The yellow garden spider, Argiope aurantia Lucas (Araneae: Araneidae), is widespread in North America, whereas the Florida garden spider, Argiope florida Chamberlin & Ivie (Araneae: Araneidae), is restricted to scrub habitats in the southeastern USA. Both orb-weavers occur syntopically in native oak scrub at the Archbold Biological Station in Highlands County, Florida. We hypothesized that female spiders of both species might have a high degree of dietary niche overlap because their orb-webs are located near each other, despite the fact that they may differ in body size and web characteristics. We located 180 side-by-side pairs of mature A. aurantia and A. florida, and assigned them at random into 3 treatment groups of 60 pairs each. Argiope aurantia in group 1 were more heavy-bodied and longer legged than A. florida. In group 2, the orb-webs of both species were hung at the same height aboveground, but A. aurantia webs were larger with fewer sticky spirals than those of A. florida. Over the course of 5 d, the 2 spider species in group 3 captured almost the same number of aerial arthropods (80 for A. aurantia; 76 for A. florida). We identified most prey (94%) to genus or species. Dietary niche overlap at the ordinal level was extensive (70%), but it shrank dramatically when calculated at the generic/ species level (20%). We concluded that not only were A. aurantia consuming substantially larger prey than A. florida, but also that they were feeding on different kinds of aerial insects moving through gaps in oak scrub.
Spiders are among the dominant predators on earth (Wise 1993; Bond et al. 2014; Cushing 2017). Although they cannot fly, the ability of most species to disperse long distances by ballooning on the wind allows spiders to colonize all continents except Antarctica (Foelix 2011; Bond et al. 2014). The immense diversity of spiders both in species number (some 47,700 species in > 100 families are currently recognized) and in body size permits exploitation of a wide range of habitats and arthropod prey (Bradley 2013; Foelix 2011; World Spider Catalog 2018). A primary adaptation of spiders is their widespread use of webs for prey capture; the complex architectural geometry and astonishing material properties of silk extend the functional phenotype of these predators by orders of magnitude (Venner & Casas 2005 and references therein; Zschokke et al. 2006; Blamires 2010).
Predation by orb-weaving spiders has received considerable scientific attention. Numerous field studies show that the majority of captured prey are small relative to the size of host spiders, plus the spectrum of insects and spiders taken from orb-webs is not a random sampling of the number and taxonomic diversity of aerial insects and spiders in the environment (Nentwig 1982, 1985,1987; Venner & Casas 2005; Zschokke et al. 2006; Blamires 2010). A prime example of such research focuses on araneid spiders in the genus Argiope (Araneae: Araneidae). These spiders are convenient because they construct large, flat, long-lived orb-webs that are easily detected hanging vertically in old fields, grasslands, and shrubby habitats (Levi 1968; Enders 1973, 1976). They are diurnally active, and the prey they wrap in silk may be identified readily to taxonomic order in the field, or removed for further study (Horton & Wise 1983; Stowe 1986). For example, surveys of predation by mature female yellow garden spiders, Argiope aurantia Lucas (Araneae: Araneidae), at 6 temperate localities in the eastern US found that 4 orders of insects accounted for three-quarters of 1,430 prey: Hymenoptera (38.5%), Orthoptera (16.3%), Coleoptera (10.9%), and Lepidoptera (10.8%) (Uetz et al. 1978; Brown 1981; Howell & Ellender 1984; McReynolds & Polis 1987; Blackledge & Wenzel 1999).
At the subtropical Archbold Biological Station in Highlands County in south-central Florida, numerous A. aurantia often occur close to Argiope florida Chamberlin & Ivie (Araneae: Araneidae), the Florida garden spider (Carrel 2008). These colorful congeners are similar in size, phenology, and orb-web structure; however, their geographic distributions are vastly different: A. aurantia ranges widely across North America from southeast Canada south to Guatemala, whereas A. florida is highly restricted to relatively xeric sand scrub from eastern North Carolina to south Florida (Levi 1968, 2004; Justice et al. 2005.).
We hypothesized that, even though their orb-webs are located more or less at the same height side-by-side in oak scrub at Archbold, the 2 species might have significant differences in their trophic niches that reflect subtle differences in the size of mature females and their webs. Alternatively, knowing that competitive interactions usually are rare or nonexistent between these spiders (Wise 1993), the 2 Argiope species might show extensive prey niche overlap, sensu Pianka (1973).
Furthermore, we took advantage of the fact that a remarkably high proportion of the insect and spider fauna found at Archbold have been identified taxonomically over the course of decades of diligent work by many entomologists, and now are deposited in the Station's Collection of Arthropods. Hence, we were uniquely equipped to assess dietary niche overlap at 3 taxonomic levels (order, family, and genus/ species), which to our knowledge has never been reported in a study of this kind.
Materials and Methods
STUDY AREA
The Archbold Biological Station is located 12 km south of the town of Lake Placid in Highlands County, Florida, near the southern terminus of the sandy Lake Wales Ridge (27.1833333°N, 81.3500000°W). The predominant vegetative associations in the study area, comprising 280 ha of the Station that is very flat (elevation 38–42 masl), are scrubby flatwoods, which are dominated by low (1–2 m high) shrubby oaks (Quercus inopina Ashe, Quercus chapmanii Sargent, Quercus geminata Small; all Fagaceae) and palmettos (Serenoa repens (Bartram) Small, and Sabal etonia Swingle ex Nash; both Arecaceae). Interspersed among the scrubby flatwoods to varying degrees are 2 other vegetative associations: sand pine scrub, with widely scattered stands of sand pine (Pinus clausa (Chapman) Sarg.; Pinaceae), and an understory of xerophytic shrubs and flatwoods, with open stands of south Florida slash pine (Pinus elliottii var. densa Little & Dorman; Pinaceae) and an understory and ground cover of mesic grasses, herbs, saw palmetto (S. repens), and assorted shrubs (Abrahamson et al. 1984; Menges 1999; Deyrup & Deyrup 2012). The entire study area had regrown after it was consumed 5 yr earlier by a wildfire of unparalleled high intensity that produced a burn of unprecedented homogeneity at Archbold; all shrubs and forbs had been burned to ground level (Abrahamson & Abrahamson 2002).
In Sep 2006, both orb-weavers were numerous (60–80 newly matured females of each per ha) across most of Archbold. We located 180 pairs of A. aurantia and A. florida females whose orb webs were only 5 to 10 m apart. Each pair was assigned at random to 1 of 3 treatment groups: spider body size, orb-web characteristics, or prey determination. No spiders were harmed in this study.
SIZE OF ARGIOPE SPP.
We used 2 body size measures as indicators of foraging success in A. aurantia and A. florida females. Spider wet weight is known to reflect lifetime foraging success, whereas length of the first leg reflects foraging success prior to the final molt (Herberstein & Heiling 1998 and references therein).
We carefully removed 60 spiders of both Argiope species from their webs and took them to the laboratory where they were anesthetized with carbon dioxide gas and weighed to the nearest 0.1 mg on an electronic balance. Subsequently, the length of the first right leg (tarsus to coxa) was measured to the nearest 0.1 mm using a caliper (Model 144MM, General Tools and Instruments, Secaucus, New Jersey, USA). All 120 spiders were returned to their webs. Interspecific differences in mean body sizes were evaluated using Student's t-test (GraphPad QuickCalcs 2017).
ORB-WEBS OF ARGIOPE SPECIES
We used the method of Herberstein and Tso (2000) to characterize the size and shape of 60 orb-webs of each Argiope species in the field. Using a metallic tape measure, we determined the height to the center of the hub of the orb-web above ground (h), the horizontal diameter of the orb-web (dh), the length of the upper vertical radius (ru), the length of the lower vertical radius (rl), the length of the upper hub radius (Hru), and the length of the lower hub radius (Hrl) to the nearest mm. We then counted the number of sticky spiral threads in the upper and lower halves of the web (Su and Sl, respectively). Using the formulas presented by Herberstein and Tso (2000), we calculated the capture area (in cm2) and the mesh height, and the average distance between adjacent sticky spiral threads (in mm), for each web. Interspecific differences in mean web measurements were evaluated using Student's t-test (GraphPad QuickCalcs 2017).
PREY OF ARGIOPE SPECIES
We marked the location of 60 pairs of webs occupied by A. aurantia and A. florida females in the native scrub with stake flags so that we could quickly find them at dawn and dusk. Initially, we removed all prey hanging in each web in late afternoon, then we returned the next d after dawn (7: 00 AM to 9: 00 AM) and before dusk (5: 00 PM to 7: 00 PM) for 5 consecutive d, and carefully removed with forceps all arthropods trapped in a web or being eaten by the resident spider. Specimens were preserved in 70% isopropyl alcohol, returned to the lab, and identified to species. Following the period of daily prey removal, we fed each spider by gently tossing an assortment of beetles and crickets into a web in order to approximate the nutritional state it would have had if left undisturbed.
We measured the body length of each prey item to the nearest 0.1 mm under a dissecting microscope using an ocular micrometer (Model SZX12, Olympus Corporation, Shinjuku, Tokyo, Japan). Appendages such as antennae and ovipositors were excluded. We also measured the width of the thorax or abdomen, whichever was wider. We estimated dry body mass and consumable dry body mass to the nearest 0.1 mg using taxa-specific regression equations (Sample et al. 1993; Sabo et al. 2002; Straus & Aviles 2018). Differences in captured prey at the ordinal level were evaluated statistically using the Chi square test (GraphPad QuickCalcs 2017).
We determined the trophic niches of the 2 spiders using the Schoener overlap index, one of the simplest and most attractive measures of niche overlap (Krebs 1989). This index measures the percentage similarity of prey between 2 species, the actual area of overlap of their resource utilization curves. We did this at 3 taxonomic levels: order, family, and species.
Results
SIZE OF ARGIOPE SPIDERS AND THEIR ORB-WEBS
Argiope aurantia females were consistently larger than A. florida females according to the 2 measurements we made. As shown in Table 1, living A. aurantia females were typically 2.2 times heavier than A. florida females. In addition, the first right legs of A. aurantia averaged 1.3 times longer than those of A. florida.
Both spiders hung their webs in the scrubby vegetation at essentially the same height aboveground (ca. 1.1 m on average to the hub). But according to the other 7 measures (Table 1), the 2 species produced significantly different orb-webs.
Orb-webs of A. aurantia on average were 1.3 to 1.5 times wider on the horizontal and vertical axes than A. florida orb-webs, and the calculated difference in prey capture area was even greater (1.9 times in favor of A. aurantia).
On average, there were 50 and 65 sticky spirals in the upper and lower halves of A. florida orb-webs, which was 1.5 times more than those in a typical A. aurantia web (34 and 42, respectively). This meant that the average mesh height in A. aurantia (5.1 mm) was 2.2 times greater than that in A. florida orb-webs (2.3 mm). Our findings mirror those of Uetz at al. (1978), who reported mean mesh heights of 4.6 mm and 2.7 mm for A. aurantia and A. trifasciata (Forskål), the banded garden spider, respectively, sympatric in an Illinois prairie.
PREY OF ARGIOPE SPIDERS
Over the course of 5 d we collected about the same number of arthropods from the 2 species: 80 items for A. aurantia and 76 for A. florida. We removed two-thirds more prey from spiders late in the d than at dawn, strongly suggesting that both species were primarily diurnal hunters. There was no statistically significant difference between the 2 species in number of prey captured at dawn versus dusk (χ2 = 0.194; df = 1; P = 0.66).
Table 1.
Comparison of spider sizes, orb-webs, and prey parameters for 2 Argiope species (N = 60 females per group). Values for each spider species represent the mean ± SD.
Prey trapped by A. aurantia were substantially larger than that of A. florida (Table 1). On average, insects in A. aurantia webs were 1.6 times longer and 2.4 times heavier than those in A. florida webs.
We were able to identify taxonomically 94% (= 146 of 156) of the specimens to genus or species, respectively (Table 2). Inspection of the prey list revealed some interesting differences between the spiders. Small homopterans were entirely absent from webs of A. aurantia, whereas spiders and neuropterans were missing from A. florida webs. Large robber flies (Asilidae), dragonflies (Aeshnidae and Libellulidae), and leather-colored bird grasshoppers (Schistocerca alutacea Harris; Orthopera: Acrididae) were present only in A. aurantia webs, but medium-size blister beetles (Meloidae) and horseflies (Tabanidae) were retrieved only from A. florida webs.
Of the 15 insects found in both spiders' webs, half were orthopterans. Six species were represented by more than singletons: acorn weevils (Cucurlio sp.; Curculionidae), female fire ants (Solenopsis invicta Buren; Hymenoptera: Formicidae), southern yellowjackets (Vespula squamosa (Drury); Hymenoptera: Vespidae), toothcercus shortwing grasshoppers (Melanoplus forcipatus Hubbell; Orthoptera: Acrididae), spotted-winged grasshoppers (Orphulella pelidna (Burmeister); Orthoptera: Acrididae), and glassy-winged toothpick grasshoppers (Stenacris vitreipennis (Marschall); Orthoptera: Acrididae).
Insects known to be endemic to Florida scrub were exceedingly uncommon in the diets of both spider species. Only 6 specimens (4% of all prey) belonging to 3 species were collected: 1 elongate June beetle (Phyllophaga elongata (Linell); Coleoptera: Melolonthidae) in an A. florida web; 1 nigropicta grasshopper (Aptenopedes nigropicta Hebard; Orthoptera: Acrididae) each in A. aurantia and A. florida webs; and 2 and 1 toothcercus shortwing grasshoppers (M. forcipatus), respectively, in A. aurantia and A. florida webs. These results are the opposite of our initial hypothesis that A. florida, being found mostly in xeric, scrubby habitats in the southeastern United States, may tend to prefer scrub-adapted insects in their diet.
DIETARY NICHE ANALYSES FOR ARGIOPE SPIDERS
As shown in Table 3, the taxonomic diversity of prey of the 2 Argiope species was very similar. There was not a significant interspecific difference in the number of prey captured at the ordinal level (χ2 = 9.39; df = 5; P = 0.0946). However, there was a highly significant difference between the spiders when consumable dry biomass was estimated (χ2 = 189.33; df = 5; P < 0.001). Two orders of insects set them apart; A. aurantia obtained more biomass from dipterans and less from coleopterans than expected by chance, whereas the opposite held true for A. florida, where adult females ate more coleopterans and fewer dipterans than expected by chance.
Table 3.
Contrast between taxonomic composition of prey captured by adult female Argiope spiders (N = 60 each). Other taxa include Araneae, Hemiptera, Lepidoptera, and Neuroptera. The table provides the following information: observed cell totals, (expected cell totals), and [Chi-square statistic for each cell].
Because we were able to identify almost all prey items to genus or species, we calculated the dietary overlap of the 2 spiders at 3 taxonomic levels using Schoener's similarity index. The dietary niche overlap dwindled from 70% of 10 orders shared, down to 39% of 36 families held in common, all the way to 20% of 76 genera or species found at least once in orb-webs of both species (Table 4). Thus, taxonomic scale mattered greatly; what seemed to be a high degree of commonality with a coarse niche filter was completely reversed when a much finer filter was applied. We conclude there was little overlap in the kinds of prey eaten by the spiders, even though they were trapping aerially active arthropods side-by-side in what appeared to be the same scrubby habitat.
Discussion
The sizes of Argiope used in our study and their orb-webs (Table 1) were very similar to values reported by other authors. Mean body lengths of A. aurantia and A. florida presented were almost identical to values published by Elgar (1991), McReynolds and Polis (1987), and Blackledge (2011). The average diameter of A. aurantia orb-webs we measured in the Florida scrub was 3 to 8 cm wider than webs of this species in Tennessee and Mississippi (McReynolds and Polis 1987; McReynolds 2000). Comparable data for A. florida were not available. The orb-webs of both Argiope were quite high in the scrub; the hubs typically were slightly more than a m aboveground. The reason for this is that spiders in most instances suspended their webs from 1.5 to 2 m tall woody vegetation, placing the orbs across flight paths of scrub insects (Justice et al. 2005; Carrel & Deyrup 2014). Other studies report hub-heights for A. aurantia averaging 64 to 66 cm, about two-thirds of our findings (Enders 1974; McReynolds & Polis 1987; McReynolds 2000). The lower elevations in these studies almost certainly is because they were performed in old fields and other north temperate habitats dominated by relatively low growing grasses and herbaceous vegetation.
Table 4.
Dietary niche overlap in 2 Argiope species at 3 taxonomic levels.
Despite having an orb-web with almost twice the capture area (Table 1), A. aurantia captured almost the same number of aerial prey as A. florida. Perhaps A. aurantia webs are less efficient per unit capture area than those of A. florida, because webs of the former have only two-thirds as many sticky spirals on average, and the gaps between sticky silken threads are more than twice as great as the latter species (Table 1). Small (4–6 mm long) beetles and wasps that got caught in A. florida webs likely would fly through A. aurantia orbs unharmed. Unfortunately, the relationship between design of webs and prey capture is complex and unclear. A review of data for 17 species of orb-weavers showed there is no significant correlation between mean mesh height and mean length of prey captured (Herberstein & Heiling 1998). Incidentally, the rate of prey capture by both Argiope females (about 1 item per spider every 4 d) was identical to the rate we reported previously for mature red widow spiders, Latrodectus bishopi Kaston (Araneae: Theridiidae), in late winter before the palmettos begin to flower in the scrub at Archbold (Carrel & Deyrup 2014).
Most prey (72% by count) caught by A. aurantia females in our study belonged to 3 insect orders (Orthoptera > Coleoptera > Hymenoptera [Table 3]), which was quite different than the ordinal ranking reported for this lspecies' prey in the 5 studies mentioned in the introduction (Hymenoptera > Orthoptera > Coleoptera = Lepidoptera). We suspect this difference is explained by the striking contrast in the abundance of flowering plants at the study sites. Florida scrub in Sep is largely lacking in flowering herbs and shrubs; inflorescences of Palafoxia feayi Gray (Asteraceae), a lanky endemic perennial, and few others are scattered across the landscape. This contrasts dramatically with the northern temperate sites used by other authors: old fields, grasslands, and roadsides used by them were replete with verdant stands of goldenrods, asters, and many blooming herbs, such as smartweed, Polygonum punctatum Elliott (Polygonaceae), all of which are very attractive to bees and wasps. If one discounts the hymenopterans, then the rankings of their studies and ours become similar: orthopterans and coleopterans are top-ranked prey of Argiope.
We extracted 8 silk-encased blister beetles, Epicauta spp. Dejean (Coleoptera: Meloidae), from webs of A. florida but none from A. aurantia. These field-derived data validate the findings of Smedley et al. (1995) that A. florida readily feeds on living blister beetles flipped into their webs by experimenters, whereas almost all other spiders either ignore or reject meloids unharmed after an attack. Extensive field observations of more than 5,000 webs made over the past 12 yr at Archbold add additional support; we recorded 15 instances of Epicauta spp. being eaten by mature Florida garden spiders, but none by yellow garden spiders (J. Carrel, unpublished observations).
Orthopterans, particularly short-horned grasshoppers (Acrididae), were the most important taxon in the diets of both Argiope. They were common in the spiders' prey base (comprising 28% and 14% of all items captured by A. aurantia and A. florida, respectively) (Table 3). Furthermore, they contributed 45% and 39% of the estimated biomass consumed by female A. aurantia and A. florida, far more than any other insect order (Table 3). Odonates, although uncommon in the prey base, because their large size contributed disproportionately to the consumable biomass of both spiders (13% and 15% for A. aurantia and A. florida, respectively).
The “rare, large prey” hypothesis espoused by Blackledge (2011) argues that orb-weavers in general derive the bulk of their energy from a small subset of the largest possible insects flying in their habitats. He supported his argument by compiling a meta-analysis of published data on the diets of 31 species of orb-weavers. He defined large prey as “insects at least 66% as long as the spiders capturing them.” When applied to our data, we found that 64% and 41% of prey items of A. aurantia and A. florida, respectively, were large insects; they comprised 92% and 79% of the respective spider's consumable biomass. Hence, we reject the rare, large prey hypothesis for Argiope living in subtropical Florida scrub.
Dietary niches of spiders have been documented mainly for species feeding in agroecosystems to evaluate their potential as biological control agents. Nyffler (1999) reviewed 40 published studies of 5 families of orb-weavers and 4 families of hunters performed in Europe and the US. He found 10 insect orders and Araneae make up the bulk of prey of these spiders, regardless of hunting mode. As noted by Wise (1993), pair-wise studies of types of prey used by sympatric spiders are uncommon. Horton and Wise (1983) were among the first to document wide overlap in types of prey used by the orb-weavers A. aurantia and A. trifasciata. They calculated the Schoener overlap index for number of prey in each order to be considerable (≥ 0.69) over the course of 2 yr. McReynolds & Polis (1987) subsequently obtained similar results with these 2 orb-weavers. Nentwig (1985) also reported extensive niche overlap, Schoener index 0.8 to 0.9, for prey captured by 4 tropical orb-weavers. Recently, Wirta et al. (2015), using DNA barcode technology, identified 21 lepidopteran and 117 dipteran prey in the gut contents of a lycosid and 2 thomisid spiders living sympatrically in the depauperate high tundra of northeastern Greenland. They found all 3 spider species were wide generalists, overlap in prey use was high (Schoener index 0.58–0.63), and the results remained qualitatively unchanged regardless of whether prey identity was considered at the level of species, genus, or family. Perhaps if they had examined the entire insect and spider fauna, which they note consists of an additional ≥ 90 species of insects and spiders, then their results might be more taxon sensitive. By identifying almost all the prey captured by A. aurantia and A. florida, we confirmed that niche overlap is very sensitive to taxonomic scale; what appeared to be extensive overlap at the ordinal level, as reported by others, with pairs of Argiope species, dwindle to almost no overlap at the species level. Thus, these 2 spiders living syntopically in Florida scrub are consuming prey, not only of different sizes, but of different species. They are generalist predators in a diverse universe of prey species, none of which is especially likely to be captured by the spiders.
Acknowledgments
We thank the Archbold Biological Station for providing research facilities. Funding for this work came in part from the Development Fund at the University of Missouri-Columbia and the Archbold Biological Station.
