Spatial and temporal structure of the spider community in the clay semi-desert of western Kazakhstan

: The spatial and temporal structure of spider communities was studied in the clay semi-desert of the north-western Caspian Lowland, western Kazakhstan (49°23' N, 46°47' E).The soils and vegetation are complex, being composed of a mosaic of desert and steppe plant communities. Besides the native associations, there are plantations of different tree species.The ground-dwelling spider assemblages in the native habitats are the most diverse. The number of species inhabiting forest plantations is three times as small. Gnaphosidae is the leading family in the ground layer.They show high abundance and diversity levels during the whole season. Thomisidae, Lycosidae, Philodromidae, and Salticidae are abundant as well. The species diversity of herbage-dwelling spiders in different open native habitats is very similar. The spectrum of dominant families (Thomisidae, Oxyopidae, Araneidae, and Salticidae) and the seasonal dynamics of their ratio in desert and steppe associations have much in common. Spider assemblages of native and artificial habitats are characterised by change from multispecies polydominant spring-summer communities to impoverished imbalanced autumn ones. Seasonal changes in the species structure of mature spider groupings in native habitats are well pronounced, while the impact of seasonal conditions is even stronger than between-habitat differences. Complexes of typical species with different levels of habitat preference are revealed.


2006
).The sharp disparity of heat and moisture causes the very low humidity of the territory.The evaporative power reaches 1000 mm, which is 3 times the total rainfall.In addition, the meteorological conditions of the region are characterised by long-term fluctuations with regular cyclic reiterations of drought and moist periods (RODE 1959, LlNDEMAN et al. 2005, SAPANOV 2006).
Another characteristic feature of the study area is a well pronounced complex pattern of soils and vegetation, with a combination of typical desert, semidesert and steppe habitats.The co-existence of such Spiders of steppe and semi-desert regions of the Palaearctic, unlike those of the temperate zone, are still poorly studied.There is some faunistic information (e.g.Ponomarev 1981, 1988, 2005, 2008, Ponomarev &, Tsvetkova 2003, Ponomarev &TSVETKOV 2004a, 2004b, POLCHANINOVA 1992, 1995, 2002, KOVBLYUK 2006, EFIMIK et. al. 1997, ESYUNIN Sc EFIMIK 1998, ESYUNIN et al. 2007, TU-NEVA Sc ESYUNIN 2003), but very little attention has been paid to such ecological aspects as the structure of populations, their dynamics, and the mechanisms of community function in arid and semi-arid conditions (ESYUNIN 2009).This paper is focused on studying the spatial and temporal structure of spider assemblages in the clay semi-desert in the Volga and Ural rivers' interfluve.contrasting biotopes is caused by pronounced microrelief and, consequently, differences in moisture, soil substrates and their properties (RODE & POLSKIKH 1961).
Microelevations are occupied by plant communities of the desert type, with Kochia prostrata, Artemisia pauctflora , and Salsola laricina on saline soils.The groundwater is saline.Forb-grass vegetation ( Stipa spp., Festuca valesiaca , Agropyron cristatum , etc.) on dark chestnut and meadow chestnut soils with fresh groundwater occupies microdepressions (down to 0.4 m deep); they represent steppe habitats.This mosaic of elements constitutes most of the territory.Large depressions (down to 1-1.5 m deep, area of 1-100 hectares) with steppe plant communities take up about 10-15 % of the area.These large depressions are best supplied with water, due to runoff from the surrounding area.Besides these mentioned native associations, there are 50-year-old plantations composed of different tree species.
Material for this work was collected by the author (April-October 2004-2005) and Dr. K.G.Mikhailov (June-September 1984) in three native habitats (desert associations of microelevations, and steppe associations of large depressions, and microdepressions) and three artificial ones: oak ( Quercus robur) forest belts, oak patch in a park, and elm (Ulmus pumila) forest belts.The collections in the latter habitat took place only in 1984.In recent years, the vitality of the forest-belt has become very poor; the trees are very sparse so the conditions in it have approached those of an open habitat.
Traditional collecting methods were used: pitfall trapping (one transect -10 traps), hand-sorting of soil and litter samples (0.25 x 0.25 m, 10 samples) and sweeping (one sample -4 x 25 sweeps, 3 times a day, at 00:00, 8:00 and 16:00).Sampling was carried out every 7-10 days.Pitfall traps were set in microelevations, microdepressions and woody plantations.Soil and litter samples were taken in all studied habitats.As the plantations had a rather poor and scattered herbaceous layer, sweeping was undertaken only in native habitats.
The material includes a total of 15000 pitfall days, 570 soil and litter samples, and 268 sweeping samples.
One of the most important features of the spider population in the clay semi-desert is its strongly pronounced seasonality and vertical stratification.Thus, I analysed the structure of spider complexes separately by layer, i.e. ground and herbaceous layers, and seasons, i.e. spring, summer and autumn.When calculating the ratio of families, I considered spiders of all instars.With respect to the seasonal changes in species compositions I used mature spiders only, although I suggest that the differences revealed might reflect certain phenological trends.
Taxa with a relative abundance of >5 % were considered predominat.The habitat preference of species was calculated using Pesenko's coefficient (if.) (PESENKO 1982), which represents a mathematical transformation of the share of a species in a single biotope to its share in all other biotopes: R. = (n../N -n./N)/(n../N.+ n./N), y ij j i y j 1 where n.-number of specimens of /-species in samples from y-biotope with total volume A.; n. -number of specimens of /-species in all other biotopes with total volume N. Single records of species were omitted from the calculation.The choice of this coefficient was based on the variety of the collecting methods used, which caused the heterogeneity of the data obtained and the difficulties in their unification.Using relative indices (not absolute ones) simplifies the interpretation of data and makes miscellaneous information comparable.
The value of the coefficient ranges from -1 (absolute avoidance) to +1 (absolute preference).Statistical data analysis was performed using Statistica 6.0.

Results
About 20000 spider specimens were captured and studied, with about 7000 of these spiders being mature.Altogether, 172 species from 88 genera and 21 families were recovered.Taking into account the scant information published previously, the spider fauna of the Dzhanybek Station amounts to 184 species from 93 genera and 22 families.A checklist and the distribution of species between the study habitats has been made available elsewhere (PlTERKINA 2009, PlTERKINA & MIKHAILOV 2009).Since the time of these mentioned papers some taxonomical changes have taken place or some identifications were refined, thus some species names may not coincide.Namely, Ero sp.turned out to be Ermetus inopinabilis Ponomarev, 2008, Theridion cf. uhligi Martin, 1974-T. uhligi , Thanatus constellatus Charitonov, 1946-T. oblongius cuius (Lucas, 1846), and Eresus cinnaberinus (Olivier, 1789) -E.kollari Rossi, 1846.
Species structure of spider communities and its seasonal dynamics The communities of ground-dwelling spiders in the native habitats -microelevations and microdepressions -are the most diverse (about 90 species).The number of species inhabiting forest plantations is three times as small (about 30 species) (Tab.1).
The activity of spiders in the open habitats fluctuates from 20 to 70 ind./ 100 pitfall-days, with the highest numbers in spring and summer.The amplitude of its fluctuation is much higher in the forest plantations (from 3-4 to 100 ind./100 pitfall-days).The density of the spider population, based on soil and litter samples, reaches its highest values in autumn (up to 117 ind./m 2 ).
Gnaphosidae is the dominant family in the native associations.They exhibit high abundance and diversity levels (about 50 %) during the whole vegetation season, this being quite typical of arid and semi-arid landscapes.The proportions of Lycosidae and Salticidae are less, but also stable.Linyphiidae predominate in spring and autumn, Oxyopidae in summer, Titanoecidae in spring and summer, Thomisidae in summer and autumn.The dominant complex of the tree plantations is less diverse.The proportion of Gnaphosidae is significantly lower than in native habitats (about 20-30 %), while the abundance of Thomisidae is high and stable during the entire vegetation season (about A 1 -Pearson r Complete Linkage 30-50 %).Pisauridae show a peak in their abundance in spring and autumn, whereas Liocranidae peak in summer.
Seasonal change in species dominance is well pronounced and the species set is relatively stable across different years (Tab.1).For example, in the desert habitats, T. veter anica , Haplodrassus cf.soerenseni , E. eltonica , D. ros trains, Z. orenburgensis predominate in spring populations in both years of study.The stable summer dominants are P. braccatus , H. horridus , Oxyopes c£ xinjiangensis , D. rostratus and Z. orenburgensis.The autumn populations are rather imbalanced.Cheracteristic is a high level of predominance of 1-2 species that can change in different years (Z.orenburgensis , X. marmoratus or D. rostratus).The dominant complexes of oak plantations have much in common and include several species abundant during the whole vegetation season (Z.gallicus, O. pratkola, X. luctator) (Tab.1).The species diversity of herbage-dwelling spiders in the open native habitats is very similar: about 50 species (Tab.2).The abundance ofhortobiotic spiders fluctuates with a high amplitude, reaching its maximum in summer (about 100 ind./ 100 sweeps).The spectrum of predominating families and the seasonal dynamics of their proportion in desert and steppe associations have much in common.The seasonal change of the predominant complexes of species is also well-pronounced (Tab.2).In spring and, especially, summer, the sets of abundant species are not stable in different years.On the contrary, the autumn populations of all habitats are very similar.They are mainly formed by two species, Xysticus marmoratus and X. striatipes.Co-dominance of Cheiracanthium cf.virescens adds originality to the autumn assemblages of microelevations, E. michailovi to those of microdepressions, and H. lineiventris to those of large depressions (Tab.2).
Clustering the mature spider complexes for separate seasons (Fig. 1) yielded interesting results.Two large clusters were revealed among grounddwelling spiders: assemblages of native biotopes and of forest plantations (Fig. 1 A).Within them, the populations were not united by habitat, as one would expect, but by season.The cluster of open habitats includes populations of microelevations and microdepressions during spring, summer and autumn.Microclimatic conditions in woody plantations were presumably comparatively smoother, even though no direct abiotic measurements were taken.The cluster of artificial forests appears to be less differentiated.The same tendency is also obvious when clustering the herbage-dwelling spider complexes: three pronounced clusters united spring, summer and autumn assemblages of microelevations, microdepressions and large depressions respectively (Fig. IB).
Habitat preferences of species Spider assemblages of desert associations are the most specific.The share of species collected only in microelevations is highest (24 %), whereas it is half this in the other biotopes.Most of unique species, with few exceptions, exhibit low abundance levels and hardly play coenotic roles (Tab.3).

Microelevations (desert habitats)
Oh CD S .1 As many as 25 species occur in all native habitats, another five in all forest plantations.Two species, Lathy s stigmatisata and Xysticus ninnii , are ubiquitous and inhabit all studied habitats.However, finding the species in a particular habitat does not necessarily indicate habitat preference.In order to estimate preference level, Pesenko's coefficient (F.) was used.A complex of species, including taxa both with high (F.>0.7) and relatively low (0.3>/y>0.7) levels of habitat preference, was revealed for eac ii habitat (Tab.4).

Discussion
It is well known that the denser the vegetation the greater is density of spiders, and the greater the diversity of vegetation the greater the spider species diversity (DUFFEY 1962).But the spider assemblages of both the ground and herbaceous layers of open native habitats (microelevations, microdepressions and large depressions) are very similar not only in species diversity but also in density.This was rather surprising as the low, sparse and rather poor desert plant communities look much more miserable compared to the dense forb-grass vegetation of steppe habitats.This reveals a complex of species well adapted to the extreme conditions of desert associations.On the contrary, the communities of forest plantations appear to be significantly impoverished.The poorness of soil fauna under Dzhanybek plantations was demonstrated for other arthropods as well (CHERNOVA 1971, KRIVOLUTSKII 1971, etc.).
Calculating the level of habitat preference (if .)revealed complexes of typical species for each habitat (Tab.4).In spite of mosaic structure and a comparatively small size of desert and steppe elements (some tens of square meters) in complex Northern Caspian semi-desert, the spider groupings formed on them are rather specific and contain sets of species associated with the particularities of the substrate (soil) and vegetation of those elements.The complexes of typical species of native habitats -microelevations and microdepressions -are the richest (35-40 species).
In addition, there is a complex of species which can inhabit several types of native habitats with similar probability levels (except for woody plantations).These are Trichoncoides cf.pis cat or, G. bituberculata, A. v-insignitus, A. cursor , P. histrio, Z. segrex, etc.
Complexes of typical species of woody plantations are poor and include 12-15 species, although the level of habitat preference is very high (Tab.4).Most of them are nemoral species.Populations in the plantations are very likely composed of highly eurytopic species ( D.pusillus , Z.gallicus , P mirabilis ) and typical dwellers of intrazonal associations ( S .zimmermanni , T. schineri ) with a small participation of forest species (O.praticola) which could be introduced with plant material.
On the other hand, the structure of spider assemblages is heavily determined by macroclimatic conditions and their seasonal changes.The analysis of seasonal features of population structure shows that the spring and summer spider assemblages of both ground and herbaceous layers are characterised by high species diversity levels and a relatively high number of predominating species, as opposed to the impoverished, imbalanced autumn populations (Tab.1-2).The same pattern was recovered by ESYUNIN (2009) for spiders of steppe and steppe-like habitats in the Ural Mountains.
Clustering the spider complexes for separate seasons confirmed the prevailing role of seasonal differences in species proportions for mature spider groupings of native habitats when comparing between-habitat differences (Fig. 1).The populations of native associations were not united by habitats, but by seasononality.A similar trend has been also shown by ESYUNIN (2009) for the spider populations of steppe-like habitats in the Ural Mountains.
It is interesting to note that such a tendency was The research was carried out in the environs of the Dzhanybek Research Station of the Russian Academy of Sciences (49°23'N, 46°47'E), located on the border between the Western Kazakhstan Province of the Republic of Kazakhstan and the Volgograd Province of the Russian Federation.The area studied is a flat, nearly undrained plain in the north-western Tatyana V. PITERKINA, Laboratory of Synecology, Institute of Ecology and Evolution, Russian Academy of Sciences, Leninskii Prospect 33

Table 4 :
Pesenko's coefficient of a habitat preference {Fij) of spiders.Species are grouped according to their preference to a certain habitat.Within the groups species are ranked in order of decreasing the values of Fij.Grey background: high level of habitat preference (0.7 < F..< 1.00); bold: relatively low level of habitat preference (0.3 < F..< 0.7).Habitats as in Fig.snout-beetles (Coleoptera, Curculionoidae) investigated at the Dzhanybek Station during the same period.These phytophagous insects showed that the influence of between-habitat differentiation on the structure of their populations -which was determined by their close links with the plants on which they forage(KHRULEVA et al. in press) -was much stronger than seasonal changes.Spiders being a group of mobile generalist predators are more likely to be influenced by abiotic factors.

Table 3 :
Unique species per type of habitat.