2007 SOUTHEASTERN NATURALIST 6(4):597–614 Scirtid Beetles (Helodes pulchella), Leaf Litter, and Treeholes: Is There Evidence of Facilitation in the Field? John Q. Burkhart1,2, Leslie Smith1,3, Shawn Villalpando1,4, and Christopher J. Paradise1,* Abstract - The primary resource of temperate forest treeholes is leaf litter, and different insects specialize on particular stages of decay. Helodes pulchella (scirtid beetle) takes part in a processing chain by shredding leaf litter and creating material for other consumers. We hypothesized that variation in scirtid density and resources influences the insect community. To test this, we manipulated scirtid beetle and resource densities in field mesocosms. We used a two-factor design (3 litter levels by 3 scirtid densities), and monitored insect communities from April 2004 to June 2005. We detected no statistically significant effects of scirtids on leaf decay. However, during the first season, species richness was higher in mid-summer in the presence of low scirtid density compared to treatments with high scirtid densities or those with no scirtids. The dominant species was Aedes triseriatus (eastern treehole mosquito) and its abundance was unaffected by either scirtid or leaf litter. However, mean pupal mass of mosquitoes was greater at low scirtid densities. Facilitation is suggested by the combination of high mosquito densities and large pupal mass. Densities of Culicoides guttipennis (ceratopogonid midge) were higher at intermediate and high resource levels without scirtids, compared to treatments with any scirtids, suggesting a negative interaction between midges and scirtids. We demonstrated strong, and not always positive, effects of scirtid beetles on insect communities in water-filled treeholes. Introduction The state and availability of resources may have wide-ranging effects at both the population and community levels (Hunter and Price 1992, Naeem 1988, Srivastava and Lawton 1998, Walker et al. 1997), and may be dependent upon processing chain interactions (Daugherty and Juliano 2002, Paradise 1999, Paradise and Dunson 1997). Processing chains, and other facilitative interactions, have been hypothesized to affect community diversity and individual populations (Bruno et al. 2003, Hacker and Gaines 1997, Jones et al. 1997). During feeding, processors condition resources for consumers that specialize on more advanced decay states (Heard 1994a). For instance, leaf shredders, by processing leaf litter, leave behind smaller particulate organic matter (Daugherty and Juliano 2002). A commensalism results as more resources, distributed in a wider range of particle sizes, are made available to 1Department of Biology, Davidson College, Davidson, NC 28035-7118. 2Current address - Department of Biology, University of South Alabama Mobile, AL, 36688- 0002. 3Current address - Graduate School of Oceanography, University of Rhode Island, MERL Room 4, Box 9, 11 Aquarium Road, Narragansett, RI 02882. 4Current address - ASU Box 12581, Appalachian State University, Boone, NC 28607. *Corresponding author - chparadise@davidson.edu. 598 Southeastern Naturalist Vol. 6, No. 4 consumers through the actions of the processor (Heard 1994b, Paradise 1999, Paradise and Dunson 1997). Processing chain amensalisms also may occur if resources are abundant, if consumers are dependent upon physical or microorganism processing, or if processors are highly efficient feeders (Heard 1994a). Under these conditions, consumers are less dependent upon processors, and foraging by processors has a net effect of reducing resources for consumers (Heard 1994a). Processing-chain interactions have been reported in streams, pitcher plants, Heliconia bracts, carcasses, treeholes, and bromeliads (Bradshaw 1983, Dieterich et al. 1997, Heard 1994b, Jonsson and Malmqvist 2005, Naeem 1990, Paradise 1999, Paradise and Dunson 1997, Schoenly and Reid 1987, Seifert and Seifert 1979, Srivastava 2006). While several of these experiments have shown that processors can increase the growth, mass, or survival of other consumers, studies of the effects of a processor on diversity and composition may lead to increased understanding of ecological communities (Bertness and Calloway 1994). Helodes pulchella Guerin (scirtid beetle), a common resident of treehole communities in the southeastern US, is a leaf shredder and a potentially important processor (Barrera 1996, Paradise 2004). Particulate or dissolved organic matter increases as scirtids shred leaf litter because not all particles are consumed by scirtids; unconsumed particles become available for filterfeeders and browsers (Daugherty and Juliano 2002, Paradise and Dunson 1997). As scirtids process litter, they also may increase the surface area available for microorganism growth. This can be important because microbes significantly add to the diet of detritivores in treeholes (Walker et al. 1991). In addition, fecal production by scirtids could increase resources for other detritivores (Daugherty and Juliano 2003). Finally, dead scirtids themselves, as they decay, could provide resources to microorganisms and other insects (Daugherty et al. 2000). Southeastern US treeholes are dominated numerically by larvae of Aedes triseriatus Say (eastern treehole mosquito; Harlan and Paradise 2006, Lounibos 1983), a filter-feeder and browser that consumes small particulate matter and microbes (Merritt et al. 1992). Other common insects include Culicoides guttipennis Coquillet (ceratopogonid midge), Mallota posticata Fabricius (syrphid), and Telmatoscopus albipunctatus Williston (psychodid). These species are detritivores and consume detritus in various stages of decay (Barrera 1988). In the laboratory, scirtid beetles facilitate both eastern treehole mosquitos and ceratopogonid midges, and positive impacts are greatest when resources are limiting (Paradise 1999, 2000; Paradise and Dunson 1997), as predicted by theory (Heard 1994a). The purpose of this study was to test the hypothesis that scirtid beetles increase treehole insect abundance and community diversity. These increases are resource-dependent, as differences in leaf litter abundance have large effects on diversity and individual species abundances (Paradise 2004, Schoener 1986) and because processing-chain interactions can change as resources increase (Paradise 1999). We predicted that a processing-chain commensalism between scirtids and other treehole detritivores would lead to increases 2007 J.Q. Burkhart, L. Smith, S. Villalpando, and C.J. Paradise 599 in insect diversity and abundance. Further, we predicted increases when resources are limiting and scirtids are present, compared to when resources are limiting and scirtids are either absent or very abundant (Heard 1994b). When present at high densities, scirtids decrease growth rates and survival of detritivores (Paradise 1999), and the net effect of high densities of scirtids when resources are limiting is to reduce resource availability to detritivores. At the community level, we thus predict lower diversity and overall abundance (but higher relative abundance) of common species in treatments with high densities of scirtids in comparison to treatments with lower densities, within the same resource level. To test these predictions about processing-chain interactions, we conducted a field mesocosm experiment in which leaf litter and scirtid beetle densities were varied independently of one another. To test the effects of these factors on colonization and community development, we monitored populations within mesocosms over time. Methods We created mesocosms using 7.62-cm internal diameter PVC pipe cut into 11-cm lengths. We use the term mesocosm to reflect the fact that these containers are similar in size to natural treeholes. These mesocosms are known to support the full range of treehole insects in our study area (Harlan and Paradise 2006). We affixed fiberglass window screening to the inside of the pipe by overlapping it beyond the two ends. The screen created a textured inner surface and darkened the interior, promoting insect oviposition and allowing scirtids to crawl to the top for air. We used silicon caulk to seal on an end cap, which held the overlapping screen in place. At the top, we used a PVC coupling that was sawed in half to seal the overlapping screen in place. We used caulk to seal this coupling and the screen to the top of the mesocosm. The total capacity of each mesocosm was approximately 540 ml. We attached mesocosms in pairs to a frame with expandable polyurethane foam (Great Stuff, Dow Chemical Co., Midland, MI). Frames were approximately 50 cm wide by 35 cm high by 25 cm deep and were constructed of 1.5-cm PVC pipe. We flushed mesocosms several times in the laboratory over three weeks to remove any volatile chemicals. The frames were tied to trees at a height of less than 1 m throughout a 30-acre hardwood forest on the Davidson College Ecological Preserve (DCEP, Davidson, NC; 30°30'30"N, 80°49'45"W) in September 2003. We glued fi- berglass window screen (2-mm mesh) to the top of each frame, about 25 cm above the mesocosms, to reduce the amount and size of natural debris that entered mesocosms. Finally, we wrapped each frame and the trunk of its tree in 2.5-cm mesh chicken wire to exclude most vertebrates, and cut a hinged door into the cage to allow researchers access (Paradise 2006). We randomly assigned one of ten treatment combinations to each mesocosm. We used three levels of leaf litter crossed with three levels of scirtid density, all of which were within the ranges found in treeholes (Paradise 2004; Paradise, unpubl. data), with four replicates of each treatment 600 Southeastern Naturalist Vol. 6, No. 4 combination. Leaf-litter levels were 1, 5, and 10 grams of dried Quercus rubra L. (red oak) leaves/L (hereafter low, intermediate, and high, respectively). We collected medium-sized scirtid larvae (mostly 2nd instar) from a large basal treehole present on the DCEP and added them at three densities: 0, 26, and 100 individuals/L (hereafter none, low, and high, respectively). The tenth treatment consisted of four mesocosms supplied only with water, giving a total of 40 mesocosms. The latter treatment was used to estimate debris accumulation, which was limited to dust and small particles that could fit through the 2-mm mesh covering the cages. The mass of particulate matter in these mesocosms could be used to subtract from total litter mass of other mesocosms to determine litter decay. We collected leaves in September 2003 and added them to mesocosms in early October 2003, along with a small aliquot of filtered treehole water and ≈500 ml of distilled water (which filled the mesocosms). We then covered mesocosms with plastic sheeting held in place with cable ties, to prevent oviposition. In late October 2003, scirtids were added at the appropriate densities. Mesocosms were covered with no-see-um netting (0.5-mm mesh), which allowed for gas exchange between the water column and the atmosphere but prevented colonization by insects. We timed the addition of leaf litter and scirtids to simulate the exposure of scirtids to leaf litter shortly after natural leaf fall. The mesocosms remained in this state until they were opened in early March 2004. In early April 2004, we began monitoring for colonization of insect larvae. Mesocosms were sampled every two to three weeks to determine when colonization began, to monitor scirtid survival and maintain constant water levels, and to eliminate any larvae of Toxorhynchites rutilus Coquillett (predatory mosquito), as we were interested only in bottom-up effects. Monitoring consisted of extracting 50 ml of water from a mesocosm using a baster from just below the water surface. A second 50-ml sample was taken subsequently from the bottom. In addition, observation of mesocosms allowed us to detect any predatory mosquito larvae swimming to the surface. We performed three complete censuses each in 2004 and 2005, which involved removing all water and leaf litter from each mesocosm using basters and forceps. All leaves were individually examined for insect larvae. The water was spread out into pans, and carefully examined for insect larvae. All were identifi ed to the lowest taxonomic level possible and counted. Some unknown larvae were brought to the laboratory for examination or to rear to adulthood for identifi cation. All other material, except for predatory mosquito larvae, was then placed back into its original mesocosm. After the second census in July 2004, we restocked scirtids, as needed, to their nominal levels. We collected mosquito pupae from mid-June to mid-August 2004 every three to seven days. This period coincided with emergence of the first generation of mosquitoes to inhabit mesocosms. Pupae were collected by catching them with a plastic pipet and returning them to the laboratory. We froze, weighed, sexed, and identified emergent adults to species. We took 2007 J.Q. Burkhart, L. Smith, S. Villalpando, and C.J. Paradise 601 only up to three pupae from any single mesocosm during a collection bout, as collecting all adults might negatively affect future colonization. Thus, individual biomass of emergent mosquitoes was used as an indicator of habitat productivity. Most (>90%) of the mosquito pupae collected were eastern treehole mosquitos, so we used only those individuals for statistical analysis. The remaining larvae were of Aedes albopictus (Skuse) (Asian tiger mosquito), and those individuals were not analyzed. Any predatory mosquito larvae observed were removed during these collections. We terminated the experiment after we completed the final census in early June 2005. We brought all leaf litter and coarse particulate matter back to the laboratory. As we were not able to determine decay of leaves during the experiment, this was our only method to assess leaf-litter condition, even if mesocosms also contained debris that had fallen in during the experiment. However, we determined that the material that collected in the four mesocosms without leaf litter was all <1 mm in size. We rinsed leaf litter from other mesocosms through a # 35 sieve, which allowed us to estimate the loss of mass from the initial stock. We determined the species richness (minus scirtids), Shannon-Weiner diversity (minus scirtids), percent dominance of the most common species, and densities of scirtids and the two most common insect larvae (eastern treehole mosquitoes and ceratopogonid midges). Because species richness estimates are biased by rare species, high values indicate presence of numerous rare species. To provide another indication of effects on species composition, we also calculated and performed statistical analysis on diversity (as -H = Σ (pi * ln (pi)), where pi = the proportion of species i individuals in the community) and percent dominance. Each of these might be influenced by presence of processors. We analyzed the proportion of leaf-litter mass lost using a two-factor analysis of variance (ANOVA) in MINITAB (Version 13.31 for Windows), and graphically examined the change in leaf-litter dry mass. For mass of eastern treehole mosquito adults, we used a nested repeated measures ANOVA, using individual mosquitoes as a random-effects factor nested within mesocosms, and leaf-litter and scirtid densities as fully crossed fixed-effects factors. For species richness, diversity, and densities of scirtids, mosquitoes, and ceratopogonid midges, we used profile analysis to test for effects of scirtid density and leaf litter level over time. Scirtids mostly were absent from mesocosms in 2005, so we analyzed response variables from 2004 only. In profile analysis, differences and averages at consecutive time points become transformed variables in two-way MANOVAs with leaf litter and scirtid densities as fixed-effects factors (von Ende 2001). Profile analysis allowed us to compare responses of any one variable over time without the variance problems associated with repeated measures ANOVAs that use time as a factor (von Ende 2001). Tests for interactions between time and other factors were done by comparing differences at sequential sampling points, while tests for main effects were performed by using the averages of successive time points. We adjusted α to 0.01 for the profile analyses (experiment-wise 602 Southeastern Naturalist Vol. 6, No. 4 α of 0.05 ÷ 5 profile analyses). All data were tested for univariate normality and homoscedasticity, and transformations were used where appropriate. Results Leaf litter and scirtids We found a significant leaf-litter effect on change in dry mass. Intermediate and high leaf-litter treatments lost about half of their dry mass, but low leaflitter treatments lost 75–85% of their dry mass (Fig. 1; F = 31.09; df = 2, 26; P < 0.001). This is a valid estimate of mass lost, as there was no coarse particulate matter recovered from mesocosms with no leaf litter. We found no effect of scirtid density or the interaction between scirtid and leaf-litter densities on proportion of leaf mass lost (scirtids: F = 0.14; df = 2, 26; P = 0.87; interaction: F = 0.27; df = 4, 26; P = 0.89). These results indicate that changes in leaf-litter dry mass were caused primarily by scirtid-independent processing. We found a significant leaf-litter effect on scirtid densities. Decline in scirtid densities was highest at low resources and high scirtid densities (100 individuals/L). Only in high leaf-litter mesocosms did scirtid densities remain near nominal by the time of the first census, while densities were lower than nominal in low and intermediate leaf-litter treatments (Table 1, Fig. 2). At the time of our first census (June 2004), we found few scirtids remaining in low and intermediate leaf-litter mesocosms, and thereafter scirtid densities did not vary greatly in those treatments, and remained at a consistently Figure 1. Proportional loss of dry mass from initial dry mass measured in October 2003 to final dry mass measured in June 2005. N = 4 for each mean in each graph. Sc/L = scirtid beetles per liter. 2007 J.Q. Burkhart, L. Smith, S. Villalpando, and C.J. Paradise 603 low level. Prior to that, however, sampling confirmed that numerous scirtids had survived the winter in the mesocosms. Insect communities We cumulatively found larvae or pupae belonging to 10 dipteran species. Major families observed were Culicidae (mosquitoes), Ceratopogonidae (biting midges), Psychodidae (moth flies), and Syrphidae (hover flies). Mosquito species included the eastern treehole mosquito, the Asian tiger mosquito, Orthopodomyia signifera (Coquillett) (white-lined moquito), and the predatory mosquito. For species richness tallies, we Table 1. Statistical results of profile analysis MANOVAs on 2004 censuses. All densities were log-transformed. For scirtids, only low and high densities of scirtids were used as levels in the MANOVA because we attempted to track changes in densities of the original cohort added. Pvalues < 0.01 were deemed significant (marked with *). LL = leaf-litter density, scirtid = scirtid density, and λ = Wilk’s λ. Response variable Source of variation λ df P Log H. pulchella density LL 0.43 4, 34 0.005* Scirtid 0.94 2, 17 0.616 LL x scirtid 0.84 4, 34 0.551 LL x date 0.94 4, 34 0.907 Scirtid x date 0.98 2, 17 0.863 LL x scirtid x date 0.89 4, 34 0.738 Log mosquito density LL 0.97 4, 52 0.949 Scirtid 0.94 4, 52 0.816 LL x scirtid 0.71 8, 52 0.304 LL x date 0.89 4, 52 0.549 Scirtid x date 0.98 4, 52 0.973 LL x scirtid x date 0.82 8, 52 0.724 Log ceratopogonid midge density LL 0.69 4, 52 0.042 Scirtid 0.77 4, 52 0.142 LL x scirtid 0.84 8, 52 0.801 LL x date 0.87 4, 52 0.445 Scirtid x date 0.62 4, 52 0.010* LL x scirtid x date 0.84 8, 52 0.787 Insect species richness LL 0.75 4, 52 0.105 Scirtid 0.82 4, 52 0.261 LL x scirtid 0.66 8, 52 0.183 LL x date 0.92 4, 52 0.721 Scirtid x date 0.63 4, 52 0.010* LL x scirtid x date 0.62 8, 52 0.119 Insect diversity LL 0.74 4, 52 0.088 Scirtid 0.75 4, 52 0.107 LL x scirtid 0.89 8, 52 0.930 LL x date 0.88 4, 52 0.481 Scirtid x date 0.71 4, 52 0.067 LL x scirtid x date 0.94 8, 52 0.990 604 Southeastern Naturalist Vol. 6, No. 4 counted all mosquitoes as a single species unless we positively identified more than one species (early instar eastern treehole mosquito and Asian tiger mosquito are difficult to distinguish). The insect community was dominated numerically by mosquitoes, with ceratopogonid midges as the second most commonly found taxonomic group. Mosquito larvae (not counting predatory mosquito) were found in 100% of all counts. When present, mosquitoes made up 69.4% of all larvae. In 85% of counts, across all combinations of leaf litter and scirtid density, mosquitoes were the dominant taxon. Mosquito populations peaked in June and steadily declined through the summer and fall of 2004 (Figs. 3a, b, c). However, we found no effect of leaf-litter or scirtid densities on the densities of mosquitoes (Table 1). Figure 2. Scirtid (H. pulchella) densities over time in each of nine treatment combinations. a. Scirtid densities in low leaf-litter mesocosms. b. Scirtid densities in intermediate leaflitter mesocosms. c. Scirtid densities in high leaf-litter mesocosms. N = 4 for each mean in each graph. 2007 J.Q. Burkhart, L. Smith, S. Villalpando, and C.J. Paradise 605 Although both eastern treehole mosquito and Asian tiger mosquito pupae were collected, Asian tiger mosquito sample sizes were insufficient for parametric analysis (<9% of 570 pupae collected). Eastern treehole mosquito adult mass demonstrated a sex-specific response to scirtid treatment. We Figure 3. Mosquito (eastern treehole mosquito and Asian tiger mosquito) and ceratopogonid midge densities over time in each of nine treatment combinations. Because eastern treehole mosquitos and Asian tiger mosquitos are difficult to identify to species in early instars, they were pooled together. a. Mosquito densities in low leaf-litter mesocosms. b. Mosquito densities in intermediate leaf-litter mesocosms. c. Mosquito densities in high leaf-litter mesocosms. d. Midge densities in low leaflitter mesocosms. e. Midge densities in intermediate leaf-litter mesocosms. f. Midge densities in high leaf-litter mesocosms. N = 4 for each mean in each graph. 606 Southeastern Naturalist Vol. 6, No. 4 found a significant leaf-litter by scirtid interaction for mass of adult males collected from mid-June to mid-August 2004, but no significant effects on female mass for the same period (Table 2). Low scirtid densities at low and high leaf-litter produced significantly larger males than high scirtid densities at low and high leaf-litter (Fig. 4). This does not appear to be a mosquito density-dependent effect; low scirtid/high leaf-litter mesocosms in June had statistically equivalent mosquito densities as high scirtid/high leaf-litter mesocosms, and yet produced larger male mosquitoes. Ceratopogonid midge larvae were found in 86.7% of samples, and when present, this midge made up 28.7% of all larvae and was the dominant species in 15% of all counts. Midge densities were significantly affected by the scirtid density by date interaction (Table 1). There was a trend towards higher midge densities with increasing leaf-litter levels (Figs. 3d, e, f). At both intermediate and high leaf-litter levels, midge densities were highest when no scirtids were present, and at high leaf-litter, there was an inverse relationship between scirtid and midge densities (Fig. 3f). Midge densities in different scirtid density treatments converged over time, leading to the significant interaction with date. Psychodids were found in 22.2% of counts, and appears to have two generations per year, as they were present in June and September, and almost completely absent from mesocosms in July. Densities of this larva were less than 1.0 individual/L, except in mesocosms with intermediate leaf-litter levels and high scirtid densities in September (2.5 ± 1.6 (SE) individuals/L). Very few psychodids were found in low leaf-litter mesocosms, regardless of scirtid density. Larvae syrphids were found only in June 2004 in low densities, in only 2.9% of all samples. They were never found in intermediate leaf-litter mesocosms, and only rarely in low leaf-litter mesocosms, with the highest densities in high leaf-litter mesocosms (0.03 ± 0.01 [SE] inidividuals/L in low leaf-litter and 0.26 ± 0.18 [SE] individuals/L in high leaf-litter mesocosms). There was no systematic trend in syrphid distribution across scirtid densities. Other larvae belonging to the families Dolichopodidae (Systenus sp.), Ephydridae, and Syrphidae (Myiolepta sp.) also were observed, all in less than 3% of all samples. Table 2. Results of repeated-measures ANOVA on mean pupal mass. Pupal data transformed and pooled by month. LL = leaf litter level, and scirtid = scirtid density. Statistically significant effects are marked with *. Response variable Effect F df P Log male mass LL 0.12 2, 50 0.886 Scirtid 1.55 2, 50 0.218 LL * scirtid 3.27 4, 50 0.015* Individual (LL * scirtid) 0.58 50, 95 0.983 Log female mass LL 0.76 2, 74 0.472 Scirtid 1.11 2, 74 0.333 LL * scirtid 1.51 4, 74 0.205 Individual (LL * scirtid) 2.12 74, 107 <0.001* 2007 J.Q. Burkhart, L. Smith, S. Villalpando, and C.J. Paradise 607 Figure 4. Mean adult eastern treehole mosquito mass in each of nine treatments. a. Adult male mass. Letters above bars denote statistical equivalence as determined by post-hoc pair-wise comparisons. b. Adult female mass. There were no significant effects of scirtid density (Sc/L) or leaf-litter level on female mass. 608 Southeastern Naturalist Vol. 6, No. 4 Mesocosms typically contained between one and four insect species at any time. No mesocosm contained more than 50% of the regional species pool, except one mesocosm with high leaf-litter and high scirtid density in June 2004 that contained 5 species. The median local species richness was 2, and mesocosms contained 2 or 3 species 84.2% of the time. Scirtid density and date interacted significantly to affect insect species richness Figure 5. Insect species richness and diversity over time in each of nine treatments. a. Species richness in low leaf-litter mesocosms. b. Species richness in intermediate leaflitter mesocosms. c. Species richness in high leaf-litter mesocosms. d. Diversity in low leaf-litter mesocosms. e. Diversity in intermediate leaf-litter mesocosms. f. Diversity in high leaf-litter mesocosms. N = 4 for each mean in each graph. 2007 J.Q. Burkhart, L. Smith, S. Villalpando, and C.J. Paradise 609 (Table 1; Figs. 5a, b, c). In July, there was approximately 1 more species per mesocosm on average in intermediate and high leaf-litter mesocosms with low scirtid densities, than in mesocosms with the same amount of leaf litter and either no scirtids or high scirtid densities (Figs. 5b, c). In previous and subsequent censuses, this trend was either absent or reversed. There were no leaf-litter effects on species richness and no significant effects of any kind on diversity in 2004 (Table 1). Diversity was low, being on average less than 0.8, presumably due to the dominance by mosquitoes and midges. Diversity was slightly lower in the presence of scirtids in September than in their absence, but was not statistically significant (Figs. 5d, e, f). Discussion Ecological theory suggests that processing chains may play a large role in structuring communities (Bruno et al. 2003, Heard 1994a). We thus predicted positive, facilitative impacts of scirtid beetles on treehole detritivore populations and diversity under conditions of low resources and low processor density; facilitation is most likely to occur when resources are limiting and efficiency of processing is low due to low density of shredders (Hacker and Gaines 1997, Heard 1994a, Paradise 1999, Paradise and Dunson 1997). Scirtids may be necessary to change leaf litter to a state available for other detritivores to consume (Daugherty and Juliano 2002, Heard 1994b). Positive effects of scirtid beetles on growth of eastern treehole mosquitos and ceratopogonid midges under conditions of low leaf-litter resources have been observed (Paradise 1999, Paradise and Dunson 1997). Scirtids did not have a significant effect on litter mass in this study, even though experimental evidence suggests that scirtids can increase availability of fine particulate organic matter (Daugherty and Juliano 2002, 2003). However, it is likely that their activity facilitated growth of microbes, on which other detritivores feed (Kaufman et al. 2000). This may have been one of the mechanisms that produced the observed effects. Studies of the effects of scirtid foraging on microorganism populations are needed to confirm this. Differences in scirtid activity during the winter and spring may have been enough to alter the response of communities with different scirtid densities. Despite the decline in scirtid densities over time, we demonstrated scirtid effects on ceratopogonid midge densities, the mass of adult eastern treehole mosquitos, and species richness. In the case of both species richness and adult eastern treehole mosquito size, our predictions were consistent with our findings; with ceratopogonid midge, our predictions did not match our observations. Densities of ceratopogonid midges were highest in mesocosms in which scirtids were absent, the effect being more pronounced in the high leaf-litter condition. The effect occurred during peak midge densities in May and June (Harlan and Paradise 2006, Paradise 2004), while scirtids had not yet declined in density, suggesting a strong negative interaction between scirtids and midges. Midges either did not survive in the presence of high densities 610 Southeastern Naturalist Vol. 6, No. 4 of scirtids, or females avoided ovipositing in those habitats, both of which would explain their absence in high scirtid-density mesocosms. If females did not avoid ovipositing in habitats with scirtids, then midge larvae either competed with scirtids for resources or were consumed by scirtids. Midges tend to crawl in and around sediment and detritus (Barrera 1988, Paradise 2004), creating a situation in which these small, early instar larvae could be accidentally consumed by foraging, shredding scirtids. However, Paradise and Dunson (1997) found no clear effect of scirtids on survival of midges in laboratory mesocosms, suggesting oviposition avoidance in the field, not low survival of hatching larvae. Midge oviposition appears also to be quite sensitive to other conditions, including presence of the top predator (L. Smith and C.J. Paradise, unpubl. data). Eastern treehole mosquito pupae were larger in the presence of low densities of scirtids, although this was not consistent across all leaf-litter levels. Mosquito mass is important for the survival of the mosquitoes in which both females and males experience greater fitness benefits from achieving a larger size either due to their ability to produce more eggs, or by increasing their longevity (Benjamin and Bradshaw 1994, Livdahl 1984, Lounibos et al. 1993). Adult mass attained by mosquitoes is density-and resource-dependent (Walker et al. 1997), yet we found some of the largest mosquitoes in treatments with low levels of leaf litter and high densities of mosquitoes—exactly where we would expect to see reductions in mosquito mass in the absence of facilitation. Large mosquitoes emerging from containers with low levels of leaf litter, high densities of mosquitoes, and low densities of scirtid processors provides strong evidence of facilitation. Treatment conditions in which we expected facilitation to occur, either from theory or past laboratory experiments (Daugherty and Juliano 2002; Paradise 1999, 2000; Paradise and Dunson 1997), but where we did not observe facilitation here, can likely be attributed to other biotic interactions also affecting growth and density of mosquitoes in this complex community. These exceptions aside, we still demonstrate a facilitative effect on mosquito growth under some conditions. Communities were dominated by eastern treehole mosquito and ceratopogonid midges, and dominance was not affected by either leaf litter or scirtid density. Diversity was predicted to increase with either resources or facilitative effects of processors, reflecting, in part, a predicted decrease in dominance and an increase in diversity of available resources (Armbruster et al. 2002, Hacker and Gaines 1997, Jones et al. 1997). Here, diversity was not affected by either scirtid densities or leaf litter, due to the continued high dominance of mosquitoes and midges, and also by the lack of effect of scirtids on leaf litter. However, species richness was affected by the scirtids, at least during some times. We found between 1 and 5 species in any one community, which is similar to other studies in both natural treeholes and mesocosms (Harlan and Paradise 2006, Paradise 2004, Yanoviak 1999). Small differences in local species richness, in communities in which maximum local richness is ≤5, represent large 2007 J.Q. Burkhart, L. Smith, S. Villalpando, and C.J. Paradise 611 proportional changes in species richness (Yanoviak 1999). In July, richness in treatments with intermediate or high leaf-litter and low scirtid densities remained high, while richness in other treatments declined. By September 2004, richness had dropped in the low scirtid-density treatment and had risen proportionately in the no and high scirtid-density treatments. This result was probably caused by rare species (most likely M. posticata, or T. albipunctatus) colonizing or persisting in low scirtid-density mesocosms through July and emerging before September. Individuals of these same or different species then colonized no and high scirtid-density mesocosms later and were counted in September. The increased presence, but low densities of rare insects explains the patterns in species richness, dominance, and diversity. In communities dominated by one or two species, an increase in the number of rare species might have no impact on diversity if dominance is maintained by one or two common species. The maintenance of rare species in treehole metacommunities probably comes about from adaptations to extreme habitats, habitat generality, variation in life cycles, and high vagility (Holyoak et al. 2005). For instance, the psychodid T. albipunctatus, one of the more frequently observed rare species, can be found in a wide range of habitats (Hribar et al. 2004), and is not an obligate treehole breeder, allowing a metapopulation to be maintained even if it is outcompeted within treeholes by dominant species. Other rare species also may be facultative treehole breeders or may be found only at certain times of the year, such as the syrphid M. posticata. To determine the factors that affect species richness in treeholes, it is critical to determine when a given species is more likely to be found. In addition to effects of scirtid densities, we predicted positive effects of leaf litter on treehole insect communities based on theory and previous research (Fish and Carpenter 1982, Léonard and Juliano 1995, Naeem 1990, Paradise 2004, Sota 1996). Specifically, we expected to observe higher densities and diversity of insect larvae, because greater availability of litter leads to greater habitat heterogeneity and space. Habitat heterogeneity significantly impacts species richness in phytotelmata communities and accounts for a high proportion of variation in species richness within these systems (Armbruster et al. 2002, Paradise 2004, Yanoviak 1999). This effect was not as strong within our mesocosms because natural phytotelmata have a wider range of heterogeneity, including variation in factors such as water volume and container size, factors that we held constant (Armbruster et al. 2002, Kitching 2000, Paradise 2004, Sota et al. 1994). The wider range of resource conditions in natural habitats may then interact with the effects of scirtids on species richness that we demonstrated. Previous research (Paradise 1999, Paradise and Dunson 1997) showed that scirtid beetles can have a significant positive impact on populations of detritivores within the communities of which they are a part. We found that positive effects of scirtid beetles in naturally colonized container communities were more prevalent when scirtid densities were low, as compared 612 Southeastern Naturalist Vol. 6, No. 4 to when they were high or when scirtids were absent from the community. Effects were neither always present nor facilitative, as predicted, and leaflitter effects were absent or moderated by scirtid presence. In our experiment, scirtid beetles positively affected species richness and negatively impacted the ceratopogonid midge population, although the effects were temporally variable. Scirtid beetles also positively affected the size of the dominant species in the treehole, the eastern treehole mosquito. Thus, we conclude that scirtids do affect treehole communities, possibly by creating conditions favorable to increased richness of rare species and increased mass of the dominant species. The effects we observed were conditional and mediated by other factors, such as resource abundance, lifecycles of the other treehole inhabitants, time of year, and dominance by single species. Acknowledgments We thank Duncan Berry, Charlie Chrisawn, Nicole Harlan, Lauren Harshaw, Benjamin Kegan, Stella Kenyi, Benjamin Kittinger, Lindsay Nakaishi, and Kate Williams for assistance in the field and with data management, and Davidson College for permission to work on the Davidson College Ecological Preserve. This research was supported by NSF grant DEB-0315208 to C.J. Paradise, NSF-REU grant DBI- 0139153 to the Davidson College Biology Department, and a Davidson College Faculty Study and Research grant to C.J. Paradise. The experiments conducted comply with the current laws of the United States of America. Literature Cited Armbruster, P., R.A. Hutchinson, and P. Cotgreave. 2002. Factors influencing community structure in a South American tank bromeliad fauna. Oikos 96:225–234. Barrera, R. 1988. Multiple factors and their interactions on structuring the community of aquatic insects of treeholes. Ph.D. Dissertation. The Pennsylvania State University, University Park, PA. Barrera, R. 1996. Species concurrence and the structure of a community of aquatic insects in treeholes. Journal of Vector Biology 21:66–80. Benjamin, S.N., and W.E. Bradshaw. 1994. Body-size and flight-activity effects on male reproductive success in the pitcher plant mosquito (Diptera: Culicidae). Annals of the Entomological Society of America 87:331–336. Bertness, M.D., and R. Calloway. 1994. Positive interactions in communities. Trends in Ecology and Evolution 9:191–193. Bradshaw, W.E. 1983. Interaction between the mosquito Wyeomyia smithii, the midge Metriocnemus knabi, and their carnivorous host Sarracenia purpurea. Pp. 161–189, In J.H. Frank and L.P. Lounibos (Eds.). Phytotelmata: Terrestrial Plants as Hosts for Aquatic Insect Communities. Plexus, Medford, NJ. 293 pp. Bruno, J.F., J.J. Stachowicz, and M.D. Bertness. 2003. Incorporating facilitation into ecological theory. Trends in Ecology and Evolution 18:119–125. Daugherty, M.P., and S.A. Juliano. 2002. Testing for context dependence in a processing-chain interaction among detritus-feeding aquatic insects. Ecological Entomology 27:541–553. Daugherty, M.P., and S.A. Juliano. 2003. Leaf-scraping beetle feces are a food resource for treehole mosquito larvae. American Midland Naturalist 150:181– 184. 2007 J.Q. Burkhart, L. Smith, S. Villalpando, and C.J. Paradise 613 Daugherty, M.P., B.W. Alto, and S.A. Juliano. 2000. Invertebrate carcasses as a resource for competing Aedes albopictus and Aedes aegypti (Diptera: Culicidae). Journal of Medical Entomology 37:364–372. Dieterich, M., N.H. Anderson, and T.M. Anderson. 1997. Shredder-collector interactions in temporary streams of western Oregon. Freshwater Biology 38:387–393. Fish, D., and S.R. Carpenter. 1982. Leaf litter and larval mosquito dynamics in treehole ecosystems. Ecology 63:283–288. Hacker, S.D., and S.D. Gaines. 1997. Some implications of direct positive interactions for community species diversity. Ecology 78:1990–2003. Harlan, N.P., and C.J. Paradise. 2006. Do habitat size and shape modify abiotic factors and communities in artificial treeholes? Community Ecology 7:211–222. Heard, S.B. 1994a. Processing-chain ecology: Resource condition and interspecific interactions. Journal of Animal Ecology 63:451–464. Heard, S.B. 1994b. Pitcher-plant midges and mosquitoes: A processing chain commensalism. Ecology 75:1647–1660. Holyoak, M., M. Leibold, and R.D. Holt. 2005. Metacommunities: Spatial Dynamics and Ecological Communities. University of Chicago Press, Chicago, IL. 520 pp. Hribar, L.J., J.J. Vlach, D.J. DeMay, S.S. James, J.S. Fahey, and E.M. Fussell. 2004. Mosquito larvae (Culicidae) and other Diptera associated with containers, storm drains, and sewage treatment plants in the Florida Keys, Monroe County, Florida. Florida Entomologist 87:199–204. Hunter, M.D., and P.W. Price. 1992. Playing chutes and ladders: Heterogeneity and the relative roles of bottom-up and top-down forces in natural communities. Ecology 73:724–732. Jones, C.G., J.H. Lawton, and M. Shachak. 1997. Positive and negative effects of organisms as physical ecosystem engineers. Ecology 78:1946–1957. Jonsson, M., and B. Malmqvist. 2005. Species richness and composition effects in a detrital processing chain. Journal of the North American Benthological Society 24:798–806. Kaufman, M.G., E.D. Walker, D.A. Odelson, and M.J. Klug. 2000. Microbial community ecology and insect nutrition. American Entomologist 46:173–184. Kitching, R.L. 2000. Food Webs and Container Habitats: The Natural History and Ecology of Phytotelmata. Cambridge University Press, Cambridge, UK. 446 pp. Léonard, P.M., and S.A. Juliano. 1995. Effect of leaf litter and density on fitness and population performance of the treehole mosquito Aedes triseriatus. Ecological Entomology 20:125–136. Livdahl, T.P. 1984. Interspecific interactions and the r-K continuum: Laboratory comparisons of geographic strains of Aedes triseriatus. Oikos 42:193–202. Lounibos, L.P. 1983. The mosquito community of treeholes in subtropical Florida. Pp. 223–246, In J.H Frank and L.P. Lounibos (Eds.). Phytotelmata: Terrestrial Plants as Hosts for Aquatic Insect Communities. Plexus, Medford, NJ. 293 pp. Lounibos, L.P., N. Nishimura, and R.L. Escher. 1993. Fitness of a treehole mosquito: Influences of food type and predation. Oikos 66:114–118. Merritt, R.W., R.H. Dadd, and E.D. Walker. 1992. Feeding behavior, natural food, and nutritional relationships of larval mosquitoes. Annual Review of Entomology 37:349–376. Naeem, S. 1988. Resource heterogeneity fosters coexistence of a mite and a midge in pitcher plants. Ecological Monographs 58:215–227. Naeem, S. 1990. Resource heterogeneity and community structure: A case study in Heliconia imbricata phytotelmata. Oecologia 84:29–38. 614 Southeastern Naturalist Vol. 6, No. 4 Paradise, C.J. 1999. Interactive effects of resources and a processing-chain interaction in treehole habitats. Oikos 85:529–535. Paradise, C.J. 2000. Effects of pH and resources on a processing-chain interaction in simulated treeholes. Journal of Animal Ecology 69:651–658. Paradise, C.J. 2004. Relationship of water and leaf-litter variability to insects inhabiting treeholes. Journal of the North American Benthological Society 23:793–805. Paradise, C.J. 2006. Experimental design. Davidson College Biology Department. Available online at http://www.bio.davidson.edu/people/chparadise/treehole/ methods.html. 15 June 2006. Paradise, C.J., and W.A. Dunson. 1997. Insect species interactions and resource effects in treeholes: Are helodid beetles bottom-up facilitators of midge populations? Oecologia 109:303–312. Schoener, T.W. 1986. Resource partitioning. Pp. 91–126, In J. Kikkawa and D.J. Anderson (Eds.). Community Ecology: Pattern and Process. Blackwell Scientific, Oxford, UK. 432 pp. Schoenly, K., and W. Reid. 1987. Dynamics of heterotrophic succession in carrion arthropod assemblages: Discrete seres or a continuum of change? Oecologia 73:192–202. Seifert, R.P., and F.H. Seifert. 1979. A Heliconia insect community in a Venezuelan cloud forest. Ecology 60:462–467. Sota, T. 1996. Effects of capacity on resource input and the aquatic metazoan community structure in phytotelmata. Researches on Population Ecology 38:65–73. Sota, T., M. Mogi, and E. Hayamizu. 1994. Habitat stability and the larval mosquito community in treeholes and other containers on a temperate island. Researches on Population Ecology 36:93–104. Srivastava, D.S. 2006. Habitat structure, trophic structure, and ecosystem function: Interactive effects in a bromeliad-insect community. Oecologia 149:493–504. Srivastava, D.S., and J.H. Lawton. 1998. Why more productive sites have more species: An experimental test of theory using treehole communities. American Naturalist 152:510–519. von Ende, C. 2001. Repeated-measures analysis: Growth and other time-dependent measures. Pp. 134–157, In S.M. Scheiner and J. Gurevitch (Eds.). Design and Analysis of Ecological Experiments. Oxford University Press, Oxford, UK. 432 pp. Walker, E.D., D.L. Lawson, R.W. Merritt, W.T. Morgan, and M.J. Klug. 1991. Nutrient dynamics, bacterial populations, and mosquito productivity in treehole ecosystems and microcosms. Ecology 72:1529–1546. Walker, E.D., M.G. Kaufman, M.P. Ayres, M.H. Riedel, and R.W. Merritt. 1997. Effects of variation in quality of leaf detritus on growth of the eastern treehole mosquito, Aedes triseriatus (Diptera: Culicidae). Canadian Journal of Zoology 75:706–718. Yanoviak, S.P. 1999. Effects of leaf-litter species on macroinvertebrate community properties and mosquito yield in Neotropical treehole microcosms. Oecologia 120:147–155.