Research Article |
Corresponding author: Stoyan Stoyanov ( stoyans@abv.bg ) Academic editor: Jesus Maldonado
© 2020 Stoyan Stoyanov.
This is an open access article distributed under the terms of the Creative Commons Attribution License (CC BY 4.0), which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited.
Citation:
Stoyanov S (2020) Cranial variability and differentiation among golden jackals (Canis aureus) in Europe, Asia Minor and Africa. ZooKeys 917: 141-164. https://doi.org/10.3897/zookeys.917.39449
|
Golden jackal (Canis aureus) expansion in the last decades has triggered research interest in Europe. However, jackal phylogeny and taxonomy are still controversial. Morphometric studies in Europe found differences between Dalmatian and the other European jackals. Recent genetic studies revealed that African and Eurasian golden jackals are distinct species. Moreover, large Canis aureus lupaster may be a cryptic subspecies of the African golden jackal. Although genetic studies suggest changes in Canis aureus taxonomy, morphological and morphometric studies are still needed. The present study proposes the first comprehensive analysis on a wide scale of golden jackal skull morphometry. Extensive morphometric data of jackal skulls from Europe (including a very large Bulgarian sample), Asia Minor, and North Africa were analysed, by applying recently developed statistical tools, to address the following questions: (i) is there geographic variation in skull size and shape among populations from Europe, Anatolia and the Caucasus?, (ii) is the jackal population from the Dalmatian coast different?, and (iii) is there a clear distinction between the Eurasian golden jackal (Canis aureus) and the African wolf (Canis lupaster sensu lato), and among populations of African wolves as well? Principal component analysis and linear discriminant analysis were applied on the standardized and log-transformed ratios of the original measurements to clearly separate specimens by shape and size. The results suggest that jackals from Europe, Anatolia and the Caucasus belong to one subspecies: Canis aureus moreotica (I. Geoffroy Saint-Hilaire, 1835), despite the differences in shape of Dalmatian specimens. The present study confirmed morphometrically that all jackals included so far in the taxon Canis aureus sensu lato may represent three taxa and supports the hypothesis that at least two different taxa (species?) of Canis occur in North Africa, indicating the need for further genetic, morphological, behavioural and ecological research to resolve the taxonomic uncertainty. The results are consistent with recent genetic and morphological studies and give further insights on golden jackal taxonomy. Understanding the species phylogeny and taxonomy is crucial for the conservation and management of the expanding golden jackal population in Europe.
Canid, C. anthus, C. aureus, C. aureus moreotica, C. lupaster, morphology, skull morphometry, taxonomy
The golden jackal (Canis aureus Linnaeus, 1758) is one of the most widely distributed canid species and is found in many areas of Europe, Asia and Africa (
Jackal expansion in the last decades has triggered research interest in Europe. Many aspects of golden jackal’s ecology, diet, population density, genetics, legal implications of range expansion and management have been studied thoroughly in Europe. However, jackal phylogeny and taxonomy are still controversial. As many as 13 subspecies of golden jackal have been distinguished historically, but taxonomic revision is needed (
Craniometric differentiation of golden jackal in Europe has been so far poorly studied. While genetic studies are increasing, recent papers on cranial morphometry are still scarce and describe local populations. Morphometric analyses of museum specimens have shown that jackals from Dalmatia appeared to be morphologically well distinct from their counterparts from the Balkan Peninsula and Africa, with the greatest similarity to the jackals from Asia Minor (
Morphometric relationships of the European golden jackals with jackals from the Asiatic part of the species’ range have not yet been determined. Moreover, none of the studies so far have analysed morphometrically jackal populations on a larger scale. Consequently, the understanding of historic development of jackal populations in Europe is lacking (
Bulgarian territory is considered the core area of golden jackal distribution in Europe with the highest population density (
I morphometrically compared a total of 285 skulls of Eurasian golden jackal (Canis aureus) from Europe and Asia Minor and African wolf (Canis lupaster sensu lato) from North Africa. Most of the jackal skulls were collected in Bulgaria. This sample included 198 jackal skulls from subadult and adult golden jackals. Juvenile specimens were defined as individuals with fully developed second dentition, but less than 10 months of age; subadults as individuals more than 10 months, when they reach sexual maturity, but less than two years of age; and adults as two years and older. I determined the age in consideration of upper incisive teeth wear (
Map of Eurasian golden jackal range (
Fourteen skull measurements, following
Skull measurements employed in the analyses (following
All measurements were tested for normality by QQ plots and the Shapiro-Wilk test. I applied multivariate analyses in order to explore the most significant variation in size and shape of the skulls. Shape in general tends to provide more reliable information than size on the morphology of organisms (
The advantages of ratios are that their computation is simple, and that one can easily interpret them in geometric terms of shape variation (
For clear separation of shape and size, the PCA was applied on the standardized (dividing each measurement by geometric mean) and log-transformed ratios of the original measurements (
Geometric interpretation of PCA and LDA was made by using graphical tools developed by
For detailed mathematical descriptions and statistical frameworks of the applied methods see
The skull samples used in this study were obtained from individuals that died in vehicle collisions, due to natural causes or as a result of legal hunting. I also measured museum specimens. No animal was killed for the purpose of this study.
Shapiro-Wilk tests and QQ plots showed that all measurements did not deviate significantly from a normal distribution. However, for most of the following statistical methods the assumption of normally distributed data is not strongly suggested. PCA revealed that there was a clear separation between the predefined taxonomic groups. Four clusters could be differentiated projecting the data along isosize and the first principal component in shape space: European golden jackals, including Anatolia and the Caucasus, Dalmatian jackals, and two groups of African wolves (in the broad sense) (Fig.
Principal component analysis. Separation between taxonomic groups along isometric size and first principal components in shape space. Ellipses show 95 % confidence interval for each group. Specimens of Canis aureus are divided in two groups – Europe and Asia Minor, and Croatia (the Dalmatian coast). A country origin of European specimens is marked with different symbols B country origin of African specimens is presented.
There are no clear differences in skull shape between the taxonomic groups revealed by the second shape principal component plotted against isosize (Fig.
Principal component analysis. The first two principal components in shape space account for 53.6 % of the variance. Ellipses show 95 % confidence interval for each group. Specimens of Canis aureus are divided in two groups – Europe and Asia Minor, and Croatia (the Dalmatian coast). A separation between taxonomic groups along isometric size and second principal component B separation between taxonomic groups along first two principal components in shape space.
The “PCA ratio spectrum” allows the interpretation of principal components in shape space (Fig.
The results from PCA suggested that we could find the best separation of groups by employing LDA. The analyses were applied twice. First, I tried to discriminate the three taxonomic groups: Eurasian golden jackals, African golden jackals and African wolves. Next, I conducted analyses by including specimens from Dalmatia as a separate group, following the results from PCA and assumptions about the differences between Dalmatian jackals and their counterparts from the Balkan Peninsula and Africa, found by morphological and genetic studies so far. The LDA showed that skulls of Canis aureus, Canis anthus s. str. and Canis lupaster s. str. could be clearly distinguished (Fig.
For practical reasons, characters that would allow quick and easy identification of most specimens might be useful, for instance in field work. One or two ratios would be preferable, as these are easily calculated and differences in proportions can sometimes even be estimated by eye (
The ratios that best separated taxon groups revealed by the LDA ratio extractor. The measure δ indicates how well shape discriminates in relation to size. A value of δ close to unity means that separation is mainly due to size, whereas a value close to zero indicates separation is mainly due to shape.
The skulls of Dalmatian jackals are relatively broader overall, with a broader braincase, larger palatal and zygomatic breadth, and a shorter condylobasal length, compared to the skulls of jackals from Europe and Asia Minor. The differences are mostly in shape, but not in size of skulls. The ratio Iob/Palb very well separates the Eurasian from the African jackals, the latter also having a slightly longer upper carnassial (P4). These two groups are almost identical in size but with a different skull shape. Canis lupaster s. str. is well separated by Canis aureus, having a bigger Iob/Rb ratio and smaller diameter of the auditory bulla (Bull), in comparison to the length of the upper carnassial (P4). The skulls of Canis lupaster s. str. are bigger and broader, with a more elongated shape. The differences are both in size and shape of skulls. The ratios Skb/Palb and Lp4/Mlm1 best separate the Canis lupaster group from the Canis anthus group.
Assessment of the LDA performance by cross validation. Number of specimens classified in each group.
Groups | Classified as: | ||
---|---|---|---|
Original | Canis aureus | Canis anthus s. str. | Canis lupaster s. str. |
Canis aureus | 236 | 1 | 3 |
Canis anthus s. str. | 3 | 14 | 2 |
Canis lupaster s. str. | 5 | 0 | 21 |
Groups | Canis aureus | Canis anthus s. str. |
---|---|---|
Canis anthus s. str. | 4.583 | – |
Canis lupaster s. str. | 3.901 | 4.104 |
Assessment of the LDA performance by cross validation. Number of specimens classified in each group.
Groups | Classified as: | |||
---|---|---|---|---|
Original | Canis aureus (Europe & Asia Minor) | Canis aureus (Croatia – Dalmatia) | Canis anthus s. str. | Canis lupaster s. str. |
Canis aureus (Europe & Asia Minor) | 216 | 3 | 1 | 2 |
Canis aureus (Croatia – Dalmatia) | 5 | 13 | 0 | 0 |
Canis anthus s. str. | 3 | 0 | 14 | 2 |
Canis lupaster s. str. | 5 | 0 | 0 | 21 |
Groups | Canis aureus (Europe & Asia Minor) | Canis aureus (Croatia – Dalmatia) | Canis anthus s. str. |
---|---|---|---|
Canis aureus (Croatia – Dalmatia) | 3.574 | – | – |
Canis anthus s. str. | 4.852 | 4.575 | – |
Canis lupaster s. str. | 3.907 | 5.074 | 4.270 |
The results suggest that there is no clear differentiation among Eurasian jackals in skull size and shape. Although the sample size of Bulgarian jackals included in the analysis is the largest analysed to date, they form a homogenous cluster, but with large individual variability. Furthermore, there were hardly any differences in skull shape between the Bulgarian jackals and the specimens from Turkey, Greece, Hungary and the Caucasus. The Bulgarian jackal skulls encompass all other specimens from Hungary, Greece, Turkey and the Caucasus on the plots, as was revealed by PCA and LDA. The amount of geographical variation among the Eurasian jackals is comparable with sex and age differences within the entire Bulgarian subpopulation. However, the golden jackal skulls from Bulgaria showed also weak differentiation in size and shape, depending on the age and sex of the animals, despite their considerable individual variability (
Only the Dalmatian jackals showed differences in shape, but not in size. Their skulls were easily separated by linear discriminant analyses and appeared to be broader and with shorter condylobasal length. Such differences were found also by other morphometric studies (
The sample included only two museum specimens from Greece with their measurements published by
All specimens from the Caucasus and Anatolia also fall into the cluster of Eurasian golden jackals and did not differ from Bulgarian, Hungarian and Greek skulls. These results were expected and confirm the hypothesis about jackal penetration in Eastern Europe from Anatolia or from the Caucasus in two ways, that correspond to the potential paths at the end of Pleistocene and Holocene: along the northern Black Sea coast and through the Bosporus (
Still, the question about differences between the Dalmatian jackals and the European population remains. It is clear, however, that there is no reason to consider these morphological differences as evidence for the existence of more than one subspecies on the Balkans and adjacent European countries. The subspecies Canis aureus ecsedensis (Kretzoi, 1947), or its synonym Canis aureus hungaricus Ehik, 1938 (
Both PCA and LDA revealed clear differences between Eurasian golden jackals and the two groups of African wolves (Canis lupaster sensu lato). The results from PCA and LDA suggested the existence of significant morphological variation within Canis lupaster (in the broad sense). The population of African wolves was separated in two very distinct clusters both in size and shape of skulls. Although, there are significant differences in size between populations of Canis lupaster, with East African individuals being smaller than North and West African ones (Viranta 2017), it seems that this is not a clinal variation, and at least two different morphotypes exist (
African golden jackals (here referred to as Canis anthus s. str.) could be easily separated from the Eurasian specimens by their skull shape and length of upper carnassial (P4). Differences in skull shape and dental morphology could be explained by their food preferences. Longer carnassial teeth are usually correlated with a more carnivorous diet (
A recent comprehensive study of African and Eurasian golden jackals, based on mitochondrial and nuclear genome sequences, has found strong support to merit the recognition of Canis anthus as a genetically distinct canid species that diverged approximately 1.3 million years ago from related grey wolves (
Multivariate analyses revealed that jackal specimens of Canis aureus sensu lato, included in this study, formed three very clearly distinct clusters in shape space: European jackals, including Anatolia and the Caucasus, African golden jackals and African wolves. There was no pronounced geographic variation in skull size and shape among the specimens from Europe and Asia Minor. These results support the opinion of
I am grateful to Stoyan Vassilev for giving me the possibility to measure all skulls from his own collection and to Nikolay Spassov who provided the access to the collection of the National Museum of Natural History, Sofia. Special thanks to Matthew Hayward and Christopher Glasby for language editing and suggesting minor improvements in the manuscript. I would like to thank both anonymous reviewers for the comprehensive reviews and the relevant comments, critiques and recommendations, which I believe significantly improved the manuscript.
Origin and catalogue number of measured skulls
Data type: list of specimens.
Explanation note: The list includes information about museum ID, source, catalogue number, collector's name (if available), species identification, origin, sample locality, and sex of all individuals, used in the analyses.