THE ACTIVITY OF THE HAIR FOLLICLE 451 The site of pieces of skin has been altered, either by rotating through 180 ø grafts in the flank of rats or by transposing flaps (10). Such follicles always retain their periodicity and do not take up the rhythm of the sites into which they are put. So, on a rotated graft, the wave of growth begins dorsally and moves ventrally, whereas on the adjacent flank it moves normally from belly to back. The hair is about twice as long on the back as on the belly and the length produced is characteristic of the original, not of the new, site. These experiments also dispose of two other hypotheses about the control of hair growth. Since the nerves are completely cut and hair growth occurs long before they could possibly have regenerated, it is clear that follicular activity is not dependent on nervous connections. And since, by translocation of skin, islands of active, well vascularized follicles can be produced within areas of resting follicles having no demonstrable blood supply, it is clear that activity is not induced by a wave of vasculariza- tion. The above results do not, however, conclusively demonstrate the existence of an intrinsic control mechanis•n they could be explained on the alternative hypothesis that there is a characteristic regional latent period in response to some systemic factor, and that this response is retained even when the position of the skin is altered. A conclusive experiment must involve the complete removal of the follicle from the animal. Ideally, it would be nice to demonstrate cyclic activity in follicles in tissue culture but it is extremely difficult to culture adult skin and neither we, nor as far as we know any one else, have been able to maintain the follicles alive for long enough to test the hypothesis. It is, however, possible to take pieces of skin and to graft them into animals of different age in which the skin cycles are in a different phase. In the circumstances in which new activity at the site is due in the donor before it is due in the recipient, such transplanted follicles have been shown to continue in phase with the donor (11). Follicles can be put out of phase with their neighbours either by grafting, or by plucking the "club" hairs. If a piece of skin in the flank of a rat is removed during the resting stage of the hair follicles and immediately sutured in its original position, the passage of the hair growth wave is delayed in comparison with that on the adjacent body (10). If club hairs are plucked during the resting phase, activity of the follicle is induced and eruption of hair usually, but not always, occurs about 12 days later (12). The fact that in both these experiments follicles can be put out of phase with their neighbours suggests that intrinsic factors play a major part in the control of their periodicity. The study of unrotated grafts also provides evidence against the hypothesis, discussed by some authors (13), that hair growth waves are

JOURNAL OF THE SOCIETY OF COSMETIC CHEMISTS propagated. Eruption of hair dorsal to the margin of the graft occurs at the same time as on the adjacent flank, and does not await the retarded growth-front on the graft itself (10). Systemic factors The influence of systemic factors has been clearly demonstrated by experiments in which homografts have been exchanged between animals of different ages (11), and in which rats have been joined in parabiotic pairs (14). When eruption of hair on the recipient animal is due in advance of that in the donor, the activity of follicles on the grafts can be brought forward into line with those of the host. If rats of different ages are joined in parabiotic pairs, there is clear evidence of a cross action which tends to bring the waves of hair growth into phase with each other indeed, even animals of the same age tend to synchronize their moult cycles when they are parabiotically joined. These facts provide strong evidence that systemic factors normally play some part in the control of follicular activity. It is possible that as yet undiscovered hormones are responsible for synchronization. At the same time there is ample experimental evidence that many known hormones affect the activity of the hair follicle. Steroid hormones of the ovaries (15, 16), the testes (15, 16), or the adrenal cortex (16, 17, 18) all prolong the resting phase, i.e. they delay the initiation of follicular activity. Hence either gonadectomy (15, 16) or adrenalectomy (18, 10) will shorten the resting phase and, in that sense, initiate a new active phase. Similarly thyroid hormone shortens the resting stage, and inhibition of the thyroid by propylthiouracil lengthens it {20). Removal of the pituitary reduces the length of the resting phase, so that any affect on the thyroid would appear to be completely overridden by the resulting lowered secretory activity of the adrenals and gonads (1½, 20). One in- teresting fact is that all hormonal treatments, whether they cause advance- ment or retardation of activity, affect the dorsal follicles more than the ventral ones. Another is that even when the resting stage is greatly prolonged by treatment with propylthiouracil, it is still possible to induce activity by plucking the club hairs {20). These pieces of evidence suggest that all the hormones may act through a common mechanism and that, when delay is induced, the potentiality of the epidermal part of the follicle is not affected. The hypothesis that hormones act indirectly through the dermis is consistent xvith, though not proved by, the facts. It is possible that hormones could affect the rate of growth of the hair, the duration of the active phase, and the rate of shedding of club hairs independently of any affect on the initiation of anagen, that is to say on the overall periodicity of the follicle. Oestrogens, for example, have been sho•vn to reduce the rate of hair growth and to delay the shedding of