THE INFLUENCE OF MICROHABITAT AND PREY AVAILABILITY ON BURROW ESTABLISHMENT OF YOUNG

The survival of an animal largely depends on its ability to locate and use a suitable habitat. The suitability of a habitat will depend on such things as prey availability, microhabitat characteristics and the interaction of these,but our understanding of the interaction of these factors and how they effect the animal’s choice of habitat is poor (Krebs, 1978). For spiders, there have been many studies which show the importance of habitat characteristics and prey abundance in the selection of foraging and web sites (e.g., Savory, 1930; Enders 1977; Riechert, 1976; Riechert and Tracy, 1975). However, most studies deal with adult spiders (some exceptions being the work of Waldorf, 1976; Enders, 1977 and Hallander, 1970) and our understanding of which factors may influence habitat selection in newly dispersing young spiders is limited. These factors are particularly important for burrowing wolf spiders (Geolycosa) since the selected burrow site is generally used throughout the life of the spider (Wallace, 1942)’. In this paper I test selected hypotheses about the interactions among burrow establishment, prey availability, and several microhabitat characteristics in two species of burrowing wolf spiders. This paper is not concerned with the relationship between the burrow site characteristics and survival of the spiderling, which is best studied in the field (see Reichert, 1976 and Reichert and Tracy, 1975 for


Introduction
The survival of an animal largely depends on its ability to locate and use a suitable habitat.The suitability of a habitat will depend on such things as prey availability, microhabitat characteristics and the interaction of these, but our understanding of the interaction of these factors and how they effect the animal's choice of habitat is poor (Krebs, 1978).
For spiders, there have been many studies which show the importance of habitat characteristics and prey abundance in the selection of foraging and web sites (e.g., Savory, 1930;Enders, 1977;Riechert, 1976;Riechert and Tracy, 1975).However, most studies deal with adult spiders (some exceptions being the work of Waldorf, 1976;Enders, 1977 andHallander, 1970) and our understanding of which factors may influence habitat selection in newly dispersing young spiders is limited.These factors are particularly important for burrowing wolf spiders ( Geolycosa ) since the selected burrow site is generally used throughout the life of the spider (Wallace, 1942) Geolycosa spp.are obligate burrowers which establish a burrow shortly after leaving the mother (Wallace, 1942;McCrone, 1964;pers. observ.)and, with the exception of short foraging sorties and the reproductive wanderings of mature males, live their entire lives within a burrow (Wallace, 1942;McCrone, 1964;McQueen, 1978;Humphreys, 1975;pers. observ.).Generally, newly-dispersing spiderlings construct burrows in the vicinity of the maternal burrow (McQueen, 1978).(1) Prey/ no prey (prey) -indicating whether food was provided during the experimental period.

Specimens
(2) Vegetation/ no vegetation (vegetation) -indicating whether small bits of grass were provided in the experiemental container.
(3) Crevice/ no crevice (crevice) -indicating whether a depression in the burrowing surface was provided.
(4) Burrow/ no burrow (burrow) -the "dependent" variable, indicating whether a burrow was constructed.Any burrow which was large enough to contain the spider was scored as an established burrow.
For each experiment I tested two hypotheses concerning a threeway contingency table defined by the variables prey, burrow and one of the other two variables (

Results
The first set of hypotheses tested the independence of burrowing with the variables vegetation and crevice respectively within the levels of the variable prey (Table 2).The hypothesis that burrowing is independent of vegetation within a prey group (fed or unfed) was rejected for G. turricola but not for G. micanopy (Table 2).
Both fed and unfed G. turricola showed a higher number of burrowers in the group that was provided with vegetation (86% and 43% burrowing in the fed and unfed groups respectively).For G.
micanopy, the number of burrowers is approximately the same for the fed spiders with and without vegetation.A higher percentage of the unfed G. micanopy burrowed when no vegetation was present (23% and 40% for the group with and without vegetation respectively).More spiders burrowed in the presence of a crevice than when no crevice was available for both fed and unfed G. micanopy.
The second set of hypotheses tested the independence of burrowing and prey given the level of each of the other two variables.In every case, the hypothesis of independence was rejected (Table 2).In nearly every case, the percentage of spiderlings constructing burrows was lower in the unfed group.This is true for each level of the two variables, vegetation and crevice except for the unfed G. micanopy who were provided a crevice.In that case, more spiderlings constructed burrows than did not.

Discussion
The results indicate that, of those variables considered, prey was most strongly associated with the establishment of a burrow.In nearly every case, groups of spiderlings which were given food had a 1984] Miller -Geolycosa 127 higher percentage of burrows than those that were not.Although the experimental design allows consideration of only the importance of the presence of prey, higher rates of burrow establishment in the presence of food may imply either a nutritional advantage or a response to prey availability.
With respect to nutritional differences, even though burrowing behavior is innate and may be observed as early as in late preemergent spiderlings which have not fed (pers.observ.), the successful establishment of a burrow may be more likely if a spiderling is able to obtain food prior to dispersing.Many lycosid spiderlings disperse after only a short time on the mother (e.g., about seven days, Higashi and Rovner, 1975) and do not feed prior to dispersal (Foelix, 1982).My observations of field populations of other Geolycosa indicate that spiderlings may cling to the mother for up to three weeks and some young remain in the maternal burrow for at least one molt after leaving the mother's abdomen as Engelhardt (1964) found in Trochosa spp.A possible advantage of this extended association with the mother is that those spiderlings that remain may receive nurishment by sharing prey captured by the mother, as in Sosippus floridanus (Brach, 1976), or by cannibalism.Hallander (1970) Eberhard, 1971), vegetation structure (e.g., Enders, 1975) or temperature (e.g., Riechert and Tracy, 1975) and prey characteristics such as prey availability (e.g., Kronk and Riechert, 1979;Enders, 1977;Morse, 1981) are known to influence positioning of webs and the location of foraging sites in spiders.
Geolycosa burrows function in both thermoregulation (Humphreys, 1975) and prey capture (Gertsch, 1942, pers. observ.),so the placement of the burrow may be related to these functions.
However, the major thermoregulatory attribute of the burrow is its depth and not its location relative to the surrounding vegetation (Humphreys, 1975).Geolycosa regulate body temperature by moving up or down the tunnel.The selection of a burrow site, therefore, is more likely to be related to prey availability and microhabitat factors relating to ease of construction or which provide some protection from predators.
The results show a difference between G. turricola and G.
micanopy in the relationship between vegetation availability and frequency of burrow establishment.Within a feeding state, the number of burrows established is independent of vegetation for G.
micanopy but not for G. turricola.This difference reflects the turret construction habits of the two species.Geolycosa turricola nearly always constructs a conspicuous turret from whatever material is available, whereas G. micanopy shows considerable variation in turret construction and often has burrows with no turret (Wallace, 1942).The different relationship between burrowing frequency and vegetation is probably not a result of a preference for vegetation material used in the experiments, since there appears to be no specificity for turret material in Geolycosa (Wallace, 1942).
Nearly all Geolycosa observed in the lab readily used artificially constructed burrows.Field observations of dispersing G. turricola (Miller and Miller, in prep.)indicate that over one half of burrows constructed by dispersing G. turricola spiderlings were built within a surface crack or depression.The results presented here also indicate a preference for burrowing when a surface irregularity is present.
Surface cracks and crevices could provide protection from predators and thermoregulatory advantage during the initial phase of

Table 1 .
Observed cell frequencies of burrow establishment for Geolycosa turricola and G. micanopy under experimental conditions.B = burrow established, NB = no burrow established.

Table 2
Sand was used as the burrowing medium in all cases.Spiderlings which were to receive food were given newly-hatched crickets at the beginning of the experimental period.A metal probe was used to observed cannibalism in Pardosa pullata of the