INTRODUCTION

Cerdocyon thous (Linnaeus, 1766), the crab-eating fox, is a member of the family Canidae and occurs in a wide range of the Brazilian territory, from dense forests to open grasslands, with exception of the federal states Acre and Amazonas (^{Berta, 1982}; ^{Machado & & Hingst-Zaher, 2009}; ^{Lucherini, 2015}). Adult animals weigh from five to eight kg and measure approximately 65 cm from head to tail (Berta; ^{Emmons & Feer, 1997}). C. thous is a nocturnal and crepuscular species, omnivorous, a generalist and opportunist (Berta; ^{Langguth, 1975}).

Nevertheless, the species is affected by habitat fragmentation and road mortality and threatened by hunting from farmers and the proximity to domestic dogs and their potentially contagious diseases (^{Cheida et al., 2011}; Lucherini).

^{Moura et al. (2007}) point out that the interaction of wild animals with their natural habitat or their environment in captive conditions can be better understood, based on morphological studies. Basic knowledge of morphological characteristics in wild species might be crucial for conservation strategies. Therefore, the anatomy of wild animals, including the skull, needs to be studied.

The skull is a part of the axial skeleton, forming the head skeleton and enclosing the brain, sensory organs and digestive and respiratory structures. It consists of two parts: the neurocranium and the viscerocranium. The neurocranium surrounds and protects the brain and is composed of flat and irregular bones. The viscerocranium comprises the facial bones, consisting of irregular bones and the mandibula (^{Nickel et al., 1986}).

^{Machado (2006}) states that the shape of the skull is one of the most important criterions for establishing racial patterns and evolution models for species. Some works on comparative osteometry in mammalian species with focus on phylogeny and evolution have been published by ^{Christiansen (1999}), ^{Lyras & Van Der Geer (2003}), ^{Ferguson & Larivière (2004}) and ^{Goswami (2006}).

Domestic cats previously have been the object of analyses in order to establish craniometric patterns (^{Sarma et al., 2002}) and domestic dogs have been studied even more profoundly (Machado; ^{Evans & De Lahunta, 2013}; ^{Shimming & Pinto e Silva, 2013}). The skull type in domestic dogs is classified employing the skull index, which allows to distinguish three morphological categories: brachycephalic, mesaticephalic and dolichocephalic skulls (Evans & De Lahunta).

Literature on skull morphology in wild animals is rare. ^{Christiansen & Adolfssen (2006}) studied the skulls of 56 wild carnivores, including felids, canids and ursids. Nevertheless, the authors focussed on aspects related to the mastigatory force and did not concentrate on craniometric patterns. Furthermore, ^{Santos et al. (2017}) realized a study with seven maned wolves (Chrysocyon brachyurus), verifying whether the craniometric data collected allows a classification of the skull type after ^{Getty (1986}). According to ^{Hofmann-Appollo (2009}), various wild canid species present an intermediate skull type, being classified as mesaticephalic and therefore resembling the domestic dogs.

The objective of this study was to realize craniometric measurements in C. thous , patronize the craniometric values for this species and compare these characteristics to the craniometric patterns in domestic dogs in order to conclude whether the classification of skull types (dolicho-, mesati- and brachycephalic) is appropriate for C. thous as well.

MATERIAL AND METHOD

A total of 11 skulls of C. thous were used in this study, irrespective of age and sex. The samples were collected from road-kill specimens with the approval of the Biodiversity Information and Authorization System (SISBIO, permit n° 49266-1) and belong to the anatomical collection of the Wild Animal Research Laboratory of the Federal University of Uberlândia. Craniometric points, parameters and indices that were applied in this study, were based on works on craniometry in domestic dogs (^{Simoens et al., 1994}; ^{Onar & Günes¸, 2003}; ^{Janeczek et al., 2008}; ^{Shimming & Pinto e Silva, 2013}) and are described in Tables I to III and illustrated in Figures 1 to 5.

All measurements were performed with a manual calliper Starrett® 125 MEB in total three times, on three different occasions by the same researcher in order to minimize the risk of measurement errors. The data were organized in spreadsheets and statistical analysis was performed (mean and standard deviation) using the software BioEstat 5.3®. The classification of the skull type of C. thous followed the definition of Evans & De Lahunta for domestic dogs.

Table I Description of nine craniometric points. 

Table II Description of 12 craniometric parameters: linear measurements between craniometric points of the skull (from ‘a’ to ‘m’) and of the occipital area (from ‘n’ to ‘r’). 

Table III Description of six craniometric indices: ratios between craniometric parameters. 

Fig. 1 Photograph of the skull of C. thous, dorsal view. BR (Bregma), E (Eurion), I (Inion), N (Nasion), PR (Prosthion), R (Rhinion), Z (Zygion), b (Length of the skull), c (Length of the neurocranium), d (Length of the viscerocranium), e (Width of the neurocranium), f (Zygomatic width) and j. 

Fig. 2 Photograph of the skull of C. thous, lateral view. B (Basion), I (Inion), N (Nasion), PR (Prosthion), b (Length of the skull), c (Length of the neurocranium), d (Length of the viscerocranium) and g (Length of the base of the skull). Scale: 3 cm. 

Fig. 3 Photograph of the skull of C. thous, ventral view. B (Basion), PR (Prosthion), g (Length of the base of the skull), h (Condylobasal length) and m (Palatal length). Scale 3 cm. 

Fig. 4 Photograph of the skull of C. thous, caudal view. B (Basion), I (Inion), a (Height of the skull), n (Width between jugular processes), o (Width between occipital condyles), p (Width of the foramen magnum), q (Height of the occipital triangle) and r (Height of the foramen magnum). Scale: 3 cm. 

Fig. 5 Photograph of the mandible of C. thous, lateral view. P (Pogonion) and l (Length of the mandible). Scale: 3 cm. 

RESULTS AND DISCUSSION

Tables IV and V show the results (mean ± standard deviation) of parameters measured in 11 skulls.

The length of the skull of C. thous in this study was 131,66 ± 14,86 mm, while Santos et al. present a skull length of 126.76 ± 5.29 mm in their analysis of skulls of seven C. brachyurus, maned wolves, although C. thous is a considerably smaller canid (Cheida et al.). In an analysis of 25 skulls of domestic mesaticephalic dogs of unknown breed, ^{Shimming & Pinto e Silva (2013}) mention a skull length of 177.72 ± 16.22 mm. Machado found a skull length of 135.19 ± 38.61 mm in eight mesaticephalic dogs of different breeds and 154.56 ± 21.22 mm in eight brachycephalic boxers.

Table IV Craniometric measurements of C. thous, in millimetres (mm). 

Table V Craniometric indices in C. thous, in millimetres (mm). 

The specimens of this study presented values similar to what Machado found in mesaticephalic domestic dogs. Dogs with the mesaticephalic skull type generally present a scissors bite occlusion, meaning that the maxillary incisors are located slightly rostral to the inferior incisors, allowing the dog a powerfull bite to tear its prey apart and for defence (Evans & De Lahunta). According to ^{Dyce et al. (1997}) brachycephalic breeds present a predisposition for a shorter upper jaw, maxilla, and dolichocephalic breeds have a higher incidence of a longer than normal upper jaw (both cases resulting in problems when catching prey).

In C. thous the zygomatic width was 71.10 ± 7.36 mm, whereas Santos et al. observed 121.87 ± 6.30 mm in C. brachyurus. In mesaticephalic dogs ^{Shimming & Pinto e Silva (2013}) report 95.44 ± 7.85 mm and Machado measured 106.89 ± 28.11 mm zygomatic width. In contrast, in brachycephalic dogs 139.68 ± 9.63 mm were observed (Machado).

According to Evans & De Lahunta the skull index is 39 mm in dolichocephalic dogs, 52 mm in mesaticephalic and 81 mm in brachycephalic dogs. The skull index for C. thous , observed in this study, was 53.17 ± 2.54 mm, resembling the skull index of mesaticephalic dogs as described by Evans & De Lahunta. Similar values were also observed in the German Shepherd breed by ^{Onar (1999}) and in domestic dogs of unknown breed by Shimming & Pinto e Silva %, describing a skull index of 51.44 mm and 53.83 ± 3.36 mm, respectively.

On the other hand, Santos et al. present a skull index of 96.17 ± 3.89 mm for C. brachyurus. This raises some questions, since C. brachyurus and C. thous belong to the same subtribe Cerdocyonina. However, probably the discrepancy between the indices of the two species is a result of methodological errors in the work of Santos et al., as their calculations might be based on wrong measurements, as they report a cranial length for the maned wolf inferior to what is described in this study for C. thous , even though C. thous is the smaller species. The results of this study were not compared with the work of Machado, since the cited work employs only the craniofacial index for defining the type of skull.

CONCLUSION

The craniometric points described in the crab-eating fox, were similar to what is established for the domestic dog and therefore the craniometric indices were calculated in the same way as is common for the domestic dog. This study shows that the craniometric characteristics of C. thous follows what is described for the domestic dog and its skull type can be classified as mesaticephalic.