would expect to find cones if anywhere, there was no sign of a cross section of a cone.

The best presentation of the retina of the mouse can be given by a diagram (fig. 10). Fig 9 is a drawing of a section through the whole eye, showing the size of the lens in comparison with the rest of the eye.

FIG. 10 A Section from central region of retina of mouse, showing rods without admixture of cones.

B Retina of mouse, showing entrance of optic nerve.

C First section of retina of eye of mouse, dorsal cutting, showing choroid coat, pigment cells and rods, no cones being present.

D Section of retina of eye of rabbit, showing rods and cones. a, nuclear layer; b, cones; c, rods; d, pigment.

A search through the sections of the eye failed to reveal structural signs of a fovea in any part of the eye. There was to be found no point in which the inner layers of the retina are sacrificed for the benefit of the rod layer, nor was there discovered a differentiation of the rods in any particular region.

V. SUMMARY

We may now bring together for convenience of reference the conclusions that have been drawn from the several researches: 1. The mouse distinguishes differences in grays and in brightness of lights with a considerable degree of accuracy. The discrimination of the albino mouse is not so good as that of the mouse with pigmented eyes. Illumination of the environment is influential in determining choice of light or dark objects. Black and white are preferred to grays.

2. Red and blue objects, which appear of the same intensity to the human eye, are discriminated between by the black mice, the percentage of error being less than in the case of the grays. Red and yellow are preferred to blue and green.

3. Albino mice do not show any discrimination between red and white lights. With black mice a very bright red and a white of low intensity are distinguished with greater difficulty than colors which are to the human eye of equal brightness. Discrimination between green and blue light is not apparent.

4. Perception of form is very poorly developed. The eye does not seem to be suited to any distinct perception of outlines. 5. The distance of objects is perceived within a range of 15 cm. 6. The mice fail to profit by estimating the depth of objects. The black mice make more errors in this respect than the albinos.

7. Our anatomical investigation shows mice to be lacking in retinal cones, confirming what has been surmised by Allen and by Morgan, and stated by Slonaker. We do not think it follows, as Morgan would suggest, that the absence of cones in mice, bats. hedgehogs and such nocturnal animals implies inability to distinguish colors. It is quite possible that the rods in the mouse are

adapted to the distinguishing of such few color contrasts as may be of importance in its life and habits.

8. There is no fovea or other structurally differentiated portion in the eye of the mouse. The range of vision is very wide, all parts of the retina being equally sensitive, a condition which is enjoyed at the sacrifice of distinct perception of form.

9. There is possible for the mouse a small field of binocular vision. This is not used for estimating distance, as there is no convergence of the eyes. It is of service rather as a means of orientation.

10. The kinaesthetic sense is more important than vision in determining the actions of the mouse. The latter is of use in indicating the presence and general direction of an enemy. Food is found largely through the sense of smell. In other words, smell is an active sense; vision is a protective or passive sense, while the behavior of the animal is largely the result of motor habits, formed through kinaesthetic sensations.

CONTRIBUTIONS FROM THE ZOOLOGICAL LABORATORY OF THE MUSEUM OF COMPARATIVE ZOOLOGY AT HARVARD COLLEGE, E. L. MARK, DIRECTOR.

NO. 212.

REACTIONS OF FROGS TO CHLORIDES OF AMMONIUM, POTASSIUM, SODIUM, AND LITHIUM

LAWRENCE W. COLE

The experiments herein recorded were designed to test reactions of the common leopard frog, Rana pipiens Schreber, to solutions of the chlorides of ammonium, potassium, sodium, and lithium. As these reactions were obtained from frogs whose brains had been destroyed, they are, strictly speaking, spinal-cord reactions and dependent, therefore, on spinal nerves. Whether or not they can be related to taste will be discussed toward the close of this paper. The subject was suggested to me by Professor G. H. Parker and the work was done under his direction.

A series of preliminary experiments showed that the lower limit. of the susceptibility of the frog's skin to these salts is not far from a solution of m./2 strength. Even the most responsive specimens of R. pipiens reacted only once or twice to m./2 solutions of the chlorides of sodium or of lithium. They reacted very frequently, however, to ammonium and potassium solutions of that strength. 3 m. solutions were the strongest ones tried, for larger amounts of some of these chlorides do not dissolve completely in water.

Merck's salts were used in the preparation of the solutions and each solution was titrated against one of potassium chloride taken as a standard until results within 2 per cent of accuracy were obtained. The titration was done in silver nitrate with potassium chromate as an indicator.

Freshly prepared brainless frogs were suspended by the lower jaw from a hook on a lever apparatus such as had already been used by Parker and Metcalf ('06) for similar experiments on earthWorms. Essentially the same method had previously been em