The adder (Vipera berus) in Southern Altay Mountains: population characteristics, distribution, morphology and phylogenetic position

Ethics statement

This study was reviewed and approved by the Ethical Committee of the Institute of Zoology, Chinese Academy of Sciences (Permit Number IOZ14060). All procedures performed in this study were in accordance with the instructions and permission of the Ethical Committee and the Chinese Wildlife Management Authority.

Study area and field surveys

The study was conducted in the Southern Altay Mountains in the region bordering Russia, Mongolia and Kazakhstan. The elevation ranges from approximately 1,000 m in river valleys to 4,374 m at the summit. The study area has a typical temperate continental climate with short, cool summers and long, cold winters. The forest type is the boreal coniferous forests and the main tree species include Siberian spruce Picea obovata, Siberian larch Larix sibirica, Siberian pine Pinus sibirica, Siberian fir Abies sibirica, European aspen Populus tremula, and birch Betula pendula. The natural vegetation on the mountains varies with different altitude and aspect. The following vegetation belts can be recognized: coniferous forest, broad-leaved forest, mixed forest, shrubby steppe, subalpine shrub and meadow, alpine meadow, and nival belt.

We chose the Kanas river valley as the main survey area while we conducted a complementary research in Daxiaodonggou Forested Area (Fig. 1). The Kanas river valley in the northwest of Southern Altay Mountains is relatively close to the distribution range of the adder; the valley is strictly protected by three protection stations within the core area of Kanas National Nature Reserve. Daxiaodonggou Forested Area is located in the northwest-central part of the mountains, about 130 km from the main survey site. A combination of 20 line transects and random searches were conducted to investigate the abundance and distribution of adders in the two survey areas (Fig. 1). In Kanas river valley, we set up 6 line transects around each protection station (a total of 18 transects), which were spaced 1–2 km apart to assist spatial independence of snake observations. Transects began at the east of Kanas Lake and ended on the edge of Bai Lake. Each transect began where the forest or steppe edge met the river, aiming to representatively sample all habitats and exposures. The length and direction of transects were varied due to safety hazards from the terrain or vegetation (see Data S2). To investigate whether the adder was distributed in Daxiaodonggou Forested Area, we arbitrarily established 2 line transects to extend between the valley and hillside. Opportunistic sightings (e.g., travelling between protection stations along the riding track) were also recorded during fieldwork.

Except during rainy or foggy weather, we surveyed each transect at least twice between 10:00 and 19:00 h during a 170-day period (21 June–17 September 2014, 21 June–9 September 2015, see Data S2) by walking slowly (ca. 1–1.5 km/h). We searched mainly in open ground and under vegetation for adders. On all 18 line transects in Kanas river valley, we identified and recorded adders for the forward journeys only, to avoid counting the same individuals twice. At least 10 days were allowed between successive surveys on the same transect.

To standardize sampling, we used a simple encounter rate (number of sightings per km walked ± SE) as a proxy for relative species abundance. We could not distinguish black individuals between sexes based on the sexual dichromatism in the dorsal pattern due to melanism, so we did not determine the sex ratio of the targeted adder population in this study. We collected 9 specimens (3 males and 6 females) in Kanas river valley. We captured one adder from Daxiaodonggou Forested Area to determine the occurrence of the adder, and then released it at the point of capture. All the collected specimens were preserved in 95 or 100% ethanol and deposited in the herpetological collections of the Institute of Zoology, Chinese Academy of Sciences, Beijing, China. Collection information and voucher numbers are shown in Data S3.

Morphological description and comparison

For the collected specimens, we described the morphological features in terms of body dimensions, scalation and color pattern. The measurements of eight characteristics traditionally used in Vipera systematics were recorded: apicals (refers to the apical scales, similarly hereinafter for the other different scales), supralabials, sublabials, circumoculars, crowns, loreals, ventrals, and subcaudals.

We also extracted the corresponding morphological data of related taxa from the available literature: V. b. berus from Northwestern, Southwestern and Eastern Europe (Nilson, Andrén & Szyndlar, 1994; Milto & Zinenko, 2005); V. b. sachalinensis from Northern Korea and Sakhalin (Nilson, Andrén & Szyndlar, 1994); V. b. bosniensis from Balkan (Nilson, Andrén & Szyndlar, 1994); V. altaica from Eastern Kazakhstan (Tuniyev, Nilson & Andrén, 2010); V. renardi from the eastern lowland of Central Asia (Nilson & Andrén, 2001; Tuniyev, Nilson & Andrén, 2010). Considering the sexual dimorphism present in pholidosis characteristics, both sexes were studied separately for ventrals and subcaudals.

We tested whether there were significant differences in the number of scales among different taxa from different geographical regions using a one-way analysis of variance (ANOVA), and we also performed a post-hoc Student–Newman–Keuls (SNK) multiple comparison tests to compare selected pairs of data. If data did not meet the criteria for homogeneity of variance, two-tailed unequal variance t-test followed by Bonferroni correction was performed (Ruxton, 2006). All statistical approaches were completed using R v3.1.2 (R Core Team, 2013). P < 0.05 was regarded as statistically significant. However, because of the small sample size, we did not conduct statistical analyses on the ventrals and subcaudals from different sexes.

Molecular phylogenetic analyses

Genomic DNA was extracted from tissue samples of each specimen using TIANamp Genomic DNA Kit (Tiangen Biotech, Beijing, China), following the manufacturer’s instructions. We amplified and sequenced partial mitochondrial cytochrome b (Cyt b) gene with the primers (Ursenbacher et al., 2006):

L14724Vb (5′-GATCTGAAAAACCACCGTTG-3′),

H15548Vb (5′-AATAGAAAGTATCATTCTGGTTTAAT-3′),

L15162Vb (5′-CTCCCATGAGGACAAATATC-3′),

and H15914Vb (5′-CCAGCTTTGGTTTACAAGAAC-3′).

Polymerase chain reaction (PCR) was performed in a 40 µL final volume with 4 µL DNA template, 20 µL of Ex Taq polymerase (RR001A, TaKaRa), 4 µL (20 mmol/L) of each of the two primers and 8 µL of double distilled water. PCR programs to amplify the two parts of Cyt b started with a cycle of denaturing at 94 °C for 10 min, followed by 35 cycle of 94 °C for 45 s, 50 °C for 60 s, and 72 °C for 90 s, and a final extension at 72 °C for 10 min. PCR products were sequenced by Sino Geno Max (Beijing, China), using ABI 3730XL sequencers. Sequences were deposited in GenBank (accession numbers KU942378, KX345249, KX345250, KX345251, KX345252, KX345253, KX345254, KX345255 and KX345256).

Cyt b sequences were visualized with MEGA 6.0 (Tamura et al., 2013), aligned with the ClustalW option included in this software and double-checked by eyes (Larkin et al., 2007). The number of haplotypes and polymorphic sites was obtained with DnaSP v5.0 (Librado & Rozas, 2009). Only one haplotype was recognized in all 9 samples and compared to other 33 sequences downloaded from GenBank. We used V. seoanei, V. altaica, and V. renardi as outgroups (see Table S1). Sequence differences were calculated using MEGA 6.0 based on Kimura’s two-parameter model (Tamura et al., 2013). The best-fit model of nucleotide substitution was calculated by the Akaike Information Criterion (AIC) approach in Modeltest v3.7 (Posada & Crandall, 1998). Then, we conducted the Bayesian inference (BI) of phylogeny with MrBayes v3.2 (Ronquist & Huelsenbeck, 2003) using the best-fitted model. Four chains were run simultaneously and each Markov chain was started from a random tree and run for 1 × 10⁷ cycles, with sampling every 1,000th cycle. The convergence was estimated by measuring the standard deviation of the split frequency among parallel chains. Chains were considered to have converged when the average split frequency was lower than 0.01. The initial 25% trees were discarded as burn-in values.

Field surveys

In the two survey sites, we recorded a total of 34 sightings of adders, of which 31 were from Kanas river valley (17 and 14 in 2014 and 2015, respectively) and 3 from Daxiaodonggou Forested Area. There were 29 sightings of subadults and adults (henceforth adults), 1 juvenile trampled by a horse with body length of 270 mm, 1 newborn in Kanas valley, and 3 adults in Daxiaodonggou Forested Area. However, we did not mark the encountered adders in the field survey, so the same individual was likely counted multiple times. In Kanas river valley, a total of 105 km of line transects was surveyed and the overall encounter rate was 0.15 ± 0.05 sightings/km. Melanistic adders were relatively rare with only 17% (n = 5) of the total number of adults. The adders were found mostly in coniferous forest (18 out of 34 sightings, 53%) at elevations ranging from 1,200 m to 2,300 m (see Data S1). There was only one snake species (i.e., V. berus) observed in Kanas valley which is about 80 km from the nearest population of V. renardi (Fig. 1), but Elaphe dione and Coluber spinalis occurred in Daxiaodonggou Forested Area.

Morphological characteristics

The largest specimen (voucher number IOZ015018) examined by us was a female measuring 510 mm in total length and 54 mm from the posterior margin of cloacal opening to the tip of tail (tail length); the latter was equal to 10.6% of body length. The largest male (IOZ015020) had a total length of 490 mm, including 68 mm tail length (13.9% of total length). The males averaged at a total length of 449.3 ± 21.5 mm (n = 3) while the mean length of the females was 487.8 ± 6.7 mm (n = 5). Some morphological characteristics were identical in all collected specimens: 2 large supraoculars and 1 large frontal plate on top of head, parietals divided, 1 canthal and 1 supranasal scale on each canthus rostralis, 1 scale between nasal and rostral on each side, 2 apicals in contact with rostral, and a single row of scales between supralabials and eye. However, individual variation was evident in some scalation characters as follows: 4 to 10 intercanthals, 5 to 14 intersupraoculars, 2 to 5 loreals on each side, normally 9 to 10 supralabials and 9 to 11 sublabials on each side, and 1 to 3 preventrals. The eyes were surrounded by 18 to 25 circumoculars; the rostral bordered by 2 supralabials. 144 to 150 ventrals and 31 to 38 subcaudals in females, 140 to 146 ventrals and 37 to 40 subcaudals in males, all subcaudals in double rows, 20 to 21 rows of dorsal scales around midbody. Detailed morphological data of each specimen with the corresponding voucher number can be found in Data S3.

Dorsal pattern consisted of a continuous winding zig-zag band in females and males. Lateral body blotches were present except in melanistic males, ventral side black, the tip of tail pigmented by yellow or orange. Preoculars and frontal were separated from nasal and supraoculars on each side, respectively. Females, whether or not they were melanistic, had light dots on ventrals and reddish-brown colored throat. Non-melanistic specimens had two dark oblique bands on the dorsal head and light spots on some supralabials, apicals, supranasals, and supraoculars.

Sexual dichromatism in the dorsal pattern was well expressed. All non-melanistic females had a dark brown zig-zag band on light brown background. Compared with that, males showed a much higher color contrast in dorsal pigmentation, usually with black zig-zag turns and greyish ground color.

Morphological comparison

The mean values and standard errors of morphological characteristics, which were calculated for 9 specimens in this study and extracted from literature for the samples from other regions, are shown in Fig. 2 and Table 1. The specimens examined by us were distinguishable from V. altaica and V. renardi in the character of apicals, crowns, and loreals (all P < 0.05). At the subspecies level, the results demonstrated geographical variation in morphological characteristics. Taken as a whole, the adders from Southern Altay Mountains exhibited a tendency toward increased number of apicals. They significantly differed from the sachalinensis subspecies in apicals, sublabials and crowns, and bosniensis in sublabials and loreals (all P < 0.05). When compared with other specific populations of the nominate subspecies from different regions, our materials only showed a difference with Northwestern Europe and Southwestern Europe populations in supralabials and crowns, respectively; they had no difference with adders from Eastern Europe. In addition, the maximal number of crown shields of head and minimal number of loreals and sublabials were recorded in our specimens.

Males and females expressed sexual dimorphism in the number of ventrals and subcaudals (Table 1). The adder from Southern Altay Mountains was similar to V. b. berus populations from Northwestern Europe and Eastern Europe and V. b. bosniensis population from Balkan in both female and male values; they showed much increased number of subcaudals compared with the nominate subspecies in Southwestern Europe and sachalinensis subspecies. Our samples also differed from V. b. sachalinensis in female’s ventrals. V. altaica resembled our specimens in ventrals (the average number is 145.9 in V. altaica, Tuniyev, Nilson & Andrén (2010) pooled both sexes when analyzing morphological data), but it had a much lower average number of subcaudals, especially in females.

Molecular phylogenetic analyses

The alignment of the mtDNA Cyt b (1,023 bp) of the adders contained 119 variable sites and 61 parsimony informative sites. All samples in our study shared the same haplotype, which was identical to that from Perm in Russia (see Data S2).

The Bayesian phylogenetic tree clearly separated three major lineages of the adder, the Italian clade, Balkan clade and Northern clade, each receiving high statistical support (Fig. 3). The three currently accepted subspecies did not represent distinct genetic lineages; only the Balkan clade conformed to the bosniensis subspecies. The majority of haplotypes of V. b. berus grouped within the Northern clade while the haplotypes of the populations from Italy and adjacent areas were clustered together within the Italian clade. The adder from Southern Altay Mountains was grouped into the Northern clade and clearly separated from the other two clades. Moreover, the BI reconstruction was not able to separate the sachalinensis subspecies from the nominate subspecies.

The Kimura two-parameter distances among all Vipera berus were displayed in Table 2. The genetic distance between the Southern Altay group and the Northern clade was 0.003 ± 0.001, which was lower than any other pairwise distances. However, a low degree of genetic differentiation (0.005 ± 0.002) was also observed between our specimens and the sachalinensis subspecies.