggest that the “‘phe- 
nomenal acuity” attributed to birds 
is simply a phenomenal power of 
fixation to which the geometrical form 
of the fovea contributes, or a phe- 
nomenal ability to detect angular 
movement in the visual field, to which 
again foveal refraction is contributory. 
There is no evidence of extraordinary 
acuity in birds which cannot equally 
well be attributed to one or other of 
these alternatives; there is no evidence 
that a distant moving object which is 
responded to is therefore resolved in 
the physical sense; there is no evidence 
that birds respond at all to a station- 
ary object at distances such that the 
human eye cannot recognize its 
nature. 
Though it may be difficult to prove 
experimentally that the above inter- 
