Journal of Mammalogy, 82(3):799±813, 2001 ALLOZYME VARIATION AND SYSTEMATICS OF THE GENUS APODEMUS (RODENTIA: MURIDAE) IN ASIA MINOR AND IRAN MILOSÏ MACHOLA N,* MARIA GRAZIA FILIPPUCCI,PETR BENDA,DANIEL FRYNTA, AND JOVANA SA DLOVA Institute of Animal Physiology and Genetics, Academy of Sciences of the Czech Republic, VeverÏõ 97, 60200 Brno, Czech Republic (MM) Downloaded from https://academic.oup.com/jmammal/article/82/3/799/2372794 by guest on 30 September 2021 Department of Biology, University of Rome II, ``Tor Vergata,'' Rome 00010, Italy (MGF) National Museum, Prague 11579, Czech Republic (PB) Department of Zoology, Charles University, Prague 12844, Czech Republic (DF, JS) Starch gel electrophoresis at 36 presumptive loci was used to study genetic variation and systematic status of 110 wood mice (genus Apodemus) from 19 sites scattered across An- atolia, Armenia, and Iran. Seventeen loci were monomorphic and ®xed for the same allele among populations, whereas 19 loci were found to be polymorphic or discriminant among samples. The following species were determined in the material: A. ¯avicollis, A. uralensis, A. hermonensis, and a taxon provisionally called Apodemus cf. hyrcanicus. The study ma- terial was compared with previously analyzed samples from western Anatolia, increasing the total material to 245 specimens from 31 localities. In general, the pattern of variation and level of genetic differentiation within and among species were comparable between western and eastern samples. Intraspeci®c genetic distances were low, ranging from 0 to 0.051, but interspeci®c distances were an order of magnitude higher. Similarly, neighbor- joining trees showed negligible differentiation between populations of individual species and no sign of intraspeci®c structuring. A. uralensis seems to prefer humid sites, whereas A. hermonensis and A. ¯avicollis also occur in drier places. Individuals referred to Apo- demus cf. hyrcanicus were limited to lowlands south of the Caspian Sea. Problems asso- ciated with the systematic relationships and taxonomy of A. falzfeini±fulvipectus±hermo- nensis±arianus and A. cf. hyrcanicus from northern Iran are brie¯y discussed. Key words: Apodemus, Asia Minor, electrophoresis, Iran, systematics, wood mice The systematics and biology of the genus Mezhzherin and Zagorodnyuk 1989; Mezh- Apodemus have received growing attention. zherin and Zykov 1991; Mezhzherin et al. During the last 25 years, efforts of numer- 1992; Vorontsov et al. 1989) and elucida- ous scientists have bene®ted from molecu- tion of the systematic identity of wood mice lar methods (especially allozyme-based), from Israel (Filippucci et al. 1989). In the which have led to major advances in our former case, A. sylvaticus uralensis was knowledge of the systematics of wood mice shown to be a distinct species, conspeci®c from Europe and some other parts of the with A. microps from Slovakia (KratochvõÂl Palearctic region. Among the most recent and Rosicky 1952), with uralensis being achievements have been a substantial re- the older synonym (Mezhzherin 1990). analysis of the genetic differentiation and Several other taxa, previously thought to be taxonomy of the entire genus within the for- subspecies of A. sylvaticus (charkovensis, mer USSR (Mezhzherin 1987, 1990, 1991, mosquensis, ciscaucasicus, and tscherga) 1996; Mezhzherin and Mikhailenko 1991; also were placed within A. uralensis (Mezh- zherin 1990; Mezhzherin and Mikhailenko * Correspondent: [email protected] 1991). Subsequently, occurrence of 4 spe- 799 800 JOURNAL OF MAMMALOGY Vol. 82, No. 3 cies was reported for Caucasus and Trans- mus from Asia Minor and Iran. Our aim caucasus: A. uralensis, A. fulvipectus, A. was to uncover the systematic identity of ponticus, and A. hyrcanicus (Mezhzherin et animals under study, to compare genetic al. 1992; Vorontsov et al. 1992). In Israel, variation and intra- and interspeci®c differ- the presence of 3 species was demonstrated: entiation of the studied samples with pre- A. mystacinus, A. ¯avicollis, and newly de- viously published data from western Turkey scribed A. hermonensis (Filippucci et al. (Filippucci et al. 1996), and to discuss some 1989). On the other hand, A. sylvaticus, pre- of the taxonomic implications. A. mystaci- Downloaded from https://academic.oup.com/jmammal/article/82/3/799/2372794 by guest on 30 September 2021 viously believed to inhabit the entire Mid- nus, although also found in the study area, dle East, has not been found in Israel (Fi- was not considered. lippucci 1992; Filippucci et al. 1989). Despite recent progress in our knowledge MATERIALS AND METHODS of the systematics and zoogeography of A total of 110 individuals of Apodemus wood wood mice, there are still geographic areas mice were collected from 19 sites scattered from which only incomplete or confused in- across Anatolia, Armenia, and Iran (Fig. 1; Ap- formation is available. Among these areas pendix I gives details on the collecting sites, are Asia Minor and Iran. Usually, occur- sample sizes, and population codes). Voucher rence of 6 species of Apodemus has been specimens were deposited as skulls and skins at credited to this region: A. mystacinus, A. the Department of Zoology, Charles University, ¯avicollis, A. sylvaticus, A. uralensis, A. Prague, Czech Republic (D. Frynta, in litt.). ponticus, and A. fulvipectus (Filippucci et Of the 19 sampling localities, 1 (Burdur) was located in western Anatolia, 1 (Seyfe) was lo- al. 1996). However, information on the tax- cated in the north-central Asian Turkey, 10 sam- onomic status and geographic distribution ples were taken from eastern Asia Minor, and of these taxa has been based mostly on the remaining 7 samples were collected from fragmentary data. Moreover, A. uralensis northern (3), southwestern (1), and southern Iran has only been reported from UludagÆ and the (3). Data were compared with and supplemented Rize region (Steiner 1978a, 1978b), where- by data published in Filippucci et al. (1996), as A. ponticus and A. fulvipectus have been who reported on 20 populations of A. sylvaticus, thought to be restricted to areas adjacent to A. ¯avicollis, A. uralensis, and A. hermonensis the Caucasus and northern Turkey, respec- from Turkish Thrace and western Anatolia. tively (Musser and Carleton 1993). Thus, the total number of sites was increased to Recently, Mezhzherin (1997), in his re- 31 and the number of animals analyzed to 245. Muscle and kidney tissue from each specimen vision of Apodemus from northern Eurasia, were snap-frozen in liquid nitrogen and pre- reported A. mystacinus, A. ponticus, A. ur- served at 2808C until processed. Homogenates alensis, and A. hyrcanicus for Asia Minor for electrophoresis were obtained from tissue and Iran. All those taxa, except hyrcanicus, crushed in distilled water immediately before have been considered to be distributed electrophoresis. throughout Asian Turkey and, in the case of Standard horizontal starch gel electrophoresis uralensis, as far south as Israel. At the same (Harris and Hopkinson 1976) was carried out on time, results from a genetic and morpholog- the homogenates, and 24 enzymes coded for by ic survey of mice from western Anatolia 36 presumptive loci were analyzed (Appendix (Filippucci et al. 1996) demonstrated that II). Details of the electrophoretic procedures the distribution of A. sylvaticus seemed to were described by Filippucci et al. (1988). Sam- ples of A. uralensis, A. hermonensis, A. ¯avi- be limited to a small area in a northwestern collis, and A. sylvaticus with known genotypes Asia Minor on the southern coast of the were used as standards and run alongside un- Black Sea. A. ¯avicollis, A. uralensis, and known samples. Isozymes were numbered in or- A. hermonensis also were shown to be pre- der of decreasing mobility from the most anodal sent in the area studied. one. Allozymes were numbered according to We present new genetic data for Apode- their mobility, relative to the most common al- August 2001 MACHOLA N ET AL.ÐALLOZYME VARIATION IN APODEMUS 801 1% (P1%) and 5% (P5%) criteria, mean observed heterozygosity (Ho), and mean ``unbiased'' ex- pected heterozygosity (He) based on conditional expectations (Nei 1978). The amount of genetic divergence between populations was estimated with Nei's (1978) un- biased genetic distance and the chord distance introduced by Cavalli-Sforza and Edwards (1967). Our rationale for using 2 indices was Downloaded from https://academic.oup.com/jmammal/article/82/3/799/2372794 by guest on 30 September 2021 because they are based on different evolutionary models. Although Nei's indices of genetic dis- tance have been the most frequently used mea- sure used in electrophoretic studies, problems are inherent in these procedures (no matter if ``standard'' or ``unbiased'' indices are used) in cases of different levels of polymorphism among populations (cf. Hillis 1984). On the other hand, the chord and arc indices of Cavalli-Sforza and Edwards incorporate some realistic assumptions about the nature of evolutionary change in gene frequencies, thus avoiding the undesirable prop- erties of the Nei's measures (Swofford et al. 1996). However, relative superiority of respec- FIG. 1.ÐMap of Turkey (top) and Iran (bot- tive methods clearly depends on which under- tom) with collecting sites, indicated with black lying model is more evolutionarily appropriate circles (numbers are the same as in Table 1): 1, within and among populations of wood mice, as Burdur; 2, Seyfe; 3, GuÈzyurdu; 4, Sumelas; 5, well as within the genus Apodemus as a whole. Damar; 6, Kabaca; 7, YalnõzcËam Pass; 8, BagÆ- Neighbor-joining trees (Saitou and Nei 1987) dasËan; 9, Sõrbasan; 10, DogÆubayazõt: 11, Hak- were constructed from both Nei's and Cavalli- kaÃri; 12, Yerevan; 13, Gholaman; 14, Baba Has- Sforza's indices; support for each node was ex- an; 15, Abshar; 16, Sivand; 17, Vali Abad; 18, amined with the bootstrap procedure (Felsen- Asalem; 19, Now Kandeh (see Appendix I for stein 1985) using 1,000 pseudoreplications. more details). Localities indicated with black PHYLIP, version 3.5c (Felsenstein 1995) was squares are those studied by Filippucci et al.