AMPHIBIOUS MAMMALS 447 



some of its fibers even lead into the ciliary processes for anchorage. 

 There is a massive sphincter, equalled among other mammals only by 

 that of Lutra, and so arranged that the pupil can close to a short and 

 narrow sUt, about one millimeter by four — and does so, as a rule, the 

 moment the head comes above water. The slit is vertical excepting in the 

 bearded seal (Phoca barbata) , where it is set almost horizontally — really, 

 diagonally with the lower end toward the temple. Possibly in barbata it 

 is normal to the water surface when the animal rears up, just as the oppo- 

 sitely-slanted palpebral fissure of turtles becomes parallel to the water at 

 such times. The walrus forms an exception in that its pupil is always a 

 broad horizontal oval like that described (by some) for the manatee, 

 which is comprehensible in view of the similar sedentary feeding habits 

 of the walrus. 



It is reasonable enough for the pupil to close down when the eye is 

 suddenly exposed to somewhat brighter light upon being lifted into the 

 air — the same phenomenon is seen in the sea-snakes, for example. But 

 why close so far, and why to a slit? Why are a nocturnal retina and a 

 tapetum necessary for vision to continue, out of water, in the seals? They 

 are diurnal and arhythmic in habit. There is no explanation of the matter 

 in the literature, but at least there is a clue : Years ago, Lindsay Johnson 

 puzzled over the astonishing degree of astigmatism which he found in 

 both seals and sea-lions. Out of water, and under the influence of a 

 cycloplegic drug — that is, one which dilates the pupil and paralyzes the 

 accommodation — they showed four diopters of myopia in the vertical 

 meridian and thirteen diopters in the horizontal, resulting in nine 

 diopters of astigmatism against the rule (/. e., with a vertical axis) as 

 though the animals were wearing four-diopter spherical spectacles with a 

 nine-diopter cylinder superimposed, the axis of the cylinder upright. All 

 of this refractive error resides in the cornea, hence of course disappears 

 in water. 



In the preceding chapter we learned the virtues of a stenopaic aperture 

 (pp. 255-6). The ideal one is the pinhole; but no vertebrate pupil which, 

 when dilated, is a very large circle (as in the seals) can easily close to a 

 very small pinhole. The nearest approach it can make is a slit. A slit will 

 focus an object-point as a line which will be parallel to the slit. A cylin- 

 drical, astigmatic lens will, at its second focal plane (see Fig. 13, p. 28), 

 image a point as a line perpendicular to the axis of the cylinder. So, if a 

 slit pupil lies parallel to an astigmatic axis, the combination will image a 

 point as a point, and will thus eliminate the astigmatism of the whole 



