152-207) go to show that animals do learn by this perception of what the experimenter does, and its results. In the same section (pp. 207-5), he finally concludes that there is no natural tendency to learn by perception; still less by reflective," as distinct from "simple" imitation. From this I gather that, while HOBHOUSE assumes that there is no natural tendency to imitate, still this mental act may be acquired in much the same way that habits of attention may be acquired by animals (p. 204). When on p. 259 he discusses "Articulate Ideas," he mentions a case of reflective imitation in the Rhesus monkey. He even puts it more strongly: "To transfer the act and apply it to himself and his own needs, was, at lowest, a strong case of reflective 'imitation.' But the use of the term imitation in this connection is really misleading. At most my act served as a hint."'

Coming finally to PORTER's' excellent piece of work on the Psychology of the English Sparrow, we find this problem still untouched: There is some proof of ability to profit by the experience of others, or of imitation. However, before any description of the real nature of this imitation can be given, additional and varied experiments are needed."

This short and incomplete survey of the field serves to show what inadequate experimental treatment this most important subject has had. That the subject is important, is evidenced by the fact that every investigator, at the end of the discussion of his results, mentions imitation. Yet few experiments have been especially designed to bring out the positive facts-if such there be. And no sane reader would deny utterly, on the basis of the few records we have, that animals can learn by imitation. Most of us have been too busy, either in ascribing habit formations to lower and lower orders of animals, or in describing the mental processes in general in the higher animals, to give enough time and thought to a complete study of any one of the more typical mental acts. We are not criticizing any of the

1 JAMES P. PORter. A Preliminary Study of the Psychology of the English Sparrow. American Journal of Psychology, Vol. XV, p. 345.

experimental work here. Such work as has been done is abundantly necessary-the ground must be broken-but we do plead for long and careful studies in more restricted lines than that represented by simply taking an animal and watching its general behavior. It is time to put the animal in such situations that some one mental act may be exhibited to the exclusion of others. JOHN B. WATSON.

LITERARY NOTICES.

Bethe, A. Allgemeine Anatomie und Physiologie des Nervensystems. Leipsig, G. Thieme, 1903, pp. vii-485, 95 figs., 2 plates.

This volume is a series of monographs on different phases of the same subject rather than a textbook of comparative neurology, as the substance of many of the chapters represents the author's own research. The first three chapters are devoted to a historical account of the structures of the nervous elements; and in the comparative description of the nervous system (chapters 4 and 5), the structure and relation of these elements is given a more prominent place than would be possible in a textbook.

In all forms possessing a nervous system the neurofibrillae are the essential nervous elements. The motor and sensory fibrillae of invertebrates are connected by networks in the ganglion cells, and are also continuous with each other in the neuropil. The nerve fibers of vertebrates are composed of strands of fibrillae imbedded in perifibrillar substance; at RANVIER'S nodes the latter substance is interrupted, but the fibrillae are always continuous. Fibrillar networks are formed chiefly in the sensory ganglion cells; in most other cells the fibrillae pass directly through; they may enter by one dendrite and pass out by another without entering the neurite of the cell. BETHE maintains that motor and sensory fibrillae are put into direct connection by the "GOLGI nets" which envelop most of the central ganglion cells.

The peripheral networks of cells and non-medullated fibers are dealt with in chapters 6 and 7. The entire nervous system of the medusae is composed of these structures. They are also found in the integument, vascular system and digestive tract of mollusks, arthropods, worms and vertebrates. That these networks and not the muscles are the true conducting elements of the medusae is shown by the fact that two distinct bands of muscle may be made to contract by stimulating one of them. If the entire central nervous system of a gastropod mollusk is removed, stimulation of the body wall at one point will produce a general muscular response. In a similar manner the applica

tion of a stimulus to the intestinal wall of the frog will produce contraction of the muscles of the stomach and oesophagus.

Chapter 8 gives a detailed account of the staining reactions of the nervous elements. NISSL's plates and the neurofibrillae owe their peculiar staining powers to two specific substances. The "NISSL substance" is soluble in aqueous solutions of HCl and ammonia; the "fibrillar substance" is soluble in acid alcohol and many of the ordinary fixing reagents. It may be fixed by corrosive sublimate and is insoluble in water, chloroform and ether. Good preparations of the fibrillae may therefore be obtained by fixing in ether and substituting ether for alcohol as a dehydrating reagent. Two methods of this kind are described. Centrosomes, intracellular canals, pigmentation, and changes in structure produced by poisons, or due to other abnormal conditions, are briefly discussed in chapter 9.

Nerve degeneration and regeneration form the subject matter of chapters 10-12. The first evidence of degeneration is the disappearance of the fibrillar substance and consequent loss of staining power. Degeneration is not necessarily due to the lack of continuity between. fiber and cell but a wound which does not affect the conductibility of the fibrillae may produce degenerative changes. The conductibility of fibers may be interrupted by application of pressure or of injurious gases but degeneration will not ensue. The cutting of a peripheral nerve, however, may lead to the degeneration of both the central portion of the fiber and its ganglion cell.

After the lapse of two months or more the normal structure and functional activity of an isolated and degenerate perfpheral nerve may be completely restored (auto-regeneration); if severed a second time the distal portion only will degenerate; but if the two ends of such an auto-degenerated nerve are grafted together, a union, both structurally and functionally perfect, will be established.

The peripheral nerves of the chick (chapter 13) are formed as chains of cells which may be observed before the processes of the neuroblasts within the spinal cord have become prominent. These chains of nerve cells differentiate from their substance the axis cylinders, and later become the sheath cells of the fibers. The dendrites of ganglion cells are also developed from a syncytium of nerve cells, and not as outgrowths of single neuroblasts.

The nature of nervous transmission is the subject of chapter 14. If by compression, the perifibrillar substance is practically eliminated from a certain portion of a fiber without injury to the neurofibrillae, the conductibility of the fiber is not affected. The neurofibrillae must,

then, be the conducting elements. When through degeneration, prolonged pressure, or the application of distilled water, a nerve fiber is rendered non-conductile, staining shows that the fibrillar substance has disappeared. Upon the removal of the abnormal conditions or upon the regeneration of a fiber, it is found that the return of functional activity is accompanied by the reappearance of the fibrillar substance. This substance is therefore connected with nervous transmission. A constant current of 0.05-0.2 milliamperes acting for 10 minutes upon a nerve will produce a polarization of the fibrillar substance and render the nerve non-conductile. If at once fixed and stained, that portion of the nerve near the anode will be much lighter, that portion near the kathode much darker than normal. If after polarization the fixation is delayed until the conductibility of the nerve returns, it will be stained with equal intensity throughout its course as in the case of normal fibers. The change which takes place during polarization is a chemical one. The fibrillar substance is set free at the anode and accumulates at the kathode. Darkly stained granules representing free fibrillar substance were observed at the anode among the fibrillae of the nerve. This chemical change takes place only in living and functional nerves and is therefore evidently a vital process essential to nervous transmission. According to BETHE the nervous impulse is produced by a chemico-physical process. A condition of increased affinity for the fibrillar substance passes wavelike along the fibrillae and the molecules of fibrillar substance are drawn toward the point of stimulation. Coincident with this chemical change, a negative electric current is produced. Either this current or the progressive movement of the fibrillar substance may be instrumental in transmitting the stimulus. BETHE believes that the chemical changes are of most importance in producing nervous impulses; the changes which take place are not oxidation processes, but merely fluctuations in the chemical affinity of the neurofibrillae.

Chapter 15 discusses the peculiar properties of the central nervous system, such as tonus, and inhibition of reflexes; on the ground of his well known experiment upon the brain of Carcinus, the author reasserts that both tonus and the transmission of reflexes are not dependent on the ganglion cells. In chapters 16-18 a review is given of the various phenomena characteristic of nervous reflexes. An interesting account of the effects of various poisons on the nerve elements follows (chapter 19). Most poisons affect first the elements of the central nervous system, because the fibers there are not protected by the thick sheaths which surround most peripheral fibers. Narcotics inhibit the