characters of the receptive, retentive or connective and projective organules of the brain respectively, and to gain more light upon the subject I examined the cerebral structures concerned in the sense of smell, taken as a type of special sensation.

Physiological experiment and the clinical observation of disease in the neighbourhood of the uncinate and hippocampal gyri point to these being intimately associated with the sense of smell. In selecting this sense rather than that of sight for the purposes of argument bearing upon the cortical structures concerned in the evolution of thought from impressions received, I have been guided by the circumstance that these impressions converge upon the chief site, so far as physiologists are at present agreed, of general tactile sensibility. The few remarks I shall make as regards the deeper anatomy of smell must be regarded as suggestive rather than demonstrational; more would not be germane to our present task. Foster and Sherrington suggest1 that the uncinate gyrus may not be the site of "full olfactory sensations," and that these may find completion in the cornu ammonis. It is not necessary to follow in detail all the different tracts, even so far as they are known, which lead to the supposed centres of smell. It is enough to know, as the authors just quoted state,2 that a portion of these reaches the front end of the hippocampal gyrus, often called the "uncinate gyrus." There are, however, certain features in the histology of the cornu ammonis which suggest that it may bear somewhat the same relation to the uncinate gyrus as a portion of the temporo-sphenoidal lobe, that the cerebellum does to the cerebrum.

Sir William Turner3 mentions the fact that Mr. A. B. Stirling recognised a granular layer in the hippocampus which so resembled the similar structure in the cerebellum, that it might appropriately be called the "rust-coloured layer of the hippocampus." In some sections of the cornu I have also met with cells which bear a considerable likeness to the Purkinjean bodies of the cerebellum, though they are smaller than these. If a longitudinal section be made of the cornu ammonis, so as to get the full twist of the spiral to the end of the grey matter, cells may be seen, especially at the very extremity, which exhibit the resemblance I speak of. I show specimens which contain these. The sections also demonstrate very well the fact of the invariably vertical position, relatively to the surface from which they spring, of the pyramidal cells of the 1 Op. cit., p. 1181. 2 Op. cit., p. 1180.

3 "Introduction to Human Anatomy," 1882, part i. p. 291.

cornu, and of the smaller and deeper pyramidoid cells of this structure. The true significance of this comparative histology of the cornu is as little known as that of the cerebellum, but if we are justified in regarding the cerebellum as subordinate to the cerebrum, it appears feasible on histological grounds to suggest that the cornu may also be subordinate to the more superficial areas in the temporo-sphenoidal lobe, some of the characters of which have been briefly sketched. What, then, may be assumed to be the histological character of the sensory area on which the afferent tracts of the sense of smell converge? As we have seen, there are two portions of the lower temporo-sphenoidal section referred to which differ considerably in histological features from a third part. This third part has all the characters of marshalled pyramids of the ordinary type, which are found in portions of the cortex which we know have distinctly motor properties. It would seem, therefore, rational to assume that the sensory function of the temporo-sphenoidal lobe is exercised by those parts which have not the features of the motor cortex; that, in short, the essentially receptive cortex, the cortex for special sensibility, is not that distinguished by pyramids of the ordinary type.

It comes, therefore, to be a choice between two portions. Of these, one, as we saw, has a structure which is apparently peculiar to this portion of the temporo-sphenoidal lobe; the other part has the polymorphous character of those irregular cells which lie chiefly below, but also to some extent among, the pyramids. It would therefore seem rational to assume, and I am aware that I am assuming much, that the part peculiar to this portion of the temporo-sphenoidal lobe is more likely to be associated with its special sensory function, than that portion which has more general characteristics.

I have selected this lobule, as I have already stated, and one of the senses running up to it, rather than the centre of vision, because the structure of it and the mechanism of smell seem so peculiar, and because it is situated in what is avowedly considered the most sensory region of the brain. But peculiarities are also to be found, as has been stated, in the occipital lobe, and we must remember that mind is not the product of one, but of all the senses of a correlation of the senses by means of which unity, coherency, is evolved from many sources of impression, although not, as Descartes supposed, in the pineal gland. I feel, therefore, disposed to regard the characteristics I have referred to as a chief type of the receptive organule, but not the only type, and desire

tentatively to express this opinion with all the diffidence which an acknowledged uncertainty should and, I trust, does inspire.

The retentive or connective organule.-It may be remembered that I ventured to define mind last year as "a mode of vital motion manifested by some cells of the cerebral cortex, and as the more or less permanent product or effect of such motion which we call memory." Memory or retentiveness, we know, is not an exclusively cerebral property. The behaviour of the decapitated frog is sufficient evidence to the contrary; but physiologists have also shown that the tendency of progress in development is to increase cerebral at the expense of spinal retentiveness. In man, spinal memory is probably reduced to simple reflex action, which is a receptive-motor rather than a retentive-motor phenomenon. For our present purpose, therefore, it is not necessary to widen the subject by inquiring whether there be or not a histological basis to spinal memory. As a study in comparative mnemonic histology, however, the inquiry might not be without interest.

Cerebral memory may be impressional and conclusional; that is, the automatic record of impressions received through the senses, or the record of the results of that inward digestion of such impressions by the recepto-retento-motor process we call thought or reasoning, in the exercise of which more cerebral organules than one take part. The histological characters of the storehouse of memory or the retentive organule need not, however, vary with the method of filling it. If, therefore, there be reason in regarding the histological type of the receptive organule to be such as has been described, and the histological type of the motor organule, on which I shall say a few words presently, to be of the pyramidal type, it seems feasible to argue, or assume or imagine, as you will, that the polymorphous cell, with its universality of axonic direction and equal universality of dendritic receptivity, is the type of that retentive and connective organule, the repletion and utilisation of which is the mainstay of mental life, as the greater or less storage of supplies-potential energy-is the necessary condition for the exercise of any power whatever, or kinetic energy.

The retentive organule in its histological features has some likeness to the cells of the subcortical centres. These may be regarded as an extension of the spinal cord into the brain, and the lower nuclei, we have reason to believe, may be the seat of simple reflex actions like the cord. These are, however, no more independent of the cortex than the spinal cord is, and perhaps no less. The vegetative centres, although beyond the power of

will, are not, as we have seen, beyond the power of emotion, and emotion is a cortical, not subcortical, phenomenon, although the motion inseparable from it is, so to speak, involuntary, that is, the executive energy of emotion is projected peripherally before it reaches the seat of will. Between this, however, and simple reflex action there is the important distinction, that the consciousness is fully aware of the efferent event in the case of emotion, which it need not be in the case of simple reflex action.

The projective organule.-I have used the term projective here

FIG. 35.-Pyramidal cells from the fronto-parietal region of a
cat's foetus at term. (3 obj., no eye-piece.)

instead of motor, because I have employed a term intermediate between sensory and motor to distinguish the mnemonic sensorimotor source whence mental motion probably derives a considerable portion of its impulse. I have already mentioned Bevan Lewis's belief that the largest cells of the projective or motor type-the giant pyramids are to be met with in the parieto-frontal region of greatest provocable movement (Fig. 35), as also the practically unvarying direction of the primary axons of pyramidal cells generally. A nerve cell, be it receptive, retentive or connective, or projective, is of necessity a mixed cell. It receives, transmutes, or