Tri-trophic interactions among plants, herbivorous insects, and their natural enemies are a major component of most terrestrial ecosystems. These 3 groups comprise a grand majority of the Earth's described species (Strong et al. 1984). As a consequence, human disturbance of terrestrial ecosystems has a strong potential to produce a cascading impact on biodiversity and community structure across trophic levels. For example, human-mediated disturbance has been shown to have significant consequences for insect communities by influencing species diversity, insect abundance, and community structure (Hamer et al. 1997; Denys & Schmidt 1998; Rambo & Faeth 1999; Botes et al. 2006; Jones & Paine 2006; Comor et al. 2008; Tscharntke et al. 2008). Moreover, disturbance induced changes in plant communities and plant architecture can have cascading effects on plant-feeding insects and their predators (Rambo & Faeth 1999; Scheirs & De Bryun 2002; Thies et al. 2003; Moreau et al. 2006; Tylianakis et al. 2007).

Urbanization is an obvious form of human-mediated disturbance that has been shown to have both rapid and extreme impacts on arthropod abundance, distribution, and community composition (reviewed by McIntyre 2000). For instance, Rango (2005) showed that community composition of arthropods on creosote bush differed substantially between urban and fringe deserts, and that one-third of the species found in fringe deserts were absent from urban sites. Although a number of studies have examined how urbanization affects insect communities, relatively few have addressed how urbanization impacts tritrophic interactions. Examining how urbanization affects interactions among plants, phytophagous insects, and natural enemies is essential for understanding the additive and non-additive effects that these disturbances have on biodiversity.

Florida scrub is a rare ecosystem that has been heavily impacted by urbanization (Peroni & Abrahamson 1986). Fire suppression, habitat fragmentation, and land clearing are the major threats of concern (Myers 1990). Florida scrub has a high level of endemism, especially of plants, and covers areas along the Lake Wales Ridge of central Florida (Estill & Cruzan 2001). Primarily due to habitat loss, this ecosystem is home to approximately 22 federally listed threatened and endangered plant species. Currently, Florida scrub is highly fragmented and consists of isolated habitat islands encompassing both protected and non-protected areas along the Lake Wales Ridge. In non-protected areas, agriculture and housing developments have created a patchwork mosaic of scrub habitat. Thus, urbanization may have significant effects on species diversity and abundance due to reduction in population sizes through habitat loss and reduction in dispersal among isolated patches.

The purpose of this study was to examine how urbanization influences tri-trophic interactions among the endemic Florida scrub plant Palafoxia feayi A. Gray (Asteraceae), a gallmaking midge, Asphondylia sp. (Diptera: Cecidomyiidae), and the associated parasitoid guild. We compared plant architecture, number of galls, and parasitoid species diversity between pristine and urban sites situated in and surrounding the Archbold Biological Station in Lake Placid, Florida. In particular, we addressed the following questions: (1) Do plants growing in urban vs. pristine sites differ in architecture? (2) If so, does this difference correspond to a change in the incidence of gall formation? (3) Is parasitoid species diversity and abundance increased or decreased in urban sites?

MATERIALS AND METHODS

Surveys for plants, galls, and insects were conducted during May and Jun 2008 in 2 areas near Lake Placid, Florida—the Archbold Biological Station (ABS), 27°11′ 20″ N, 81° 20′ 26″ W, and vacant lots in the Placid Lakes housing subdivision (PL), 27°15′ 22 N, 81°22′ 57″ W. These areas differ in degree of urbanization. The Archbold Biological Station contains some of the last remaining contiguous tracts of Florida scrub. Sites at ABS are pristine and are only disturbed by frequent natural fires and prescribed burns. In contrast, Florida scrub in the Placid Lakes subdivision is subject to land clearing associated with housing development. Many of the urban sites used in this study were vacant lots that had been cleared for prospective development, although some of the sites were relatively undisturbed patches of scrub that were frequently subdivided by paved roads. As a result, urban scrub sites are severely fragmented by roads and development of surrounding lots.

At each of these locations we surveyed 10 m by 10 m plots separated by a minimum of 100 m. Plots were haphazardly selected and only plots containing P. feayi were included in the surveys. Initially, plants were surveyed to assess whether the presence of galls was correlated with plant architecture. Ten plots at ABS and 20 plots at PL were surveyed. Each plant with a minimum stem length of 30 cm was measured for a number of architectural features: main stem length to the nearest cm, the number of side branches, and the number of galls. Plants of this size were used because individuals were more easily located in the scrub and were more likely to have flowered (Asphondylia primarily forms galls in floral tissue). This survey provided enough plants to test for a correlation between plant architecture and gall presence. Subsequently, 20 plots at ABS and 15 plots at PL were surveyed for the number of plants and the number of galls per plant. This second survey allowed us to test whether there was a difference in the average number of galls per plot in pristine versus urban sites.

We examined parasitoid diversity by collecting 100 galls from plants at both ABS and PL. The galls were individually placed into 88.7-mL plastic cups and covered with Glad Press'N Seal® wrap. Galls were checked daily and emergence of insects was recorded. Insects were placed in 1.5mL microcentrifuge tubes containing 70% ethanol and stored in a freezer at -2O°C for later identification. If a gall appeared infected with fungus or became severely desiccated, it was dissected immediately to salvage as many collectable, identifiable insects as possible. After a period of 30 d all galls were dissected to determine the abundance of the remaining insects. Parasitoids were identified to genus with keys in Goulet & Huber (1993) and Gibson et al. (1997).

We examined how plant architecture influenced the presence of galls by using linear regression to examine how plant height and the number of side branches related to the number of galls per plant. We also examined whether plant architecture differed between sites by using a t-test to compare the length of the main stem of plants at ABS and PL and a Mann-Whitney U-test to compare the number of side branches per plant at these sites. For comparisons of urban and pristine plots, we used t-tests to determine whether the average number of plants per plot, the average number of galls per plot, and the average gall density per plot differed between sites. Finally, we compared parasitoid species diversity between ABS and PL via the Shannon-Wiener index (a joint measure of species richness and evenness). A chi-square test was used to determine if parasitoid abundance differed with degree of urbanization. All statistical analyses were performed with JMP 5.01 (SAS Institute).

RESULTS

A total of 1473 plants was surveyed—587 (104 from ABS and 483 from PL) plants were examined for plant morphology and number of galls, and an additional 886 plants for surveys of gall occurrence only (423 from ABS and 463 from PL). Plots from PL had more plants and more galls, but the average gall density per plot did not differ with urbanization (Table 1). Overall, both plant height and the number of side branches were positively correlated with the number of galls per plant; however, neither characteristic explained a large proportion of the variation associated with the presence of galls (Fig. 1). Plants from ABS and PL did not differ in height (ABS = 105.91 cm ± 4.22 (SE), PL = 100.02 cm ± 1.96; t = 1.267, df = 585, P = 0.21), but did differ in the number of side branches per plant (Mann-Whitney U-test, U = 24448, P < 0.0001). PL plants had slightly more and a greater range of side branches (range 1 to 34, median = 3.00) than ABS plants (range 1 to 9, median = 2.00). As a result, the average number of galls per individual plant was slightly greater for PL than ABS (PL = 0.57 ± 0.04 (SE), ABS = 0.42 ± 0.06; t = 2.105, df = 1471, P = 0.0354).

TABLE 1.

SURVEY RESULTS FOR ASPHONDYLIA GALLS ON PALAFOXIA FEAYI IN PRISTINE AND URBAN PLOTS (MEAN ± SE). STATISTICAL SIGNIFICANCE BASED ON T-TESTS.

Wasps obtained from P. feayi galls were from 4 genera—Galeopsomyia (Eulophidae), Rileya (Eurytomidae), Tenuipetiolus (Eurytomidae), and Torymus (Torymidae)—and all genera were represented at both sites (Table 2). Of the 100 galls collected from each site, parasitoids were more frequently found in galls from PL (38 galls from PL and 24 galls from ABS, χ2 = 4.58, df= 1,P = 0.032). The rearings and dissections produced 161 hymenopterans, 53 from ABS and 108 from PL. Although more parasitoids were reared from galls at PL, the average rate of successful parasitism per gall was not different between sites (t = 0.988, df= 49, P = 0.328). Similarly, for plots that yielded parasitoids, Shannon-Wiener diversity indices were similar between ABS and PL (ABS = 0.665 ± 0.16 (SE), PL = 0.566 ± 0.232; t = 0.360, df = 9,P = 0.72).

DISCUSSION

The Florida scrub ecosystem is considered to be one of the most endangered ecosystems in Florida and North America (Scott 2003). Urbanization and conversion to agricultural lands have fragmented Florida scrub into a patchwork consisting of a few large, pristine sections and a number of small areas interspersed among urbanized patches. Understanding how the native scrub subsisting in this mosaic has been impacted by urbanization is an important first step to make informed decisions in scrub conservation. Here we compared how urbanization may influence tritrophic interactions among an endemic scrub plant, a gallmaking midge, and the associated parasitoids.

Comparisons of plants in urban versus pristine scrub indicated substantial differences in plant architecture. Plants growing in urban sites did not differ in overall height from plants growing in pristine scrub, but urban plants did have significantly more and a greater range of side branches. Unlike other Palafoxia species, P. feayi is a perennial and continues growing by initiating side branches on already existing stems that have finished flowering (McIntyre unpublished data). One potential reason for the difference in the number of side branches is that plants growing in housing subdivisions are less likely to experience fire. Although P. feayi can resprout after fire, all aboveground stems of the plant are destroyed. Plots located at the Archbold Biological Station are regularly subjected to natural and prescribed burns to maintain the Florida scrub ecosystem (Ostertag & Menges 1994). Hence, plants have to re-initiate above ground growth more often at ABS. The higher density of P. feayi at urban sites suggests that the plants may potentially avoid mortality associated with fire.

Fig. 1.

Linear regression of the number of galls produced by the gall midge Asphondylia in comparison to plant height and the number of side branches for Palafoxia feayi. Only plants taller than 30 cm that had galls were included (n = 145 plants).

Although there were clear differences in the architecture of plants with the degree of urbanization, this did not transcend into differences at higher trophic levels. At the individual plant level, we detected a significant correlation between gall density and plant height and gall density and number of side branches; however, this pattern only explained a small proportion of the variation. Plants at urban sites had more side branches and a higher density of galls per plant than plants in pristine scrub. At the plot level, however, gall density did not differ with degree of urbanization. Urban plots had more plants with a greater number of side branches, but the average density of galls per plot was not significantly different from pristine sites. This was surprising given that differences in the number of plants and the number of side branches has direct implications for the number of galls that are possible at a plot. Because each side branch terminates in a flower head that is the primary site of Asphondylia gall formation, this result suggests that Asphondylia has not responded to the increase in availability of potential oviposition sites or that other factors may limit midge populations at these sites.

TABLE 2.

DISTRIBUTION OF HYMENOPTERANS REARED OR DISSECTED FROM EQUAL NUMBERS (100) OF GALLS ON PALAFOXIA FEAYI FROM PRISTINE AND URBAN FLORIDA SCRUB. (MEAN ± S.E.; N.S. = NOT SIGNIFICANT).

One possible factor that may limit midge populations in urban areas is suppression by parasitoids. Indeed, more parasitoids were reared from galls collected in disturbed sites and these sites also exhibited a higher proportion of parasitized galls (38% parasitized versus 24% in pristine sites). These results, however, should be interpreted with caution as studies have shown year to year variation and among site variation in parasitoid attack rate (e.g., Heard et al. 2006). The question remains whether these modest differences in parasitism translate into consistently higher mortality rates of the gallmakers across years.

Differences in the degree of urbanization did not appear to influence species richness of parasitoids attacking Asphondylia. The parasitoid communities in urban and pristine sites were similar. All 4 species were found at both sites and Shannon-Wiener indices of diversity were similar. Whether these parasitoid species are specialists on Asphondylia utilizing P. feayi is unclear. Species from the same 4 genera are found on Asphondylia borrichiae that uses sea oxeye daisy Borrichia fructescens along Florida's Gulf Coast (Stiling et al. 1992; Rossi et al. 2006). If parasitoids that attack Asphondylia are recruited from other gallmakers in the local community, then there may be a stable species pool of parasitoids in scrub habitats that can use galling larvae on P. feayi. If this were the case, changes caused by urbanization would need to influence the entire host-parasitoid insect community in order to potentially change parasitoid species diversity on urban P. feayi.

Urbanization has direct impacts on plant community structure and plant architecture of the Florida endemic P. feayi. These impacts, however, do not appear to significantly influence utilization patterns by the galling midge Asphondylia and its associated parasitoid community.