Gamebirds such as pheasants, partridge and quail (Galliformes) are keystone species in Himalayan ecosystems for their role in maintaining forest dynamics by digging and seed dispersal / predation syndromes, as well as their meat and eggs being an important protein source for numerous species. However, gamebirds are often threatened by non-sustainable harvest for food, plumage and the live bird trade, as well as destruction and degradation of the forests on which they largely depend (Fuller & Garson 2000).

Studies of Himalayan avian community structure (e.g., Elsen et al. 2017, Srinivasan et al. 2018) have been restricted to passerines rather than larger gamebirds. Although community ecology studies of gamebirds are available (e.g., Brooks et al. 2001), these have been in relatively homogenous regions such as lowland tropical forest, which contrasts sharply with heterogeneous montane environments such as the Himalayas. Nevertheless, both tropical lowland and Himalayan forests can support high species diversity (Cai et al. 2018).

Pheasant communities in the Himalayas are logistically challenging to study due to remote sites, high altitudes leading to hypoxia, and the elusive nature of pheasants in the wild. Consequently there are significant gaps in data for many species, including Blood Pheasant Ithaginis cruentus and White Eared Pheasant Crossoptilon crossoptilon (McGowan & Kirwan 2019, McGowan et al. 2019). Nonetheless, there have been some in-depth studies of these two species, among others, in these high-altitude environments (e.g., Li 1981, Lu 1986, Jia et al. 1999, 2004, 2005, Lu et al. 2006).

The objective of this study was to assess Galliform community structure and beta diversity (simply defined as similarity of species composition between regions) at two sites in the Chinese Himalayas c.125 km apart. Additionally, we describe aspects of the biology of Ithaginis cruentus and Crossoptilon crossoptilon.

Figure 1.

Map of study region showing location of Langdu and Gehuaqing, Yunnan, China.

Methods

Study area.—The study took place in two areas within the Three Parallel Rivers UNESCO World Heritage Site in Diqing Autonomous Prefecture, Yunnan, China: (1) near the Hong Shan part of the World Heritage Site, near Langdu village in Shangri-La county (28°14.87′N, 99°58.28′E); (2) in the southern part of Baimaxueshan National Nature Reserve near Gehuaqing village in Weixi county (27°35.74′N, 99°17.43′E). Hereafter these sites are referred to as Langdu and Gehuaqing, respectively (Fig. 1; Buzzard et al. 2018). These two sites were selected as each comprised a connected series of intact natural habitats over a range of elevations, believed to be representative of the two regions.

The sites presented differing conservation management regimes as Langdu was not subject to forest guard patrols, although hunting was officially illegal, whereas Gehuaqing benefitted from active conservation management, including regular patrols by forest guards. Langdu was primarily inhabited by Tibetan herders who raise mainly yaks (Bos gunniens) and yak-cow hybrids, plus some horses (Equus caballus) and mules. Gehuaqing was primarily inhabited by farmers of the Lisu Minority who mainly raise pigs and engage in apiculture, but are well known as skilled hunters.

Elevation at Langdu spanned 4,000–4,800 m, whereas at Gehuaqing habitat sampled ranged from 3,050–3,600 m. In both areas, habitat was primarily forest dominated by conifer trees including spruce (Picea spp.), firs (Abies spp.) and juniper (Juniperus spp.), as well as hardwoods including birch (Betula spp.), oaks (Quercus spp.), and rhododendrons (Rhododendron spp.). Distinct vegetation changes occurred with elevation and aspect (whether the slope is north- or south-facing). As elevation increased, trees and shrubs diminished in height and alpine meadows occurred at their upper limits, above which rock and scree were the dominant land cover.

Camera trapping .—To document presence we used heat- / motion-activated camera traps (Bushnell Trophy Cam) for >3 years, spanning October 2011–January 2014. We deployed cameras in all seasons, at 34 sites (22 at Langdu, 12 at Gehuaqing) separated from each other by at least 500 m along wildlife trails or routes likely to be used (Table 1). For subsequent pictures of the same species, we considered independent captures as those that occurred at least one hour apart (Rovero & Marshall 2009). We calculated relative abundance index values by dividing the value of independent captures by the number of trap-days where the species was confirmed as present. We estimated the age of Crossoptilon crossoptilon by sending camera-trap images to aviculturists familiar with the species' developmental growth stages, and polled them to estimate ages of offspring in the images.

Results

Beta diversity.—Five of the eight species of Galliformes recorded were pheasants, with an additional two to three species of large passerines (Table 2). Despite habitat similarity between the two sites, beta diversity showed little overlap, with only a single Galliform (Temminck's Tragopan Tragopan temminckii) shared between the two sites. Five species (Chinese Grouse Bonasa sewerzowi, Buff-throated Partridge Tetraophasis szechenyii, Tibetan Snowcock Tetraogallus tibetanus, Blood Pheasant and White Eared Pheasant) were found only at Langdu, and two (Koklass Pheasant Pucrasia macrolopha and Lady Amherst's Pheasant Chrysolophus amherstiae) were exclusive to Gehuaqing (Table 2). A sixth species, Sclater's Monal Lophophorus sclateri, was visually detected at Langdu (PJB unpubl.) but not captured by a camera-trap.

BLOOD PHEASANT Ithaginis cruentus

A total of 36 independent photographs of Blood Pheasants were taken at six different camera-trap sites representing primarily a component of Rhododendron forest, including (n = 1 unless otherwise noted) mixed Rhododendron forest (n = 2 sites), mixed Rhododendron–conifer forest, Rhododendron–oak scrub, Rhododendron forest scrub and mixed conifer–broadleaf forest. Microhabitat characteristics varied and comprised forest (sparse to dense, at times mesic) to mesic clearings (covered by moss and some sticks), understorey completely lacking to present at edges, with substrate (moss, dried leaves and twigs, pebbles and boulders). Topography was level to sloping and Ithaginis occurred at 4,028–4,407 m.

TABLE 1

Location, altitude, habitat and number of trap-days of 20 camera traps at Langdu (L) and Gehuaqing (G), Yunnan, China.

This species shows sharp elevational migration, being present only mid spring to early autumn (April–October), retreating to lower elevations in winter. While peak occurrence was in August–September, records declined dramatically to just one in October, followed by none in November–March (Fig. 2).

TABLE 2

Species presence, abundance and trap-days at Langdu (22 sites) and Gehuaqing (12 sites), Yunnan, China.

Ithaginis was active at dawn (06.00 h) until after nightfall (19.00 h). Bimodal peaks of activity were 08.00–10.00 h, with a stronger peak in late afternoon / early evening (16.00–19.00 h; Fig. 3). It was active during temperatures from -3^°C to 22^°C (mean = 7.5^°C), and snow was visible on the ground in some photos. A limited sample size (n = 7), suggests the species was not more active during any particular phase of the moon, with three photos (43% combined) during both new / near new moon and half-moon cycles, and a single photo (14%) during a full moon.

Flock size was 1–4 birds, with a mean of 1.2 (n = 36). The total ratio of adult males to females was 32:14 (2.3 males / female; per Karanth et al. 2011). The commonest social group was solitary adult males (n = 21, 58%) followed by lone adult females (n = 7, 19%), male–female (presumably bonded) pairs (n = 4, 11%), adult male ‘pairs’ (n = 2, 6%) with single records (3%) of adult female pairs and a quad of three adult males and one adult female.

WHITE EARED PHEASANT Crossoptilon crossoptilon

Fifteen independent photographs of White Eared Pheasant were taken at seven different camera-trap sites representing primarily Rhododendron or conifer forest, including (n = 1 unless otherwise noted) Rhododendron–oak scrub (n = 2), mixed Rhododendron–conifer forest, mixed conifer forest, mixed conifer–broadleaf forest, mixed forest and high-elevation meadow. Microhabitat characteristics comprised primarily forest (open to dense), lacking or with a dry leaf litter / herbaceous understorey, and substrate from dried leaves and twigs, to pebbles and boulders on bare ground bordering forest edge. Open habitats were mesic, from moss-covered clearings to herbaceous slopes with rocky outcrops. Topography was level to sloping and Crossoptilon was recorded at 4,028–4,359 m.

Figure 2.

Seasonal presence of Blood Pheasant Ithaginis cruentus and White Eared Pheasant Crossoptilon crossoptilon at Langdu, Yunnan, China.

Figure 3.

Activity patterns of Blood Pheasant Ithaginis cruentus and White Eared Pheasant Crossoptilon crossoptilon at Langdu, Yunnan, China.

Similar to Ithaginis, this species displayed a sharp elevational migration, being present only in April–October, moving to lower elevations in winter. Peak occurrence was during August, with at most one record each in the months of April–June and October (Fig. 2).

Crossoptilon was active from morning (08.00 h) until after nightfall (19.00 h). While a sharp peak of activity occurred in mid afternoon (15.00–16.00 h), it was primarily active in the morning (08.00–12.00 h; Fig. 3). The species was active during temperatures of 5–24^°C (mean = 9.6^°C), and snow was visible on the ground in some photos.

Photos were obtained between 27 July and 21 August of three different broods each of 2–4 young that were estimated to be four months of age (E. Benhardt, J. Berger, K. Landig, J. Pfarr in litt. 2019) foraging with 2–4 adults, suggesting hatch dates in late March–late April. Additionally, we obtained a photo of a copulation between a pair in the presence of three other adults on 11 August.

Flock size was 1–13 birds, with a mean of 3.5 (n = 15). Total ratio of adults to subadults was 43:10 (4.3 adults / subadult). The most common social group was lone adults (n = 8, 53%) followed by both (n = 3 each, 20%) groups of 4–5 adults, and groups of adults (2–4) with subadults (2–4), with a single group of ≥ 13 adults (n = 1, 7%).

Discussion

Beta diversity .—At a site c.1,600 km north of Langdu in Qi-Lian Xian (Qinghai province), an extremely similar community of six species of gamebirds was recorded (Li 1981), comprising three of the same species (Bonasa sewerzowi, Tetraogallus tibetanus, Ithaginis cruentus), two congeners (Chestnut-throated Partridge Tetraophasis obscurus, Blue Eared Pheasant Crossoptilon auritum) and a partridge (Tibetan Partridge Perdix hodgsoniae), instead of Tragopan temminckii. Similar to our findings at Langdu, Ithaginis and Crossoptilon were the most abundant species (Li 1981).

In contrast, despite being just c.125 km distant and sharing similar physiognomic characteristics and forest attributes, gamebird species composition between Langdu and Gehuaqing was dramatically different, with only Tragopan temminckii shared between the two sites. Unless conservation management at Gehuaqing is ineffective, the factors responsible for the very different species composition were probably not a product of conservation management regimes, as Langdu, without any formal forest guards, harboured three times the number of Galliformes as Gehuaqing.

One of the most relevant factors distinguishing Langdu from Gehuaqing was a difference of several hundred metres in elevation, with the Langdu site at 4,000–4,800 m and Gehuaqing at 3,050–3,600 m. This change in altitude results in different temperature gradients and plant species composition, which are strong factors determining avian community structure, as has been shown elsewhere in the Himalayas (Elsen et al. 2017, Srinivasan et al. 2018). Others (e.g., Thiollay 1996) have also found elevation to play a role in predicting species turnover in other montane regions.

One must also consider the cultural differences between Langdu and Gehuaqing (Li et al. 2016). Tibetans in Langdu are traditionally transhumance herders and Buddhists, and hunting is considered morally questionable. In contrast, gathering forest products and hunting is important in the Lisu culture, suggesting that they might have hunted some Galliform species to local extinction in Gehuaqing. In north-west Yunnan, such cultural differences had a significant influence on Musk Deer Moschus spp. distribution, for example, more than the protected status of an area (Li et al. 2016).

Ithaginis cruentus .—We report novel information on temporal presence, activity patterns and population sex ratios. Additional findings are compared with other studies, below.

From spring to early autumn Ithaginis was present at 4,000–4,400 m, whereas Lu et al. (2006) found this species at 3,400–3,700 m in south-west Shiqu county (Sichuan province). However, both of these ranges fall within the overall elevational range (3,200–4,700 m) provided by Delacour (1951) and MacKinnon & Phillipps (2000).

Delacour (1951) indicated that flocks of 10–20 splinter into monogamous pairs during the breeding season. Others have indicated Ithaginis occurs in small to large flocks numbering five individuals to as many as 70 post-breeding (Madge et al. 2002) or 8–19 with a mean of 10.67 (Lu et al. 2006). In contrast, the majority of our records were of lone individuals, followed by a small number of male–female pairs. This is probably due to our records being outside winter when larger flocks form.

Our population sex ratio data are novel being slightly more than two males per female during the breeding season. Others have found breeding pairs to associate with young bachelor flocks (Jia et al. 1999), or polyandrous groups of two males and a female rearing broods together (Ludlow & Kinnear 1944). One theory for the evolution of polyandry is a response to an abundance of males in a given population (Willson & Pianka 1963), such as the disparate sex ratio we observed of >2 males/female.

Crossoptilon crossoptilon .—Similar to Ithaginis cruentus, we report novel information on temporal presence and activity patterns for this species. Additionally, we report some notes on reproductive period; these and additional findings are compared with other studies.

Lu (1986) indicated that C. crossoptilon is found in coniferous forest during spring, whereas in the summer McGowan (1994) stated that it occurs in alpine meadows, and Madge et al. (2002) reported it in subalpine birch and Rhododendron scrub above the treeline. In contrast to these authors reports of the species being restricted primarily to one or two specific habitats in spring and summer, we found Crossoptilon associated with at least six different habitats during these seasons, some of which (e.g., subalpine coniferous and mixed forests) were thought to be used only in winter by McGowan (1994) and Madge et al. (2002). Jia et al. (2005) found this species to be negatively associated with distance to nearest permanent water and herb cover, but positively with shrub cover, tree cover and tree height.

While most prior references indicate C. crossoptilon occurs below 3,900 m (Delacour 1951, Lu 1986, MacKinnon & Phillipps 2000), we found it only above 4,000 m to nearly 4,400 m, i.e. closer to the max. summer altitudes of 4,300 m (Madge et al. 2002) or 4,600 m (McGowan 1994).

Very little to no reliable information is available for C. crossoptilon reproduction in the wild (McGowan 1994). It is presumed that monogamous pairs (Madge et al. 2002) split off from larger groups in spring (Delacour 1951) to nest and lay eggs in May–June (McGowan 1994). However, Lu (1986) noted small flocks during the breeding season, rather than just solitary birds and monogamous pairs (Delacour 1951). We observed the same pattern, with a copulating pair in a small flock, broods of young foraging with multiple adults, and flocks of up to 13 birds (mean = 3.5).

The largest flock we observed was 13, consistent with others (e.g., MacKinnon & Phillipps 2000, Madge et al. 2002). However larger flocks of up to 30 are found in winter (Madge et al. 2002), and historically flocks of several hundred (Lu 1986) to 1,000 birds (Wang et al. 2012) have been reported, although these estimates are primarily from Buddhist monasteries where the birds are provided regular food by the monks in these regions.

Conservation implications .—While humans are the primary predators of Galliformes, other potential predators of adult and young birds and their eggs include several species recorded by our camera-traps (Buzzard et al. 2018). Macaques (Macaca sp.), Leopard Cat Prionailurus bengalensis, Yellow-throated Marten Martes flavigula, Masked Palm Civet Paguma larvata, domestic dog Canis lupus familiaris and Wild Boar Sus scrofa were present at both sites, and Yunnan Snub-nosed Monkey Rhinopithecus bieti only at Gehuaqing. Additionally, local people report that Snow Leopard Panthera uncia occurs at Langdu, although this has not been confirmed (Buzzard et al. 2017).

All of the gamebirds recorded in this study are currently classified as Least Concern (LC) by BirdLife International / IUCN (2018), with the exception of two Near Threatened species (Bonasa sewerzowi and Crossoptilon crossoptilon) found only at Langdu. C. crossoptilon was not only the most abundant species in our study, but was also observed to be breeding well. Although the species was treated as Vulnerable 25 years ago (Collar et al. 1994), it was considered abundant nearly 70 years ago (Delacour 1951). McGowan (1994) indicated the primary threats to be forest destruction and hunting.

The robust population of Near Threatened species such as C. crossoptilon, coupled with the fact that Langdu harboured much higher abundance, as well as three times as many species, of Galliformes, is surprising given that conservation management is weak at Langdu compared to Gehuaqing, with no formal forest guard patrol in place. One possible explanation is the majority of Langdu occupants are Tibetans more focused on herding, whereas Gehuaqing inhabitants include the Lisu Minority who are traditionally hunters (Li et al. 2016). Supporting this, all but one of the photos at Gehuaqing were taken during crepuscular periods (n = 3) or at night (n = 5). Game species often shift their activity to darker photoperiods to avoid hunters who are more active by day (e.g., Brooks et al. 2001). In cases of communities actively harvesting wildlife (e.g., Gehuaqing Lisu), it is assumed that hunting has only increased over the decades of economic development since the 1970s. Additionally a collapse of traditional wildlife management practices (e.g., refuges, rotation and hunting seasons) may have been a contributing factor. Thus increased availability of firearms and markets led to a decline in biodiversity, which may have recovered somewhat since legal bans on guns and sales of wildlife came into effect in the 1990s.

Comparing species-presence data from a nearby site in north-west Yunnan, all of the species of pheasants photo-trapped herein, plus Silver Pheasant Lophura nycthemera, were also recently camera-trapped at Baima Snow and Wuliang Mountains (XL unpubl.). Of these seven species, nearly a century ago Beebe (1936) mentioned anecdotally that Tragopan temminckii, Crossoptilon crossoptilon and Chrysolophus amherstiae were present at sites in north-west Yunnan, while Ithaginis occurred nearby. It is surprising that Beebe (1936) did not mention the other three species recorded, as they were not rare in our study, comprising a combined 46% of photos (Lophura nycthemera n = 19, Tetraogallus tibetanus and Pucrasia macrolopha n = 7 each). Another historical example from the region is provided by Andrews & Andrews (1918), who repeatedly remarked how rare game was in the region, especially birds; nonetheless they recorded Tragopan temminckii and Chrysolophus amherstiae, plus Red Junglefowl Gallus gallus and Lophura nycthemera. Although not quantitative, the differences in species turnover over time provided in these simple analyses help to provide insight into historic hunting pressure in the region.