Appendix A. Criterion for choosing abundant GP and criteria used to assign the abundant GP
to functional groups based upon laboratory feeding trials.
Criterion used to select the abundant predators for statistical analyses and laboratory feeding trials
Figure A1 shows the rank-abundance curve used to define the abundant predators: the spider genera Xysticus, Gnaphosa, Drassyllus, Phrurotimpus, Lathys, Titanoeca, Ariadna and Schizocosa, the spider family Erigonidae, and the centipede orders Lithobiomorpha and Scolopendromorpha.
|
|
| FIG. A1. Abundance (mean ± SE) of GP (spiders and centipedes) in the leaf litter of the forest floor. The rank-abundance curve was built by averaging the mean densities across the two years of the study (2001 and 2002). Because the Salticidae is a very speciose spider family and the same species is rarely found in two different samples (e.g., Jiménez-Valverde and Hortal 2003), we pooled all genera within the Salticidae to form a single group. Notice that if we were to remove the group Salticidae, which has many rare genera and species, there would be a substantial gap between Schizocosa and Latrodectus. Thus, taxa from Schizocosa to the left were categorized as “abundant GP.” Note that a density of c. 1 m-2 (dashed line) separates the “abundant” from the “rare” GP. Only the abundant GP were included in multivariate models testing responses of specific functional groups. All groups listed are spiders except for the Order Pseudoscorpionida, the centipede Orders Lithobiomorpha and Scolopendromorpha, and the scorpion Vaejovis. |
Criteria used to assign the abundant GP to functional groups based upon laboratory feeding trials
Methods
Trials were conducted in containers 12 cm in diameter and 8.5 cm high, with a cotton wick to provide moisture and two leaves to provide attachment points for the webs of the web builders. A single predator was allowed to settle in each container for at least 72 hours before ants were introduced. In the first set of trials, conducted from 11 May to 12 July 2001, we offered 5 individuals of the two most abundant ant genera (Aphaenogaster and Prenolepis) to individuals of each of the most abundant, large GP (spider genera Ariadna, Drassyllus, Gnaphosa, Schizocosa, Titanoeca and Xysticus; and centipede Orders Scolopendromorpha and Lithobiomorpha; Fig. A1). Number of replicates ranged from 3 to 42 depending on predator availability. Ants and predators were used in the order in which they were collected. In 2001 predators and ants were used only once. We considered that the predator had eaten an ant if the latter was dead and either dismembered or emptied of haemolymph. To ensure that ants would not kill or fight each other, all ants within each trial belonged to the same nest. Trials were run for 7 days and those in which the predator died from ant predation or other causes were discarded, yielding 205 valid trials.
The following year, from 12 June to 14 August, we offered a single individual of two other abundant ant genera, Camponotus and Crematogaster. Only the most abundant Camponotus species (Species 1) was used. As these workers are >5 mm long, we used only GP longer than 4 mm. We used only one ant individual in order to decrease the rate of predation by ants on the spiders, which was observed in 2001. Ants were used only once, but predators were re-used. They were allowed to acclimate to the container for 24 h before the ant was introduced; containers were checked 24 h later for evidence of predation on the ant. We conducted valid trials with 247 predators (range for each GP 8-69), which were tested with a variable number of ants, depending upon their availability.
Results
Most of the abundant GP were able to feed on at least two species of ants (Fig. A2), although there was high variation in the rate of feeding. Only Drassyllus and Scolopendromorpha were obvious non-ant feeders. During the trials in 2001 (5 ants offered to each predator at once), the spiders Drassyllus and Schizocosa were killed by the ants at a relatively high frequency (32% and 28%, respectively). During the 2002 trials, when only a single ant was introduced, only one Schizocosa individual was killed (by Camponotus). Taking the results of the feeding trials into account (Fig. A2) and information on foraging behavior of the groups (Gertsch 1979, Foelix 1996, Uetz et al. 1999), the different groups of abundant GP (Fig. A1) were assigned to functional groups as follows: (1) Active ant-feeding GP: the spiders Gnaphosa and Schizocosa, and the Lithobiomorpha centipedes. (2) Sit-and-wait ant-feeding GP: the spiders Ariadna, Titanoeca and Xysticus. (3) Active non-ant-feeding GP: the spider Drassyllus and Scolopendromorpha centipedes. (4) Small GP of unknown ability to feed on ants: the spiders Lathys, Phrurotimpus and Erigonidae.
|
| FIG. A2. Results (Mean ± SE) of laboratory trials in which four ant species were offered to the eight genera of abundant GP with the largest individuals. The trials with Aphaenogaster and Prenolepis were run during 2001 and were slightly different than the trials for the other two ant species, which were run in 2002. Predation rates for Aphaenogaster and Prenolepis represent the proportion eaten out of five ants simultaneously offered to each predator. In the case of Crematogaster and Camponotus, different numbers of ants were sequentially offered one at a time to each predator depending upon ant availability. In this latter case, predation rates were calculated as the proportion eaten (see text above for more details). |
LITERATURE CITED
Foelix, R. F. 1996. Biology of spiders. Oxford Univ. Press, Oxford, UK.
Gertsch, W. J. 1979. American spiders. Van Nostrand, New York, New York, USA.
Jiménez-Valverde A., and J. Hortal. 2003. Las curvas de acumulación de las especies y la necesidad de evaluar la calidad de los inventarios biológicos. Revista Ibérica de Aracnología 8:151–161.
Uetz, G. W., J. Halaj, and A. B. Cady. 1999. Guild structure of spiders in major crops. Journal of Arachnology 27:270–280.