Sibling mice species of the genus Sylvaemus Ognev, 1924 (Mammalia, Rodentia) in the Ukrainian Carpathians

У регіоні Українських Карпат трапляються три види роду Sylvaemus: мишак жовтогрудий (S. tauricus), мишак лісовий (S. sylvaticus) та мишак уральський (S. uralensis). Питання таксономії роду та ідентифікації видів у деяких частинах ареалу залишаються дискусійними, а у регіоні Українських Карпат взагалі є недостатньо вивченими. Нами досліджено понад 250 зразків мишей роду Sylvaemus, здобутих у регіоні Українських Карпат, з них морфометрично проаналізовано 216 зразків за 4 екстер’єрними та 11 краніальними ознаками. Показано, що за лінійними розмірами тіла лише S. uralensis ідентифікується з високою вірогідністю. Для пари видів tauricus–sylvaticus найменш мінливою серед лінійних розмірів тіла є довжина задньої стопи, що можна розглядати діагностичною ознакою, хоча для неї також характерне деяке перекривання значень. Для правильної ідентифікації видів необхідно застосувати краніометричні ознаки. Серед вивчених нами 11 краніометричних ознак найменш мінливими для пари tauricus–sylvaticus є довжина верхнього ряду молярів (M13), ширина (CRB) та висота (CRH) мозкової капсули. Змішані вибірки дорослих особин трьох видів можна розділити з мінімальним або практично без перекриття значень за відношенням довжини верхнього ряду молярів до ширини мозкової капсули, кондилобазальної довжини черепа та довжини слухового барабану. Аналіз ступеню перекривання значень окремих ознак у дорослих особин показав, що пара видів S. tauricus–S. sylvaticus найбільше розходяться за такими ознаками, як довжина верхнього ряду молярів (M13), ширина (CRB) та висота (CRH) мозкової капсули, кондилобазальна довжина черепа (CBL) і довжина слухового барабану (BUL). Подібна тенденція була виявлена і для пари S. sylvaticus–S. uralensis. Розроблено регіональний діагностичний ключ для ідентифікації дорослих особин, який дозволяє вірогідно ідентифікувати 93,5% зразків. Завдяки перевизначенню зразків показано, що S. tauricus має найбільш широку висотну та біотопну преференції, S. sylvaticus трапляється переважно у вологих заплавних біотопах (ліси, чагарники), майже виключно уздовж річкових долин, по яким проникає далеко в гори, а S. uralensis представлений нечисленними знахідками з рівнинних заплавних біотопів.


Zoltán Barkaszi
Three species of the genus Sylvaemus occur in the region of the Ukrainian Carpathians: the yellow-necked field mouse (S. tauricus), the long-tailed field mouse (S. sylvaticus) and the pygmy field mouse (S. uralensis). Issues of the genus's taxonomy and identification of species have remained controversial in some parts of the geographic range, while in the Ukrainian Carpathians they have been studied scarcely at all. We studied about 250 mice specimens from the Ukrainian Carpathians belonging to the genus Sylvaemus, among which 216 were analysed morphometrically based on 4 external and 11 cranial characters. Results indicate that by linear body dimension only S. uralensis can be differentiated with high probability. For the pair of species tauricussylvaticus, the hind foot length is the least variable among linear body characters, which might be considered diagnostic, although values of this character also tend to overlap. To identify species correctly, it is necessary to use craniometrical characters. For the pair of tauricus-sylvaticus, the least variable among the 11 studied characters are the upper molars length (M13), braincase width (CRB), and braincase height (CRH). Mixed samples of adult specimens of the three species can be differentiated with minimal or practically no overlap by using the relation of the upper molars length to braincase width, condylobasal length, and auditory bulla length. Analysis of characters' uniformity in adult specimens showed that S. tauricus and S. sylvaticus differ from one another the most by the upper molars length (M13), braincase width (CRB), braincase height (CRH), condylobasal length (CBL), and auditory bulla length (BUL). A similar tendency was revealed for the pair of S. sylvaticus and S. uralensis. A regional identification key was developed for differentiation of adult mice, which allows identifying reliably 93.5% of specimens. Results of the revision of samples suggest that S. tauricus has the widest altitudinal and habitat preferences, S. sylvaticus occurs mainly in humid floodplain biotopes (shrubs,

Introduction
Field mice of the genus Sylvaemus are common species in the mammal fauna of Europe. The species richness of the genus increases in Europe eastward: in Western Europe sympatrically occur S. tauricus and S. sylvaticus, in Central Europe there is a third species -S. uralensis, while in Eastern Europe a fourth one -S. witherbyi -appears (Orlov et al., 1996). Accordingly, in the region of the Ukrainian Carpathians three species co-occur (Barkaszi, Zagorodniuk, 2016), namely the yellow-necked field mouse S. tauricus, the long-tailed field mouse S. sylvaticus, and the pygmy field mouse S. uralensis. Moreover, the Carpathian region represents the south-western range edge of the latter (Kryštufek et al., 2008). Due to the exceptionally high morphological similarity between the three species, their differentiation and thus the genus's taxonomy is rather a complex issue, which in many parts of the geographic range have yet remained debatable (Chelomina et al., 2007). Zimmermann (1962) first proposed to divide the Palearctic genus Apodemus into three subgenera: Apodemus, Sylvaemus, and Alsomys. Among them, Apodemus is sympatric with the two other subgenera, but it co-occurs in Eastern Europe only with Sylvaemus.
Morphological similarity between the field mouse species is related to the specifics of their evolution. Earlier it was suggested that the common ancestor of S. tauricus and S. sylvaticus appeared in Europe in the late Pliocene, most likely from eastern regions, and diverged rapidly due to allopatric speciation (Michaux et al., 2003). During the Quaternary, S. sylvaticus survived the glaciations in the Iberian Peninsula, wherefrom it recolonized almost the whole rest of Europe in the end of the last glaciation, while S. tauricus recolonized Europe, including northern Spain, during the Holocence from an Italo-Balkan refugium, where in this time S. sylvaticus suffered a serious "bottleneck effect" (Michaux et al., 2005). Further studies showed that S. tauricus likely survived the last glacial maximum in at least two refugia
Series "Biology", issue 31, 2018 located on different banks of the Danube (Bugarski-Stanojević et al., 2008), while S. sylvaticus also could recolonize Europe from a second (northern) refugium as well possibly located in the Dordogne or Carpathian region (Herman et al., 2017). The third species, S. uralensis, is represented by two races, a European and an Asian, respectively (Chelomina et al., 2007). Paleontological data confirm the results of molecular phylogeographic research, although indicate sufficient differences between the species by the dynamics of colonization of their range (Knitlová, Horáček, 2017).
The existing phenotypic similarity between field mouse species is presumably the result of a long evolutionary process under similar ecological conditions (Jojić et al., 2014). Dzeverin and Lashkova (2012) estimating the tempo of divergence in species of the genus Sylvaemus concluded that the group is in the state of an evolutionary stasis and its divergence, which occurred by both the size and form of the skull, has been slowed down by stabilizing selection.
According to Zagorodniuk and Kavun (2000), the increase of general dimensions, i.e. phylogenetic growth, can be considered the main vector of changes in the evolutionary line of Sylvaemus. Therefore, age-related variation and fixation of different stages of ontogenetic development in adults are the basis for the emergence of differences between closely related species. The growth of the group in general is repeated in the postnatal ontogenesis of the large species meaning that on early stages of ontogenesis such species will represent an ontogenetic equivalent (morphological copy) of older age stages of its smaller sibling.
Views of researchers of the fauna of Ukraine and of the Ukrainian Carpathians on the identification of Sylvaemus species based on morphological characters -both external and cranial -remain controversial. Due to the absence of a clear scheme of species diagnostics, each researcher prefers using different "key characters" (e.g., presence/absence and form of the chest spot, hind foot length, tooth row length, etc.). However, such approach often led to the amassment of mixed samples and distorted conclusions on distribution, habitat preferences and other ecological features of field mice in the region.
A plenty of studies were conducted into the morphological variation of field mouse species in order to shed light on the taxonomic structure of the genus Sylvaemus, and many attempts were made to develop a convenient scheme for species identification. Such research become more complicated by the fact that typical "tauricus-related" characters can appear in populations of S. sylvaticus as well due to low species divergence (Tchernov, 1979). Odontological characters, which are relatively convenient and useful in species diagnostics of voles, are rather polymorphic within the genus Sylvaemus and each variant can appear in any of the species, thus this criterion is usually unreliable (Tchernov, 1979).
Controversial or incorrect species identification of Sylvaemus specimens in the past certainly distorted the real picture of distribution of field mice in the region. We can be convinced of this if analyse the works of authors of the mid-20th century (subsequent authors often cited these works without critical analysis or revision of data). Some of them considered that S. sylvaticus has low abundance in the region (Sokur, 1952;Kolyushev, 1953;Tatarinov, 1956) and occurs only in lowland and piedmont areas (Sokur, 1952;Kolyushev, 1953). Another view, according to Shnarevich (1959), is that the long-tailed field mouse in Bukovina occurs mainly in piedmont and mountain areas. Other zoologists (Tatarinov, 1956;Turyanin, 1959) considered S. sylvaticus as one of the most common and widely distributed rodents in the region. Obviously, each of these researchers were using different diagnostic characters and approaches to species identification, which eventually led to controversial conclusions on the distribution of field mouse species in the Ukrainian Carpathians.
The distribution of S. uralensis in the region is the least explored. This species is known from practically all parts of Ukraine, yet it is considered more common for the eastern regions (Naglov, 1995;Mezhzherin et al., 2002). In the Ukrainian Carpathians, the presence of S. uralensis had been long denied, although Turyanin (1959) was practically the first who draw attention on the occurrence of "large and small forms of S. sylvaticus" in the fauna of Transcarpathia (i.e. Zakarpattia Oblast). However, the first record of S. uralensis in the Ukrainian Carpathians was reported only in 1980 (Polushina, Voznyuk, 1980), and later the existence of an isolated highland population of the species on Sheshurska polonina of the Chornohora range was reported as well (Kyselyuk, 1993). Few records of S. uralensis are known from the region of the Ukrainian Carpathians, although the pygmy field mouse is a common species in adjacent lowland regions of neighbouring countries (Cserkész, 2005;Cichocki et al., 2011;Čanády et al., 2014).
The aim of the present research was to analyse the morphological features of field mice of the Ukrainian Carpathians and to develop a regional key for species identification. Based on the revision of samples, we also aimed to clarify the distribution and habitat preferences of field mice in the region.

Material and methods
About 250 specimens of field mice of the genus Sylvaemus collected in the region of the Ukrainian Carpathians were analysed from the mammal collection of the National Museum of Natural History, NAS of Ukraine (Kyiv) and from the author's working collection. Data on linear body dimensions and measurements of cranial structures were obtained from 216 specimens, among which 164 were identified eventually as S. tauricus, 32 as S. sylvaticus, and 6 as S. uralensis. Record localities of specimens are shown on Fig. 1.   Fig. 1

. Record localities of the studied mice specimens of the genus Sylvaemus
The relative age of specimens was determined according to the scheme "juvenilis-subadultusadultus" based on a set of characters including general body and skull dimensions, skull sculpture and level of tooth crown wear (Delany, Davis, 1961;Adamczewska-Andrzejewska, 1967).
Morphological variation of specimens was studied using study skins, linear body dimensions (body length L, tail length Ca, hind foot length Pl, and auricle length Au) and 11 cranial characters. The latter were selected after an extensive survey of former publications in which these characters turned out to be the most promising for species diagnostics. The analysed cranial characters are as follows: IM3, upper tooth
The scheme of cranial measurements is shown on Fig. 2. Maps were created by QGIS software using Google, personal and public domain layers.

Variation of external morphological characters
As siblings, field mice have a highly similar external look by both colouration and details of body structure. In addition, different age groups of one species morphologically often resemble other age groups of another species (e.g., see Cserkész, 2005;Stetsula, 2012).
Practically there are no non-metric external characters, which allow reliable identification of field mice specimens. During species identification, the attention is usually paid first to the presence and outlines of the chest spot, which is, according to the "traditional" view, similar to a collar in S. tauricus or to a tie or a blurred drop in S. sylvaticus, while S. uralensis has no chest spot at all (Fig. 3). Yet in case of field mice from the Ukrainian Carpathians this character seems to be highly variable and the chest spot can be very alike in S. tauricus and S. sylvaticus (Fig. 4).
Species identification based on linear body dimensions is also problematic. Most of the external metric characters is relatively highly variable (Table 1) and in case of correct preliminary age determination body dimensions allow to reliably differentiate only specimens of S. uralensis (Fig. 5). For the pair of species tauricus-sylvaticus, the hind foot length is the least variable among linear body characters, which might be considered diagnostic, although values of this character also tend to overlap.
Thus, using only external characters for species identification in field mice leads to unreliable results hence it is necessary to involve craniometric characters as well (see further). For instance, using the hind foot length in combination with upper molars length decreases the overlap of values and increases the diagnostic weight of the hind foot length (Fig. 6).

Variation of craniometric characters
Among the 11 studied craniometric characters, the least variable for the pair of species tauricussylvaticus are the upper molars length (M13), braincase width (CRB), braincase height (CRH) ( Table 2). In
Series "Biology", issue 31, 2018 case of S. uralensis, comparison of coefficients of variation with those in other species would be incorrect due to the small number of specimens in the pygmy field mouse sample.
Some other characters with relatively higher levels of variation also can be used for species identification as additional criteria, and their diagnostic value is higher when used in combination with the least variable characters. In particular, mixed samples of adult specimens of the three field mouse species can be discriminated with minimal or practically without overlap of values using the relation of upper molars length to braincase width (Fig. 7), condylobasal length (Fig. 8), and auditory bulla length (Fig. 9). The probability that values would overlap in the pair of tauricus-sylvaticus increases in the space of such characters as M13/NAL (Fig. 10), M13/ROH (Fig. 11) and M13/FIL (Fig. 12).
Data analysis showed that combinations of some other craniometric characters (i.e. not in comparison to M13) can also be used in species identification as additional criteria, particularly in cases when skulls are damaged, especially their diagnostically relevant structures. However, these characters show a greater overlap of values, for example, CRB/FIL and CRH/CRB in the pair of tauricus-sylvaticus, while in case of CRH/BUL, NAL/CRB and CRB/ROH extreme values overlap in all three species.  Total overlapping of values was observed when analysing the relation of nasal bones length to rostral height (NAL/ROH) and to incisive foramina length (NAL/FIL) thus such combinations of characters are unsuitable in species diagnostics.

General trends of variation and approaches to species diagnostics
Previous research into the geographical variation of species of the genus Sylvaemus showed that general body dimensions of S. sylvaticus in Europe increase westward, while in S. uralensis and S. tauricus body size increases eastward (Zagorodniuk, 1996). The geographic range of all three species overlap in Central Europe, thus in the Carpathian region we can observe an overlap in the species' morphological
Series "Biology", issue 31, 2018 characters as well, which complicates species identification. Considering the relatively sufficient variation of field mice even on small geographic distances, the development of regional identification keys is extremely important for reliable species diagnostics.
According to our results, external (both metric and non-metric) characters allow to discriminate reliably only S. uralensis from the other two species, while craniometric characters with the lowest coefficients of variation can be used for identification of adult specimens of all three species. Analysis of characters uniformity in adult specimens showed that S. tauricus and S. sylvaticus differ from one another the most in upper molars length (M13), braincase width (CRB), braincase height (CRH), condylobasal length (CBL), and auditory bulla length (BUL). A similar tendency was revealed for the pair of species S. sylvaticus-S. uralensis, although we draw attention again to the small number of S. uralensis specimens in the studied sample (Table 3). Besides, there is also a tendency that characters of length are reliable for diagnostics more often than characters of width and height. Considering all of the revealed differences between the field mouse species, we propose a regional diagnostic key for their identification. Using this key, we could reliably identify 202 of the 216 examined in details mice, i.e. 93.5% of specimens. In particular, 12 of 164 specimens (7.3%) of S. tauricus were reidentified from the sample of S. sylvaticus, 5 of 32 specimens (15.6%) of S. sylvaticus were re-identified from the sample of S. tauricus, while all 6 specimens (100%) of S. uralensis were re-identified from the sample of S. sylvaticus. The absence of field mice originally identified as S. uralensis in the collections is might be related to the fact that the species status of the pygmy field mouse had not been long accepted in the region, and detailed revision of the Carpathian field mice samples was not conducted before.
Problematic specimens, which we could not reliably identify, were those with highly damaged skulls in which, respectively, we could not examine diagnostically important characters. Thus, the proposed scheme works the best for specimens with the fullest set of characters, especially craniometric characters.
In case of identification of damaged and problematic specimens, when some of the characters proposed in the key are unavailable, and in order to increase the reliability of identification it is worth considering the studied specimen in a space of two characters, particularly of those shown on Figs 7-12.

Distribution of field mice in the region
Analysis of specimens from personal and museum collections showed that the general distribution (Fig. 13) and habitat preferences of field mice in the region of the Ukrainian Carpathians do not differ substantially from those in other parts of the geographic range (Marsh, Harris, 2000;Kuncová, Frynta, 2009), including Ukraine (Mezhzherin et al., 2002;Hoofer et al., 2007).

Fig. 13. Distribution of three species of Sylvaemus in the region of the Ukrainian Carpathian based on re-identification of specimens
Data suggest that S. tauricus has the widest altitudinal and habitat preferences among field mice of the studied region and occurs from the lowlands up to poloninas (i.e. subalpine meadows). The yellownecked field mouse is a common and abundant species of forest habitats including deciduous, coniferous, and mixed forests, but also occurring in shrubs, clear cuttings, timberline habitats, etc. On the contrary, S. sylvaticus mainly occurs in humid floodplain habitats entering far into the mountains along river valleys. The pygmy field mouse, S. uralensis, is represented by few records from lowland floodplain habitats (banks of the Tisza river), and its occurrence in highland biotopes of the Ukrainian Carpathians (Kyselyuk, 1993), in our opinion, requires a revision.