Revue Suisse deZoologie, vol. hors serie: 223-234; aout 1996 Life cycle, habitat selection and home ranges of Arctosa cinerea (Fabricius, 1777) (Araneae: Lycosidae) in a braided section of the Upper Isar (Germany, Bavaria) Volker FRAMENAU1 2. Martin DIETERICH-. Michael REICH2 & Harald PLACHTER- 1 University ofMelbourne Department ofZoology, Parkville, Victoria 3052, Australia; 2 Universitat Marburg Fachbereich Biologic Fachgebiet Naturschutz, D-35032 Marburg, Germany. Life cycle, habitat selection and home ranges of Arctosa cinerea (Fabricius, 1777) (Araneae: Lycosidae) in a braided section of the Upper Isar (Germany, Bavaria). - Investigations on basic population characters of the lycosid spiderArctosa cinerea were conducted from June 1994 to July 1995. The spiders were individually marked and recaptured to estimate abundance and home range size. Maximum spider densities found were 0.3 ind./m2. During their larval stages and during the first four weeks as adults, the spiders prefer gravelbanks where they hide under stones during the day. As they get older they prefer sandy substrates for the establishment of their burrows. A. cinerea has a two year life cycle. The species is diplochronous with reproductive periods in autumn and spring. The adult sex ratio is biased towards females (1 : 1,8). Maximum home ranges were .260 m2 for males and .140 m2 for females. 1 1 Key-words: Lycosidae - Arctosa cinerea - life cycle - sex ratio - habitat selection - home range - river - floodplains. INTRODUCTION The floodplains of braided alpine rivers provide a highly dynamic habitat. Recurring flood events often produce gravel banks bare of any vegetation. These are inhabited by a specifically adapted arthropod community, characterized by a large number of predators. Particularly ground beetles (Carabidae), rove beetles (Staphy- linidae) and spiders (Araneae) are abundant on gravel bars (Kuhnelt 1943, Plachter 1986«, h, Hering 1996, Manderbach & Reich 1996). As there are no vertical structures to fix a web. the spider community on the gravel bars mainly consists of ground living species. Most abundant are lycosid and linyphiid spiders (e.g. Boumezzough 1983, Bigot & Favet 1985, Droschmeister 1994). Manuscript accepted07.12.1995. Proceedingsofthe Xlllth International Congress ofArachnology. Geneva. 3-8-IX.1995. 224 VOLKERFRAMENAU, MARTIN DIETERICH, MICHAELREICH & HARALD PLACHTER One of the riparian inhabitants, the lycosid spider Arctosa cinerea (Fabricius, 1777), can be found in the whole palaearctic region (Platnick 1993). A cinerea is up mm to 20 in size and therefore one of the largest predacious arthropods on alpine gravel banks and one ofthe largest spiders in Central Europe. A cinerea not only occurs on riverbeds and lakesides, but also on sandy beaches and coastal dunes, where the spiders construct burrows to hibernate and to breed. In Germany, A. cinerea has also been found in gravel pits (Sacher 1992). Stochasticity, either demographic, environmental or catastrophical, plays a major role for the population dynamics of most species and for the development of certain ecosystems (Ewens etal. 1987, Shaffer 1987, Reice 1994). Local extinction of populations is one of the characteristics in such ecosystems (Luken 1990, Hovestadt et <//. 1992). Braided floodplains are characterised by floods, which mainly occur during the later spring in Central Europe. The aim of this study is to demonstrate how life cycle and habitat association ofA. cinerea are adapted to recurring flood events and what spatial needsA. cinerea has within its habitat (home range size). STUDY AREA AND METHODS The study site (fig. 1) was located in the braided floodplain of the Upper Isar (Germany, Southern Bavaria) between Wallgau and the Sylvenstein Reservoir. The floodplain is between 300 and 500 m wide and the Isar has a slope of 5-6 %6. Since Fig. 1 Study areaatthe UpperIsar. ARCTOSA C1NEREA (FABRICIUS, 1777) (ARANEAK: LYCOSIDAE) m 1919, up to 25 3 s-> ofthe Isar water has been diverted into the Walchensee to pro- duce hydroelectric power. However, flood events with peak flows of up to 160 m3-s-' still occurregularly (Bayerisches Landesamt fiirWasserwirtschaft, 9SK). 1 The study site, located 820 m above sea level at river-km 242,2, can be dis- tinguished into three parts according to vegetation structure and isolation: • The "gravel bar'* (750 m2), bare ofvegetation. • The "gravel islands" (2.850 m2), characterized by an early successional com- munity on sandy substrate dominated by Calamagrostis pseudophargmites (Poaceae), Rinanthus glacialis (Scrophulariaceae) and willow (Salix eleagnos). Willow cover was partly dense in the eastern section ofthe largest island while the western parts including two smaller islands lacked any vegetation. • The "hinterland" (5.600 m2), covered by a later successional community dominated by willows (Salix eleagnos, S. purpurea) and the German tamarisk (Myricaria germanica). The ground layer was mainly covered by mountain aven (Dryas octopetala) and thyme (Thymus serpyllum). The vegetation covered 30 - 50%. The whole study area was devided into grids of 5x5 m marked by wooden pegs. Each point where a spider was found, was determined with a precision ofabout m. The grids also allowed to draw a precise map of substrate structure and vege- I tation in orderto analyse the habitat qualities favoured by the spiders. Habitat structure Sediment was categorised into 5 classes according to the relative cover ofrocks (0> 20 mm) and fine sediment (0< 2 mm). Two categories were to be found in two different conditions depending on their firmness (table 1 ). The larger parts of the sedi- ment fraction produce hollows were the night activeA. cinerea hides during the day. Table 1 Categories ofsediment structure cover [%] ofstones mm sediment largerthan 20 category in diameter main characteristics 1 100 large amountofhollows and crevices, no sandy substrate (no abilitytoretain moisture) 2 31-99 gaps are filled with sandy substrate 3.f 6-30 firm condition: stonescan notbe picked up without resistance 3.1 6-30 loosecondition: stonescanbe picked up without resistance 4.f 1-5 firmcondition 4.1 1-5 loosecondition 5 sand only 226 VOLKER FRAMENAU, MARTIN DIETER1CH, MICHAEL REICH & HARALD PLACHTER Six categories were defined in terms of vegetation cover 0%, 1 - 5 %, 6 - 15 %, 16 - 30%, 31 - 60% and 61 - 100%. A higher resolution was used as vegetation became sparse, becauseA. cinerea is especially known forhabitats with little vegetation cover. One of the above categories for sediment structure and vegetation cover was assigned to every in2 of the "gravel bar" and the "gravel islands". In order to analyse the preferences ofA. cinerea with respect to a certain category, the number ofcaptures within each category was compared to the relative share of this category in the whole area. The preference for sediment structure was examined separately, both for spiders on the surface (walking orhiding under stones) and forthose within a burrow. Individual markings and sampling method To estimate abundance and to assess movement patterns, adults spiders were individually marked with bee tags glued onto their prosoma using a cyane-acrylate based adhesive. In addition, olderjuveniles with a prosoma width of more than 4 mm (approx. last twojuvenile instars) were individually marked with a xylene and toluene free paint. Different colours were applied to the prosoma and one leg segment according to a predetermined code. Between June 15th and October 12th, 1994 the "gravel islands" and the "gravel bar" were sampled at weekly intervals. Starting in August the "hinterland" was also examined weekly to find those spiders which had retreated from the bank to hibernate in burrows. In 1995 the whole study site was checked on June 15th and July 15th. In order to prevent a spider leaving the grid without being counted, each 5 x 5 m grid was examined by walking in a spiral from the outer to the inner part. Rocks were turned, small gravel and sandy substrate was combed with the fingers. Burrows were easy to locate eitheras an open hole or as a small, "hill" ofsand, ifshut. Abundance home range and activity A The Jolly index was used to estimate the abundances of cinerea foreach day of the investigation (Jolly 1965). This allowed for the determination of changes in population density caused by death, birth, immigration and emigration. Home ranges were determined by connecting the outermost capture points. The area ofthe resulting polygon is an estimate for the home range ofeach individual ("minimum area method") (Odum & Kuenzler 1955, Southwood 1979). In contrast to Kuenzler ( 1958) who considered only those individuals caught at least four times, all individuals captured three or more times were included in the home range esti- mates ofA. cinerea. The activity of different individuals was compared by the "average daily distances" (ADD) measure. ADDs were calculated separately for the whole study period and forthree week intervals. RESULTS Overall, 393 adult spiders were marked in summer 1994. The recapture rate was 82%. The maximum number of recaptures was 10. On the average, every spider ARCTOSA CINEREA (FABRICIUS, 1777) ARANEAE! LYCOSIDAE) ( 227 was captured 4,6 times. Winter mortality appeared to be high. While 19 marked 1 spiders were collected in the study area in October 1994. only six of those were still found in June ofthe following year. In July 1995 only 5 nymphs and no adult spiders were recorded (fig. 2). June94 July94 Aug. 94 Sep. 94 Oct. 94 June95 July95 Fig. 2 Numberofspiders caught duringthe study c o winter winter (diapause) (diapause) PFT K't K?t FT spiderlings / adults AMJJASOND JFMAMJJASOND JFMAMJJASO year: 1 2 3 Fig. 3 Life cycle ofArctosacinerea Femalescarryingacocoon are indicated by K|. Kt. 228 VOLKER FRAMENAU, MARTIN DIETERICH, MICHAELREICH & HARALD PLACHTER Life cycle and sex-ratio A cinerea has a two year life cycle (fig. 3). Cocoons are carried by females in June and July/August. The hatched nymphs hibernate in different stages. Their final moult is in August ofthe following year. Male spiders reach maturity about one week earlier than females. The adults hibernate again. Females then breed in two clearly separated periods in June and from mid July to early August. As an exception to this rule, one female was found breeding from August 22nd to September 7th, 1994 (fig. 2). Only one of 17 females was found to breed in both periods. Since there are always two overlapping generations in a population, adult spiders are found throughout the whole year. The average life span was 41 days forboth, adult males and females. The maximum observed life span for an adult female was 314 days. One male survived at least 297 days. The sex ratio of adult A. cinerea is female biased, and constant from early September until the end ofthe year (6 619 9 1:1,8). In the lastjuvenile instars 113 out of454 individually markedjuveniles were males, 178 females (6 6/9 9 1 : 1,57). The sex of 163 ofthese nymphs was indeterminable. The surplus offemale adults increases after hibernation because males die after mating, whereas the females care for their brood. Habitat selection and abundance At the Isar A. cinerea prefers vegetation free habitats. As soon as there is a sparse plant coverthe number ofspiders decreases (fig. 4a). percentageonsurfaceofstudysite percentageonsurfaceofstudysite percentageofcaptures(incl.95% percentageofcaptures(incl.95% confidenceinterval) confidenceinterval) 1 t 6-15% 16-30% 31-60%61-100% 1 2 3.1 3.1 4.1 4.1 S vegetationcover categoryofsedimentstructure a) b) categoryofsedimentstructure c) Fig.4 Habitat association of Arctosa cinerea at the study site: (a) captures relative to vegetation cover; (b) captures ofvagrant spidersrelativeto sediment structure (forexplanation ofsediment categories see table 1); (c) captures in burrowsrelative to sedimentstructure i ARCTOSA CINEREA (FABRICIUS, 1777) (ARANEAE: LYCOSIDAE) 229 The demands with respect to sediment structure are more complex. Alter their last moult in August the adult spiders live as vagabonds for about four weeks and then begin to dig a tube. The spiders hibernate in these burrows. Whereas the females remain in their burrows after hibernation, the males wander around. Different sediment types are preferred during the vagrant and the burrowing stages. Throughout the vagrant period, a clear preference for sediments with a high percentage of stones and rocks (category 2) is obvious, whereas those sediments consisting of smaller particles (3.f, 3.1, 4.f and 5) are avoided (fig. 4b). The vagrant spiders spend the day under rocks without digging burrows. Only occasionally, a small hole dug by the spider can be observed under a rock (7% of all observations). Some spiders were found running on the surface (20%). The same spider was never caught twice under the same rock. Spiders in burrows show a different distribution. More burrows were found in areas with substrates consisting of smaller pieces (categories 3.1, 4.f, 5) (fig. 4c). In some cases spiders moved between burrows. About 10% of the burrows were not exclusively used by one spider, but rather by several individuals over the sampling period. Q Q < O c T—O -i trt T3 "O <D D) ro > TO d d 99 d d 99 (n = 99) (n = 173) (n = 117) (n = 207) (p < 0,001) (P < 0,001) Fig. 5 (a) Home range sizes and (b) average daily distances (ADDs) ofArctosa cinerea (median, 1st and 3rd quartile and range between maximum and minimum). Significance ol differences between 9 9 and 66 is indicated (U-test). 230 VOLKER FRAMENAU, MARTIN DIETERICH, MICHAELREICH & HARALD PLACHTER Over the whole study site maximum abundance was 0.03 ind./m2 (September 19th, 1994), while by early October the population declined to 0,014 ind./m2 The . decrease from 269 to 133 individuals within this period corresponds to an average mortality of5,4 ind./day. If only the most favoured habitat structures are considered, abundances increase to 0,3 ind./m- (areas with sediment category 2) and 0,13 in./m2 (vegetation free areas). Locomotory activity and home range Significantly larger home range of males (average: 210 m2 maximum: 1.260 , m2) than those of females (average: 130 m2, maximum: 1.140 in2) indicate a higher mobility of male spiders (fig. 5a). Some spiders remained in the same burrow during the entire observation period (home range regarded as m2). One single marked male was found on an isolated gravel bank 2 km down- stream from the study area after floods in May and June 1995. It was not included in the estimation ofhome rangesize. However, this recapture shows the potential for dis- persal through floods. While the maximum average daily distance (ADD) was recorded for a female, the average ADD is larger for males (fig. 5b). Lower female activity in July as a result ofbrood care and higher male activity during a reproduction period in late August/early September account for the significant differences in the ADDs between males and females during these two periods (fig. 6). In October, most spiders have moved into a burrow and only few individuals remained active. The spiders mainly moved east- and northwards leaving the areas nearthe wateredge (Framenau etal. 1996). 45 n =43 40 Q Q 35 n =136 < 30 n =34 COD c 25 n =55 ro uS n =7 T3 20 n =132 "COO 15 n =6 T n =170 n =100 0OTO3) 10 I n =16 n =43 > 5 ra tfcf 99 cfcf99 tftf99 cfcf99 dd99 cftf99 13.6.-30.6. 1.7.-24.7. 25.7 -7.8. 22.8.-4.9. 5.9.-25.9. 25.9.-16.10. Fig. b Average daily distances (ADDs) covered during three week sampling periods in 1994 (median, 1st and 3rd quartile and range between maximum and minimum). • indicates significant differencesbetween ? ? and 66 (U-test, p<0,001 ). ARCTOSA C1NEREA (FABRICIUS, 1777) (ARANEAE: LYCOSIDAE) 231 DISCUSSION Life cycle, phenology and sex-ratio The exact life cycle is known for only a few sider species (Foelix 1992). For wolfspiders, life cycles of up to four years have been described (e.g. Conley 1985). Smaller lycosid spiders inhabiting gravel bars (e.g., Pardosa wagleri (Hahn), Pirata knorri (Scopoli)) have a one year life cycle (Framenau, unpublished data). However, the life cycles of spiders are not necessarily correlated to body size only (Schaefer 1976). Climatic factors also can play a major role. Pardosa lugubris (Walckenaer) hibernates two winters in Sweden and Scotland, but only one in England and the Netherlands (Edgar 1971 ): The two year life cycle of A. cinerea corresponds to those recorded for other Arctosa species and larger lycosids in Central Europe. Arctosa perita (Latreille) needs two years to complete its life cycle (Schaefer 1976) just like all Trochosa species in central Europe (Engelhardt 1964). The two breeding periods in combination with a reproductive activity ofmales in fall can be explained by a diplochromous activity with two reproductive phases in autumn and in spring. Females copulating in fall breed in June ofthe following year, while those mating in spring breed in July/August. Diplochronous activity is recorded only for a few spider species. Just 9 (=3,2%) out of 277 species analysed by Schaefer (1976) for the temperate zone fall into this category. The origin of the diplochronous activity is discussed in detail in Tretzel (1954), Heydemann (1960), Engelhardt (1964) and Schaefer (1976). Schaefer (1976) also described diplochrony for A. cinerea. However, his results, based on pitfall collections from the coast of the Baltic Sea, differ somewhat from those in the continental alpine region. Along the shores ofthe Baltic Seafemales carrying a cocoon were already found in May, whereas at the Isar they did not appear before June. In June and July Schaefer detected juveniles only. Adults were again found in August and September, which is in accord with the data from the Upper Isar. At this time the males of the new generation show their reproductive activity. In October no spiders were caught along the Baltic Sea. In coastal and riparian habitats diplochrony can be interpreted as a risk spreading strategy: With one part of the reproductive period in autumn it is guaranteed that there will always be some inseminated females present in spring. Should a flood prevent reproductive activity in spring, these females can be the founders of the next generation. Female biased sex ratios are well known for social spiders, e.g. the thomisid Diaea socialis Main (Rowell & Main 1992) or the ridiid Anelosimus eximius (Keyserling) (Aviles & Maddison 1991 ). However, compared to findings in non-social spiders (Heydemann 1962, Humphreys 1988), the female biased sex ratio ofA. cinerea is surprising. Schaefer (1987) explains female biased sex ratios within spider popu- lations by a shorterlifespan ofmale spiders. But this does not explain the female biased ratios in the last larval stages recorded forA. cinerea along the UpperIsar. Habitat selection The preference ofA. cinerea for vegetation free habitats confirms observations by other authors (e.g. Bochmann 1942, Knulle 1953, 1959). In effect, most Arctosa 232 VOLKER FRAMENAU, MARTIN DIETERICH, MICHAELREICH & HARALD PLACHTER species are known from vegetation free habitats, in particular seashores and river- banks (Knulle 1959. Dondale & Redner 1983, Tanaka 1991). Some occupy heath and lichen habitats in high mountains orarctic tundra (Dondale & Redner 1983). A. cinerea depends at least on two different kinds ofsediment structure. Banks covered with large gravel offer many hiding places for the vagrant stages and crevices under rocks serve as a shelter against the sunlight, reduce temperature fluctuations and protect A. cinerea against predators (e.g. the pompilid wasps Anoplius infuscatus and A. concinnus (Framenau, unpublished data)). During the burrowing stage A. cinerea prefers sandy substrates. The sand enables the establishment ofburrows. Presumably due to its large size, the density of A. cinerea, even in the most preferred microhabitats (0,3 Ind./m2) is low compared to about 3 ind./m2 for Trochosa ruricola (De Geer) (Hackman 1957) and 1,4 ind./m2 for Pardosa agricola (Thorell) (Albert & Albert 1976). The habitat association ofA. cinerea shows that this species strictly depends on the dynamics of the river. Floods, on one hand cause high mortality, on the other hand they guarantee the existence of vegetation free gravel bars and sandy areas which otherwise would soon fall into succession (Muller 1991). Locomotory activity and home range size Locomotory activity in wolf spiders mainly reflects the search for suitable microhabitats and sexual activity (Cady 1984; Enders 1975). Kuenzler (1958) found home ranges for lycosids to be smaller than those recorded forA. cinerea. The average home range size of Hogna timuqua (Wallace) is 41,7 m2 for males and 9,6 m 2 forfemales. Large home range sizes in A. cinerea are mainly caused by movements in autumn. The spiders retreat from the water edge before hibernation. This can be inter- preted as a search for a more protected area for the winter diapause (Framenau etal. 1996). Acknowledgement This work is part ofthe research project "Ecology and Conservation of Alpine Rivers". Funding was provided by the German Ministry for Science, Education, Research and Technology (support-no. 033 9530) and the Donors Association for the Promotion of Science and Humanities in Germany. The Government of Southern Bavaria (Oberbayern) gave permission to collect spiders in the "Karwendel und Karwendelvorgebirge Nature Reserve". REFERENCES Albert. A.M. & Albert. R. 1976. Abundance and biomass of Pardosa agricola (Thorell) (Araneae, Lycosidae) on a shingle bank ofthe River Lune (Lancashire). Bulletin ofthe BritishArachnologicalSociety3 (9): 237-242. Aviles, L. & Maddison, W.P. 1991. When is the sex ratio biased in social spiders? Chromosome studies of embryos and male meiosis in Anelosimus species (Araneae: Theridiidae).JournalofArachnology 19: 136-145.