Developmental Morphology of the Skin and Hair Follicles in Normal and in 'Ragged' Mice by j . SLEE 1 From the A.R.C. Animal Breeding Research Organization, Edinburgh WITH TWO PLATES R A G G E D (Ra) is a semi-dominant mutant gene which was first reported by Carter & Phillips (1954). The adult morphology, the genetics, and the embryology of the mutant mice were described by Slee (1957 a, b). It was found that adult ragged heterozygotes (i?a+) had sparser coats than normal, many of their hair follicles being incompletely developed and non-functional. Ragged homozygotes (RaRa) were almost naked. Most of their pelage hair follicles were either absent, or abnormal and non-functional. Ra-\- embryos could be identified from 16 days' gestation by the retardation of their sinus hair-growth. RaRa embryos were characterized from 13 days' gestation by the occurrence of a generalized subcutaneous oedema which persisted until birth, and also by retardation in the development of their sinus hairs and follicles. Pelage follicle primordia appeared at the normal time (14 days' gestation) in Ra-\- and RaRa embryos but subsequently developed slowly in RaRa embryos, especially when the oedema was pronounced. This paper describes the development of the skin and the pelage follicles in normal and in ragged mice, from the 14th day of gestation until the 21st day after birth. These data are linked with the previously described anomalies of adult ragged mice in an attempt to establish the causes of the anomalies and their times of first appearance. Moreover, it was hoped to test the concept of Falconer, Fraser, & King (1951), which supposed that each of the hair-fibre types in the adult mouse was associated with a specific time group of developing follicles in the embryo or neonatal mouse. On this theory, the deficiency of zigzag hairs found in adult Ra-\- mice might be associated with early postnatal follicle defects, whereas the slight excess of guard hairs and awls should stem from a higher than normal follicle density in the late embryo. The near-nakedness of adult RaRa mice could derive from anomalies at many stages of follicle development. From the observations of Griineberg (1943), Falconer et ah (1951), and Gibbs (1941) it was thought that the initiation of hair follicle primordia would begin 1 Author's address: A.R.C. Animal Breeding Research Organization, Field Laboratory, Dryden Mains, Roslin, Midlothian, Scotland. |J. Embryol. exp. Morph. Vol. 10, Part 4, pp. 507-29, December 1962] 508 J. SLEE — D E V E L O P M E N T OF SKIN AND in the 14-day embryo and continue until 6 or 8 days after birth. The growth of early initiated follicles would proceed simultaneously with the initiation of later ones, but, after initiation was completed, follicle development was expected to continue until the completion of the first hair-cycle at about 3 weeks of age. The rate and time of follicle initiation and development were compared in normal and ragged mice. The differentiation and growth of skin layers (epidermis, dermis, and adiposus) were also studied, being relevant to some anomalies found in adult RaRa mice, viz. epidermal hyperplasia and absence of the adipose layer. Furthermore, Slee (1957a) found that in normal mice neighbouring follicles were usually in the same stage of the hair-cycle at any time, whereas adjacent hair follicles in Ra-\- mice tended to be out of phase with each other. Consequently the patterns of hair-growth were ill-defined. Information on the cause and development of this anomaly was expected to assist in an understanding of the normal hair-cycle. Since some form of hair-growth cycle occurs in many, perhaps in nearly all, mammals, the problem of its control is worth investigation. MATERIAL AND METHODS The data described in this paper come from observations on 34 embryos and 103 suckling mice of ages from 13 days' gestation until 21 days' post-partum. The mice were drawn from a small, but not deliberately inbred, population. At each age -\--\-/Ra-\- comparisons were made between littermates. RaRa mice were in some cases compared with Ra+ littermates, but + + littermates were rarely available since throughout most of the experiment viable RaRa mice were not obtainable from intercross matings and had to be produced by backcross matings of the type Ra+% x RaRad (Slee, 19576). The mortality of RaRa mice, especially those showing high-grade expression of the gene, was higher than normal both before and after birth. Therefore the RaRa mice used tended to be of the 'low-grade' type which showed little subcutaneous oedema. Two to five mice of each genotype ( + + , Ra+, and RaRa) were killed at each of the following ages: 17- and 18-day embryos, and 0, 1, 3, 4, 5, 7, (9), 12, (15), 18, (21) days' post-partum. At the parenthesized ages no RaRa mice were available. Two skin samples (from the neck and sacrum) were taken from the dorsal surface of each mouse and used for the preparation of whole mounts or sagittal sections. For the whole mounts the skin was shaved, scraped free of connective tissue, fixed in Bouin's fluid, and stained in Delafield's haematoxylin. For the sagittal sections the skin was shaved, fixed in Bouin's fluid, sectioned at 7 n, and stained with Delafield's haematoxylin and eosin. In addition, 10 whole embryos aged from 13 to 16 days' gestation were sectioned transversely at 10 /i and stained as above. The skin preparations were used to obtain the following measurements: follicle density, hair-fibre density, follicle length, rate of follicle development, and thickness of epidermis, dermis, and adiposus. Follicle density was measured H A I R FOLLICLES IN MICE 509 on whole mounts of skin up to 4-6 days after birth, after which the density and size of hairs and follicles were too great for accurate counts to be made. Therefore, follicle density was also measured independently on the sagittal skin sections throughout the whole age range, by counting the number of follicles per standard microscope field. This method included small follicle primordia which might not be scored on the skin whole mounts. Follicle development was assessed by the appearance of such features as sebaceous glands, hair canals, hair shafts, &c. Follicle lengths and the mean thickness of the skin layers were measured directly on the sagittal sections. Measurements of the stratum corneum layer of the epidermis were discarded because varying amounts of this layer were lost during the preparation of the skin. Subsequent references to 'epidermis' are therefore not inclusive of the stratum corneum. The dermis layer of the skin was defined from the underside of the epidermis to the adipose layer. The adiposus extended from its junction with the dermis to the panniculus carnosus muscle layer. RESULTS 1. Skin morphology Epidermis The data here amplify the descriptions of normal mouse epidermis development due to Gibbs (1941) and Hanson (1947). Comparable information on the development of the epidermis in 40 Neck ragged mice is juxtaposed. 35 The changes in thickness of the epidermis which occur during hair- ho follicle development and during the Sc 25 progress of the first hair-cycle are | 20 V ~ 15 shown in Text-fig. 1. In general there 10 was no significant difference between 17 01 3 5 7 9 11 13 15 17 19 21 Pre -natal Post-natal Aqe in days ragged and normal mice before birth Sacrum and up to 4 days afterwards. However, after 4 days RaRa epidermis and to some extent Ra+ epidermis both decreased in thickness significantly less than normal. The cytological diiferentiation of normal mouse epidermis varies con17 11 13 15 17 19 21 01 siderably. It develops from the early Pre-natal Post-natal Aqe in days embryonic periderm to form about TEXT-FIG. 1. Mean epidermal thickness (ex4 rows of Malpighian cells in the d udtag 15-day embryo. Then it thickens and becomes more proliferative, and the other typical cell layers appear before birth. Moving proximally ft, ' L - • " " ~ * * - » 510 J. SLEE —DEVELOPMENT OF SKIN AND from the stratum corneum the following layers can be distinguished: stratum lucidum, stratum granulosum (containing keratohyalin granules), and the Malpighian layer and stratum intermedium which comprise together 2-5 rows of columnar cells. As the epidermis decreases in thickness some time after birth, it also dedifferentiates. During this process the stratum granulosum and stratum intermedium eventually disappear and the number of cell rows in the Malpighian layer is reduced to one or two. The morphology of the Malpighian cells also alters in that their nuclei become smaller, more irregularly shaped, and pyknotic during dedifferentiation, and their orientation becomes roughly pavement-like rather than columnar. Maximum dedifferentiation is normally established by 15-18 days after birth when the epidermis consists only of stratum corneum and 1-2 rows of Malpighian cells. This is typically the state of adult epidermis during the resting stage of the hair-cycle. Summarizing then, as the epidermis reaches maximum thickness in the late embryo and neonatal mouse, differentiation of the separate layers is most complete: as its thickness diminishes dedifferentiation occurs, during which the stratum granulosum and stratum intermedium virtually disappear. In ragged mice the cycle of differentiation and dedifferentiation is similar, but the process of dedifferentiation takes place later and more slowly than normal and never becomes so complete, especially in ragged homozygotes. Even as late as 18-21 days after birth stratum intermedium cells and keratohyalin granules are usually visible in ragged mice. A striking abnormality which occurs in a few RaRa mice is that of localized hyperplasia of the epidermis. Usually seen during the proliferative phase of the epidermal cycle, i.e. in late embryos or 0-4-day-old suckling mice, these regions of hyperplasia are caused by gross thickening of the stratum granulosum and the Malpighian layer, with an increase in the number of cell rows. Compared with an average epidermal thickness of 30-35 /x, the hyperplastic regions may be 140 fi thick in extreme cases. In contrast, some RaRa embryos showed abnormally thin and cytologically undifferentiated epidermis. But in these cases the thin epidermis was associated with (and possibly caused by) a high-grade of subcutaneous oedema. In the affected zones the dermis/adipose region was disrupted and occupied by oedematous fluid. The morphology of Ra-\- epidermis was generally similar to normal. Dennis The dermis does not become histologically distinguishable from the adipose layer until about 1 day after birth. From this age until 21 days, dermis thickness, in normal and ragged mice, fluctuates from 65 to 125 /x. These fluctuations seemed random and there were no consistent differences between the three genotypes + + , Ra+, and RaRa. Adiposus Throughout most of the period of these observations there were very large HAIR FOLLICLES IN MICE 511 and highly significant differences between + + , Ra+, and RaRa mice in the thickness of the adipose layer (Text-fig. 2). Although the + + adiposus achieved a far greater thickness than in Ra+ mice, the latter maintained its thickness for longer so that the two coincided 450 in the later stages. RaRa adipose Neck 400 thickness was nil or very small 3,350 throughout the period of the obserZ 30° vations. | 250 I 2 00 2. The initiation and density of 150 hair follicles 100 Early embryos 50 Estimates of follicle density on 3 5 7 9 11 13 15 17 19 21 17 Aqe in days Pre-notol Post-natal the body were made during the early stages of follicle development TEXT-FIG. 2. Mean thickness of the adipose layer. The bar lines denote ± one standard by counts of follicle primordia on the error, where these were large enough to appear perimeters of successive transverse on this scale. Adiposus thickness in the sacral sections of whole embryos. Three region varied in a fashion closely similar to that of the neck region, in all genotypes. oedematous embryos (putatively RaRa, see Slee, 19576) and three nonoedematous embryos (putatively Ra+) between the ages of 14 and 16 days were compared in this way. In each case follicle density was significantly lower in the RaRa embryos (means: 7-61,9-16, and 12-00follicles per section) than in their nonoedematous littermates (corresponding means: 9-66,16-00, and 15-63). Although no allowance was made for the mechanical stretching effect of subcutaneous oedema upon the skin, these results did suggest that early pelage follicle initiation proceeded slowly in high-grade RaRa embryos. However, there was no evidence that follicle initiation began any later than normal. In Ra-\- and RaRa embryos, as in normal embryos, pelage follicles first became visible at 14 days' gestation. Late embryos and suckling mice Follicle density changes were estimated from whole mounts of skin, up to 4 days' post-partum, on the type of material shown in Plate 2, figs. L-N. Independent data obtained from sagittal sections over the whole period of development are shown in Text-fig. 3. Both sets of data indicated that differences in follicle density between + + and Ra-\- mice were small and often of doubtful significance. Although + + embryos showed a higher follicle density than Ra-\embryos at 17 days' gestation, at some stages after birth Ra+ mice seemed to achieve a density as high as or higher than normal. This was clearer in the data from skin whole mounts, where for the first 2 or 3 days after birth Ra+ follicle density was significantly greater than normal. RaRa mice showed generally a lower follicle density than + + or Ra-\- mice, but at a few stages the differences were not statistically significant. After birth follicle density was always much 512 J. SLEE —DEVELOPMENT OF SKIN AND greater on the sacrum than in the neck region for all three genotypes. This is in agreement with earlier data on adult follicle density in these mice. Fluctuations in follicle density will result from the interaction of two opposing processes: (i) the initiation of new follicles, tending to increase density, (ii) growth of the animal with consequent expansion of the skin, causing decreased follicle density. The steep rise in density near birth must indicate a peak rate of follicle initiation at this time, especially in the sacral region. The decline in density from about 5 days after birth will be due to body growth, but it also indicates that follicle initiation has slowed down or ceased after this age. Since + + and Ra-{mice grow at similar rates between birth and weaning (Slee, 19576) direct comparison between them is valid. It therefore appears that their rates and times of follicle initiation are fairly similar, although Ra-\- mice may be rather slower before birth and slightly faster immediately afterwards. RaRa mice, however, grow more slowly than normal and so their rate of follicle initiation must be more subnormal than is apparent 7 9 11 13 15 17 19 21 from the graphs of follicle density. Pre-notol Post-natal Aqe in days Both direct observations and the TEXT-FIG. 3. Mean follicle density estimated density trends in Text-fig. 3 suggest from sagittal sections of skin. The bar lines that the initiation of new follicle denote ± one standard error, wherever these were large enough to appear on this scale. primordia had ceased in all genotypes soon after 5 days' post-partum. This is rather earlier than was reported by Gibbs (1941). The tendency for Ra-\- follicles to become denser than normal for a transient period soon after birth may be related to their subnormal size. Each of these follicles occupies a smaller surface area of skin than the larger normal follicles, and so a higher density may be attainable for physical reasons. When later a large number of new primordia are initiated in normal mice, the situation seems to become equalized or reversed (Text-fig. 3). Follicle group formation Different phases of follicle initiation in the mouse are associated with typical patterns at each stage of development. This phenomenon, which occurs in most mammals, has been described in mice by Gibbs (1941). She pointed out that follicle groups were transient in the mouse. Groups of 3 (trios) appeared by the 513 HAIR FOLLICLES IN MICE 4th day after birth; then as more follicles were initiated the groups became obliterated. In the present work it was thought that the changes in follicle density and rates of initiation might be associated with changes in the group patterns. Group formation was, therefore, studied on the skin whole mounts used for follicle density estimates (Plate 2, figs. L-N). Age Stage 0 o o o oo o o 0 oQo ooOoo 14-17 day embryo n oOo o oOoo ooOO000 ooO°° ooQoo o ooOOo° o o O o o oooOoo ooOOooo 0O000O00 oooOOO° 0 0 17 day embryobirth Birth-3 days' post-partum 4 onwards TEXT-FIG. 4. A diagrammatic representation of the normal stages of follicle group formation and subsequent obliteration during development of the follicle population. From these observations the normal process of follicle group formation can be summarized (Text-fig. 4). There was an initial period when the early initiated follicles were uniformly spaced, a period near birth when clearly defined groups were visible, and a period during which the initiation of many new follicles in between existing groups caused the formation of rows of follicles in which separate groups were not distinguishable. Groups seemed to be formed rather earlier in these mice than was reported by Gibbs. Ragged mice showed some differences. In i t a + mice, especially immediately after birth when follicle density tended to be high, there was less-clear grouping of follicles. The groups in RaRa mice were always less clear than normal, but those groups that were distinguishable often persisted much longer. This seems to confirm that the rate of follicle initiation was subnormal. 5584.4 L1 514 J. SLEE —DEVELOPMENT OF SKIN AND 3. Hair-follicle morphology The development of the normal hair follicle has been well described by, for example, Stohr (1903) and Pinkus (1958). The sequence of development in the young mouse is very similar to that of the adult hair-cycle (Dry, 1926; Chase, Rauch, & Smith, 1951). In essence it comprises the following stages: (i) the formation of a primordium from the basal layer of the epidermis, (ii) downgrowth with development of the follicle shaft and bulb and invagination of the dermal papilla, (iii) differentiation of an inner sheath and hair cone, (iv) formation of the hair fibre and the accessory sebaceous gland, (v) hair proliferation, (vi) shortening of the follicle accompanied by distal or upward movement of the completed hair fibre, (vii) follicle quiescence. This series of events is equivalent to the first hair-cycle and it spans the period from the 14-day embryo, when follicle initiation first commences, to the 3-weeks-old weanling, when all follicles have become quiescent. The development of ragged and normal follicles is compared below. 0 1 3 5 7 9 11 13 15 17 19 21 Pre-natal Post-natal Aqe in days The abnormal growth retardation of Ra-\- follicles and the almost TEXT-FIG. 5. Mean follicle length estimated on totally arrested development of sagittal sections of skin. The bar lines denote ± one standard error wherever these were RaRa follicles are shown in terms large enough to appear on this scale. of mean follicle length in Textfig. 5. All of the postnatal between-genotype differences shown were highly significant except between + + and Ra-\- after 15 days. In + + mice peak follicle length (the period of maximum hair proliferation) was reached at 9 days in the neck region and at 12 days on the sacrum. Ra-\- follicles, growing more slowly than normal, attained a lesser mean maximum length, and were later than normal in reaching their peak (viz. 13 days sacrum, 15 days neck). It may be significant that normal follicles reached their maximum length earlier at the neck than at the sacrum, whereas in Ra-\- mice this sequence was reversed. If we assume that the retardation in Ra-\- follicle growth was greater at the neck than at the sacrum, this evidence falls into line with data on the adult ragged mouse, where hair-follicle abnormalities were typically most severe in anterior regions. HAIR FOLLICLES IN MICE 515 During the period of retraction in follicle length, after 12 days, the delayed progress of Ra-\- follicles again becomes apparent. They commence to shorten slightly later than normal follicles and so their mean lengths coincide with normal during this period. At the time of maximum follicle length (9-13 days) a few of the largest Ra+ follicles were longer than any normal follicles although the mean length of all Ra+ follicles was subnormal. These large follicles had apparently been initiated early and were at no time retarded in development. It seems likely that they were guard-hair-producing follicles; because such follicles are normally initiated first (Falconer, Fraser, & King, 1951), and because in adult Ra+ mice the guard-hair fibres tend to be longer than normal (Slee, 1957a). At most periods of development Ra-\- follicles were clearly shorter than normal; but their length distribution was more dispersed, since there were many more small follicles than normal and a few more very large follicles. Apart from their decreased size and growth rate, other important abnormalities of ragged hair follicles were observed. These are best indicated by a day-to-day comparison with the normal. Initiation of the earliest Ra-\- and RaRa follicles commences at the normal time—in the 14-day embryo. Between 14 and 17 days' gestation there is little difference from normal except in those RaRa embryos which show high-grade subcutaneous oedema (Slee, 19576). Here there is some retardation of early follicle development from the young primordial stage. By 17-18 days the normal embryo has some follicles with bulb, papilla, and inner-sheath development. There are fewer such in Ra-{- and none in RaRa embryos. In normal and Ra-{- embryos the follicles of the neck region are already more advanced morphologically than those at the sacrum. This difference persists normally, but it is confounded somewhat in Ra-\- mice and markedly in RaRa mice with the tendency for all follicle anomalies—including growth retardation—to be more severe in anterior regions. After birth many new follicles are initiated. In Ra+ and especially in + + mice, follicle development becomes out of phase in that the new follicles are behind the early initiated follicles in their size and stage of development. In RaRa mice the new and the older follicles are all still in primordial stages, and follicle development soon begins to fall markedly behind normal (Plate 1). In some individual RaRa mice many of the follicle primordia tend to be abnormal in structure. The epithelial cells forming the outer margins of these follicles are not grouped in the even radial arrangement found in normal mice, but they are scattered in the region of the primordial cell mass imparting to it a rather diffuse and indefinite structure (Plate 2, figs. 0, P). No dermal papilla anlagen are visible at the bases of these RaRa follicle primordia. Often the mesoderm is disrupted by the presence of subcutaneous oedema, and possibly this prevents the formation of the papilla which in turn could affect the organization of the follicle primordium. 516 J. SLEE —DEVELOPMENT OF SKIN AND Between 1 and 3 days after birth + + follicles develop fast, hair fibres are formed in some follicles, and eventually no primordia remain. In Ra+ mice development of the largest (earliest initiated) follicles proceeds normally, but the smaller follicles which were initiated postnatally lag behind comparable normal follicles (Plate 1). Ra-{- follicles are, therefore, more out of phase than normal in their stages of development. In RaRa mice the follicle population is still limited to primordia of varying sizes. Between 5 and 7 days after birth + + follicles become closely in phase in their stage of development, as the later initiated follicles grow sufficiently fast to catch up those initiated early. All normal follicles are fully functional hair-producers by this stage. Late initiated Ra-\- follicles are, by comparison, grossly retarded, whereas the early follicles are normally developed. The Ra+ follicle population therefore becomes out of phase in its development (Plate 1). Many large follicles become curved or bent in shape, and may be orientated at abnormally acute angles in relation to the skin surface (Plate 1, fig. E). These distortions seem to be associated particularly with regions where the skin thickness is abnormally small (see Text-fig. 2). From 7 to 12 days + + follicles remain uniform in length, orientation, shape, and phase of growth. Hair proliferation and follicle size are maximal. The Ra~\- follicle population presents a wide variety of types from normally developed hair-producing follicles to others so retarded as to consist merely of strings or clumps of epithelial cells. It is clear that the follicles which are grossly retarded at this stage will never become functional, since the phase of follicle-growth is almost over (Text-fig. 5). These incomplete follicles, none of which possesses differentiated inner sheaths, hairs, or sebaceous glands, obviously represent the non-functional follicles which are found in adult Ra+ mice (Slee, 1957a). In RaRa mice there are still many follicle primordia remaining at this stage. There are also incomplete follicles similar to those in Ra+ mice. The largest of these follicles possess differentiated shafts and bulbs but no inner sheaths, hairs, or sebaceous glands. They are still much smaller than normal. Between 12 and 15 days hair-growth ceases and the follicles begin to retract upwards. This process commences rather earlier in normal follicles than in even the functional follicles of Ra-{- mice. The accompanying decrease in skin thickness also tends to occur slightly later in Ra-\- mice. The Ra+ follicles of distorted shape become even more grossly curved and bent during this stage of development (Plate 2,fig.J). Those follicles which were well grown, but later than normal in commencing to shorten, appear to be compressed as the skin thickness diminishes. They maybe out of phase with the skin contraction even though this is slightly delayed. In RaRa mice the largest of the follicle-like structures also become curved or bent in some degree (Plate 2, fig. K). The most advanced of these structures have by this stage developed hair canals which are frequently plugged with keratin. No hair-producing follicles were seen. Between 18 and 21 days the normal follicles completely retract and enter the HAIR FOLLICLES IN MICE 517 resting stage closely in phase with each other. The largest Ra+ follicles follow suit; those Ra+ follicles which are slightly retarded but functional behave similarly, but enter this stage slightly later than normal. Many of the incompletely formed follicles also undergo length retraction (but to a lesser extent than normal) during this stage. The largest of the incomplete follicle structures in RaRa mice show a tendency to retract slightly but very much later than normal. The numerous incomplete follicles show little change in morphology. The RaRa follicle population is now so grossly abnormal as to be difficult to describe in terms of the normal stages of development. Setaceous gland formation Hair formation Birth A«. Ir. An», ^ 18 Birth An. In ,jOyS ,. 18 eck — A* NSocrum 1-RaRa J TEXT-FIG. 6. A diagram showing trends in follicle development as they differ between normal and ragged genotypes. The points are approximate mean values based on counts obtained from sagittal sections of skin. Text-fig. 6 shows in quantitative terms the degree of retardation in structural development suffered by ragged follicles. The two characteristics used—hairfibre and sebaceous gland formation—are both closely related to the stage of development of the follicle as a whole. In general these data confirm the descriptive information given above. The whole range of morphological abnormalities in the hair follicles of young ragged mice can now be assessed and summarized as follows: (i) Follicle initiation commences at the normal time in Ra-\- and RaRa mice. Some early initiated Ra-\- follicles are normal in their rate and final extent of development. (ii) The next initiated Ra+ follicles are still functional, but their growth is retarded to varying degrees so that they are out of phase with each other in development and in completion of each stage of hair-fibre formation. (iii) The slow growth-rate of Ra+ follicles is associated with a subnormal mean maximum size and a delay in the time of reaching maximum size. 518 J. SLEE — DEVELOPMENT OF SKIN AND (iv) The shape and orientation of many Ra-\- follicles and the largest RaRa follicles are abnormal, (v) Later initiated Ra-\- follicles and nearly all RaRa follicles are grossly retarded in development, and they remain as incomplete non-functional aggregates of epithelial cells, (vi) In general the retardation of RaRa follicles is more severe than that of Ra+ follicles, but most of the consequent abnormalities are qualitatively similar. One additional RaRa anomaly is apparent failure of dermal papilla formation in some early follicle primordia. DISCUSSION This discussion is based on the present observations and also on data from two previous papers on ragged mice (Slee, 1957 a, b). Much of the earlier material can only be discussed now in the context of the juvenile abnormalities. Juvenile follicle abnormalities and the adult syndrome The basic abnormality of the pelage in ragged mice is a general restriction of hair-follicle development and hair-growth in young suckling mice. In Ra-\mice this restriction is most severe in late initiated follicles (postnatal) and it does not affect the early initiated follicles. The result is that some follicles remain incompletely developed, and the rest of the population becomes out of phase in its development. Growth restriction is more extreme in anterior than in posterior regions and everywhere its intensity is much greater in RaRa mice than in Ra+ mice. These abnormalities link directly with those of adult ragged mice, viz. in Ra+ mice: incomplete development of many follicles, virtual absence of zigzag hairs in anterior regions of the body and subnormal density of zigzags in posterior regions, structural hair defects, the existence of transitional hair types, and outof-phase progress of adjacent follicles through the hair-cycles; in RaRa mice: incomplete development or absence of nearly all follicles, causing virtual hairlessness. According to a hypothesis of Falconer, Fraser, & King (1951) hair follicles initiated in the normal 14-17-day embryo later produce guard hairs, those initiated between 17 and 19 days produce awls and auchenes, and those initiated after birth produce zigzags. Since it is the late initiated follicles whose growth is most curtailed in juvenile Ra-{- mice, one would expect, on the hypothesis of Falconer et al., that the zigzag hairs would be affected most in the adult mice. This was, in fact, the case. Moreover, the existence of transitional fibre types in the adult is consistent with the time sequence of follicle development being upset in the juvenile mouse. On the same theory, the fact that adult Ra+ guard hairs tended to be longer than normal and the awls and zigzags shorter than normal again fits with the juvenile data, since a few of the early initiated Ra-{- follicles were larger than normal at some stages, whereas all the HAIR FOLLICLES IN MICE 519 other functional follicles were smaller than normal. Despite its general sufficiency, the hypothesis of Falconer et al. does seem at variance with the present data in one respect. It was found in adult Ra-\- mice that the density of guard hairs and awls tended to be greater than normal. However, the follicle density of Ra+ mice is never greater than normal before birth, when guard hair and awl follicles would supposedly be initiated, but may slightly exceed normal immediately after birth when zigzag follicles would ex hypothesi be initiated. One must conclude that the high density of guard hairs and awls in adult Ra+ mice does stem from the supranormal postnatal follicle density. Probably the Falconer-Fraser-King hypothesis is correct for the normal mouse but in Ra-\mice the time sequence of follicle development, although in essence the same, may have been slightly disorganized by changes in the relative growth rates of different follicle types. Consequently, discrete periods of follicle initiation are not precisely associated with production of the same fibre types as normal. Nevertheless, it is almost certain that guard hairs, awls, and zigzags are, in Ra+ mice, still related to the same order of follicle initiation: early, moderate, and late as in normal mice. It is probably true of nearly all mammals that the earliest formed follicles become largest and produce the largest types of fibre. Such differences in follicle size, associated with their time of initiation, may actually determine the type of fibre which each type of follicle shall produce. We can now conclude, with Falconer et al., that certain discrete periods of follicle initiation are each associated with the production of a certain fibre type in the normal mouse. In the Ra-\- mouse the order of formation of different follicle/fibre types is unaltered, but the times allotted for each period of follicle initiation are assumed to have been slightly disorganized. If the sequence of follicle initiation and the production of a heterogeneous fibre population are interrelated processes it is not surprising that their disruption in i?a+ mice causes changes in fibre-type proportions, defects of hair structure, and the production of transitional hair types. All these anomalies occur consistently in adult Ra-\- mice. The abnormalities of the RaRa mouse are so extreme that the relationship between follicle and fibre-type anomalies cannot be meaningfully discussed. Partial follicle agenesis or severe follicle-growth retardation in the juvenile mouse is paralleled in the adult by virtual nakedness, except for the existence of a few fibres unclassifiable into the normal morphological types. Causes of the anomalies in hair-follicle development The pelage defects of ragged mice were generally more severe in some body regions than in others. For example, Ra-\- juvenile follicle development was more retarded in the neck region than on the sacrum. As a probable consequence Ra+ adult mice lacked almost all zigzag hairs on the head, neck, and anterior regions of the body; whereas on the posterior dorsum the frequency of zigzags was almost normal. Between anterior and posterior areas there seemed to be an 520 J. SLEE —DEVELOPMENT OF SKIN AND evenly graded increase in density of zigzag hairs. In RaRa mice the situation was similar in that rather more of the few hair fibres produced existed in posterior regions of the body. It may be significant that the least abnormal areas are those which, according to Dry (1926), cease hair-growth last during production of the first pelage in the normal mouse. For example, in the head and neck regions maximum follicle size is normally reached by 9 days after birth (present data) and hair-growth ceases 17 days after birth (Dry); whereas on the sacrum maximum follicle size is not reached until 12 days and hair-growth ceases at 19 days. Present data suggest that follicle initiation commences at similar times on the neck and sacrum, but that follicle development proceeds more slowly and finishes later on the sacrum. It seems, therefore, that the effects of the Ra gene are least severe where there is most time available for follicle- and hair-growth. Moreover, the Ra gene acts least strongly against the earliest initiated follicles (the guard-hair- and awl-producers) which again have the most time available for development. These facts suggest that a primary effect of the Ra gene is to retard rather than prevent the development of many hair follicles. This effect might then interact with the time factor so that follicles with least time for development—which need to grow fastest—would be most severely affected. Nevertheless, follicle-growth retardation may be so extreme as virtually to prevent all development (as in many RaRa follicles). In such cases the effects are indistinguishable from those of growth prevention. If we assume that retardation of follicle-growth is a major effect of the Ra gene, the next step is to ask how this effect is mediated. Two possible processes could be postulated: (i) mechanical interference with follicle-growth, (ii) slowing down of the rate of division of the follicle-forming cells. Presumably any interference of the first type would be visible histologically. Then the only possible suppressor of follicle development would be the presence of subcutaneous oedema. In high-grade RaRa embryos between 15 and 17 days' gestation, subcutaneous oedema can be widely distributed (Slee, 1957&). Skin depths of up to 600 fx between the epidermis and the panniculus muscle layer may be entirely permeated by oedematous fluid, causing complete disruption of the mesodermal layer. Despite the apparent isolation of the epidermis, follicles are initiated normally in these areas. This may be possible because the initiation of follicle primordia is caused by migration of the epidermal Malpighian cells, and not at first by an increased rate of mitosis (Balinsky, 1950). However, one would expect there to be clear effects on follicle development and, in fact, follicles developing over sites of oedema are usually atypically small even for RaRa embryos. Many such follicles do not possess dermal papillae, perhaps as a result of the permeation of the dermis by oedema. The development of the follicle blood-supply is also likely to be impaired. But there are serious objections to regarding oedema as the main cause of growth retardation in ragged follicles. First, oedema has not been seen (macroscopically or microscopically) in any postnatal Ra+ mice, and it appears very rarely and HAIR FOLLICLES IN MICE 521 inconspicuously in Ra-\- embryos. Nevertheless, consistent coat defects, stemming from defects of postnatally formed follicles, are typical of all Ra-{- adult mice. Secondly, oedema in RaRa embryos is of variable extent in different individuals and is hardly visible in certain anatomical regions, e.g. the head. Despite this, similar skin and follicle anomalies occur in all body regions in all adult RaRa mice. Thirdly, if oedema acted as a physical agent to interfere with folliclegrowth one would expect that the few follicles which did mature would occur in isolated groups at points of local failure of the mechanical interference. In fact, the functional follicles occur singly and they are evenly spaced from each other. It is concluded that subcutaneous oedema does not play a primary role in retarding the development of ragged mouse follicles, but it may intensify the defects of early follicles in high-grade RaRa mice. We are left now with the possibility that follicle-growth in ragged mice is retarded directly by an abnormally slow rate of cell-division in the follicleforming cells—presumably the epithelial cells of the external sheath. As to the cause of such a decrease in cell-division rate, there is little direct evidence. However, an important point is that the follicle-sheath epithelial cells are indistinguishable cytologically and histologically from the germinative epithelial cells of the epidermis. These cells make a continuous layer from which small aggregates of cells form follicle primordia and later develop into diverticuli as they form follicle external sheaths. There is evidence that these cells, when in the epidermis of ragged mice, appear to divide normally. First, epidermal thickness, which is related to the speed of cell-division, is always as great as or greater than normal in Ra+ and RaRa mice. Secondly, the rate of cell-division in neonatal RaRa mice is not held down to any specific level since it can be artificially increased by topical applications of testosterone propionate (Slee, unpublished). Thirdly, regions of epidermal hyperplasia, obviously involving high rates of cell-division, occur in some RaRa mice of all ages. This evidence tells against the possibility of a general epidermal defect in ragged mice. If we accept that the epithelial cells of ragged mice can divide at least as fast as normal when they are situated in the epidermis, but not when they are in hair-follicle external sheaths, then we must look for physiological differences between these regions in the control of cell-division. There is some evidence that the unknown stimulus which initiates the growth of hair follicles (and presumably stimulates celldivision) is specific to the whole germinative epithelium cell type, whether these cells are in the epidermis or in the hair follicles (see Chase, 1954; Slee, 1957fl). Therefore, if the follicle sheath and epidermis cells react differently, one suspects differences at the site of response rather than a failure of one stimulus at the systemic level. The main difference between epithelial cells of the epidermis and of the hair follicles is that the latter become part of a differently organized structure. As a part of this structure, the dermal papilla may have roles, first as an organizer and then as a channel for follicle nutrition. Therefore, if it could be shown that the dermal papillae were absent or defective in many Ra-\- and 522 J. SLEE —DEVELOPMENT OF SKIN AND RaRa follicles, there might be an acceptable explanation for the abnormalities of follicle development. However, morphological evidence of papilla abnormalities is scanty. Some RaRa follicle primordia lacked papillae, especially in mice affected by oedema, and the cells of these follicles were often diffusely scattered and disorganized (Plate 2, fig. O). But most RaRa follicles and all Ra-\- follicles appeared normal in these respects, even those which later became non-functional, and it seems that if papilla defects are responsible for all follicle abnormalities of ragged mice they must be mainly undetectable by ordinary histological procedures. For the present the problem is unresolved, but some defect specific to the hair follicles themselves seems implicated. Results of the restrictions in hair-follicle-growth Many of the obvious effects of the hair-follicle growth retardation have already been made clear. They can be summarized as follows: (i) failure to complete development in some follicles, (ii) delay in the completion of development in other follicles, (iii) disorganization of the orderly phases of follicle development which are normally associated with the production of different fibre types, (i) results in the absence of nearly all fibres in RaRa mice. In Ra-{mice, where the late initiated follicles only are affected, many zigzag fibres are absent, (ii) causes the functional Ra+ hair follicles to be out of phase in their development and therefore out of phase in their progress through subsequent hair-cycles, (ii) and (iii) cause hair-structure defects and result in the production of transitional types of fibre. In addition, the partial failure of hair-follicle growth may have some effects on skin thickness. Observations have shown (Text-fig. 1) that the epidermis in young ragged mice (especially RaRa) tends to be thicker than normal during the later stages of follicle development. This is strong evidence for a theory developed by Mottram (1945), Bullough (1952), Chase (1954), and Slee (1957a). On this theory the epithelial cells of the hair-follicle sheaths and of the germinative layer of the epidermis compete for a substrate supplying energy necessary for cell-division. So a system of priorities might exist whereby down-growing hair follicles during the stages of hair proliferation would obtain substrate at the expense of the epidermal cells. This process could cause the decrease in epidermal thickness which normally occurs during the advanced stages of follicle-growth. In juvenile ragged mice, where follicle-growth is slow or arrested, competition for substrate would be reduced and epidermal cell-division consequently less restricted. It is therefore suggested that the thicker than normal epidermis generally found in 5-21-day-old ragged mice and the regions of gross epidermal hyperplasia sometimes seen in RaRa mice result from the restricted growth of their hair follicles. The second aspect of skin morphology which seems to be associated with hairfollicle development is the thickness of the adipose layer. Chase, Montagna, & Malone (1953) first showed clearly that, in normal adult mice, adiposus thickness HAIR FOLLICLES IN MICE 523 varied closely with the hair-follicle-growth cycle. Throughout my observations on adult and juvenile mice of both normal and ragged genotypes it was noticed, in agreement with Chase et ai, that a thickened adipose layer was always associated with hair follicles in the down-growing proliferative stage. Conversely, a thickened adiposus was never found when the follicles were not in this stage. This positive association persisted between genotypes and through developmental stages within genotypes. For example, a comparison of Text-figs. 2 and 5 shows that follicle-growth is clearly related to adipose thickness throughout the period of juvenile follicle development. Within + + and Ra-\- genotypes adipose thickness is greatest when follicle length is maximal; in RaRa mice it is always small, in line with their rudimentary follicle development. Between genotypes the subnormal adipose thickness of Ra-\- and RaRa mice parallels their subnormal follicle development, even to the extent that the delayed retraction in Ra-\- follicle length with respect to normal is exactly matched by a synchronous delay in the time of decrease of adipose thickness. If these two phenomena are associated and synchronous it is simplest to suppose that one causes the other, rather than to postulate independent causal systems. Therefore it is suggested that follicle down-growth normally causes expansion of the adipose layer. The converse, that adipose expansion normally induces follicle down-growth, seems less likely because in ragged mice many follicles with retarded development were restricted in size more than could result from restraint due to lack of skin expansion. Such follicles did not even stretch the full depth of available dermis. Normally, development of the whole follicle population may cause expansion of an adipose layer which is refractory, in that a full complement of developing follicles is required to cause its full expansion. Growth retardation of some follicles, as in Ra-\- mice, might then be sufficient to make the degree of adipose expansion subnormal; near agenesis of the follicle population as in RaRa mice could result in virtual absence of any adipose layer. It is further argued that the subnormal depth of skin consequently available for those ragged follicles which do develop fully results in these follicles being constricted. They therefore become abnormal in shape and orientation (Plates 1 and 2). These morphological abnormalities seem especially marked after 15 days' post-partum, when the hair follicles are retracting and skin thickness diminishing. It seems as if the adiposus normally contracts as retraction of the follicle population permits it to do so. In Ra-\- mice where the follicles are out of phase in development some follicles begin to retract before others. Apparently adiposus contraction commences with retraction of the first follicles, so that the later retracting follicles, already distorted by their previous growth in a restricted space, become further compressed by the adiposus contracting ahead of their retractive phase. Similar anomalies were found in the few large RaRa follicles which occur. Most of these relationships between subnormal skin depth and abnormal follicle morphology occur also in adult ragged mice and similar explanations can be offered there. 524 J. SLEE —DEVELOPMENT OF SKIN AND Control of the mouse hair-cycle In many rodents hair-growth occurs in a cyclic fashion at regular intervals of time. Each cycle consists of a period of follicle-growth and hair proliferation followed by a rest period during which the follicles are quiescent and there is no hair-growth. A complete cycle lasts 28 days in the mouse. Follicles in any one region of the body enter the hair-cycle together and follow it in phase with each other, but follicles in any other region will enter the hair-cycle again in phase with each other but at a slightly different time from the first region. This arrangement, by which follicles enter synchronously within body regions but sequentially between regions upon any specific stage of the hair-cycle, causes hair-growth waves. Neighbouring hair follicles therefore tend to be in phase through the hair-cycles, except near the margin of an advancing wave front. The waves of hair-growth, which are usually bilaterally symmetrical, spread anteroposteriorly across the body according to a consistent pattern, which has been well described for the mouse by Dry (1926). The question of how these growth cycles are initiated and controlled is of fundamental importance and has been extensively considered, e.g. by Durward & Rudall (1949) and Whiteley (1958). After a study of adult ragged mice Slee (1957a) concluded that the controlling mechanism must involve a systemic agency and a degree of follicle autonomy. There is one particular feature of the ragged syndrome which may be important for an understanding of the normal hair-cycle. This is the fact that neighbouring hair follicles can be persistently out of phase with each other throughout adult hair-cycles, apparently as a result of becoming out of phase during their initial development in the suckling mouse. This original lack of synchrony seems to arise by retardation of the later initiated follicles. If this is accepted, it carries the possible corollary that the synchrony of normal adult hair follicles within body regions depends upon their becoming in phase during early development, where, as we have seen, the late initiated follicles grow sufficiently rapidly to catch up with those initiated early. Moreover, on this theory, normal differences between body regions in their time of entry into different phases of the adult hair-cycles could result from differences in the original time of completing juvenile follicle development in these regions. This would provide a mechanism for the hair-growth waves described above. There is not much evidence here for or against this last possibility, but it is probable that normal follicles on the anterior dorsum do develop slightly ahead of those on the posterior dorsum. This fits with the antero-posterior direction of the hair-growth wave found in subsequent cycles. At this point the evidence of Johnson (1958) from the rat, and of Nay & Fraser (1954) who used naked mice, must be considered. They showed that the duration of the hair-growth cycles varied between different body regions. Developing the ideas of Nay & Fraser we can now see that the regional differences in hair-cycle duration, which they observed, could interact with regional differences in the time of cycle commencement to produce waves HAIR FOLLICLES IN MICE 525 of hair-growth whose pattern varied in successive cycles. Some such changes do occur in the normal mouse (Dry, 1926; Wolbach, 1951). Moreover, if the original wave patterns depended solely upon pre-set synchronies built up during early development one would expect the patterns of later hair generations to become irregular. They do (Fraser & Nay, 1955), and they would not do so if the cycles were self-equilibrating. It is now reasonable to assume that hair follicles are potentially independent in their cyclic behaviour. Although they are normally set off in their development synchronized within body regions, this synchronization does not seem to be continuously imposed by systemic factors. This is shown by the fact that individual follicles may persistently behave independently of their neighbours due to the action of the ragged gene, or as a result of fibre plucking (David, 1934). It is therefore suggested that synchronization may depend entirely upon neighbouring follicles reaching a specific stage of development at the same time during their original development, and that this degree of synchrony tends to remain for several hair-cycles unless interfered with surgically or genetically. Although some general systemic stimulus may still be required to start each hair-growth cycle, these overall conclusions emphasize the individual hair follicle as an autonomous unit. To some extent this is aetiologically different from the conclusions reached by Ebling & Johnson (1959) after their elegant experiments with the rat. They supposed that the reaction times of different follicles to stimuli initiating a hair-cycle are 'determined by graded local thresholds to a systemic stimulus'. I suggest that follicles' reactions to any systemic stimulus are individually determined and that their times of reaction depend upon their order of completing their original development. Since this order varies between body regions, their times of reaction are in phase within regions and out of phase between regions. If this is accepted, the postulate of graded local thresholds can be discarded. We now have a possible explanation for the form of the rodent hair-cycle, but the question of how the cycles are initiated and terminated remains unanswered. It seems certain that some systemic stimulus is required. Moreover, the relationships found between epidermal thickness and follicle-growth in ragged mice reinforce the conclusions of Chase (1954) by indicating that the stimulus does not affect the hair follicles as such but is effective upon a whole cell type. It increases the mitosis rate in at least three components of the germinal epithelium: the epidermal Malpighian layer, the follicle external sheath, and the follicle bulb. There is some evidence that the systemic stimulus may involve a complex endocrine mechanism (Mohn, 1958). Or it can be argued that the removal of a mitosis inhibitor may start each hair-cycle (Chase & Eaton, 1959); this suggestion fits with the evidence of Bullough & Laurence (1960), indicating that the removal of an inhibitor may cause an increased mitosis rate in injured epidermis. Whatever the stimulating mechanism, the hair follicles only react to it when they are ready. This moment of reaction for each follicle may depend on its internal physiology becoming responsive at a certain time after the previous cycle. 526 J. SLEE —DEVELOPMENT OF SKIN AND Obviously there is considerable follicle autonomy, but recent experiments of Ebling & Johnson (1961) with rats have shown that although individual follicle reaction times persist to a degree they can slowly be overruled by some form of systemic stimulus. This was apparent when the two were set at variance by grafting skin between rats at different stages of the hair-cycle. These results are not opposed to the view developed above that the reaction time of the individual follicle is normally set by its original time of development, but they do indicate that it is not always irrevocably determined. We are left with the concept of an autonomous follicle response mechanism interacting with a systemic stimulus. Under certain conditions (e.g. in rat-skin transplants but not, apparently, in Ra-\mice) the systemic factor seems able to modify the follicle reaction times. A complete explanation of the hair-growth cycle is obviously far from being available, but the question is an important one. The rate of mitosis achieved by the hair-follicle bulb during the proliferative stage of the cycle is about the highest achieved by any tissue of the body. Knowledge of how this periodic burst of cell-division is initiated and so precisely controlled may be fundamental to an understanding of normal and malignant growth processes. SUMMARY 1. Hair-follicle initiation and development, and concomitant changes in skin morphology, were observed in normal ( + + ) mice, and in ragged heterozygotes (Ra+) and homozygotes (RaRa), from the 14-day embryo until 3 weeks after birth. 2. Follicle initiation commenced at the normal time (14 days' gestation) in Ra+ and RaRa embryos. Early initiated Ra+ follicles were normal, but development of the later follicles was retarded or arrested. The growth of most RaRa follicles was arrested early. Grossly retarded follicles (Ra+ and RaRa) were non-functional and lacked hairs and sebaceous glands. These follicle defects resulted in lack of zigzag hairs in the adult Ra+ pelage and near-nakedness of RaRa adults—as described in a previous paper. 3. The degree of growth retardation in functional Ra+ follicles was variable, so they became out of phase with each other in their development. This anomaly was thought to cause in the adult mice intermediate types of fibre and out-ofphase progress of adjacent follicles through the hair-cycles. 4. Follicle density in Ra-\- mice was subnormal before birth, but near normal after birth. RaRa follicle density was consistently subnormal. In all genotypes follicle density was greater on the sacrum than in the neck region. 5. The epidermis in ragged mice became thicker than normal after 5 days post-partum, possibly because of decreased competition from the hair follicles for nutrients essential for mitosis. The depth of the adipose layer of the skin seemed related to the extent of follicle development. It was abnormally thin in Ra-\- mice and almost absent in RaRa mice. Possibly hair-follicle-growth, here deficient, is normally required to induce expansion of the adiposus. The shallow HAIR FOLLICLES IN MICE 527 adiposus apparently impeded growth in fully developed Ra+ follicles, and their shape and orientation frequently became abnormal. 6. The discussion deals first with interrelationships and causes of the juvenile and adult abnormalities due to the Ra gene. Finally, possible controlling mechanisms for the normal rodent hair-cycle are analysed in the context of this and other work. RESUME Etude morphologique du developpement de la peau et des follicules pileux chez la Souris normale et le mutant ''ragged'' 1. L'apparition et le developpement du follicule pileux ainsi que les changements survenant simultanement dans la peau ont ete etudies chez la souris normale ( + + ) d'une part, et chez l'heterozygote (Ra-\-) et l'homozygote (RaRa) d'autre part, depuis l'embryon de 14 jours jusqua'u jeune de 3 semaines. 2. L'apparition du follicule commence au stade normal chez les embryons Ra-\- et RaRa, c'est-a-dire le 14e jour de gestation. Les jeunes follicules Ra+ sont normaux, alors que les follicules plus ages sont retardes ou arretes. La croissance de la plupart des follicules RaRa est arretee tres tot. Les follicules tres retardes {Ra-\- ou RaRa) ne sont pas fonctionnels et sont depourvus de poils et de glandes sebacees. Ce defaut de follicules fonctionnels produit chez Padulte Ra+ des zones depourvues de poils en zigzag et chez l'adulte RaRa, une nudite presque absolue ainsi qu'il a ete decrit precedemment. 3. Le retard de la croissance des follicules fonctionnels Ra-\- est variable, de sorte que leur developpement se trouve dephase. Cet etat de chose cause chez la souris adulte l'apparition de types de fibres intermediates, et le dephasage de follicules voisins dans le cycle du pelage. 4. La densite des follicules chez les souris Ra-\- est inferieure a la normale avant la naissance, mais a peu pres normale apres. La densite des follicules RaRa est nettement inferieure a la normale. Dans tous les genotypes la densite des follicules est plus elevee sur le sacrum que sur le cou. 5. L'epiderme des souris 'ragged' devient plus epais que chez un temoin 5 jours apres la naissance, peut-etre en raison de la concurrence reduite exercee par les follicules sur les facteurs metaboliques essentiels pour la mitose. L'epaisseur de la couche adipeuse de la peau semble en relation avec le developpement des follicules. Elle est anormalement mince chez les souris Ra-{- et presque inexistante chez les souris RaRa. Onpeut supposer que la croissance du follicule, ici reduite, soit normalement necessaire pour determiner l'epaississement de la couche adipeuse. La minceur de la couche adipeuse semble empecher la croissance des follicules Ra-\- totalement developpes, leur forme et leur direction devenant souvent anormales. 6. La discussion porte tout d'abord sur les relations et les causes des anomalies du jeune et de l'adulte dues au gene Ra. Enfin les mecanismes susceptibles de 528 J. SLEE — D E V E L O P M E N T OF SKIN AND controler le cycle normal du pelage chez les Rongeurs sont analyses en fonction du present travail et de la litterature actuelle. ACKNOWLEDGEMENTS Part of this work was done at the Institute of Animal Genetics, Edinburgh University; thanks are due to Professor C. H. Waddington, F.R.S., who provided laboratory facilities there, and to Dr. T. C. Carter who gave much useful advice and encouragement. I am also indebted to Mr. D. B. Hill and Miss J. M. Henderson for technical assistance, to Mr. W. D. Roberts for the text-figures, and to Mr. D. Pinkney for the photographs. REFERENCES BALINSKY, B. I. (1950). On the developmental processes in mammary glands and other epidermal structures. Trans, roy. Soc. Edinb. 62, 1-31. BORUM, K. (1954). Hair pattern and hair succession in the albino mouse. Ada path, microbiol. scand. 34, 521-41. BULLOUGH, W. S. (1952). The energy relations of mitotic activity. Biol. Rev. 27, 133-68. & LAURENCE, E. (1960). The control of epidermal mitotic activity in the mouse. Proc. roy. Soc. B. 151,517-36. CARTER, T. C , & PHILLIPS, R. J. S. (1954). Ragged, a semi-dominant coat-texture mutant. / . Hered. 45, 151-4. CHASE, H. B. (1954). Growth of the hair. Physiol. Rev. 34,113-26. & EATON, G. J. (1959). The growth of hair follicles in waves. Ann. N.Y. Acad. Sci. 83, 365-9. MONTAGNA, W., & MALONE, J. D. (1953). Changes in the skin in relation to the hair growth cycle. Anat. Rec. 116, 75-81. RAUCH, H., & SMITH, V. W. (1951). Critical stages of hair development and pigmentation in the mouse. Physiol. Zool. 24, 1-8. DAVID, L. T. (1934). Studies on the expression of genetic hairlessness in the house mouse. / . exp. Zool. 68, 501-18. DURWARD, A., & RUDALL, K. M. (1949). Studies on hair growth in the rat. / . Anat. Lond. 83,325-35. DRY, F. W. (1926). The coat of the mouse. / . Genet. 16, 287-340. EBLING, F. J., & JOHNSON, E. (1959). Hair growth and its relation to vascular supply in rotated skin grafts and transposed flaps in the albino rat. / . Embryol. exp. Morph. 7, 417—30. (1961). Systemic influence on activity of hair follicles in skin homografts. / . Embryol. exp. Morph. 9, 285-93. FALCONER, D. S., FRASER, A. S., & KING, J. W. B. (1951). The genetics and development of 'crin- kled'—a new mutant in the house mouse. / . Genet. 50, 324—44. FRASER, A. S., & NAY, T. (1955). Growth of the mouse coat. IV. Comparison of naked and normal mice. Aust. J. Biol. Sci. 8, 420-7. GIBBS, H. F. (1941). Development of the skin and hair of the mouse. Anat. Rec. 80, 61-81. GRUNEBERG, H. (1943). The development of some external features in mouse embryos. / . Hered. 34, 88-92. HANSON, J. (1947). The histogenesis of the epidermis in the rat and mouse. /. Anat. Lond. 81,174—97. JOHNSON, E. (1958). Quantitative studies of hair growth in the albino rat. I. Normal males and females. / . Endocrin. 16, 337-50. MOHN, M. P. (1958). The effects of different hormonal states on the growth of hair in rats. In The Biology of Hair Growth, New York: Academic Press Inc. MOTTRAM, J. C. (1945). Effect of change of coat on the growth of epidermal warts in mice. Nature, Lond. 155, 729-30. NAY, T., & FRASER, A. S. (1954). Growth of the mouse coat. III. Patterns of hair growth. Aust. J. Biol. Sci. 7, 361-7. PINKUS, [H. (1958). Embryology of hair. In The Biology of Hair Growth, New York: Academic Press Inc. SLEE, J. (1957a). The morphology and development of 'ragged'—a mutant affecting the skin and hair of the house mouse. I. Adult Morphology. / . Genet. 55, 100-21. (19576). The morphology and development of 'ragged'—a mutant affecting the skin and hair / . Embryol. exp. Morph. Vol. JO, Part 4 J. SLEE Plate J J. Embryol. exp. Morph. Vol. 10, Part 4 J. SLEE 2 H A I R FOLLICLES IN MICE 529 of the house mouse. II. Genetics, embryology and gross juvenile morphology. /. Genet. 55, 570-84. STOHR, P. H. (1903). Entwicklungsgeschichte des menschlichen Wollhaares. Anat. Hefte, 23, 3-66. WHITELEY, H. J. (1958). Studies on hair growth in the rabbit. /. Anat. Lond. 92, 563-7. WOIBACH, S. B. (1951). Hair cycle of the mouse and its importance in the study of sequences of experimental carcinogenesis. Ann. N.Y. Acad. Sci. 53, 517-36. EXPLANATION OF PLATES PLATE 1 FIG. A. Sagittal section of + + skin at birth (neck region), showing the different stages of development of early and later initiated follicles, x 70. FIG. B. Sagittal section of Ra+ skin at birth (neck region). Development of the later initiated follicles is already abnormally retarded. x70. FIG. C. Sagittal section of RaRa skin at birth (neck region). Follicle development is markedly subnormal. x70. FIG. D. Sagittal section of + + skin (sacral region), 5 days after birth. Due to fast growth of the late initiated follicles the whole population is almost in phase in the stage of development. The adipose layer is greatly expanded, x 70. FIG. E. Sagittal section of Ra+ skin (sacral region), 5 days after birth. Note the retarded growth of many late initiated follicles (arrowed) with the resultant out-of-phase development of the follicle population. Note also the abnormally small depth of the adipose layer and the consequently (?) acutely angled orientation of the large follicles, x 70. FIG. F. Sagittal section of + + skin (sacral region), 7 days after birth. The follicle population is now closely in phase in development, x 70. FIG. G. Sagittal section of Ra+ skin (sacral region), 7 days after birth. Adjacent follicles are out of phase in development. The incomplete late initiated follicles (arrowed) are not capable of further development. Note the sharply curved large follicles—probably distorted because of the subnormal depth of adiposus. x 70. FIG. H. Sagittal section of RaRa skin (sacral region), 7 days after birth. Note the very retarded growth of all follicles, none of which would have become functional. The largest follicle may be distorted because of the virtual absence of an adipose layer, and the consequently very small depth of skin. X70. PLATE 2 FIG. I. Sagittal section of + + skin (neck region), 18 days after birth. The follicles have retracted, in phase with each other, into the resting stage of the hair-cycle, x 70. FIG. J. Sagittal section of Ra+ skin (neck region), 18 days after birth. The two large follicles are severely curved and distorted, possibly as a result of the contraction in skin thickness. The small clumps of dark-stained Malpighian cells are fragmented, incompletely developed follicles, x 70. FIG. K. Sagittal section of RaRa skin (neck region), 18 days after birth. The large follicle is nonfunctional and distorted in shape, probably because of skin contraction. The small cell clusters are fragments of incompletely developed follicles, x 70. FIGS. L, M, N. Whole mounts of + + , Ra+, and RaRa skin respectively, taken at birth. Note that the follicle groups, clear in + + , are less clear in Ra+ and absent in RaRa skin. The abnormally retarded follicle development in Ra+ and, even more, in RaRa skin is also clear, x 70. FIG. O. Sagittal section of + + skin (neck region) from an 18-day embryo, to show a typical follicle primordium at an early stage of development. Note the orderly arrangement of the epithelial cells in the primordium and the cells of the dermal papilla immediately below, x 370. FJG. P. Sagittal section of RaRa skin (neck region) taken at birth. The follicle primordium shown here is typical of some found in high-grade RaRa neonatal mice. Note the abnormal disorganization and diffuse scattering of the epithelial cells which form the primordium, and the absence of a dermal papilla. X370. (Manuscript received 8:iii:62) 5584.4 M m