536 w. JAMES :

we also have to compare the space-value of objective angles and lines by superposing them on the same retinal tract. Neither procedure would be required if our eye-movements were apprehended immediately as distinct lengths and direc- tions in space. To compare retinal tracts, it would then suffice simply to notice how it feels to move any image over them. And two objective lines could be compared as well by moving different retinal tracts along them as by laying them along the same. It would be as easy to compare non- parallel figures as it now is to judge of those which are parallel. 1 6. General Summary. With this we may end our long and, I fear to many readers, tediously minute survey. The facts of vision form a jungle of intricacy ; and those who penetrate deeply into physiological optics will be more struck by our omissions than by our abundance of detail. But for students who may have lost sight of the forest for the trees, I will recapi- tulate briefly the points of our whole argument from the beginning, and then proceed to a short historical survey, which will set them in relief. All our sensations are positively and inexplicably exten- sive wholes. The sensations contributing to sp&CQ-perception seem ex- clusively to be the surface of skin, retina, and joints. ' Mus- cular ' feelings play no appreciable part in our feelings of form, length, direction, &c. The total bigness of a cutaneous or retinal feeling soon becomes subdivided by discriminative attention. Movements assist this discrimination by reason of the pecu- liarly exciting quality of the sensations which stimuli moving over surfaces arouse. Subdivisions, once discriminated, acquire definite relations of position towards each other within the total space. These ' relations ' are themselves feelings of the subdivisions that intervene. When these subdivisions are not the seat of stimuli, the relations are only reproduced in imaginary form. The various sense-spaces are, in the first instance, inco- herent with each other, and, by feeling alone, both they and their subdivisions are often but vaguely comparable in point of bulk and form. The education of our space-perception consists largely of two processes reducing the various sense-feelings to a com- 1 Cp. Bering in Hermann's Handb. der Physiol. iii. 1, pp. 553-4.