Aspects of the skin morphology in the sub-Antarctic fur seal Arctocephalus tropicalis

The present study examined whether sweat glands are present in the skin of sub-Antarctic fur seals, Arctocephalus tropicalis, using standard histological procedures and light microscopy. Aspects of the microanatomy and histology of skin were described to elucidate structure and function. Sweat glands were present both in the furred areas of the body trunk, as well as in naked skin areas of the flippers. The density of the apocrine sweat glands appeared higher in fur-covered areas than in naked areas, while the size of sweat glands was similar amongst naked and fur-covered skin. The superficial position of sweat glands in the dermis of naked skin areas and the large, coiled, tubular nature of the secretory portion seem to indicate sweat gland activity in contrast to the deep lying, narrow sweat glands in fur-covered areas.


Introduction
Pinnipeds are adapted to conserve heat in their normally cold marine surroundings with its higher thermal capacity.They are, however, subjected to considerable heat loading on land (exogenous and endogenous) when they haul out to breed and moult.Behavioural (Bester and Rossouw 1994) and postural (Gentry 1973) adjustments are utilized in addition to physiological thermoregulation (Castellini 2018) to combat thermal overloading.
Postural adjustments related to the degree of exposure of the essentially naked flippers are extensively used by otariids (fur seals and sea lions) under conditions of thermal loading (Bartholomew and Wilke 1956;Gentry 1973).Fur-covered skin varies little in temperature resulting from the constant external insulation (fur), but changes in skin temperature of naked flippers allow dissipation of excessive heat (Irving et al. 1962).In naked or hairless skin, by definition, hair cannot be seen by the naked eye over the general skin surface, whether there are openings present or not (Rotherham et al. 2005).Such naked skin areas are considered the major sites for heat dissipation (Irving et al. 1962;Whittow et al. 1972) through a combination of heat convection, radiation and cutaneous evaporation of sweat while on land (Odell 1974).
Sweat glands associated with the pelage occur over the general body surface of pinnipeds (e.g., Montagna and Harrison 1957;Sokolov 1960;Scheffer and Johnson 1963;Ling 1965aLing , 1968;;Ling and Button 1975;Rotherham et al. 2005;Khamas et al. 2012;da Silva et al. 2020).Sokolov (1958) thought that in pinnipeds there were no sweat glands associated with naked areas of flippers, or other hairless areas of skin.Subsequently, sweat glands were reported from naked skin of both the hind and fore flippers of the northern fur seal Callorhinus ursinus, the hind flipper (but not the fore flipper) of the New Zealand fur seal, Arctocephalus forsteri, but not from the hairless part of the front flipper of the Antarctic fur seal A. gazella (Ling 1965a).Due to the inconsistent usage and unqualified meaning of "naked" and "hairless" skin in the literature, the interpretation of published information is difficult.In California sea lions Zalophus californianus sweat glands were present throughout the dermis, and the presence and distribution of sweat glands appeared to be species specific (Khamas et al. 2012).However, there is no doubt that amongst fur seals, the northern fur seal C. ursinus (Ling 1965a), the Cape fur seal (CFS) A. pusillus pusillus (Rotherham et al. 2005), and the South American fur seal (SAFS), A. australis (da Silva et al. 2020), possess sweat glands in both fur-covered and naked skin areas.
The present study investigated (a) furred and (b) naked skin in the sub-Antarctic fur seal (sAFS), A. tropicalis, to ascertain the presence/absence of sweat glands.We briefly describe aspects of the microanatomy and histology of skin to elucidate structure and function, in terms of the thermoregulatory process, following Rotherham et al. (2005).Such information might be important to address the influence of the sub-Antarctic environment on aspects of the terrestrial phase of fur seal populations (Bester 1982).

Materials and methods
Skin samples were collected immediately after death from two male sAFS, A. tropicalis, aged 8 (standard length [STDL] = 1.31 m, mass = 37.5 kg) and 14 years ((STDL = 1.62 m, mass = 97.3kg), respectively, that were shot, as part of another study at Gough Island (40 o 20' S, 9 o 54' W) during October 1978.Pieces of skin of 100 mm 2 were cut with a scalpel to the depth of the subcutaneous adipose tissue from a number of sites (n = 12) representing (a) fur-covered (n = 6 sites, e.g., chest and neck), (b) naked (n = 5 sites, e.g., leading edge of front flipper and tip of hind flipper), and (c) thick naked skin from a pressure area on the underside of a front flipper (n = one site).Furred areas were clipped before sampling.
All tissue was fixed with Bouin's fluid (Disbrey and Rack 1970) for 48 h and stored in 70% ethanol.Further processing used standard histological procedures (Disbrey and Rack 1970) following Rotherham et al. (2005).Tissue sections were cut transversely and longitudinally to the long axis of the body, perpendicular to the free skin surface, at 7 μm with a sliding microtome.Sections were stained with Ehrlich's acid haematoxylin and counterstained with a 1% eosin solution, and studied under a light microscope.The sweat glands and layer thickness of the skin were measured (μm) using a calibrated eyepiece micrometer.The outside diameters (width + depth)/2 of tubules in the secretory portion of the sweat gland were taken from the first 10 tubules viewed in cross section from each sample site.The volume of individual sweat glands was estimated using the formula for determining the volume of a sphere (4/3πr 2 ), where the radius (r) was taken as the half of the mean of two outside diameters measured at right angles to each other.The penetration of sweat glands into the dermis relative to the free surface of the skin was measured.The numbers of common pilary canals opening on the surface of the skin per unit surface area (mm −2 ) were counted as a measure of pilosebaceous unit density.
Mean values and standard deviations (S.D.) were calculated for each parameter.But without the original data (see next paragraph), we could not test for differences within and between the fur-covered and naked skin sample sites.In the original study, measurements were taken from fur-covered [6 sites] and naked [5 sites] skin areas.Sample sizes for each parameter measured, could therefore vary between 10 (per site) and 50 (5 × 10) in naked skin areas to 60 (6 × 10) from fur-covered sites, not distinguishing from which of the two adult males the samples were taken from.The sample sizes for the calculated parameter means and S.D. must therefore be assumed to lie between 10 and 60 (this study).
The present study of the sAFS preceded that of the CFS by 25 years, using almost the exact same sample processing and measuring procedures as for the CFS (Rotherham et al. 2005).Unfortunately, the original sAFS data which were processed to produce a BSc Zoology honours project report by one of us (MvO), was subsequently lost.All that remained was (a) the text of the report without accompanying figures, (b) two tables, and (c) colour images (n = 7) of cross sections through the skin (no scale available) representing (a) a furred area on the trunk of the body, (b) a naked area from the dorsum of a flipper, and (c) a pressure area from the ventral side of a flipper.
We therefore only use the numerical data contained in the surviving two tables.We also limit ourselves to a concise qualitative description of the histology of sAFS skin without elaboration beyond that which can be seen in the few (low resolution) colour images.We then compare the information with the measurements and histological descriptions of the skin in the closely related CFS (Rotherham et al. 2005) and SAFS (da Silva et al. 2020).

Results and discussion
The skin of fur seals is composed of two principal layers: the outer epidermis and the underlying dermis.Beneath the dermis a layer of loose connective tissue commences (Fig. 1a).

Epidermis
Four cell layers-the inner stratum germinativum (or stratum basale), stratum spinosum, stratum granulosum and a stratum corneum were distinguished (Fig. 1b-d).The poor resolution of the photomicrograph prevented the possible identification of a 5th layer, the stratum lucidum (Fig. 1d) which is present in the thick epidermis of flippers in SAFS (da Silva et al. 2020).In the thicker epidermis, the stratum corneum was dense and compact (Fig. 1d), making it more resistant to sloughing (da Silva et al. 2020).
In naked areas (Fig. 1b-d) the mean epidermal thickness (μm) appeared thicker than in the fur-covered areas (Fig. 1a; Online Resource 1a), especially the stratum spinosum and the stratum granulosum.Significant differences in the thickness of skin layers between fur-covered areas and naked skin areas, as well as the thickness of skin layers amongst naked skin areas, were also described for the CFS (Rotherham et al. 2005).The epidermis of the sAFS formed ridges and furrows as it abuts the contours of the underlying dermis in fur-covered areas, with clear-cut interlocking epidermal and dermal papillae in thick epidermis of the naked flipper sections (Fig. 1b-d) as in SAFS (da Silva et al. 2020).
Although the thickness of individual epidermal layers, and total epidermal thickness, varied considerably between sample sites in the different body regions, the apparently greater layer thicknesses of naked areas are undoubtedly related to the absence of a protective fur coat.This finds maximum expression in areas which are more prone to abrasion and pressure such as the ventral side of the flippers (Fig. 1d) as in SAFS (da Silva et al. 2020).

Dermis
The dermis consisted of a thin outer pilary layer, and a reticular layer.The papillary layer had dermal papillae which projected into the epidermis as described above, and extended only slightly below the base of the papillae where it merged more or less gradually with the thicker reticular layer (Fig. 1b-c; Online Resource 1a).
The abovementioned information for the epidermis specifically (da Silva et al. 2020; this study), and both principal layers of skin in CFS (Rotherham et al. 2005) conforms to observations that the thickness of the epidermis plus dermis is inversely related to the development of the pelage in pinnipeds (Sokolov 1960).

Pilosebaceous units
Each pilosebaceous unit extended into the dermis and contained a guard hair, an underfur fibre bundle, a sweat gland and sebaceous glands, all sharing a common pilary canal which opens through the epidermis to the exterior (Fig. 1a; Online Resource 1a, c).The underfur fibres were grouped posteriorly to the guard hair with which they were associated, and the follicles of the underfur fibres were clustered together to form a bundle (Online Resource 1c) in sAFS (Condy and Green 1980; this study) as in all fur seals (Yochem and Stewart 2018).The epidermal matrices of the underfur (secondary) follicles were closer to the skin surface than those of the guard hair (primary) follicle (Fig. 1a; Online Resource 1a) and both the primary and associated secondary follicles occur in all the fur-covered body areas examined (Rotherham et al. 2005;da Silva et al. 2020;this study).
No emergent hairs were found in the naked area, only primary follicles as shown by da Silva et al. (2020).The pressure area was devoid of hair follicles (Fig. 1d), most likely because of the limited material (one photomicrograph) available, as hair follicles were found in the ventral flipper regions in SAFS, but in lower numbers (da Silva et al. 2020).The density of pilosebaceous units as measured by the density of common pilary canals (Online Resource 2-Table 1) appeared higher in fur-covered areas than in naked areas.Pilosebaceous units disperse as a seal increases in age and body surface area (Scheffer and Johnson 1963); this has also been observed in A. tropicalis (Condy and Green 1980).The seal body and appendages grow at different rates, resulting in elongate and thin flippers, with long cartilaginous extensions to the digits, especially in otariids (Cooper 2018).Therefore, the relatively sparse pilosebaceous population of the naked skin of the flippers compared to the furred skin of the torso, may ontogenically have resulted from a stretching effect, the flippers in particular having a high surface area to volume ratio (Odell 1974).

Sweat glands
Each apocrine sweat gland consists of a secretory part and an excretory duct (Rotherham et al. 2005).The secretory part was situated in the dermis and extended well below the bases of the hair follicles (Fig. 1a; Online Resource 1a).The secretory part of the tubule was coiled and twisted on itself; hence, in sections it appeared as a small cluster of cross and oblique sections of tubules (Fig. 1a; Online Resource 1a, c).The excretory duct followed a somewhat spiral course through the dermis to finally open in the pilary canal (Rotherham et al. 2005).Rotherham et al. (2005) also described, for the first time in pinnipeds, eccrine sweat glands, the excretory ducts of which opened up directly on the skin surface.Da Silva et al. (2020) described the same sweat gland structure for thick epidermis, but identified it as merocrine sweat glands.In Z. californianus, sweat glands occupied most of the dermis in the skin of the flippers, and a few scattered sweat glands were found to be merocrine in nature (Khamas et al. 2012).
Only deep-lying sweat glands were present in fur-covered areas (Fig. 1a) that appeared to penetrate to a greater depth than the shallow-lying sweat glands of the naked areas (Fig. 1c), as in CFS and SAFS.On the other hand, a wide range of mean volume of sweat glands in fur-covered (chest and neck) and naked skin areas (hind flipper tip and anterior dorsal edge of front flipper) exists (Online Resource 2-Table 2).Neither did the volume fraction of sweat glands, whether associated with hairs (i.e., apocrine) or not (i.e., eccrine/merocrine) differ significantly amongst body regions of SAFS in both the superficial and deep dermis (da Silva et al. 2020), despite the observation that in the thick epidermis the sweat glands that opened directly at the surface of the skin were larger in size and had larger lumens.By contrast, in CFS, the mean diameters of the secretory portion of apocrine sweat glands were significantly larger (around 6-7 times larger) in the naked skin areas than in furred areas, while the secretory portion of the eccrine sweat glands did not differ between such areas (Fig. 2, Table 1 [the 6th line entry being in error and should be 'apocrine secretory gland diameter'] in Rotherham et al. 2005).
The density of common pilary canals (Online Resource 2-Table 1), each of which represents a pilosebaceous unit with its associated apocrine sweat gland (Fig. 1a; Online Resource 1a), was higher, on average, in fur-covered areas than in naked areas (this study).However, no significant differences in sweat gland density were found between such areas in CFS and SAFS (Rotherham et al. 2005;da Silva et al. 2020).This is the probable result of (a) our using the density of pilosebaceous units as a proxy for apocrine sweat gland density; (b) our inability to clearly separate eccrine sweat glands from apocrine sweat glands based on the differences in secretory ducts openings (directly at the skin surface or into the common pilary canal); and (c) the differences in the relative ages (and sizes) of the fur seals used in the comparison (adult male sAFS, adult female CFS, and SAFS pups).The latter consideration is likely in view of the increased dispersal of pilosebaceous units as seals increase in age and body surface area (Scheffer and Johnson 1963;Condy and Green 1980).
The role of sweat glands, and therefore cutaneous evaporation in the absence of free water to dissipate heat, is unclear.The superficial position of sweat glands in the dermis of naked skin areas (Fig. 1c) as compared to furcovered areas (Fig. 1a), and the apparently large, coiled, tubular nature of the secretory portion (Online Resource 1b) seem to indicate sweat gland activity in contrast to the deep lying, narrow ones in fur-covered areas.In CFS the apocrine sweat glands were considerably larger in naked skin areas (Rotherham et al. 2005), and "large abundant" sweat glands were found in the flippers of C. ursinus (Bartholomew and Wilke 1956).Such size differences in sweat glands were, however, not significant amongst naked and fur-covered skin in SAFS pups (da Silva et al. 2020), and apparently not in sAFS adult males (this study).Also, in otariids, skin of the entire body has a tendency to be emissive during both hot and cold weather (Khamas et al. 2012).Nevertheless, what was assumed to be sweat appeared on the naked flippers of Z. californianus (Whittow et al. 1975), C. ursinus (Scheffer 1962) and A. tropicalis (this study).Considering that an estimated 15-22% of the heat production of Z. californianus was lost through cutaneous evaporation, the naked flipper areas and its array of sweat glands seem to be suited to facilitate such a process.

Sebaceous glands
The sebaceous glands are located in the dermis and associated with hair follicles (Fig. 1a; Online Resource 1a).Sebaceous glands were found both in fur-covered (Fig. 1a) and naked skin areas (Fig. 1b) but were absent in pressure areas on the underside of the fore flippers (Fig. 1d).An apparent increased size of the sebaceous glands in naked skin areas compared to fur-covered skin was established for the less hairy parts of otariids (Ling 1965a).The dense fur and secretions from the sebaceous glands may help water-proof and prevent water-to-skin contact, but with the progressive loss of hair in naked areas such a need is met by enlarged lipid secreting sebaceous glands (Ling 1968).

Vascularization
Vascularization was conspicuous in those areas of the dermis that were close to the epidermal layer in naked skin areas (Fig. 1b, c).Vascularization was also conspicuous in the dermal papillary ridges in the fur-covered areas in SAFS (da Silva et al. 2020).The relatively large size of the blood vessels coupled with the reduced insulation and predominance of papillary and interpapillary pegs that would increase the dermal/epidermal interface in naked skin, promotes thermoregulation (da Silva et al. 2020).

Conclusions
The microanatomy of the skin of A. tropicalis suggests that naked skin areas possess functional sweat glands which may be implicated in thermoregulation through cutaneous evaporation of sweat.The presence of fine hairs with associated cutaneous glands in naked (hairless) skin was also detected in the fur seals C. ursinus, A. forsteri. A. p. pusillus and A. australis.The absence of fine hairs in the fore flippers of both A. forsteri and A. gazella (Ling 1965b) is curious and probably reflects less extensive searching or inadequate material.Although eccrine (merocrine) secretory ducts opening at the skin surface were not detected in A. tropicalis, eccrine sweat glands are no doubt present as in CFS and SAFS (Rotherham et al. 2005;da Silva et al. 2020).The circulatory system in naked skin areas of the (hind) flippers especially appeared conducive to heat transfer to the environment (da Silva et al. 2020).
The function of sweat glands in A. tropicalis is likely to play an inferior role in thermoregulation.The temperature of the flippers in pinnipeds in and out of water (Irving et al. 1962;Whittow et al. 1972;Ohata et al. 1977;Gallivan and Ronald 1979) fluctuates dynamically.The possible precedence of consideration of water economy over the regulation of body temperature (Whittow et al. 1972) also comes into play, as does the renal efficiency of the fur seal (Bester 1975;Ortiz 2001) which can only be considered adequate when judged against the normally cool and moist marine (oceanic) environment (Bester 1975).

Fig. 1 a
Fig. 1 a Photomicrograph (no scale available) of an oblique cross section through the furred skin of a sub-Antarctic fur seal showing the two principal layers of the skin, (a) the outer epidermis, and (b) the underlying dermis, and a pilosebaceous unit comprising (c) a common pilary canal, (d) a guard hair, (e) associated under hairs, (f) sebaceous gland, and (g) sweat glands.b Photomicrograph (no scale available) of a cross section through the naked skin of a flipper tip in a sub-Antarctic fur seal showing the (a) vascularization (V), (b) a sebaceous gland (SB) and (c) the cartilaginous extension of a digit.c Photomicrograph (no scale available) of a cross section through the naked skin of a sub-Antarctic fur seal showing the (a) layered epider-