Species Identification of Common Native Arctic Mammals in Inuit Fur Clothing Based on Hair Microscopy

ABSTRACT Correct material identification is considered essential when documenting museum objects. This study examines the morphology of mammal hair and records the geographical use of common species in Inuit fur clothing collected by the National Museum of Denmark (NMD) from c. 1830–1940 in the Bering Strait region, Alaska, Arctic Canada, and Greenland. Through hair microscopy, the purpose is to test whether original identifications are correct to assess the origin of unique Inuit garments. By means of transmitted light microscopy (TLM) of stained, 1 µm thick cross-sectioned hairs and undyed, longitudinally mounted hairs, the research reveals that specific morphological structures are characteristic of the common native reindeer/caribou, musk ox, members of the seal family, domestic dog, wolf, Arctic fox, polar bear, and wolverine. Rarer animals (hare, lynx, otter, etc.) are not part of this study because of limitations in the collection. Hairs from seal species are difficult to distinguish from one another. Hairs from dog and wolf are distinguishable but have relatively similar morphology. Therefore, to confirm identification, supplementary analyses are required. The hair microscopy technique was used on 49 garments in NMD’s collections, and the results were compared to the original macroscopic species identification. The study revealed that the latter method is often erroneous when it comes to dog/wolf and wolverine fur.


Introduction
The aim of the research presented here is to assess the accuracy of the transmitted light microscopy (TLM) of hair method for identification of fur from Arctic mammals. The goal is to demonstrate if the original macroscopic identification of Arctic mammalian fur in cultural institutions such as the National Museum of Denmark is correct. The microscopy method could eventually serve to uncover the provenance of Inuit fur garments currently without connection to archival information.
When examining museum objects, identification of the object's constituents is considered an important piece of information. This provides information about the wealth of the original owner and/or the specific traditions regarding production of the object. It may also reveal information about the fauna, flora, or inorganic materials in the area where the object was created or perhaps the possibilities of acquiring imported material from other areas. A striking example of this and the focus of this article is the fur skin that nineteenth to twentieth century Inuit (in the following termed as such according to Krupnik 2016 andIssenman 1997) in the Bering Strait region, Alaska, Arctic Canada, and Greenland used to create their characteristic clothing over hundred years of traditional, excellent craftsmanship. Due to extreme weather conditions, Inuit depend on warm and water-resistant clothing. Normally, skins from native animals in their local area were used, sometimes supplemented with traded skins. Thus, the specific use of native fur skin among specific Inuit groups may indicate their geographic provenance.
From a research point of view, knowing the identity of the material may help to identify misplaced or incorrectly registered items as well as providing information about items with unknown or lost provenance. Lost identification numbers, misplaced objects, and inadequate registration are common problems in museum collections. For example, at the National Museum of Denmark (NMD), 1171 items, or 14% of the collection of Inuit skin clothing, are currently without provenance, while other pieces of skin clothing are incompletely registered (calculated by the author in the NMD's digital object registration system December 2021).
For almost 200 years, species identification of skin and fur material at the NMD was based on macroscopic identification, probably after comparison to similar objects in the collection, e.g. comparing archaeological to historical objects (Anon. 1836(Anon. -1837. Information may have been acquired from the collector, manufacturer, or user at the time the museum acquired the object. While the results of the macroscopic method and information from the previous owner provide valuable knowledge about the material of the clothing, the information has never been fully verified. Based on the collection at the NMD this article attempts to evaluate the reliability of the original information about Inuit clothing especially with regards to fur skin.

The structure of hair
Morphologically, hair consists of three distinct structural layers, made from the protein keratin: the cuticle or cuticular layer, the cortex, and the medulla (this may be absent in the hair of some mammal species) (Zhang et al. 2019, 1) (see Figure 1, left). At taxonomic levels, these hair layers differ morphologically, making it possible to differentiate between species (Tridico et al. 2014, 102). The surface of the cuticle is covered with scales, which stabilise the hair and its attachment in the hair follicle (Haarløv 1986, 148-149). The cortex has a solid and compact structure, made up of spindle-shaped cells, while the medulla has a porous structure, in some species spaced with air pockets (Chattha et al. 2011, 54). The pattern of the cuticular scales, the granular or aggregate pigment in the cortex and the medulla, as well as the relative dimensions of the medulla and cortex may indicate the species (Brunner and Coman 1974, 8, 11;Chattha et al. 2011, 54;Teerink 2003, 8). In addition, the construction of the medulla is speciesspecific (Tridico et al. 2014, 102).
Most mammals have two types of hair: primary and secondary. Although the two types of hair are formed in the same way, they vary in thickness and length within the same species. The dimensions of the hair depend on the season, food, climate, and moulting cycle (Wildman 1954, 9-10). Primary hair (or guard hair) is formed first, deepest in the dermis (Haarløv 1986, 149). Primary hairs with diameters from c. 30-60 μm may be termed intermediate hair (Rowell et al. 2001(Rowell et al. , 1673. Fully developed primary hairs are often long, tapered, shiny, and thick (Stoves 1957, 194).
Secondary hair (or wool hair) is thin, placed around the primary hair according to a special pattern, characteristic of the animal species, and of the location on the skin. Fully developed secondary hair is short and less robust than primary hair (Stoves 1957, 194). Some mammals, including humans, oxen, and horses, have no secondary hair (Haarløv 1986, 149). Secondary hair is considered to be of little diagnostic value but occasionally has particular structures that are useful for identification (Brunner and Coman 1974, 2-3). New-born animals only have secondary hair, making identification of the species by microscopy almost impossible (Teerink 2003, 4).
The distance from the root determines the outline of the growing hair. It is often round at the root and tip of the hair but can show large variations on the thickest part of the shaft, e.g. elliptical, kidney shaped, or cigar shaped (Stoves 1957, 201) (see Figure 1, right). Several studies have documented the cuticular scale pattern in various ways, either by means of casts of the hair surface (Wildman 1954;Brunner and Coman 1974;Teerink 2003) or photographed directly, focusing on the surface of longitudinally mounted hairs (Stoves 1957, 184-185). The morphology of the medulla has been classified according to various systems, usually relying on Wildman's classification from 1954 (Brunner and Coman 1974;Kondo 2000, 9;Teerink 2003;Tridico et al. 2014, 105).
In their research of mainly Australian mammals Brunner and Coman concluded, that: 'the shape of the hair cross-section is undoubtedly the most important single criterion used in the hair identification system' (Brunner and Coman 1974, 5). Thus, the cross-sections are a sufficient method to identify unknown species, without further classification of hair structures (Brunner and Coman 1974, 5, 20-21). For species identification, cross-sections at the widest point on the primary hair are of the highest diagnostical value (Stoves 1957, 201;Brunner and Coman 1974, 4), while cross-sections at other points have supplementary value (Brunner and Coman 1974, 5;Teerink 2003, 11).
Cortex thickness relative to the total diameter of hairs could provide further clues for species identification (Teerink 2003, 8), and the ratio of cortex width to medulla width could be useful according to Brunner and Coman (1974, 8). In the literature, the medullary index of hair, i.e. the ratio of the medulla diameter to the hair diameter, is used as a criterion for species identification. However, some scholars regard it as of minor importance, e.g. Tridico for discriminating between domestic dog and domestic cat (Felis catus) (2014, 102).

Morphological characterisation applied in this study
Initially, hairs with medullae, > c. 60 μm wide were classified as primary hair and hairs c. 30-60 μm wide as intermediate hair. In general, hairs without medullae, < 30 μm wide were identified as secondary hair. In this context, the total width of the hair, regardless of hair shape, meant the largest dimension. To clarify, diameter is used in the following text about hair width. The most widely accepted classification systems were chosen to categorise the hair: (a) cuticular scale pattern, (b) morphology of the medulla, (c) shape of cross-section, (d) cortex ratio/hair diameter, and (e) pigmentation: (a) This study applied Stoves' photo-technique (Stoves 1957, 184-185) and incorporated Teerink's terminology (Teerink 2003, 6-8), based on Wildman (1954, 56-58). The cuticular pattern was thus categorised according to (1) scale position, (2) scale patterns, (3) scale margins, and (4) distance between the scales. (b) This study followed Wildman's original classification of the medulla morphology (1954, 49-53) (see Figure 2). (c) With regards to hair shape, Brunner and Coman's (1974) drawings of cross-sections served as a benchmark in this study (Figure 1, right). (d) In this study, following Teerink (2003, 8) and Brunner and Coman (1974, 8) the ratio of cortex width: total hair diameter was assessed. The relative dimensions of the cortex: hair width was measured and calculated from the longitudinal mounting. (e) Since hair pigmentation and its distribution in the cortex or medulla could confirm an identification, based on other characteristics (Brunner and Coman 1974, 10-11), the location of any pigmentation was noted in this study.

Sample preparation
The sampling site on the garment was photographed. A few hair strands from the fur were sampled using sterile tweezers, disposable gloves, and polypropylene sample bags. For cross-section, hair strands were embedded as straight as possible in EPON epoxy resin. Perpendicular to the hair direction, and using a microtome, the cured epoxy was cut into sections with a thickness of 1 µm, stained with 1% toluidine blue, and mounted in a drop of Pertex Mounting Medium® on a microscope slide beneath a coverslip. For longitudinal preparation hair strands were rinsed in acetone, dried, and mounted longitudinally like the cross-sections.
Microscopy was performed with a Leica DM4 M using transmitted light (TL) (100x, 200x, 400x, and 500x magnification), allowing scale-bar, measurement, and stacking of digital photos. Depending on the species, 30-80 hairs from each skin were photographed and assessed.
From the cross-sectioned mounts, this study assessed: . Classification of hair shape . Hair diameter . Possible pigmentation From the longitudinal mounts: . Classification of cuticular pattern and medulla . Measurement of hair diameter and cortex width, for cortex ratio assessment

Empirical material
A reference hair collection, covering the native Arctic mammal species most commonly used in Inuit clothing (Hatt 1969, 7-11;Issenman 1997, 32-36), was provided by the Natural History Museum of Denmark (NHMD) from its study collection of Arctic mammal skins. The skins had been biologically identified by NHMD's curators. Samples were collected on complete skins from the lower back (as recommended by Brunner and Coman 1974, 17), the abdomen, and  (e), concavo-convex (f), kidney shaped (g), or shaped like a dumbbell (h). Likewise, the medulla has different appearances according to the outline: large, medium-size, or divided. The medulla can also be absent, e.g. in secondary hair. Illustration after Brunner and Coman 1974, 6. occasionally from the extremity area, of the following species (Table 1). Morphological differences in dorsal, abdominal, and, where relevant, extremity hair were assessed separately. In some species the ratio between the hair types differed considerably, where most secondary hairs were found on the abdomen. Statistical analyses were not performed because of the limited amount of data. In the following figures of species, the number of variables (n) sums up the different hairs listed from the entire skin.
For comparison, 52 hair samples from 49 garments from the NMD Inuit collection were selected and species identified by means of TLM. The criteria for the selection of the garments were: (a) intact accession number and archival information, (b) dating from the late nineteenth to early twentieth century, (c) representative Inuit provenance, and equal distribution of male and female clothing, (d) when possible, affiliated garments, and (e) representing the most used species, identified by macroscopic recognition.

Cross-section characteristics
In this study, the cross-section of Arctic mammalian hairs revealed obvious morphological differences regarding hair shape, presence of medullae, and the sorting in primary, intermediate, and secondary hair. The study identified eight different hair shapes, two of which had not been previously registered: hat shape and nut shape (see Figure 3).

Hair morphology in Arctic mammals
The microscopy results were listed in groups with the closest familiar relationships: ruminants, Phocidae, and carnivores (Canidae, polar bear, and wolverine). Their characteristics are specified below. This study detected no morphological differences between hairs located on different parts on the animal, i.e. the back, the abdomen, or the limbs. However, ruminants and carnivores contained more secondary hair on the abdomen than on the back and limbs.

Ruminants' hairs
In cross-section, primary hairs from R. tarandus and O. moschatus are round to oval, (Figure 4). The species differ from each other as follows: . R. tarandus have distinct wide-open cobweb medullae, almost without visible cortex.  . R. tarandus' primary hair has wide lattice medullae.
The ratio cortex width: total primary hair diameter is 1:99. . O. moschatus has medium width, unbroken or lattice medullae; ratio cortex width: total primary hair diameter is up to 2:5.
In both species, the cuticular pattern in primary and secondary hairs cannot be determined by TLM.
Pigmentation is only visible in the cortex of O. moschatus. The Appendix provides TLM photomicrographs of ruminants.

Phocidae hairs
In cross-sections of Phocidae hairs, the absence of medullae and flattened shapes prevail (Figure 5a and b). The obvious differences between the species are: . Juvenile E. barbatus have predominantly oval primary hairs. . Adult P. groenlandicus and sub adult P. hispida have oblong primary hairs. These hairs are difficult to distinguish based on their shapes in cross-section, but secondary hairs have slightly different shapes. C. cristata, E. barbatus, and P. hispida have round secondary hairs; P. groenlandicus, C. cristata and E. barbatus have round to oval secondary hairs.  Table 1). Light blue colour indicates the extent of the cortex. Dark blue colour shows different appearances of the medulla. The eight hair shapes exist in two versions: either consisting of the cortex, without medulla (column A), or of cortex and medulla (columns B1 and B2). The medulla is either narrow (ratio cortex width: hair diameter > 4:5), medium (4:5-1:5), or wide (1:5-1:99). Column B3 reveals the variation in the medullae, i.e. as compact remnants or cobwebs. Primary hairs have the greatest variation in shapes. Nut shape is only found in intermediate hairs, while round and oval shape are seen in both primary, intermediate, and secondary hairs. Eye and hat shape only occur in primary hairs.
. Juvenile C. cristata can be identified by the remnants of medulla in cross-section (as well as unbroken medullae in longitudinal section). Hairs from C. cristata are generally more oval than other species. . Adult P. vitulina and Largha seal (Phoca largha) differ from other species by having oblong primary hairs. . Eye and hat shapes are most common in hairs from young seals.
Hairs from Phocidae are easily identified in longitudinal mounts, because of the large diameter of primary hair and usually absent medullae, but the specific species cannot be identified based on this. It is impossible to determine the cuticular pattern in primary hairs by TLM, whereas secondary hairs have identical cuticular patterns (transversal, mosaic, smooth, distant). Pigmentation occurs randomly and is considered useless for identification. The Appendix provides TLM photomicrographs of Phocidae.

Carnivores' hairs: Canidae
In cross-section, the differences between hairs from Canidae are as follows (Figure 6): . C. lupus has kidney shaped primary hairs. . C. lupus familiaris and C. lupus differ from V. lagopus by not having oblong primary hairs. . V. lagopus (winter) has dumbbell shaped primary hairs. . The thinner hairs of V. lagopus reveals the identity of this species.
Hairs from Canidae are easy to identify in longitudinal mounts, where the medium width of the medullae, presence of unbroken medullae in primary hairs, and/ or a uniserial ladder in intermediate hairs discloses the family. The ratio cortex width: total primary hair diameter is 1:3. The cuticular pattern in primary hairs cannot be determined by TLM, whereas secondary hairs have an identical cuticular pattern (transversal, elongated petal, smooth, distant). Pigmentation occurs randomly in the cortex but is not useful for identification. The Appendix provides TLM photomicrographs of Canidae.

Carnivores' hairs: polar bear and wolverine
In cross-section, the primary hairs in both species are of the same size, oval with medullary remnants, but the cortices have obvious differences between the species (Figure 7): . U. maritimus has a wide cortex with sparse pigmentation. . G. gulo has a pigmented, medium-size cortex. In addition, intermediate hairs of G. gulo are nut shaped.
Both species have round secondary hairs.
In longitudinal mounts, the species are distinguishable by the ratio cortex width: total primary hair diameter; in U. maritimus up till 1:4, in G. gulo 1:2. In addition, the appearances of the medullae and the cuticular scales are different: . U. maritimus primary hairs have a simple, unbroken, medium to narrow width medulla.   The Appendix provides TLM photomicrographs of U. maritimus and G. gulo.
Hair microscopy compared with macroscopic identification Figure 8 presents the results of the microscopic species identification and the original macroscopic identification of the items from the NMD Inuit clothing  collection. For the sake of clarity, in the columns [Macro id], the binomial names are given, although in the original accession protocols, animals are registered by common names.

Identification of ruminants
The TLM analyses of the skins of 14 ruminants showed: . 11 R. tarandus matched 11 macroscopic identifications. Four of these were identified merely by longitudinal mounts. . One O. moschatus matched the macroscopic identification. . Two garments with cattle fur, Bos taurus, were identified. Both had been macroscopically identified as C. lupus familiaris.

Identification of Phocidae
The TLM analyses of 22 Phocidae showed: . P. hispida was identified in 14 samples. Of these, 12 had been identified macroscopically as Phocidae, one as P. vitulina / Phocidae, one as unidentified species. . P. vitulina was found in four samples; however, one was identified as possibly P. vitulina. Two had been identified macroscopically as Phocidae, two as P. vitulina. . P. groenlandicus was found in one sample; it had been identified macroscopically as Phocidae. . Three samples were identified as Phocidae without further specification; they had been identified macroscopically as Phocidae. Identification of Canidae, G. gulo, and U. maritimus The TLM analyses of 16 carnivores showed: . Five C. lupus familiaris were identified. Of these, one had been identified macroscopically as C. lupus familiaris, two as G. gulo. Two were unidentified. . Three V. lagopus were identified, matching the macroscopic identification. . Three Canidae were identified as C. lupus familiaris or C. lupus. Macroscopically, one had been identified as G. gulo, one as C. lupus familiaris, and one as V. lagopus. . One G. gulo matched the macroscopic identification. . Four U. maritimus matched the macroscopic identifications.
In this study of Arctic species, the characteristic morphology of primary, intermediary, and secondary hairs was assessed from stained cross-sections and unstained longitudinal mounts. Longitudinal mounting enabled identification of the cuticular patterns of secondary hairs, but not of the primary and intermediary hairs (except for C. cristata, P. vitulina, and U. maritimus). This was due to limitations in the capacity of the microscope and the large diameters of primary hairs. Cuticular patterns in secondary hairs were identifiable in most species. According to Brunner and Coman, scale patterns can often be ignored; they 'are generally only used to confirm identifications, based on other criteria' (Brunner and Coman 1974, 18). Therefore, the cuticular pattern in this study became less critical. Whenever it was possible to identify the pattern, it supported the final identification. Likewise, the random occurrence of pigmentation seemed of minor importance. Consequently, in this study hair shape and diameter, morphology of medullae, and cortex ratio were applied as the most important parameters in species identification. Hair identification is a combined process of recognition of patterns (Chattha et al. 2011, 54); however, 'it must be remembered that no system of hair identification can be fully effective' … In these cases, precise identification may not be possible (Brunner and Coman 1974, 17).
The database Alaska Fur ID Project provided important data on the hair morphology from Arctic mammals, while information was sparse in the general TLM literature. Instead, some information produced by SEM was useful for comparison (Meeks and Cartwright 2005;Galatík et al. 2011;Rast-Eicher 2016). To evaluate the relevance of the abovementioned results, the following serves as a comparison between the present study and previous research literature within the field of hair microscopy of Arctic animals.

Ruminants, hair characteristics
Only a few descriptions of cross-sectioned hairs from R. tarandus were found (Stoves 1957, 207, 214;Meeks and Cartwright 2005, 42). One study mentioned the appearance of heart shaped hair (Galatík et al. 2011), which could have developed as an artefact during preparation. The presence of intermediate hairs in O. moschatus was described in one study (Rowell et al. 2001(Rowell et al. , 1673, while nut shaped intermediate hair has remained undetected until now, as well as the illustrated cross-sections. Several authors described longitudinally mounted hairs of R. tarandus (Meeks and Cartwright 2005;Carrlee and Horelick 2011;Galatík et al. 2011¸Rast-Eicher 2016, while only two studies described this in O. moschatus (Wildman 1954, 142-144;Carrlee and Horelick 2011).

Phocidae, hair characteristics
Cross-section studies by TLM of the hair from Phocidae are scarce (Stoves 1957, 203, 210, 214;Carrlee and Horelick 2011), and the hat shaped crosssection has been undocumented until now. By means of SEM, Meeks and Cartwright claimed that hair from Phocidae had solid medulla (2005,43). In contrast, The Alaska Fur Project (Carrlee and Horelick 2011) and Furskin Identification (Galatík et al. 2011) found that there was a general absence of medullae in Phocidae. By staining the cross-section of hair with toluidine blue, this study identified Phocidae primary and intermediate hairs with absent medulla. Apart from cross-sectioned hairs from new-born and adult P. groenlandicus (Stoves 1957, 203, 214;Rast-Eicher 2016, 179), no further documentation of cross-sectioned hairs from Phocidae was uncovered. This study revealed that P. vitulina and P. largha were similar but differ from other Phocidae by having oblong primary hairs. In order to differentiate between the two species, samples require a reliable geographical provenance because the seals' natural habitats are far apart. Still, in this study, crosssection identification of specific seal species was challenged because of the species' indistinguishable hair shapes.

Hair characteristics, polar bear and wolverine
Tridico et al. demonstrated by the cross-section that the U. maritimus primary hairs contain an air-filled medulla (2014,(103)(104), in contrast to previous researchers, claiming that polar bear hairs were hollow (Wang, He, and Li 2012, 340) and lacked medulla (Morioka 2009, 578). Cross-sections of G. gulo hairs were not documented in prior literature. This study was the first to document both species' characteristics in cross-sectioned hairs, i.e. the presence of nut shaped and intermediate hairs, and medulla appearances.

Discrepancies between results of hair TLM and macroscopic identification
From the empirical material, the following was deduced regarding the original macroscopic identification of Arctic fur (see Figure 8).
Comparison of ruminants 11 of 11 R. tarandus had corresponding identification with TLM, signifying that the macroscopic identification of R. tarandus was reliable. This is consistent with the author's own experience: fur from R. tarandus is easily identified with the naked eye. Likewise, fur from O. moschatus is easily identified macroscopically. However, the analysis of one sample is not sufficient. TLM of hair was therefore preferable.
Comparison of Phocidae 18 of 22 Phocidae were macroscopically identified to the family level. The most common species P. hispida (identified in 14 samples by TLM) was not identified macroscopically at all, while P. vitulina was identified macroscopically in two samples, both confirmed by TLM. However, macroscopic identification at species level was not accomplished successfully. This observation confirms the author's experience: Phocidae fur from adult mammals is easy to identify macroscopically. Probably by training, fur from P. hispida, P. vitulina, and P. groenlandicus can be identified macroscopically. Still, TLM identification of Phocidae, especially juvenile fur, is challenging and should be confirmed by other means of identification e.g. DNA analysis or proteomic analyses.

Comparison of carnivores
Four furs were macroscopically identified as C. lupus familiaris; of these, however, by hair TLM C. lupus familiaris was only identified in one sample, C. lupus or C. lupus familiaris in one sample, and B. taurus in two samplesin this study the only examples of imported fur. Two samples, unidentified macroscopically, were identified as C. lupus familiaris by hair TLM. Four furs were macroscopically identified as V. lagopus. Of these, by hair TLM, three were identified as V. lagopus, and one as C. lupus or C. lupus familiaris. Three furs were macroscopically identified as G. gulo, one as G. gulo? By hair TLM, one sample was identified as G. gulo, two as C. lupus familiaris, and one as C. lupus or C. lupus familiaris. Thus, the macroscopic identification of Canidae and G. gulo indicated that the results were doubtful. In a DNA study of 54 previously macroscopically identified Arctic dog fur specimens from the NMD, only 31 were identified as C. lupus familiaris; the rest were from other animals (Harris et al. 2020, 5). Hair TLM of Canidae and G. gulo, if possible supported by DNA analysis, is preferred.
Four out of four macroscopically identified U. maritimus matched the results of the hair TLM. However, according to the author's experience, U. maritimus can macroscopically be confused with C. lupus familiaris and new-born Phocidae. Hair TLM is therefore preferred.

Conclusion
Species identification by the studied hair TLM is a consistent analysis method when it comes to differentiating between fur from native species of mammals used in Inuit clothing. The shape of the hairs in crosssection, the medulla morphology, if present, and the appearance of the cortex, observed in the longitudinal mounting, are the most significant factors for reliable identification.
Hairs from R. tarandus and O. moschatus are easily distinguishable from other Arctic mammals by hair TLM, as well as macroscopically. Likewise, hairs from Phocidae are easily identified. Specific species identification of Phocidae is possible by means of hair TLM but difficult because of the occurrence of similar morphological characteristics. Presumably, macroscopic identification of the seal species may be improved by training. However, inclusion of other techniques, e.g. DNA analyses or proteomic analyses, is recommended for reliable species identification of seal. Hairs from Canidae and G. gulo can be differentiated under the microscope by the shapes of cross-sections of primary hairs, and V. lagopus is distinguished by its smaller hair diameter. However, the hairs from C. lupus familiaris and C. lupus have similar shapes, making the identification more difficult. Accordingly, identification of these species needs further validation from other analyses, e.g. DNA analyses or proteomic analyses. Hairs from U. maritimus are easily distinguished in the microscope, but the fur can be difficult to identify macroscopically.
In conclusion, regarding the collection of Inuit garments at the National Museum of Denmark, TLM of fur especially from C. lupus familiaris and C. lupus is recommended because the original species identification is based on frequently erroneous macroscopic identification. Furthermore, identification of Phocidae species is a possible subject for future TLM and DNA analyses.
Above all, continued collection of well-documented reference material for hair TLM is required for consistent identification, preferably confirmed by e.g. DNA analyses. The reference material should furthermore encompass a larger variety species of all ages. Future determination of fur objects without provenance depends on a correct identification of species. Ideally, there should be at least two different, consistent analyses of the species determination. Hair TLM is recommended as one.