The harvestman Phalangium opilio L. is an abundant predator in an array of habitats, including agricultural systems. In these systems, P. opilio has been noted to feed on pest species (e.g., Colorado potato beetle larvae, various aphid species, and lepidopteran eggs). Although the potential importance of P. opilio has been recognized, little is known about this harvestman’s development or reproductive biology. We investigated the influence of diet on development time, survival, reproduction, and female longevity. We compared the effects of two diets, Helicoverpa zea (Boddie) eggs and Aphis glycines Matsumura, on development of P. opilio because this predator is known to feed on those prey in soybean agroecosystems. Individuals fed A. glycines from emergence to adulthood suffered from higher mortality, longer development time, and smaller body size than those fed H. zea eggs. Four diet treatments were used in a separate study to test their effects on reproduction: standard laboratory diet (cornmeal, bacon, and H. zea eggs), H. zea eggs only, A. glycines only, and mixed diet (A. glycines and H. zea eggs). A diet of A. glycines alone yielded higher mortality, smaller body size, and lower fecundity than the standard laboratory diet. The presence of A. glycines in a monotypic or mixed diet lengthened the preoviposition period compared with diets without this prey and yielded a smaller first clutch and smaller average clutch sizes compared with the standard laboratory diet. A. glycines is detrimental to the growth and reproduction of this predator.
Members of the order Opiliones (Arachnida) are omnivorous, but their diet consists primarily of small, soft-bodied organisms (Sankey and Savory 1974). As predators, primarily at night, harvestmen consume other harvestmen, gastropods, small earthworms, millipedes, spiders, mites, earwigs, flies, springtails, aphids, leafhoppers, and additional types of insects (Bristowe 1949, Sankey 1949, Sankey and Savory 1974, Nyffeler and Symondson 2001). The prevalence of opilionids in agroecosystems has been noted in previous studies. Predation by opilionids has been reported in agricultural and horticultural systems (Ashby and Pottinger 1974, Leathwick and Winterbourn 1984, Brust et al. 1986, Butcher et al. 1988, Drummond et al. 1990, Clark et al. 1994, Sivasubramaniam et al. 1997, Halaj and Cady 2000, Pfannenstiel and Yeargan 2002), orchards (Carroll 1982), woodlands (Adams 1984), and urban environments (Hanks and Denno 1993).
Studies conducted in the midwestern United States have elucidated the feeding behavior of opilionids in soybean fields. In Ohio soybean fields, the most common prey item for five opilionid species [Leiobunum nigripes Weed, Leiobunum vittatum (Say), Leiobunum politum Weed, Leiobunum calcar (Wood), and Hadrobunus maculosus (Wood)] was earthworms (Halaj and Cady 2000). Opilionids have been observed consuming eggs of the corn earworm, Helicoverpa zea (Boddie) (Lepidoptera: Noctuidae), in soybean fields in Kentucky. Pfannenstiel and Yeargan (2002) found that opilionids in the family Phalangiidae were the third most frequent predators on H. zea eggs in Kentucky soybean fields in 1993 and the second most frequent predators in 1994. Another study in Kentucky found that opilionids were the second most frequent predators of H. zea eggs, next to damsel bugs, Nabis spp. (Hemiptera: Nabidae) (Anderson 1996). Newton and Yeargan (2002) found that P. opilio represented ≈96% of the opilionid fauna in Kentucky soybeans during a 2-yr study; thus, it is probable that P. opilio was the opilionid most frequently observed in the earlier studies by Anderson (1996) and Pfannenstiel and Yeargan (2002).
The quality of prey can impact the fitness of a predator. In the wolf spider genus Schizocosa, prey quality was determined indirectly by measuring the effect of prey on different life history traits of the spiders (e.g., survival and development) (Toft and Wise 1999). Schizocosa species that fed on the collembolan, Tomocerus bidentatus Folsom (Collembola: Entomobryidae), exhibited high rates of survival and development. Schizocosa that were fed intermediate-quality prey, Drosophila melanogaster Meigen (Diptera Drosophilidae), initially had good rates of survival and development, but rates declined over time. Diets consisting of low-quality prey (i.e., sciarid gnats or conspecific spiders) resulted in impeded rates of development and low survival (Toft and Wise 1999). Similarly, the predator Geocoris punctipes (Say) (Hemiptera Geocoridae) survived four times longer when fed H. zea eggs than when fed Acyrthosiphum pisum (Harris) (Hemiptera: Aphididae) (Eubanks and Denno 2000). P. opilio is a polyphagous forager and, with its diverse habitat range, it encounters abroad array of prey species, including aphids. In general, aphids are considered to be low-quality prey for many arthropod predators (Bilde and Toft 1994, Toft 1995).
We have observed P. opilio feeding on the soybean aphid, Aphis glycines Matsumura (Hemiptera: Aphididae), an Asian pest species recently found in North America, including Kentucky. As noted above, P. opilio also feeds on H. zea eggs in Kentucky soybean fields (Pfannenstiel and Yeargan 2002). In the laboratory, P. opilio has been successfully reared on a mixed diet of cornmeal, bacon, and lepidopteran eggs (Newton and Yeargan 2001). In the study reported here, our primary objective was to compare the suitability of A. glycines with that of H. zea eggs as prey for P. opilio. We compared effects of H. zea eggs versus A. glycines on the development and survival of immature P. opilio in the laboratory. In a separate laboratory experiment, we evaluated P. opilio reproduction and adult female longevity on four different diets: A. glycines, H. zea eggs, a mixed diet of A. glycines and H. zea eggs, and the standard laboratory diet consisting of cornmeal, bacon, and lepidopteran eggs.
Materials and Methods
Laboratory colonies of both H. zea and A. glycines were maintained at the University of Kentucky for each experiment. H. zea was maintained using methods similar to those of Ignoffo (1965). A colony of A. glycines was kept on soybean in a greenhouse at 16:8 (L:D) and 24–30°C. The amount of prey provided always exceeded the predator's daily consumption in both experiments.
For both experiments, P. opilio were collected at the University of Kentucky's North Farm near Lexington, KY. Field-collected individuals were either large nymphs, which were reared to maturity in the laboratory, or adults. In the laboratory, adult males and females were paired and housed in large petri dishes, 15 by 4 cm (diameter by height), with mesh-covered holes (9 cm diameter) in the tops. Males and females were distinguished by the presence (on males) or absence of a prominent apophysis on the distal segment of the chelicera. Pairs were provided with food and water and were stored in incubators at 24 ± 1°C (15:9 LD). Food for each pair consisted of ≈0.2 g cornmeal (Quaker Oats), ≈0.25 g bacon (Real Bacon Bits; Hormel Foods, Austin, MN), and 50 H. zea eggs, which were the components of our standard laboratory diet. Food was provided by placing into the container a small piece of paper (≈ 3 by 3 cm) on which H. zea eggs had been oviposited and sprinkling the cornmeal and bacon bits onto the paper. Water was provided in a 9-ml vial held at a 45° angle by sculpting compound (Super Sculpey), with a cotton wick inserted into the opening of the vial. High humidity was maintained by placing open containers of water in the bottom of the incubators. Similar conditions have been used to successfully rear P. opilio (Bachmann and Schaefer 1983). Each pair was provided with ovipositional substrate that consisted of a 5 by 1-cm (diameter by height) dish filled with sterile soil and a 1 by 1 by 1-cm piece of florist's foam (green) on the soil surface. The ovipositional substrate was moistened and checked for eggs daily. Eggs were clearly visible on lifting the piece of foam, because virtually all eggs were laid at the foam-soil interface. In the event of oviposition, the ovipositional substrate was replaced with an identical new substrate, and the substrate with eggs was placed into a 8.5 by 1.5-cm (diameter by height) petri dish. Eggs were incubated under the same temperature and photoperiod described above for maintaining adults. If no oviposition occurred within 7 d, the ovipositional substrate was replaced with new substrate on the seventh day as a precaution against growth of mold or other contaminating organisms. Clutches of eggs were monitored daily and sprayed with water every other day.
Diet and Development
On eclosion, ≈20 offspring from each of 16 field-collected females were assigned to a diet treatment of H. zea eggs or a diet treatment of A. glycines. All females contributed a similar proportion of offspring to each diet treatment (i.e., 3:1 to the respective treatments). Approximately three times as many offspring were assigned to the aphid diet (231) as were assigned to the egg diet (78) because we anticipated higher nymphal mortality on the aphid diet.
Harvestmen were housed individually in 8.5 by 1.5-cm petri dishes in incubators set at the same conditions as described above. On reaching the fifth instar, they were transferred individually into 8.5 by 8.5-cm paper cartons with petri dish bottoms placed inside the base of the cartons. The tops of the paper cartons were covered with plastic wrap, and each carton top was covered with one-half of a cardboard lid to provide partial shade from the incubator lights. Harvestmen were reared from egg eclosion until adult emergence or until nymphal mortality occurred. Individuals were checked daily for molts and mortality. Water was supplied in the same manner as described above for adult P. opilio and was refreshed as needed.
Live A. glycines (≈100) were provided on an excised soybean leaf to individuals in the A. glycines diet treatment. Before introducing the soybean leaf and aphids into the P. opilio containers, the leaves were examined under a microscope, and parasitized aphids, thrips, spider mites, and other nonaphid organisms were removed. P. opilio in the H. zea egg diet treatment were fed (≈100 frozen H. zea eggs on a small piece of white paper (cut from a larger sheet on which they had been oviposited). Fresh eggs were killed (by freezing) to prevent emergence of first-instar caterpillars. Old prey were replaced with fresh prey every other day.
Mortality was recorded in each treatment, and when individuals reached adulthood, their sex and the date were recorded. Cephalothorax width of adults from each treatment was determined by measuring the distance between the third pair of coxae with an ocular micrometer.
Diet and Reproduction
Phalangium opilio were reared from the beginning of the seventh (=penultimate) instar until their death on four different diets (i.e., only A. glycines, only H. zea eggs, a mixed diet of A. glycines and H. zea eggs, or the standard laboratory diet of cornmeal, bacon, and H. zea eggs). Before the seventh instar, all individuals were reared on the standard laboratory diet. In this experiment, we were primarily interested in the effect of diet on reproduction (interval between adult emergence and first oviposition, number and size of clutches, interval between clutches) and adult female longevity. These traits likely reflected effects of diet during both the seventh instar and the adult stage. We also recorded mortality that occurred during the seventh instar and the size of adult females, which solely reflected the effect of the diet treatments during the seventh instar.
On emergence, first-instar offspring from field-collected individuals were randomly assigned to one of four diet treatments, although they did not receive those treatments until they reached the seventh instar. They were housed individually in the same environmental conditions as experimental subjects in the development experiment described earlier. Offspring used in the experiment came from a total of 22 females, with each female contributing ≈5 individuals to each treatment, for a total of ≈110 first instars per treatment (of which 37–50 individuals per treatment survived to the beginning of the seventh instar). All individuals were fed the same standard laboratory diet until the beginning of the seventh instar, i.e., amixture of cornmeal, bacon, and H. zea eggs. From the beginning of the seventh instar until their death, individuals were provided their assigned diet treatment: only A. glycines (n = 42), only H. zea eggs (n = 37), a mixed diet of A. glycines and H. zea eggs (n = 47), or the standard laboratory diet of cornmeal (≈0.2g),bacon (≈0.25g), and H. zea eggs (n = 50). The amount of A. glycines and H. zea eggs and the manner in which they were provided to P. opilio were the same as in the development experiment. In the mixed diet, aphids and H. zea eggs were provided in amounts equivalent to the monotypic aphid diet plus the monotypic H. zea egg diet. We did not attempt to quantify the amount of food consumed in any treatment, but it was apparent from observations of surviving prey that predators in the mixed diet treatment consumed both types of prey.
Food was replenished every other day. Some maintenance (i.e., checking for molts and ovipositional substrate maintenance) occurred every day; other maintenance, such as providing water and cleaning containers, occurred as needed. Between 7 and 20 d after adult emergence, males and females from the same treatment were paired and housed in the manner previously described for field-collected male/female pairs. Monitoring and recording of data (date of oviposition, clutch number, clutch size, interval between clutches, and female longevity) occurred daily until female death. Females were preserved in 70% ethanol immediately on discovery, and their cephalothorax width was determined by measuring the distance between their third pair of coxae.
For the diet and development experiment, mortality was analyzed with a 2 by 2 contingency table (dead or alive; H. zea eggs or A. glycines) using χ2 For surviving individuals, time required for development was analyzed by ranks using a one-way analysis of variance (ANOVA; SAS Institute 2000), and the size of both sexes of adults was analyzed using a two-way ANOVA (SAS Institute 2000), using Box-Cox transformed data (Sokal and Rohlf 1995).
For the diet and reproduction experiment, ANOVA was used to analyze all data except mortality. Data that did not meet the assumptions of ANOVA were transformed using Box-Cox transformation (Sokal and Rohlf 1995). Asymmetric distributions were Box-Cox transformed using the procedure TRANSREG in SAS (SAS Institute 2000). If, after Box-Cox transformation, the ANOVA assumptions were not met, the data were transformed to ranks, and an ANOVA was performed on the ranked data. This procedure, which was used for preoviposition period, female longevity, and number of clutches, is a good alternative to common non-parametric statistical tests because it maintains the power of parametric tests (Conover and Iman 1981) and allows the inclusion of additional factors that cannot be included in a Kruskal-Wallis test (Siegel and Castellan 1988). This was necessary because there were significant block effects that needed to be corrected by including a block factor in some analyses. If ANOVAs were significant, they were followed by a posthoc Bonferroni test (i.e., least significant difference [LSD] with Bonferroni correction) (Rice 1989). After transformation to ranks, preoviposition period data were analyzed with an analysis of covariance (ANCOVA) to control for the confounding effects of date of pairing; date of pairing was included as the covariate in this analysis. Mortality during the penultimate instar was analyzed with a 2 by 4 contingency table (dead or alive; four diet treatments). Linear regression analyses were used to correlate female fecundity with female size for each diet treatment and for all treatments combined.
Diet and Development
Mortality from eclosion through the last nymphal instar was significantly different for the diet treatments of H. zea eggs and A. glycines (χ2 = 30.9; df = 1; P < 0.001). P. opilio that fed on a monotypic diet of A. glycines suffered higher mortality than those fed on a monotypic diet of H. zea eggs (Fig. 1A). The diet treatments had a significant effect on the rate of development for P. opilio (ANOVA by ranks; F = 136.72; df = 1,61; P < 0.001). Individuals fed A. glycines required almost twice as long to mature as those fed H. zea eggs (Fig. 1B). There was a significant difference in the size of the newly emerged adults that had developed in different diet treatments (F = 258.21; df = 1,61; P < 0.001; Box-Cox transformation with λ = −1). Individuals fed a monotypic diet of H. zea eggs had a larger body size than those that fed on a monotypic diet of A. glycines; females were larger than males regardless of diet (Fig. 2).
Diet and Reproduction
Mortality, measured from the beginning of the seventh instar to the adult stage, was marginally different among diet treatments (χ2 = 7.69; df = 3; P = 0.053). Greatest mortality occurred in the monotypic A. glycines diet, and there were no apparent differences in seventh-instar mortality among the other three treatments (Fig. 3A).
There was a significant treatment effect on adult female size (F = 6.716; df = 3,59; P < 0.001) as aresult of diet treatment during the seventh instar. Females fed only A. glycines had a significantly smaller cephalothorax width than females fed the standard laboratory diet of cornmeal, bacon, and H. zea eggs (Fig. 3B).
Preoviposition period was significantly affected by diet (ANCOVA by ranks; F = 5.58; df = 3,83; P = 0.002). Females in both the monotypic and mixed diet treatments containing A. glycines had a significantly longer preoviposition period (time from adult emergence to first oviposition) than those in diet treatments that did not contain aphids (Fig. 4A).
Diet treatment significantly affected the number of eggs laid in the first clutch (F = 3.385; df = 3,83; P = 0.022; Box-Cox transformation with λ = 0.25). Females fed a standard laboratory diet had a significantly larger first clutch compared with either of the diets that contained A. glycines, while the monotypic diet of H. zea eggs was intermediate and not significantly different from any other diet (Fig. 4B). Statistical analysis of the effect of diet treatment on number of clutches laid per female produced equivocal results; ANOVA (by ranks; F = 2.751; df = 3,83; P = 0.048) indicated a significant effect, but means comparisons (LSD) showed no significant differences in number of clutches among diet treatments after data were Bonferroni corrected (Fig. 4C). There was, however, a trend of females that were fed the monotypic diets (i.e., only A. glycines or only H. zea eggs) to lay afewer number of clutches than females fed the mixed diet or standard laboratory diet (Fig. 4C).
Average clutch size was significantly affected by diet (F = 37.526; df = 3,83; P < 0.001; Box-Cox transformation with A = 0.25). Treatments that included A. glycines (i.e., only aphids or A glycines mixed with H. zea eggs) had significantly lower average clutch sizes than those fed the standard laboratory diet (Fig. 4D). Diet treatment significantly affected fecundity (total number of eggs laid per female; F = 3.676; df = 3,83; P = 0.016; Box-Cox transformation with λ = 0.25). Females fed only A glycines had significantly lower fecundity compared with those fed the standard laboratory diet (Fig. 4E).
Female longevity (from adult emergence to death) was not affected by treatment (F = 0.8811; df = 3,83; P = 0.456). Females lived ≈6–8 wk regardless of diet (Fig. 4F). Linear regression showed no significant relationship between female size and longevity (F = 2.05; df = 1,61; P = 0.157).
There was a significant correlation between body size and fecundity (r = 0.4443, P < 0.001), in that larger females laid more eggs (Fig. 5). When female size was included as a covariate in an ANCOVA comparing the difference in fecundity among treatments, the significance of the treatment effect disappeared (F = 1.51; df = 3,58; P = 0.221), suggesting that the observed differences in fecundity were caused by the effect of diet on female body size. The ANCOVA showed there was no significant treatment by size interaction (F = 0.17; df = 3,55; P = 0.915), indicating the slopes of the regression lines are not significantly different from each other.
Several factors affect the development and reproduction of predators, including the quality and complexity of diet. Some predators may incur low survival, long developmental periods, and lower fecundity when they feed on low-quality prey (Toft and Wise 1999). Toft (1995) found that the linyphiid spider Erigone atra (Blackwall) had increased reproductive success when fed a mixed diet compared with a monotypic diet, and he hypothesized that the effect of dietary mixing depends on the quality of the prey species being mixed. Atypena formosana (Oi), another linyphiid, also experienced higher fecundity when fed a mixed diet compared with a diet of only Collembola (Entomobryidae) (Sigsgaard et al. 2001). In general, these studies suggest that mixed diets are better than monotypic diets for reproduction, although the quality of the prey species involved can influence the effects of a mixed diet.
Phalangium opilio incurred higher mortality, longer developmental periods, and smaller adult body size when fed a monotypic diet of A. glycines. Fecundity also was compromised when fed A. glycines. A mixed diet containing A. glycines caused some adverse effects on reproduction compared with the standard laboratory diet, which did not contain aphids. P. opilio reared on the standard laboratory diet, which was the most complex diet tested, generally performed equal to or better than those fed the other diets. Compared with this diet, the adverse effects of a monotypic diet of H. zea eggs on P. opilio life history traits were less severe than those seen with a monotypic diet of A. glycines. Mixing H. zea eggs with aphids seemed to partially ameliorate the negative effects of the monotypic aphid diet. Our results seem to support the contention of Toft (1995) that effects of dietary mixing depend on the quality of the prey species being mixed.
Lepidopteran eggs generally are considered to be higher-quality prey than aphids (Eubanks and Denno 2000), but few studies have directly compared the nutritional composition of these two prey groups. Specty et al. (2003) analyzed the protein, lipid, and carbohydrate content of the aphid Acyrthosiphon pisum (Harris) and eggs of the lepidopteran Ephestia kuehniella Zeller. They found the lepidopteran eggs were two times richer in amino acids and three times richer in lipids than the aphids, whereas the aphids were 1.5 times richer in glycogen. If the biochemical composition of the prey used in our study follows this pattern, differences in nutritional composition of the prey could account for the superior performance of P. opilio on H. zea eggs compared with soybean aphids. We cannot rule out, however, the possibility that soybean aphids may possess chemicals that have toxic effects on P. opilio.
In spiders, there is a positive correlation between female body size and female fecundity, both within species (Petersen 1950, Enders 1976, Beck and Connor 1992, Wagner and Wise 1996, Arnqvist and Henriksson 1997) and across species (Marshall and Gittleman 1994, Simpson 1995). We found a similar correlation between female P. opilio size and fecundity. Our ANCOVA indicated that the negative impact of A. glycines on predator fecundity was mediated through this prey species' effect on female body size.
Overall, our study showed that A. glycines was detrimental to several fitness traits of P. opilio, including some traits that were compromised even when the aphids were mixed with lepidopteran eggs. While it is unlikely that P. opilio would typically feed on a monotypic diet in the field, this predator may consume large numbers of A. glycines in situations where these aphids are particularly abundant, as would occur in outbreak situations. If this occurs, P. opilio populations could be adversely affected. In turn, this could adversely affect the contribution of P. opilio to the natural biological control of other pests, such as H. zea.
Acknowledgments. The authors thank J. Harwood and M. Lutz for reviewing an earlier version of the manuscript and J. Moya-Loraño for statistical advice. Support for this research was provided by the USDA (CRIS 0188764). This investigation (04-08-053) was conducted in connection with a project of the Kentucky Agricultural Experiment Station.