Cannibalism, Food Limitation, Intraspecific Competition

See discussions, stats, and author profiles for this publication at: https://www.researchgate.net/publication/7437851

Cannibalism, food limitation, intraspeciﬁc competition, and the regulation of spider populations

Article in Annual Review of Entomology · February 2006 DOI: 10.1146/annurev.ento.51.110104.150947 · Source: PubMed

CITATIONS READS 213 593

1 author:

David H. Wise University of Illinois at Chicago

118 PUBLICATIONS 5,094 CITATIONS

SEE PROFILE

Some of the authors of this publication are also working on these related projects:

RESTORE Project View project

All content following this page was uploaded by David H. Wise on 09 April 2016.

The user has requested enhancement of the downloaded file. 2Nov 2005 13:14 AR ANRV263-EN51-19.tex XMLPublishSM(2004/02/24) P1: KUV 10.1146/annurev.ento.51.110104.150947

Annu. Rev. Entomol. 2006. 51:441–65 doi: 10.1146/annurev.ento.51.110104.150947 Copyright c 2006 by Annual Reviews. All rights reserved First published online as a Review in Advance on September 22, 2005

CANNIBALISM,FOOD LIMITATION, INTRASPECIFIC COMPETITION, AND THE REGULATION OF SPIDER POPULATIONS

David H. Wise Department of Entomology, University of Kentucky, Lexington, Kentucky 40546-0091; email: [email protected]

KeyWords Araneae, intraguild predation, population regulation, trophic cascades ■ Abstract Cannibalism among generalist predators has implications for the dynamics of terrestrial food webs. Spiders are common, ubiquitous arthropod generalist predators in most natural and managed terrestrial ecosystems. Thus, the relationship of spider cannibalism to food limitation, competition, and population regulation has direct bearing on basic ecological theory and applications such as biological control. This review first briefly treats the different types of spider cannibalism and then focuses in more depth on evidence relating cannibalism to population dynamics and food web interactions to address the following questions: Is cannibalism in spiders a foraging strategy that helps to overcome the effects of a limited supply of calories and/or nutrients? Does cannibalism in spiders reduce competition for prey? Is cannibalism a significant density-dependent factor in spider population dynamics? Does cannibalism dampen spider-initiated trophic cascades?

INTRODUCTION

Terrestrial food webs have a high diversity of generalist predators (32, 102, 105),

by UNIVERSITY OF ILLINOIS - CHICAGO on 08/23/07. For personal use only. among which spiders are abundant and ubiquitous (143). Cannibalism occurs in a Annu. Rev. Entomol. 2006.51:441-465. Downloaded from arjournals.annualreviews.org wide range of generalist predators (43, 100) and is perceived by many researchers to be common among spiders. Types of cannibalism can be classified according to the life-history stages and relatedness of cannibal and victim. Among spiders sexual cannibalism has been the most extensively studied, yet Elgar (31) points out that among invertebrates “...sexual cannibalism is surprisingly uncommon relative to the other types of cannibalism.” The current review focuses on the less- studied types of cannibalism in spiders. This emphasis is not meant to suggest a clear dichotomy, because some of the selective pressures molding both sexual and nonsexual cannibalism are similar. Thus, aspects of sexual cannibalism are reviewed briefly. Emphasizing research findings on nonsexual forms of spider cannibalism hopefully places in perspective what is currently known of the roles

0066-4170/06/0107-0441$20.00 441 2Nov 2005 13:14 AR ANRV263-EN51-19.tex XMLPublishSM(2004/02/24) P1: KUV

442 WISE

that cannibalism plays in the ecology of a major generalist predator, with a primary goal of identifying research questions deserving future scrutiny. Spiders are frequently generalized as highly cannibalistic, which could mean that most spider species exhibit cannibalism and/or that killing and eating conspecifics occurs frequently in those species that are cannibalistic. Given the gaps in our knowledge about the frequency of cannibalism in nature, generalizations about rampant spider cannibalism are sometimes too glibly stated, but it is true that spiders and other arachnids display a wide range of cannibalistic behaviors. Elgar & Crespi (32) list five types based upon the life-history stage of cannibal and victim: adults cannibalizing adults, adults cannibalizing juveniles, juveniles cannibalizing juveniles, adults cannibalizing eggs, and juveniles cannibalizing eggs. Among invertebrates only arachnids exhibit all five types, and all five occur in spiders (32). However, it does not follow that cannibalism occurs frequently in most spider species. Sexual cannibalism has been documented in many spider families; however, Elgar (31) points out that we have scant data on the actual frequency of sexual cannibalism in natural populations. It appears that sexual cannibalism is relatively rare among the therophosids (22, 61). Jackson (59) suggests that sexual cannibalism may not be as widespread in spiders as commonly believed and that many of the courtship behaviors attributed to the avoidance of sexual cannibalism have alternative explanations. Elgar (31), however, suggests that sexual cannibalism in spiders is widespread. Relatively few investigators have attempted to quantify the frequency of sexual cannibalism among spiders in nature. We do not have a good idea of the extent to which nonsexual cannibalism is a major spider foraging behavior. Predation by spiders on other spider species has been documented for numerous families in both nature and laboratory (143), and authors (30, 41, 44) frequently echo Bristowe’s (14) assessment that puts “spiders easily at the top of the list of spider enemies.” Even if it is true, widespread predation by spiders on other spider species does not necessarily implicate cannibalism as a major factor in spider ecology. Furthermore, it does not necessarily follow that cannibalism is frequent in nature if conspecific individuals confined in the laboratory kill and eat each other. A review of the evidence indicates that, despite the absence of extensive data on the frequency of cannibalism in many families in

by UNIVERSITY OF ILLINOIS - CHICAGO on 08/23/07. For personal use only. nature, available data from ﬁeld surveys, from accumulating evidence from ﬁeld

Annu. Rev. Entomol. 2006.51:441-465. Downloaded from arjournals.annualreviews.org experiments, and from laboratory studies in which habitat structure and densities of conspeciﬁcs and alternative prey are similar to natural conditions suggest that nonsexual cannibalism may be a signiﬁcant foraging behavior in some spider families, especially the families of wandering spiders, and among them, the wolf spiders (Lycosidae) in particular.

THEORETICAL CONSIDERATIONS

Cannibalism and Intraguild Predation The ecological roles of spider cannibalism in population and food web dynamics are understood best when compared with intraguild predation (IGP) because both 2Nov 2005 13:14 AR ANRV263-EN51-19.tex XMLPublishSM(2004/02/24) P1: KUV

CANNIBALISM IN SPIDER POPULATIONS 443

involve generalist predators feeding on other predators that share one or more nonpredacious prey species. Such a comparison also highlights how costs and benefits interact in shaping the evolution of cannibalistic behaviors in spiders. Polis (101) defined IGP as “the killing and eating of other species that use the same resources and are therefore potential exploitative competitors.” Many authors view cannibalism as a form of IGP, but the definition proposed by Polis excludes cannibalism. Separating IGP and cannibalism recognizes a crucial difference between the two behaviors, i.e., that costs of the two can be different because cannibalism can incur direct genetic costs that are impossible in IGP. On the other hand, their benefits overlap extensively, including the advantage gained by eliminating a potential competitor, an indirect effect that should be even greater for cannibalism than for IGP. The benefits ascribed to cannibalism are acquisition of a high-quality food source and elimination of a potential exploitative competitor. The possible costs are injury or death, contracting pathogens or parasites, lowering of inclusive fitness by killing relatives, and loss of sperm.

Asymmetries, Ecological Factors, and Population Consequences One can make several generalizations about factors that should influence rates of cannibalism (28, 32, 43, 100). Certain asymmetries influence the risk involved: Cannibals should prey on smaller conspecifics in order to avoid the risk of retaliation; they should avoid killing kin; and they should avoid killing a mate unless the energy or nutrients acquired outweigh the cost of losing sperm. Because cannibalism is most frequently a predator-prey interaction, its frequency should respond to the ecological factors of cannibal density, alternative prey (both abundance and relative food quality), and habitat structure, all of which can act separately or together to determine the frequencies of encounters with potential cannibals or potential prey (conspecific and heterospecific), the spectrum of foraging choices, hunger level, and the strength of exploitative competition. Many of these general attributes lead to the prediction that certain types of cannibalism should act as a strong density-dependent factor regulating population density, which has conse-

by UNIVERSITY OF ILLINOIS - CHICAGO on 08/23/07. For personal use only. quences for spider-initiated trophic cascades. Annu. Rev. Entomol. 2006.51:441-465. Downloaded from arjournals.annualreviews.org As a way to organize this review, I first discuss findings for several types of cannibalism between nonsolitary and related spiders (two or more spiders that are together because of behaviors separate from cannibalism or because they are related genetically): sexual cannibalism, sibling cannibalism, filial cannibalism, matriphagy, and cannibalism in social groups. Cannibalism between solitary spiders can be viewed more directly as foraging for prey (cannibalistic foraging) or as an agonistic interaction (interference cannibalism and cannibalistic territoriality). Effects of size asymmetries, hunger levels, and habitat features are discussed primarily under cannibalism among solitary spiders, although these factors also can influence cannibalism between nonsolitary and related spiders. Thus, as in solitary spiders, cannibalism between nonsolitary spiders also can be explained as an adaptation to food limitation. It is convenient, however, to separate types 2Nov 2005 13:14 AR ANRV263-EN51-19.tex XMLPublishSM(2004/02/24) P1: KUV

444 WISE

of cannibalism into these two broad categories because relatedness (genetic or otherwise) between cannibal and potential victim adds a unique disadvantage to the cost/beneﬁt ledger.

CANNIBALISM AMONG NONSOLITARY AND RELATED SPIDERS Sexual Cannibalism Sexual cannibalism is the killing and eating of a courting, copulating, or post- mating male by the female. Rarely do courting males cannibalize the female. The female spider is almost always the cannibal, sometimes killing her potential mate before copulating (31). The evolution of precopulatory sexual cannibalism can be explained most directly in terms of the relative value of the male as food versus his current worth as a sperm donor (92). Postcopulatory cannibalism could be explained as male investment in the young (18), yet males often attempt to escape before being cannibalized, so this cannot be a universal explanation. An excellent example of the complexity of the problem is the highly stereotyped behavior of the Australian redback male (Latrodectus hasselti), for which research has revealed likely selective advantages to both sexes (2–5). A different theory, not based upon sexual selection and parental care, postulates that the killing of courting males before copulation is the result of “aggressive spillover,” i.e., that selection among juveniles for aggressive foraging behavior due to severe food limitation has selected for genes that favor indiscriminate foraging behavior in the female, even though female fecundity supposedly is not limited by foraging as an adult (7). The most problematic feature of this theory is the postulation that severely food-limited juvenile females suddenly escape food limitation once they have matured, even though energy demands should increase, not decrease, because of the need to provide yolk to the eggs (41). Recent studies fail to support the basic assumptions and predictions of the aggressive spillover hypothesis for ﬁshing spiders in the genus Dolomedes, the spider for which the theory was originally proposed (62, 63, 69). Most studies of sexual cannibalism indicate that the female’s interests are almost by UNIVERSITY OF ILLINOIS - CHICAGO on 08/23/07. For personal use only.

Annu. Rev. Entomol. 2006.51:441-465. Downloaded from arjournals.annualreviews.org invariably the acquisition of calories or nutrients; nevertheless, it is reasonable to separate the treatment of other types of spider cannibalism from that of sexual cannibalism, for which the selective factors involved are related more to problems of parental investment than to the environmental conditions favoring nonsexual cannibalism (18). The reader is directed to the original seminal reviews of sexual cannibalism in spiders (18, 31, 92) and recent research publications on the topic that contain comprehensive reviews of the literature (3, 5, 22, 42, 84, 89, 98, 120).

Sibling Cannibalism Because of the possible genetic costs of eating close relatives, natural selection should favor preferential cannibalism of non-kin, unless the gain in energy or 2Nov 2005 13:14 AR ANRV263-EN51-19.tex XMLPublishSM(2004/02/24) P1: KUV

CANNIBALISM IN SPIDER POPULATIONS 445

nutrients is substantial. The evolution of kin discrimination in solitary spiders should be strongest among the youngest instars before they have finished dispersing and thus have a high probability of contacting each other. Cannibalism among recently dispersed (second instar) Hogna helluo (Lycosidae) wolf spiderlings was higher among pairs of non-kin than among siblings (111). That size differences were greater between than within broods could explain the result, but even if cannibalism rates differed solely because of size differences related to kinship and not because of direct recognition of kin, the potential for discrimination exists within the population because rates were lower among siblings (i.e., there is the potential for differential treatment of relatives independently of whether the individual recognizes the degree of genetic relatedness; see Reference 139 for the distinction between “kin discrimination” and “kin recognition”). A study with another species suggests that when kin discrimination in rates of cannibalism among wolf spiders does occur, it is due to size asymmetries: Cannibalism rates did not differ between pairs of siblings and non-kin Pardosa pseudoannulata spiderlings if the degree of size asymmetry was similar in both groups, but size asymmetry itself did have a large effect (58). Iida (58) argues that kin discrimination occurs within the first 24 h of dispersal owing to differing coefficients of variation in spiderling size within and between broods. The results of Hvam et al. (57) support the hypothesis that kin discrimination among newly emerged lycosid spiderlings is most likely due to differences in size asymmetries. They paired sibling and non-sibling Pardosa amentata spiderlings that differed in weight by <1% and found no evidence that kin recognition (i.e., the ability to respond to cues indicating genetic relatedness) affected the rate of cannibalism. Avoidance of cannibalism among siblings of nonsocial spiders might occur most often in species in which siblings remain in close contact when they are feeding and growing, i.e., among the subsocial solitary spiders that show extended maternal care. For example, the rate of cannibalism was much lower in experimental groupings of spiderlings of a subsocial burrowing wolf spider than in groupings of spiderlings from nonburrowing, more vagile lycosid species (1% versus 67%) (79). Kin recognition has been demonstrated in the subsocial spider Stegodyphus lineatus (Eresidae) (11). Differences between kin and non-kin pairings were marginally = by UNIVERSITY OF ILLINOIS - CHICAGO on 08/23/07. For personal use only. significant statistically (P 0.048), but kin recognition was demonstrated when

Annu. Rev. Entomol. 2006.51:441-465. Downloaded from arjournals.annualreviews.org cannibalism was measured in groups (P < 0.001).

Filial Cannibalism and Matriphagy Filial cannibalism occurs when females consume their own eggs or young. In the laboratory, female spiders consume eggs they have laid (65), but usually these eggs appear to be unfertilized. Most cannibalism of fertilized eggs by adults is not ﬁlial cannibalism but instead occurs when females invade the nests or webs of conspeciﬁc females and cannibalize their eggs, as happens with the web-building jumping spider Portia labiata (Salticidae) (20). Male S. lineatus invade nests and destroy egg sacs guarded by females, but they do not cannibalize the eggs (119). 2Nov 2005 13:14 AR ANRV263-EN51-19.tex XMLPublishSM(2004/02/24) P1: KUV

446 WISE

Females of the wandering (i.e., nonweb building) spider Clubiona cambridgei (Clubionidae) (106) and the funnel-web spider Coelotes terrestris (Agelendidae) (56) guard their egg sac against female cannibalistic intruders, a behavior that has been observed in the field and proven in laboratory experiments to increase the number of surviving young. Females of the latter two species cannot distinguish their egg sacs from those of another female; P. labiata can do so, but only if the sac has been moved to another web. A female wolf spider often adopts the egg sac of another female if given to her and carries it attached to her spinnerets (136). Absence of strong discriminatory faculties in most of these examples is not surprising, given the small probability of a female solitary spider losing contact with an egg sac that she is guarding. The probability that a female encounters recently dispersed spiderlings is greater. The female wolf spider carries hatched spiderlings on her back for several days before they disperse. She might frequently encounter her own young during and immediately after the dispersal phase. Sev- eral studies have uncovered a general inhibition against consuming spiderlings, even heterospecifics, that is highest when female lycosids carry the egg sac and disappears completely 1 to 2 weeks after the spiderlings have dispersed (6, 29, 48, 136). Filial cannibalism can occur at low rates within the first few days of dispersal. Female Pardosa milvina discriminate, although not perfectly, between their own progeny and those of other females during this period, but this discrimination disappears by 3 days postdispersal (6). In some spiders juveniles eat their mother before dispersing from the com- munal nest (33, 34, 66, 67, 118, 128–130) [termed gerontophagy (121), or more commonly matriphagy (36)]. In Amaurobius ferox (Amaurobiidae) the female is physiologically capable of producing a second egg sac; experiments show that her net reproductive output, calculated as the number of surviving midinstar juveniles, is maximized by matriphagy versus the alternative strategy of abandoning her progeny early in order to lay a second clutch (67). The amount of food provided to spiderlings through matriphagy in the Australian social crab spider, Diaea ergandros (Thomisidae), is positively correlated with the number of surviving spiderlings, probably because of a reduction in sibling cannibalism (36). In solitary species that exhibit maternal care, any tendencies for the female to

by UNIVERSITY OF ILLINOIS - CHICAGO on 08/23/07. For personal use only. cannibalize her young, and for her progeny to cannibalize each other, are sup-

Annu. Rev. Entomol. 2006.51:441-465. Downloaded from arjournals.annualreviews.org pressed during the period when the young remain with their mother. In Tegenaria atrica (Agelenidae) changing levels of tolerance and cannibalistic behavior of both the female and juveniles are correlated with alterations in the amounts of several types of chemical compounds in the cuticles of both adult and juvenile spiders (131, 132).

Social Spiders Social behavior in spiders is rare. It ranges from facultative aggregations of con- nected but individually maintained webs to groups in which adults share webbing and cooperate in brood care, and ranges from groups of genetically unrelated 2Nov 2005 13:14 AR ANRV263-EN51-19.tex XMLPublishSM(2004/02/24) P1: KUV

CANNIBALISM IN SPIDER POPULATIONS 447

individuals to species with highly inbred colonies (9, 17, 23, 134). A major as- pect of sociality is toleration, so it is not surprising that cannibalism (other than matriphagy) in spider social groups is rare and, when it does occur, appears to be related to food scarcity (23). Although many social spiders accept individuals from other groups (23), cannibalism can occur when individuals from another group enter the colony. Females of D. ergandros exhibit extended maternal care and accept unrelated juvenile immigrants into the nest in areas where nest density is high (34), although they recognize their own offspring and preferentially provide them with prey and trophic eggs. Juveniles accept unrelated immature spiders into the nest but can recognize them as nonsiblings and cannibalize them if sufﬁciently hungry (35). Subadult females that have been starved cannibalize unrelated immi- grant females and their own brothers, but they will not cannibalize unrelated male immigrants, possibly to maximize outbreeding (35).

Relationship to Population Dynamics In most types of cannibalism among nonsolitary and related spiders, the cannibal kills to obtain food, presumably because the supply of food is limited. The death of the victim, and also the degree to which cannibalism improves the growth, survival, and fecundity of the cannibal, potentially influence the population dynamics of the species. However, few, if any, comprehensive field studies have addressed these consequences for the types of cannibalism just discussed, although there is some research related to the topic, e.g., the effect of food supply on group size and emigration rates in colonial and social spiders (9, 35, 134). Most studies that relate cannibalism directly to food limitation, competition, and population dynamics have dealt with the more generalized types of cannibalistic behavior characteristic of solitary spiders that are uncomplicated by potential genetic or mating relationships between the interacting individuals. The next section focuses on these types— cannibalistic foraging, and the related behaviors of interference cannibalism and cannibalistic territoriality—but also reviews some findings for the other types of cannibalism that relate specifically to factors such as size asymmetry and hunger level. by UNIVERSITY OF ILLINOIS - CHICAGO on 08/23/07. For personal use only. Annu. Rev. Entomol. 2006.51:441-465. Downloaded from arjournals.annualreviews.org CANNIBALISM AS FORAGING FOR FOOD AMONG SOLITARY SPIDERS

Food Limitation Among Spiders In principle, killing conspecifics could have evolved as an adaptation to remove competitors for any limited resource. However, because such behavior is potentially dangerous and the results would benefit all members of the population (100), killing a conspecific primarily as a means to remove a competitor will evolve only under extensive local resource competition. In almost all cases in which a spider kills a conspecific, the victim is eaten. Extensive evidence suggests that spiders 2Nov 2005 13:14 AR ANRV263-EN51-19.tex XMLPublishSM(2004/02/24) P1: KUV

448 WISE

frequently are food limited, i.e., a relative shortage of prey limits their growth, development, reproduction, and/or survival (earlier evidence reviewed in Reference 143; for more recent evidence for food limitation in spiders, see 19, 47, 54, 69, 74, 80–83, 85, 86, 88, 103, 115, 125). Thus, it is reasonable to hypothesize that spider cannibalism has evolved primarily as a means of foraging for calories and nutrients in limited supply.

Size Asymmetry Field observations of spider cannibalism most often involve juveniles feeding on other juveniles or adults feeding on juveniles, with fewer examples of adults feeding on each other (30, 48, 149). In laboratory studies the probability of cannibalism increases with increasing size difference between paired juvenile spiders (16, 58, 78, 114). Among Pardosa agrestis (Lycosidae) juveniles that had been starved for 14 days, the cannibal was always the heavier spider (114). Cannibalism never occurred if the ratio was <2:1 and always occurred if it was >4:1. These results might appear to contradict ﬁndings of high rates of cannibalism among recently dispersed instars of the wolf spider Schizocosa ocreata (Lycosidae) (137, 138). However, Samu et al. (114), who used older instars of P. agrestis,argue that the smallest instars, because of their relatively low energy reserves, may be more risk-prone and hence more likely to attack a similar-sized spiderling.

Hunger Level Laboratory experiments with paired juvenile spiders show that starvation level strongly affects rates of cannibalism (11, 77, 78, 111, 114). For example, Samu et al. (114) paired juvenile P. agrestis with weight ratios within the range in which starved spiders showed a mixture of tendencies to cannibalize and found rates of cannibalism of 10%, 70%, and 100% for 0, 14, and 28 days of starvation, respectively. Hunger level affects the propensity to cannibalize most clearly in solitary spiders. In subsocial spiders hunger can affect rates of cannibalism (11), but not always (79), and in some social spiders cannibalism is absent even if the spiders suffer extreme hunger (23). by UNIVERSITY OF ILLINOIS - CHICAGO on 08/23/07. For personal use only. Annu. Rev. Entomol. 2006.51:441-465. Downloaded from arjournals.annualreviews.org Alternative Prey Heterospecific, nonpredacious prey tend to be safer victims than similarly sized conspecifics. Thus it is not surprising that the presence of alternative prey, whose consumption would decrease hunger level, decreases rates of cannibalism in laboratory experiments in which spiders are maintained in groups of more than two spiders (11, 112, 113, 137, 138). This expected confirmation of a seemingly simple prediction is deceptive, however, because the addition of habitat complexity, i.e., an element of realism, complicates the pattern (see below); and one experiment in a simple arena with paired juvenile Hogna helluo revealed an unexpected effect of alternative prey (111). Adding collembolans did not reduce the high rate of 2Nov 2005 13:14 AR ANRV263-EN51-19.tex XMLPublishSM(2004/02/24) P1: KUV

CANNIBALISM IN SPIDER POPULATIONS 449

cannibalism observed for starved spiders (∼90%), and increased, by almost three- fold, rates of cannibalism among recently dispersed (second instar) spiderlings. Roberts et al. (111) proposed several explanations for their finding; one of which is that hunger in the non-fed treatment caused the spiderlings to be lethargic and have lowered fighting ability, which agrees with findings of relatively low rates of cannibalism among starved second-instar wolf spiderlings in another lycosid, Pardosa lugubris (96), but is opposite of the effect of added collembolans on cannibalism in groups of second-instar S. ocreata (137, 138).

Habitat Complexity Increasing habitat complexity in a simple laboratory arena can lower rates of cannibalism of spiderlings by lycosid females (30, 107). Thus it would appear that increasing habitat complexity should decrease cannibalism between spiders, perhaps by providing hiding places for potential victims and/or by decreasing rates of contact between potential cannibals. The situation is not so simple, as experiments with immature lycosids demonstrate (107, 137). In one study, recently dispersed second-instar Schizocosa spiderlings were released into 0.3-m2 arenas with either a simple plaster-of-paris bottom or with leaf litter, crossed with a no- prey treatment or one with collembolans and pinhead crickets (137). The density of spiderlings reflected the high end of patches of S. ocreata spiderlings in nature. In the absence of alternative prey, habitat complexity had no effect on mortality from cannibalism (∼65%). In the presence of heterospecific prey, the incorporation of natural habitat structure increased cannibalism from ∼15% to 35%. The growth rate of spiderlings with leaf litter and alternative prey was 15% less than that in the simple arenas with alternative prey. The leaf-litter habitat may have provided the collembolans and crickets some protection from predation by Schizocosa, causing spiderlings in the complex habitat to be hungrier than those in the simple habitat. Thus, spiderlings in the complex habitat were more likely to cannibalize other spiderlings. The absence of a litter effect in the no-prey treatment fails to support an additional hypothesis, that leaf litter made cannibalistic ambushes more likely. Rickers & Scheu (107) conducted an experiment with Pardosa palustris (Ly- by UNIVERSITY OF ILLINOIS - CHICAGO on 08/23/07. For personal use only. cosidae) spiderlings that had first been kept together 10 to 17 days without prey Annu. Rev. Entomol. 2006.51:441-465. Downloaded from arjournals.annualreviews.org and allowed to cannibalize. Because only the largest spiderlings, i.e., the most successful cannibals, were selected for the experiment, the behaviors uncovered might differ from experiments with S. ocreata (137) owing to the possible presence of behavioral cannibalistic morphs (57, 78). The 2 × 2 × 2factorial experiment consisted of a no-prey/collembolan treatment crossed with two levels of habitat complexity (0.3 or 1.0 g of moss to mimic the range in nature) and two spider densities. Differences in habitat structure did not affect rates of cannibalism, but growth rates were higher in the treatment with the most moss when alternative prey was absent. The investigators hypothesize that increased habitat complexity decreased the energy lost in agonistic interactions, an effect not observed with S. ocreata (137). These two studies are not directly comparable because the one with 2Nov 2005 13:14 AR ANRV263-EN51-19.tex XMLPublishSM(2004/02/24) P1: KUV

450 WISE

P. palustris lacked a no-complexity treatment. Nevertheless, both failed to ﬁnd any evidence that habitat complexity decreases the rate of cannibalism among young juvenile lycosids. Increased structural complexity of the habitat could reduce cannibalism among web-building spiders by providing more sites for building webs, thereby reducing the frequency of web invasions and consequent exposure to cannibalism, as was observed in a cage experiment (112). Such an effect will be most apparent in families in which spider density exceeds the number of suitable web sites and spiders leave their webs to forage (such as erigonid Linyphiidae) (50, 51), but the effect is likely to be minimal for most web-building families because cannibalism appears to be rare among most web spiders.

Foraging Mode A strong argument for cannibalism as foraging for prey is the apparent correlation between the frequency of cannibalism and foraging mode, i.e., whether or not the spider uses a web to capture prey. Cannibalism accounts for 5% of the prey items in the diet of fishing spiders (Pisauridae) (150). Rates of cannibalism for wolf spiders range from 10% to 20% (an estimate, because conspecific and congeneric juveniles are difficult to distinguish) (30, 48, 149). Reported rates of cannibalism among web-building spiders are much lower; for example, in a review of the prey of spiders, Nentwig (91) gives no examples of cannibalism for web-building species. Because it is difficult to distinguish cannibalism from predation on closely related species, an approximate surrogate for differences in rates of cannibalism is differences in rates of predation on all spiders. A survey of the prey of spiders in cranberry bogs (10) revealed that 1% (1 of 73) of the prey of the web builders was other spiders; for wandering species the rate was 10% (12 of 115). In a review of studies reporting the prey of spiders in agroecosystems, Nyffeler (94) found overall average rates of predation on other spiders to be <1% for the web-building families and 15% for the families of wanderers. In rice fields, rates of cannibalism and IGP among lycosids were 5.5% and 3.4%, respectively; corresponding rates were 0% and 2.5% for tetragnathid orb weavers and 1.6% and 1.6% for a theridiid by UNIVERSITY OF ILLINOIS - CHICAGO on 08/23/07. For personal use only. web spinner (68). Annu. Rev. Entomol. 2006.51:441-465. Downloaded from arjournals.annualreviews.org This dichotomy can be explained by the fact that most web builders rarely leave their webs in search of prey. When they do leave, it usually is to search for anew web site or, in the case of mature males, to search for a mate (41). Thus conspecifics rarely come into contact, unless they are invading a web, in which case the invasion is more related to a web takeover than it is to predation on the owner and may rarely result in cannibalism. For example, in induced web invasions in the araneid Metepeira labyrinthea, the rate of cannibalism was small (3 of 88 encounters); most interactions involved web shaking and eventual retreat of one spider, usually the smaller one, without cannibalism (141). Another study found that web invasions involving tropical orb weavers rarely resulted in the death and eating of the loser (55). The outcome is often different when wandering spiders 2Nov 2005 13:14 AR ANRV263-EN51-19.tex XMLPublishSM(2004/02/24) P1: KUV

CANNIBALISM IN SPIDER POPULATIONS 451

meet, because when they come into foraging range of a conspecific, they are liable to be preyed upon. Signaling behaviors may reduce rates of cannibalism (discussed below), but web builders have an advantage because web vibrations alert them to an intruder. By no means have all families been thoroughly investigated with respect to the incidence of either IGP or cannibalism; and some families show a mix of foraging strategies. In a rock-face system, two species of web spiders showed high per- centages of other spiders in their diets (20% to 45%) (110), although the authors imply in their discussion that these numbers reflect IGP and not cannibalism. A major exception to proposed low rates of cannibalism in web-building families may be the erigonid Linyphiidae, which use webs close to the ground for prey capture but also capture prey by foraging off the web (49–51). Laboratory studies reveal high rates of cannibalism among groups of the erigonid Oedothorax gibbosus (Linyphiidae) housed with collembolans (135); this may reflect comparably high rates of cannibalism in nature, or alternatively, the high incidence of cannibalism may reflect the inability of these small spiders to prey on the collembolan provided, either because they are too large, as was observed in attempting to rear other genera of small linyphiids (147), or because they are unsuitable prey for other reasons (127). One field survey of several small linyphiids, including Oe- dothorax spp., failed to find evidence of predation on other spiders (1), although another study uncovered low rates of spider predation by Oedothorax insecticeps (68). Cannibalism among the young juveniles of these very small spiders would be difficult to detect in the field without intensive, systematic observations. Because adult Oedothorax species do not build webs (1) and may use lines of silk to detect prey (J. Harwood, personal communication), high rates of cannibalism in Oe- dothorax spp. would actually confirm the postulated association between foraging mode and cannibalism. More research is needed on the frequency of cannibalism among small linyphiids in nature. Linyphiid species that are both wanderers and web-spinners will likely show intermediate levels of cannibalism.

by UNIVERSITY OF ILLINOIS - CHICAGO on 08/23/07. For personal use only. Prey Quality: Calories, Nutrients, and Pathogens Annu. Rev. Entomol. 2006.51:441-465. Downloaded from arjournals.annualreviews.org One view of the nutritional demands of spiders and other generalist predators is that a shortage of calories, not nutrients, limits their populations (143). Others argue that nutrients, particularly nitrogen, are also limiting, perhaps even more so than calories (140). Spiders tend to have a higher level of nitrogen than other arthropod predators (24, 37), which means that predation on other spiders, including cannibalism, would be a particularly rich source of a potentially limiting nutrient. The balance of nutrients may also be important for spiders, which exhibit elevated rates of survival, growth, development, and fecundity on mixed diets compared with single-species regimens (96, 122, 127, 133); a mixed diet, however, can be unsatisfactory, depending on the toxicity and nutrient contents of the prey involved (12, 13, 122). If nutrient balance is important, conspeciﬁcs should 2Nov 2005 13:14 AR ANRV263-EN51-19.tex XMLPublishSM(2004/02/24) P1: KUV

452 WISE

provide the highest quality prey, because their nutrients most closely match those of the cannibal. Surprisingly, two studies with lycosid spiderlings fail to confirm this prediction but instead suggest that conspecifics are exceptionally low-quality prey. Toft & Wise (127) paired recently dispersed Schizocosa spiderlings with no alternative prey. During the first four weeks survival rate of the successful cannibals was high (>90%), but no spiders molted more than twice, all spiders had died by week 11, and at best the cannibals had only doubled their weight, compared with a fourfold increase in weight for spiders fed collembolans, fruit flies, or both. Oelbermann & Scheu (96) found a similar pattern for second-instar Pardosa lugubris. Spiderling cannibals doubled their weight in four weeks, but 90% died before reaching the next instar. These findings are not inconsistent with the results of the two experiments with lycosid spiderlings discussed in the section on habitat complexity (107, 137), because those experiments lasted 14 to 18 days, not long enough to have revealed the detrimental effects of a solely cannibalistic diet on survival and growth. The paradox persists—Why is cannibalism (and predation on closely related species) apparently common among lycosids if conspecifics are such low-quality prey? One likely explanation is that lycosids rely on cannibalism when other prey are scarce but eventually diversify their diets with heterospecific prey. An additional intriguing explanation comes from two recent experiments indicating that conspecifics may be high-quality prey, but the propensity for cannibalism among young spiderlings is a polymorphic trait. In an initial 11-week experiment, D. Mayntz &S.Toft (78) raised individual spiderlings of Pardosa prativaga on low-quality fruit flies, high-quality flies (reared on an enriched medium), and conspecifics (spiderlings in this treatment had cannibalized once before being included in the experimental design, thereby possibly overestimating the cannibalistic tendencies in the population). Overall survival of the cannibals was lower than for spiders reared on the other two diets but was higher than that observed by Toft & Wise for Schizocosa (127). The cannibals, however, consisted of two distinct groups: The majority grew slowly and exhibited the high mortality that was observed in earlier experiments, whereas 19% (5 individuals) survived for 11 weeks and grew

by UNIVERSITY OF ILLINOIS - CHICAGO on 08/23/07. For personal use only. faster than spiderlings fed high-quality fruit ﬂies. In a second experiment (78) spi-

Annu. Rev. Entomol. 2006.51:441-465. Downloaded from arjournals.annualreviews.org derlings that had fed on fruit flies and were then switched to a diet of conspecifics showed a higher growth efficiency than spiders fed solely fruit flies, indicating that conspecifics are high-quality prey. The presence of a polymorphism in cannibalistic tendency has been confirmed in an experiment with a closely related lycosid, Pardosa amentata (57). Three morphs were uncovered: Those that exhibit no cannibalism, spiderlings in which the cannibalism rate was sensitive to size asymmetries and had a latency period, and spiderlings displaying “sudden cannibalism” that occurred independently of the size ratio of cannibal and potential victim. In both studies (57, 78) most spiders showed a high latency to cannibalize, even if the size ratio was 2:1. Why are so many spiders reluctant to cannibalize 2Nov 2005 13:14 AR ANRV263-EN51-19.tex XMLPublishSM(2004/02/24) P1: KUV

CANNIBALISM IN SPIDER POPULATIONS 453

if conspecifics are high-quality prey? Perhaps the threat of retaliation is too great, given that in nature nondangerous alternative prey may appear within a few days. Perhaps the threat of acquiring a pathogen or parasite helps maintain this polymorphism (78). Contracting disease is a cost in other species (99) but has not yet been demonstrated in spiders. These studies can explain why conspecifics were low-quality prey to most of the spiders in earlier research (96, 137), but why did the earlier studies not uncover at least a few highly cannibalistic morphs that could grow, develop, and survive solely as cannibals? Species differences would not be surprising, yet one species studied earlier was in the genus Pardosa (97). Cannibalism in spiders, even in one family, the Lycosidae, is a complex behavior. Understanding its evolution and ecological significance requires that simple concepts of the spider cannibal be abandoned.

INTRASPECIFIC COMPETITION Exploitative Competition One possible indirect benefit of both IGP and cannibalism is reduction in the intensity of exploitative competition through elimination of a competitor (101). Even though prey frequently are a limited resource for spiders, exploitative competition for prey has not been documented often in web-building spiders (143). Its rarity is unlikely due to cannibalism, as cannibalism among most solitary web builders appears to be infrequent. Cannibalism appears to be common among wolf spiders; yet in this family there is evidence from field experiments that growth rates of litter-inhabiting lycosids are lower at higher densities (137, 148), which likely reflects exploitative competition for prey. This conclusion is supported by field experiments demonstrating that wolf spiders limit densities of collembolans (15, 71, 144, 148) and that increased densities of collembolans and other microbi- detritivores increase the density and individual weight of lycosid spiderlings (19, 47). There is little evidence that eliminating an interspecific competitor is the major benefit of IGP (104), and analogously, it is reasonable to argue that any reduction in intraspecific competition as a result of cannibalism in spiders is a by UNIVERSITY OF ILLINOIS - CHICAGO on 08/23/07. For personal use only. secondary, “epiphenomenal” (104) consequence of predation. Cannibalistic terri- Annu. Rev. Entomol. 2006.51:441-465. Downloaded from arjournals.annualreviews.org toriality, which is a form of interference competition that affects spider spacing, is an exception.

Interference Competition Interference, or contest, competition evolves in the context of exploitative competition for a limited resource. In spiders, contest competition can be for resources other than prey, such as web sites or the web itself. Such agonistic interactions could lead to cannibalism, since food is also a limited resource for spiders. Many researchers who study spiders appear to equate defense of the web with territoriality. Whether such behavior is territoriality depends on the deﬁnition. Despite the 2Nov 2005 13:14 AR ANRV263-EN51-19.tex XMLPublishSM(2004/02/24) P1: KUV

454 WISE

definition one might choose, defense of the web per se, and not an area in excess of the web, will not reduce exploitative competition for prey. Web builders invade the webs of conspecifics and displace them from the web, but such interactions, which often involve a series of signaling behaviors that can escalate to contact, rarely result in cannibalism (142). Wandering spiders also engage in elaborate agonistic behaviors that often lead to contact, grappling, and sometimes cannibalism, but cannibalism is rarely the outcome and usually occurs when a spider approaches another from behind (8, 60, 93). Cannibalism is likely avoided because the behaviors give the inferior contestant an opportunity to assess its chances and escape. These studies have been conducted in the laboratory, and it is not clear how frequently the behaviors occur in nature, although it is reasonable to assume that they occur any time two spiders meet face to face. These agonistic behaviors could contribute to population regulation by influencing the emigration rate from an area and by lowering the mortality rate owing to cannibalism. The most thoroughly documented example of territoriality in spiders is that of the web spider Agelenopsis aperta (Agelenidae), which reduces exploitative competition for prey by defending an area in excess of the web (108). Agonistic interactions occur between territory residents and invaders, but cannibalism is not a component of territorial defense, as injuries occur in <1% of encounters (109). Spacing patterns and behavioral observations have established that three species of burrow-inhabiting spiders exclude conspecifics from a territory in excess of the burrow (38–40, 52, 53, 75, 76, 87, 90). Cannibalism appears to be a component of territoriality in the dancing white lady spider, Leucorchestris arenicola (Heteropodidae), on the basis of spacing patterns, cannibalism rates of 3% to 8%, and reactions of burrow owners to intruders (52, 53; K. Birkhofer, personal communication). Direct evidence of the role of cannibalism in territoriality comes from field experiments with the Mediterranean tarantula, Lycosa tarantula (Lycosidae) (87). Manipula- tions of spiders in artificial burrows revealed that female spiders defended an area in excess of their burrows and that the winner of territorial disputes grew at a rate twice that of control spiders that had not engaged in a territorial contest. In a second experiment (87), approximately one third of the induced encounters between residents and intruders resulted in cannibalism, but the behavioral interactions

by UNIVERSITY OF ILLINOIS - CHICAGO on 08/23/07. For personal use only. differed from the typical predator-prey behavior of cannibalistic foraging in that

Annu. Rev. Entomol. 2006.51:441-465. Downloaded from arjournals.annualreviews.org cannibalism never occurred in the absence of escalating agonistic interactions. Whether or not an escalation resulted in cannibalism depended strongly upon the relative size of the contestants (P = 0.004), the residency status of the winner (P = 0.021), and marginally upon body condition, a measure of hunger level (P = 0.086). Thus territoriality in L. tarantula contributes to the regulation of population density through the exclusion of nonterritory owners by both agonistic interactions and cannibalism. Sexual cannibalism also likely affects the dynamics of L. tarantula populations via effects on the growth and eventual fecundity of females; a ﬁeld experiment demonstrated that strong food limitation among juvenile females was likely alleviated in the adult stage at least partly by killing and eating males (88). 2Nov 2005 13:14 AR ANRV263-EN51-19.tex XMLPublishSM(2004/02/24) P1: KUV

CANNIBALISM IN SPIDER POPULATIONS 455

POPULATION REGULATION

Cannibalistic territoriality is a special case; in principle, all types of cannibalism should regulate population density because rates of cannibalism are expected to be density dependent (28, 100, 101). Two laboratory studies have demonstrated density-dependent cannibalism between juvenile wolf spiders (107, 138), but their results fail to provide convincing indirect evidence that cannibalism might be regulating wolf spider populations in nature. Increasing the density of juvenile Pardosa palustris increased the percentage mortality by a factor of three (107), but densities were approximately 10 times higher than field densities, and the density effect may have been restricted to spiders in the presence of alternative prey (P of interaction = 0.072). This possible interaction may have been due to the relatively few spiders (5) in the low-density treatment and the high mortality rate in the absence of alternative prey at both densities (∼100% after 18 days), which incidentally suggests that conspecific P. palustris are low-quality prey. A laboratory experiment with Schizocosa ocreata employed larger arenas (0.3 m2 versus 0.004 m2), greater initial numbers of spiders per treatment (15 and 60 versus 5 and 15), and densities more similar to field conditions (50 and 200/m2 versus 1250 and 3750/m2, respectively) (138). Densities in the S. ocreata experiment approximated low- and high-density treatments in a related field experiment (137). Mortality from cannibalism was high, but the overall effect of density was only marginally significant (P = 0.084), and there was no interaction between spider density and the presence or absence of alternative prey. Furthermore, there was no hint that cannibalism was density dependent in the presence of alternative prey, the treatment more similar to conditions in nature. Thus these two laboratory experiments do not provide evidence, even indirect, that cannibalism might regulate lycosid populations. Field experiments in small (0.25 to 2.25 m2) fenced plots have found, and have failed to find, density-dependent mortality in lycosids (16, 137, 148). Buddle et al. (16) decreased the 7-day survival rate of adult and juvenile Pardosa milvina by 17% in soybean fields by increasing spider density four times over normal levels. In a forest experiment, doubling densities of juvenile Pardosa moesta and

by UNIVERSITY OF ILLINOIS - CHICAGO on 08/23/07. For personal use only. P. mackenziana had no impact on survival (15), but simply doubling spider den-

Annu. Rev. Entomol. 2006.51:441-465. Downloaded from arjournals.annualreviews.org sity also had no effect on spider survival in soybeans (16). Establishing densities of young juvenile S. ocreata above (four times), below (one fourth) and equal to the estimated mean density in the forest revealed strong density-dependent survival: After 74 days numbers in the high-density and mean-density treatments had converged, and densities in the low-density treatment had converged partially with those in the mean-density treatment (from 25% to 50% of the number in the mean-density treatment) (148). Some of the convergence of S. ocreata densities may have been due to density-dependent emigration (138), and density-dependent survival in both experiments (16, 148) could have been caused by predation by other species and not cannibalism, as other arthropod predators were not removed from the enclosures. 2Nov 2005 13:14 AR ANRV263-EN51-19.tex XMLPublishSM(2004/02/24) P1: KUV

456 WISE

In a second experiment with S. ocreata, all predators found by sifting the litter were removed before S. ocreata densities were manipulated in enclosures with more effective barriers to emigration (137). Mortality over 2.5 months was high (∼80%) and strongly density dependent, leading low- and high-density treatments (0.75 time and 3 times the estimated natural densities, respectively) to converge to a twofold difference. The absence of complete convergence suggests that emigration wasacomponent of the density-dependent response in the first study. Reducing predator densities had no effect on S. ocreata survival, implicating cannibalism as a major component of the density-dependent mortality. The laboratory experiment (137) in which leaf litter and alternative prey affected rates of cannibalism supports this interpretation, because the rate of cannibalism in the leaf-litter, collembolan treatment was similar to the rate of mortality observed in the high-density treatment of the field experiment. Two other experiments in which IGP predators were reduced but Schizocosa spp. still displayed high mortality support the interpretation that cannibalism can contribute to the regulation of densities of juvenile Schizocosa spp. (73, 145). Autumn densities of juvenile Schizocosa spp. show remarkable consistency between different years and different forests (19, 73, 137, 146, 148). It is tempting to speculate that strong density-dependent cannibalism, in the presence of shifting mortality from IGP from other spiders, centipedes, and predacious insects, contributes to this apparent stability. However, none of these field experiments quantify the relative contribution of cannibalism and IGP to lycosid mortality in the presence of other predators of Schizocosa.Nevertheless, the field and accompanying laboratory experiments do demonstrate the potential of cannibalism to regulate wolf spider populations.

TROPHIC CASCADES

Strong density-dependent cannibalism may limit spider populations below the carrying capacity set by resources (100), which should reduce the ability of spiders to exert strong trophic cascades. The existence of cannibalism, however, is evidence of severe food limitation, which suggests that the cannibal could be depressing

by UNIVERSITY OF ILLINOIS - CHICAGO on 08/23/07. For personal use only. populations in lower trophic levels enough to initiate a signiﬁcant trophic cascade.

Annu. Rev. Entomol. 2006.51:441-465. Downloaded from arjournals.annualreviews.org Meta-analyses of ﬁeld experiments with generalist predators, many with spiders, reveal that spiders can induce trophic cascades in a variety of terrestrial systems (46, 117). Accumulating evidence from agroecosystems suggests that generalist predators, including spiders, can act as effective biological control agents (95, 126), further evidence that spider cannibalism does not prevent at least some species from impacting lower trophic levels. The question then, is not whether spiders can induce trophic cascades, but rather, to what extent does spider cannibalism dampen cascades? Rarely has this question been addressed directly. Cannibalism has the potential to maintain spider populations during periods of low abundance of prey at lower trophic levels (the “lifeboat strategy”) (100), thereby promoting trophic cascades in the future when conditions change. 2Nov 2005 13:14 AR ANRV263-EN51-19.tex XMLPublishSM(2004/02/24) P1: KUV

CANNIBALISM IN SPIDER POPULATIONS 457

I restrict this part of the review to field experiments with enclosures and/or cages that include densities and habitat complexity similar to the field situation, whether it be a natural or highly managed agricultural system, and to experiments that demonstrate unambiguous effects of wandering spiders or, in one case, a web builder that exhibits significant rates of cannibalism. Spider-removal experiments reveal that ambient densities of wandering spiders, or spider complexes with large numbers of wanderers, can initiate cascades of indirect effects in grazing food webs that affect productivity in vegetable crops (124), plant damage in grassland systems (116), rates of decomposition in meadows (64), and litter decomposition in forests, where the effect can be to either inhibit (71) or enhance (72) rates of decomposition. Thus, in these systems cannibalism may dampen spider-initiated trophic cascades but does not prevent them. Polis & Strong (105) proposed that trophic cascades in grazing webs can be increased by subsidies from other webs that lead to elevated densities of generalist predators, owing to alleviation of the effects of food limitation, dampening of IGP, and a reduction in the intensity of cannibalism. Several examples of such subsidies exist for wandering spiders. Emergence of aquatic insects and their movement to adjacent terrestrial habitats can lead to increased densities of both web-building and wandering species (21, 115), which in one instance resulted in decreased damage to plants (54). Doubling the immigration rate of wolf spiders into vegetable gardens caused densities of spiderlings, but not adults, to increase, possibly because density-dependent cannibalism among the spiderlings and older juveniles affected adult numbers (123). In a subsequent experiment, adding a detrital subsidy caused all stages of wolf spiders to increase, perhaps owing to reduced cannibalism among young instars in response to increased densities of microbi-detritivores. However, the elevated lycosid numbers failed to control crop pests (47). The absence of an enhanced trophic cascade was likely due to an unusually high density of one pest, and the absence of another later in the season (47), and was not due to another possibility, i.e., that large wolf spiders might shift their feeding away from the crop-based food web to the detrital web (147). Thus, enhancing densities of prey in the fungal-based food web promises to be one technique for enhancing wolf spider numbers, owing at least partly to reduced cannibalism, as a way to increase

by UNIVERSITY OF ILLINOIS - CHICAGO on 08/23/07. For personal use only. their effectiveness in biological control.

Annu. Rev. Entomol. 2006.51:441-465. Downloaded from arjournals.annualreviews.org Wolf spiders can limit populations of some planthopper species in Spartina salt marshes (25, 45). Laboratory mesocosm experiments suggest that spider predation may initiate a trophic cascade affecting Spartina,but field experiments so far have not uncovered a limitation strong enough to induce a detectable trophic cascade in the field (26). There exists a complex interaction between plant nutrient content, amount of thatch cover, and the ability of wolf spiders to limit planthoppers. The presence of thatch enhances the ability of wolf spiders to limit planthopper densities (26), perhaps because thatch lowers rates of wolf spider cannibalism; a meta- analysis reveals that manipulating detritus has particularly strong positive effects on spider densities (70). Laboratory studies revealed moderate rates of cannibalism among the web builder Grammonota trivatatta (Linyphiidae), and predator- addition experiments in the field suggest that limited web sites, cannibalism, 2Nov 2005 13:14 AR ANRV263-EN51-19.tex XMLPublishSM(2004/02/24) P1: KUV

458 WISE

or both may limit spider numbers in the marsh. However, predation rates on planthoppers by individual G. trivatatta are much lower than those of individual wolf spiders in the system (27), which indicates that neither cannibalism among G. trivatatta nor IGP by Pardosa littoralis, which has been documented, explains the inability of this web builder to control planthopper populations. Thus, research in vegetable crops and salt marshes reveals how different environmental factors can inﬂuence rates of cannibalism among spiders, which in principle should affect their ability to initiate trophic cascades, but the studies also highlight the complex interplay of other factors that affect spider predation on herbivore populations. The extent to which cannibalism limits the strength of trophic cascades initiated by spiders remains largely unknown.

FUTURE RESEARCH

This review has uncovered several unanswered or only partially answered questions about the roles of cannibalistic foraging in the population dynamics and food web relationships of spiders: Are the frequencies, ecological correlates, and evolved behaviors of cannibalistic foraging in wolf spiders (Lycosidae), the group most intensively studied to date, representative of other families of wandering spiders? Are conspeciﬁcs high-quality prey in terms of nutrient content? Can costs other than the threat of retaliation—such as the acquisition of pathogens and parasites—explain the low rates of cannibalism exhibited by some spiders? Are behavioral polymorphisms in cannibalistic tendency generally characteristic of spider cannibalism, and if not, can their occurrence or absence be correlated with environmental variables? Do agonistic displays occur frequently in natural populations, and to what extent do they reduce mortality from cannibalism?

by UNIVERSITY OF ILLINOIS - CHICAGO on 08/23/07. For personal use only. Is spider cannibalism in nature strongly density dependent, and does it con- Annu. Rev. Entomol. 2006.51:441-465. Downloaded from arjournals.annualreviews.org tribute to population stability in spiders? Does cannibalism inhibit (by dampening trophic cascades) or enhance (as a lifeboat strategy) the effectiveness of spiders in the control of agricultural pests? Can the properties of cannibalistic encounters, from frequency to behavioral mechanisms, be accurately described from individual pairings in the laboratory? What is the actual frequency in nature of cannibalism in different spider families, and does the pattern conﬁrm the hypothesized relationship between foraging mode and rates of cannibalism? 2Nov 2005 13:14 AR ANRV263-EN51-19.tex XMLPublishSM(2004/02/24) P1: KUV

CANNIBALISM IN SPIDER POPULATIONS 459

ACKNOWLEDGMENTS Many thanks to James Harwood, David Mayntz, Klaus Birkhofer, Erin Hladilek, and Janet Lensing for their helpful comments on a draft of this review.

The Annual Review of Entomology is online at http://ento.annualreviews.org

LITERATURE CITED 1. Alderweireldt M. 1994. Prey selection and Locket (Araneae: Linyphiidae) in single- prey capture strategies of linyphiid spiders species and mixed-species diets. Ekologia in high-input agricultural fields. Bull. Br. 19:9–18 Arachnol. Soc. 9:300–8 13. Bilde T, Toft S. 2001. The value of three 2. Andrade MCB. 1996. Sexual selection for cereal aphid species as food for a general- male sacrifice in the Australian redback ist predator. Physiol. Entomol. 26:58–68 spider. Science 271:70–72 14. Bristowe WS. 1941. The Comity of Spi- 3. Andrade MCB. 1998. Female hunger can ders,Vol. 2. London: Ray Soc. 332 pp. explain variation in cannibalistic behavior 15. Buddle CM. 2002. Interactions among despite male sacrifice in redback spiders. young stages of the wolf spiders Pardosa Behav. Ecol. 9:33–42 moesta and P. mackenziana (Araneae: Ly- 4. Andrade MCB. 2003. Risky mate search cosidae). Oikos 96:130–36 and male self-sacrifice in redback spiders. 16. Buddle CM, Walker SE, Rypstra AL. Behav. Ecol. 14:531–38 2003. Cannibalism and density-depend- 5. Andrade MCB, Banta EM. 2002. Value of ent mortality in the wolf spider Pardosa male remating and functional sterility in milvina (Araneae: Lycosidae). Can. J. redback spiders. Anim. Behav. 63:857–70 Zool. Rev. Can. Zool. 81:1293–97 6. Anthony CD. 2003. Kinship influences 17. Buskirk RE. 1981. Sociality in the Arach- cannibalism in the wolf spider, Pardosa nida. In Social Insects, ed. HR Hermann, milvina. J. Insect Behav. 16:23–36 pp. 2:281–367. New York: Academic. 7. Arnqvist G, Henriksson S. 1997. Sexual 491 pp. cannibalism in the fishing spider and a 18. Buskirk RE, Frohlich C, Ross KG. 1984. model for the evolution of sexual canni- The natural selection of sexual cannibal- balism based on genetic constraints. Evol. ism. Am. Nat. 123:612–25 Ecol. 11:255–73 19. Chen BR, Wise DH. 1999. Bottom-up 8. Aspey WP. 1975. Ontogeny of display in limitation of predaceous arthropods in a

by UNIVERSITY OF ILLINOIS - CHICAGO on 08/23/07. For personal use only. immature Schizocosa crassipes (Araneae: detritus-based terrestrial food web. Ecol- Annu. Rev. Entomol. 2006.51:441-465. Downloaded from arjournals.annualreviews.org Lycosidae). Psyche 82:174–80 ogy 80:761–72 9. Avil´es L. 1997. Causes and consequences 19a. Choe JC, Crespi BJ, eds. 1997. The Evo- of cooperation and permanent sociality in lution of Social Behavior in Insects and spiders. See Ref. 19a, pp. 476–98 Arachnids. Cambridge, UK: Cambridge 10. Bardwell CJ, Averill AL. 1997. Spiders Univ. Press. 541 pp. and their prey in Massachusetts cranberry 20. Clark RJ, Jackson RR. 1994. Portia bogs. J. Arachnol. 25:31–41 labiata,acannibalistic jumping spider, 11. Bilde T, Lubin Y. 2001. Kin recognition discriminates between own and foreign and cannibalism in a subsocial spider. J. eggsacs. Int. J. Comp. Psychol. 7:38–43 Evol. Biol. 14:959–66 21. Collier KJ, Bury S, Gibbs M. 2002. A 12. Bilde T, Toft S. 2000. Evaluation of prey stable isotope study of linkages between for the spider Dicymbium brevisetosum stream and terrestrial food webs through 2Nov 2005 13:14 AR ANRV263-EN51-19.tex XMLPublishSM(2004/02/24) P1: KUV

460 WISE

spider predation. Freshw. Biol. 47:1651– Diverse Taxa. Oxford, UK: Oxford Univ. 59 Press. 361 pp. 22. Costa FG, Pérez-Miles F. 2002. Re- 33. Evans TA. 1998. Factors influenc- productive biology of Uruguayan ther- ing the evolution of social behaviour aphosids (Araneae, Mygalomorphae). J. in Australian crab spiders (Araneae: Arachnol. 30:571–87 Thomisidae). Biol. J. Linn. Soc. 63:205– 23. D’Andrea M. 1987. Social behaviour in 19 spiders (Arachnida: Araneae). Mon. Zool. 34. Evans TA. 1998. Offspring recognition by Ital. Ital. J. Zool. 3:1–156 mother crab spiders with extreme mater- 24. Denno RF, Fagan WF. 2003. Might nal care. Proc. R. Soc. London Ser. B Biol. nitrogen limitation promote omnivory Sci. 265:129–34 among carnivorous arthropods? Ecology 35. Evans TA. 1999. Kin recognition in a so- 84:2522–31 cial spider. Proc. R. Soc. London Ser. B 25. Denno RF, Gratton C, Döbel H, Finke Biol. Sci. 266:287–92 DL. 2003. Predation risk affects rela- 36. Evans TA, Wallis EJ, Elgar MA. 1995. tive strength of top-down and bottom- Making a meal of mother. Nature 376: up impacts on insect herbivores. Ecology 299 84:1032–44 37. Fagan WF, Siemann E, Mitter C, Denno 26. Denno RF, Gratton C, Peterson MA, RF, Huberty AF, et al. 2002. Nitrogen in Langellotto GA, Finke DL, Huberty insects: implications for trophic complex- AF. 2002. Bottom-up forces mediate ity and species diversification. Am. Nat. natural-enemy impact in a phytophagous 160:784–802 insect community. Ecology 83:1443– 38. Fernández-Montraveta C, Lahoz-Beltra 58 R, Ortega J. 1991. Spatial distribution of 27. Denno RF, Mitter MS, Langellotto GA, Lycosa tarentula fasciiventris (Araneae, Gratton C, Finke DL. 2004. Interactions Lycosidae) in a population from central between a hunting spider and a web- Spain. J. Arachnol. 19:73–79 builder: consequences of intraguild pre- 39. Fernández-Montraveta C, Ortega J. 1990. dation and cannibalism for prey suppres- El comportamiento agon´ıstico de hem- sion. Ecol. Entomol. 29:566–77 bras adultas de Lycosa tarentula fasci- 28. Dong Q, Polis GA. 1992. The dynamics of iventris (Araneae, Lycosidae). J. Arach- cannibalistic populations: a foraging per- nol. 18:49–58 spective. See Ref. 32a, pp. 13–37 40. Fernández-Montraveta C, Ortega J. 1991. 29. Eason RR. 1969. Life history and be- Owner-biased agonistic behavior in fe- havior of Pardosa lapidicina Emerton male Lycosa tarentula fasciiventris

by UNIVERSITY OF ILLINOIS - CHICAGO on 08/23/07. For personal use only. (Araneae: Lycosidae). J. Kans. Entomol. (Araneae, Lycosidae). J. Arachnol. 19:

Annu. Rev. Entomol. 2006.51:441-465. Downloaded from arjournals.annualreviews.org Soc. 42:339–60 80–84 30. Edgar WD. 1969. Prey and predators of 41. Foelix RF. 1996. Biology of Spi- the wolf spider Lycosa lugubris. J. Zool. ders. Oxford, UK: Oxford Univ. Press. 159:405–11 330 pp. 31. Elgar MA. 1992. Sexual cannibalism in 42. Foellmer MW, Fairbairn DJ. 2004. Males spiders and other invertebrates. See Ref. under attack: sexual cannibalism and its 32a, pp. 128–55 consequences for male morphology and 32. Elgar MA, Crespi BJ. 1992. Ecology and behaviour in an orb-weaving spider. Evol. evolution of cannibalism. See Ref. 32a, Ecol. Res. 6:163–81 pp. 1–12 43. Fox LR. 1975. Cannibalism in natural 32a. Elgar MA, Crespi BJ, eds. 1992. Can- populations. Annu. Rev. Ecol. Syst. 6:87– nibalism: Ecology and Evolution among 106 2Nov 2005 13:14 AR ANRV263-EN51-19.tex XMLPublishSM(2004/02/24) P1: KUV

CANNIBALISM IN SPIDER POPULATIONS 461

44. Gertsch WJ. 1979. American Spiders. in river shore food webs. Oikos 93:429– New York: Van Nostrand Reinhold. 274 38 pp. 55. Hoffmaster DK. 1986. Aggression in trop- 45. Gratton C, Denno RF. 2003. Seasonal ical orb-weaving spiders: a quest for food. shift from bottom-up to top-down impact Ethology 72:265–76 in phytophagous insect populations. Oe- 56. Horel A, Gundermann JL. 1992. Egg cologia 134:487–95 sac guarding by the funnel-web spider 46. Halaj J, Wise DH. 2001. Terrestrial Coelotes terrestris: function and develop- trophic cascades: How much do they ment. Behav. Process. 27:85–93 trickle ? Am. Nat. 157:262–81 57. Hvam A, Mayntz D, Nielsen RK. 2005. 47. Halaj J, Wise DH. 2002. Impact of a detri- Factors affecting cannibalism among tal subsidy on trophic cascades in a terres- newly hatched wolf spiders (Lycosidae: trial grazing food web. Ecology 83:3141– Pardosa amentata). J. Arachnol. 53: 51 In press 48. Hallander H. 1970. Prey, cannibalism and 58. Iida H. 2003. Small within-clutch vari- microhabitat selection in the wolf spiders ance in spiderling body size as a mech- Pardosa chelata Müller and P. pullata anism for avoiding sibling cannibalism in Clerck. Oikos 21:337–40 the wolf spider Pardosa pseudoannulata 49. Harwood JD, Sunderland KD, Symond- (Araneae: Lycosidae). Popul. Ecol. 45:1– son WOC. 2001. Living where the food 6 is: web location by linyphiid spiders in re- 59. Jackson RR. 1980. Cannibalism as a fac- lation to prey availability in winter wheat. tor in the mating strategy of the jump- J. Appl. Ecol. 38:88–99 ing spider Phidippus johnsoni (Araneae, 50. Harwood JD, Sunderland KD, Symond- Salticidae). Bull. Br. Arachnol. Soc. 5: son WOC. 2003. Web-location by 129–33 linyphiid spiders: prey-specific aggrega- 60. Jackson RR, Harding DP. 1982. Intraspe- tion and foraging strategies. J. Anim. Ecol. cific interactions of Holoplatys sp. Indet, 72:745–56 aNew Zealand jumping spider (Araneae, 51. Harwood JD, Sunderland KD, Symond- Salticidae). N. Z. J. Zool. 9:487–510 son WOC. 2004. Prey selection by 61. Jackson RR, Pollard SD. 1990. In- linyphiid spiders: molecular tracking of traspecific interactions and the func- the effects of alternative prey on rates of tion of courtship in mygalomorph spi- aphid consumption in the field. Mol. Ecol. ders: a study of Porrhothele antipodiana 13:3549–60 (Araneae, Hexathelidae) and a literature 52. Henschel JR. 1990. The biology of review. N. Z. J. Zool. 17:499–526

by UNIVERSITY OF ILLINOIS - CHICAGO on 08/23/07. For personal use only. Leucorchestris arenicola (Araneae: Het- 62. Johnson JC. 2001. Sexual cannibalism

Annu. Rev. Entomol. 2006.51:441-465. Downloaded from arjournals.annualreviews.org eropodidae), a burrowing spider of the in ﬁshing spiders (Dolomedes triton): an Namib Dunes. In Namib Ecology: 25 evaluation of two explanations for female Years of Namib Research, ed. MK Seely, aggression towards potential mates. Anim. pp. 115–27. Pretoria, S. Afr.: Transvaal Behav. 61:905–14 Mus. 230 pp. 63. Johnson JC. 2005. The role of body size in 53. Henschel JR. 1994. Diet and foraging be- mating interactions of the sexually canni- havior of huntsman spiders in the Namib balistic ﬁshing spider Dolomedes triton. Dunes (Araneae, Heteropodidae). J. Zool. Ethology 111:51–61 234:239–51 64. Kajak A. 1997. Effects of epigeic 54. Henschel JR, Mahsberg D, Stumpf H. macroarthropods on grass litter decompo- 2001. Allochthonous aquatic insects in- sition in mown meadow. Agric. Ecosyst. crease predation and decrease herbivory Environ. 64:53–63 2Nov 2005 13:14 AR ANRV263-EN51-19.tex XMLPublishSM(2004/02/24) P1: KUV

462 WISE

65. Kaston BJ. 1965. Some little known as- 77. Mayntz D, Toft S. 2001. Nutrient com- pects of spider behavior. Am. Midl. Nat. position of the prey’s diet affects growth 73:336–55 and survivorship of a generalist predator. 66. Kim KW, Horel A. 1998. Matriphagy Oecologia 127:207–13 in the spider Amaurobius ferox (Aranei- 78. Mayntz D, Toft S. 2006. Nutritional value dae, Amaurobiidae): an example of of cannibalism and the role of starva- mother-offspring interactions. Ethology tion and nutrient imbalance for cannibal- 104:1021–37 istic tendencies in a generalist predator. J. 67. Kim KW, Roland C, Horel A. 2000. Func- Anim. Ecol. 75:In press tional value of matriphagy in the spider 79. Miller GL. 1989. Subsocial organization Amaurobius ferox. Ethology 106:729–42 and behavior in broods of the obligate 68. Kiritani K, Kawahara S, Sasaba T, Naka- burrowing wolf spider Geolycosa turri- suji F. 1972. Quantitative evaluation of cola (Treat). Can. J. Zool. Rev. Can. Zool. predation by spiders on the green rice 67:819–24 leafhopper, Nephotettix cincticeps Uhler, 80. Miyashita T. 1991. Direct evidence of by a sight-count method. Res. Popul. Ecol. food limitation for growth rate and body 13:187–200 size in the spider Nephila clavata. Acta 69. Kreiter NA, Wise DH. 2001. Prey avail- Arachnol. 40:17–21 ability limits fecundity and influences the 81. Miyashita T. 1992. Food limitation of movement pattern of female fishing spi- population density in the orb-web spiders. Oecologia 127:417–24 der, Nephila clavata. Res. Popul. Ecol. 70. Langellotto GA, Denno RF. 2004. Re- 34:143–53 sponses of invertebrate natural enemies 82. Miyashita T. 1992. Variability in food- to complex-structured habitats: a meta- consumption rate of natural populations analytical synthesis. Oecologia 139:1–10 in the spider, Nephila clavata. Res. Popul. 71. Lawrence KL, Wise DH. 2000. Spider Ecol. 34:15–28 predation on forest-floor Collembola and 83. Miyashita T, Shinkai A, Chida T. 1998. evidence for indirect effects on decompo- The effects of forest fragmentation on web sition. Pedobiologia 44:33–39 spider communities in urban areas. Biol. 72. Lawrence KL, Wise DH. 2004. Unex- Conserv. 86:357–64 pected indirect effect of spiders on the rate 84. Morse DH. 2004. A test of sexual canni- of litter disappearance in a deciduous for- balism models, using a sit-and-wait preda- est. Pedobiologia 48:149–57 tor. Biol. J. Linn. Soc. 81:427–37 73. Lensing JR, Wise DH. 2004. A test of the 85. Moya-Laraño J. 2002. Senescence and hypothesis that a pathway of intraguild food limitation in a slowly ageing spider.

by UNIVERSITY OF ILLINOIS - CHICAGO on 08/23/07. For personal use only. predation limits densities of a wolf spider. Funct. Ecol. 16:734–41

Annu. Rev. Entomol. 2006.51:441-465. Downloaded from arjournals.annualreviews.org Ecol. Entomol. 29:294–99 86. Moya-Laraño J, Barrientos JA, Orta- 74. Lubin Y, Henschel J. 1996. The influence Ocaña JM, Bach Piella C, Wise DH. 1998. of food supply on foraging behaviour in a Limitación por la comida en las tarántulas desert spider. Oecologia 105:64–73 de Cabo de Gata (Almer´ıa). Investig. Gest. 75. Marshall SD. 1995. Mechanisms of the Medio Nat. 3:73–77 formation of territorial aggregations of 87. Moya-Laraño J, Orta-Ocaña JM, Barrien- the burrowing wolf spider Geolycosa xera tos JA, Bach C, Wise DH. 2002. Territo- archboldi McCrone (Araneae, Lycosi- riality in a cannibalistic burrowing wolf dae). J. Arachnol. 23:145–50 spider. Ecology 83:356–61 76. Marshall SD. 1996. Evidence for territo- 88. Moya-Laraño J, Orta-Ocaña JM, Barrien- rial behavior in a burrowing wolf spider. tos JA, Bach C, Wise DH. 2003. Intriguing Ethology 102:32–39 compensation by adult female spiders for 2Nov 2005 13:14 AR ANRV263-EN51-19.tex XMLPublishSM(2004/02/24) P1: KUV

CANNIBALISM IN SPIDER POPULATIONS 463

food limitation experienced as juveniles. namics of intraspecific predation. Annu. Oikos 101:539–48 Rev. Ecol. Syst. 12:225–51 89. Moya-Laraño J, Pascual J, Wise DH. 101. Polis GA. 1988. Exploitation competi- 2004. Approach strategy by which male tion and the evolution of interference, Mediterranean tarantulas adjust to the cannibalism, and intraguild predation in cannibalistic behaviour of females. Ethol- age/size-structured populations. In Size- ogy 110:717–24 Structured Populations: Ecology and 90. Moya-Laraño J, Taylor PW, Fernández- Evolution, ed. B Ebenman, pp. 185–202. Montraveta C. 2003. Body patterns as po- Berlin: Springer-Verlag. 284 pp. tential amplifiers of size and condition 102. Polis GA. 1991. Food webs in desert com- in a territorial spider. Biol. J. Linn. Soc. munities: complexity via diversity and 78:355–64 omnivory. In The Ecology of Desert Com- 91. Nentwig W. 1987. The prey of spiders. In munities, ed. GA Polis, pp. 383–437. Tuc- Ecophysiology of Spiders, ed. W Nentwig, son: Univ. Arizona Press. 456 pp. pp. 249–63. Berlin: Springer-Verlag. 448 103. Polis GA, Hurd SD, Jackson CT, Sánchez- pp. Piñero F. 1998. Multifactor population 92. Newman JA, Elgar MA. 1991. Sexual can- limitation: variable spatial and temporal nibalism in orb-weaving spiders: an eco- control of spiders on Gulf of California nomic model. Am. Nat. 138:1372–95 islands. Ecology 79:490–502 93. Nossek ME, Rovner JS. 1984. Agonistic 104. Polis GA, Myers CA, Holt RD. 1989. The behavior in female wolf spiders (Araneae, ecology and evolution of intraguild pre- Lycosidae). J. Arachnol. 11:407–22 dation: potential competitors that eat each 94. Nyffeler M. 1999. Prey selection of spi- other. Annu. Rev. Ecol. Syst. 20:297–330 ders in the field. J. Arachnol. 27:317–24 105. Polis GA, Strong DR. 1996. Food web 95. Nyffeler M, Sunderland KD. 2003. Com- complexity and community dynamics. position, abundance and pest control po- Am. Nat. 147:813–46 tential of spider communities in agroe- 106. Pollard SD. 1983. Egg guarding by Clu- cosystems: a comparison of European biona cambridgei (Araneae, Clubionidae) and US studies. Agric. Ecosyst. Environ. against conspecific predators. J. Arach- 95:579–612 nol. 11:323–26 96. Oelbermann K, Scheu S. 2002. Effects of 107. Rickers S, Scheu S. 2005. Cannibalism in prey type and mixed diets on survival, Pardosa palustris (Araneae, Lycosidae): growth and development of a generalist effects of alternative prey, habitat struc- predator, Pardosa lugubris (Araneae: Ly- ture, and density. Basic Appl. Ecol. 6:In cosidae). Basic Appl. Ecol. 3:285–91 press

by UNIVERSITY OF ILLINOIS - CHICAGO on 08/23/07. For personal use only. 97. Oelbermann K, Scheu S. 2002. Stable 108. Riechert SE. 1981. The consequences of

Annu. Rev. Entomol. 2006.51:441-465. Downloaded from arjournals.annualreviews.org isotope enrichment (δ15N and δ13C) in being territorial: spiders, a case study. Am. a generalist predator (Pardosa lugubris, Nat. 117:871–92 Araneae: Lycosidae): effects of prey qual- 109. Riechert SE. 1982. Spider interaction ity. Oecologia 130:337–44 strategies: communication vs. coercion. 98. Persons MH, Uetz GW. 2005. Sexual In Spider Communication: Mechanisms cannibalism and mate choice decisions and Ecological Signiﬁcance, ed. PN Witt, in wolf spiders: inﬂuence of male size pp. 281–315. Princeton, NJ: Princeton and secondary characters. Anim. Behav. Univ. Press. 440 pp. 69:83–94 110. Riechert SE, Cady AB. 1983. Patterns 99. Pfennig DW. 1997. Kinship and cannibal- of resource use and tests for competitive ism. Bioscience 47:667–75 release in a spider community. Ecology 100. Polis GA. 1981. The evolution and dy- 64:899–913 2Nov 2005 13:14 AR ANRV263-EN51-19.tex XMLPublishSM(2004/02/24) P1: KUV

464 WISE

111. Roberts JA, Taylor PW, Uetz GW. mosana (Oi) (Araneae: Linyphiidae): im- 2003. Kinship and food availability influ- plications for biological control in rice. ence cannibalism tendency in early-instar Biocontrol Sci. Technol. 11:233–44 wolf spiders (Araneae: Lycosidae). Behav. 123. Snyder WE, Wise DH. 1999. Predator in- Ecol. Sociobiol. 54:416–22 terference and the establishment of gener- 112. Rypstra AL. 1983. The importance of alist predator populations for biocontrol. food and space in limiting web-spider Biol. Control 15:283–92 densities: a test using field enclosures. Oe- 124. Snyder WE, Wise DH. 2001. Contrast- cologia 59:312–16 ing trophic cascades generated by a com- 113. Rypstra AL. 1986. High prey abundance munity of generalist predators. Ecology and a reduction in cannibalism: the first 82:1571–83 step to sociality in spiders (Arachnida). J. 125. Spence JR, Zimmermann M, Wojcicki JP. Arachnol. 14:193–200 1996. Effects of food limitation and sex- 114. Samu F, Toft S, Kiss B. 1999. Factors ual cannibalism on reproductive output of influencing cannibalism in the wolf spi- the nursery web spider Dolomedes triton der Pardosa agrestis (Araneae, Lycosi- (Araneae: Pisauridae). Oikos 75:373–82 dae). Behav. Ecol. Sociobiol. 45:349–54 126. Symondson WOC, Sunderland KD, 115. Sanzone DM, Meyer JL, Marti E, Gar- Greenstone MH. 2002. Can generalist diner EP, Tank JL, Grimm NB. 2003. Car- predators be effective biocontrol agents? bon and nitrogen transfer from a desert Annu. Rev. Entomol. 47:561–94 stream to riparian predators. Oecologia 127. Toft S, Wise DH. 1999. Growth, develop- 134:238–50 ment, and survival of a generalist predator 116. Schmitz OJ. 2003. Top predator control fed single- and mixed-species diets of dif- of plant biodiversity and productivity in ferent quality. Oecologia 119:191–97 an old-field ecosystem. Ecol. Lett. 6:156– 128. Toyama M. 1999. Adaptive advantages 63 of maternal care and matriphagy in a fo- 117. Schmitz OJ, Hamback PA, Beckerman liage spider, Chiracanthium japonicum AP. 2000. Trophic cascades in terres- (Araneae: Clubionidae). J. Ethol. 17:33– trial systems: a review of the effects of 39 carnivore removals on plants. Am. Nat. 129. Toyama M. 2001. Adaptive advantages of 155:141–53 matriphagy in the foliage spider, Chira- 118. Schneider JM. 2002. Reproductive state canthium japonicum (Araneae: Clubion- and care giving in Stegodyphus (Araneae: idae). J. Ethol. 19:69–74 Eresidae) and the implications for the evo- 130. Toyama M. 2003. Relationship between lution of sociality. Anim. Behav. 63:649– reproductive resource allocation and re-

by UNIVERSITY OF ILLINOIS - CHICAGO on 08/23/07. For personal use only. 58 source capacity in the matriphagous spi-

Annu. Rev. Entomol. 2006.51:441-465. Downloaded from arjournals.annualreviews.org 119. Schneider JM, Lubin Y. 1996. Infanticidal der, Chiracanthium japonicum (Araneae: male eresid spiders. Nature 381:655–56 Clubionidae). J. Ethol. 21:1–7 120. Schneider JM, Lubin Y. 1998. Intersexual 131. Trabalon M, Bagn`eres AG, Hartmann conﬂict in spiders. Oikos 83:496–506 N, Vallet AM. 1996. Changes in cutic- 121. Seibt U, Wickler W. 1987. Gerontophagy ular compounds composition during the versus cannibalism in the social spiders gregarious period and after dispersal of Stegodyphus mimosarum Pavesi and Ste- the young in Tegenaria atrica (Araneae, godyphus dumicola Pocock. Anim. Behav. Agelenidae). Insect Biochem. Mol. Biol. 35:1903–5 26:77–84 122. Sigsgaard L, Toft S, Villareal S. 2001. 132. Trabalon M, Pourie G, Hartmann N. 1998. Diet-dependent survival, development Relationships among cannibalism, con- and fecundity of the spider Atypena for- tact signals, ovarian development and 2Nov 2005 13:14 AR ANRV263-EN51-19.tex XMLPublishSM(2004/02/24) P1: KUV

CANNIBALISM IN SPIDER POPULATIONS 465

ecdysteroid levels in Tegenaria atrica spiders (Araneae, Araneidae). Oecologia (Araneae, Agelenidae). Insect Biochem. 48:252–56 Mol. Biol. 28:751–58 142. Wise DH. 1983. Competitive mechanisms 133. Uetz GW, Bischoff J, Raver J. 1992. in a food-limited species: relative im- Survivorship of wolf spiders (Lycosidae) portance of interference and exploita- reared on different diets. J. Arachnol. tive interactions among labyrinth spiders 20:207–11 (Araneae, Araneidae). Oecologia 58:1–9 134. Uetz GW, Hieber CS. 1997. Colonial 143. Wise DH. 1993. Spiders in Ecological web-building spiders: balancing the costs Webs. Cambridge, UK: Cambridge Univ. and benefits of group living. See Ref. 19a, Press. 328 pp. pp. 458–75 144. Wise DH. 2004. Wandering spiders limit 135. Vanacker D, Deroose K, Pardo S, Bonte densities of a major microbi-detritivore in D, Maelfait JP. 2004. Cannibalism and the forest-floor food web. Pedobiologia prey sharing among juveniles of the spider 48:181–88 Oedothorax gibbosus (Blackwall, 1841) 145. Wise DH, Chen BR. 1999. Impact of in- (Erigoninae, Linyphiidae, Araneae). Belg. traguild predators on survival of a forest- J. Zool. 134:23–28 floor wolf spider. Oecologia 121:129– 136. Wagner JD. 1995. Egg sac inhibits filial 37 cannibalism in the wolf spider Schizocosa 146. Wise DH, Chen BR. 1999. Vertebrate pre- ocreata. Anim. Behav. 50:555–57 dation does not limit density of a common 137. Wagner JD, Wise DH. 1996. Cannibalism forest-floor wolf spider: evidence from a regulates densities of young wolf spiders: field experiment. Oikos 84:209–14 evidence from field and laboratory exper- 147. Wise DH, Moldenhauer DM, Halaj J. iments. Ecology 77:639–52 2006. Using stable isotopes to reveal shifts 138. Wagner JD, Wise DH. 1997. Influence of in prey consumption by generalist preda- prey availability and conspecifics on patch tors. Ecol. Appl. 16:In press quality for a cannibalistic forager: labo- 148. Wise DH, Wagner JD. 1992. Evidence ratory experiments with the wolf spider of exploitative competition among young Schizocosa. Oecologia 109:474–82 stages of the wolf spider Schizocosa ocre- 139. Waldman B, Frumhoff PC, Sherman PW. ata. Oecologia 91:7–13 1988. Problems of kin recognition. Trends 149. Yeargan KV. 1975. Prey and periodicity Ecol. Evol. 3:8–13 of Pardosa ramulosa (McCook) in alfalfa. 140. White TCR. 1993. The Inadequate Envi- Environ. Entomol. 4:137–41 ronment: Nitrogen and the Abundance of 150. Zimmermann M, Spence JR. 1989. Prey Animals. Berlin: Springer-Verlag.425 pp. use of the fishing spider Dolomedes tri-

by UNIVERSITY OF ILLINOIS - CHICAGO on 08/23/07. For personal use only. 141. Wise DH. 1981. Interspeciﬁc and in- ton (Pisauridae, Araneae): an important

Annu. Rev. Entomol. 2006.51:441-465. Downloaded from arjournals.annualreviews.org traspeciﬁc effects of density manipula- predator of the neuston community. Oe- tions upon females of two orb-weaving cologia 80:187–94 P1: JRX November 2, 2005 13:47 Annual Reviews AR263-FM

Annual Review of Entomology Volume 51, 2006

CONTENTS

SIGNALING AND FUNCTION OF INSULIN-LIKE PEPTIDES IN INSECTS, Qi Wu and Mark R. Brown 1 PROSTAGLANDINS AND OTHER EICOSANOIDS IN INSECTS:BIOLOGICAL SIGNIFICANCE, David Stanley 25 BOTANICAL INSECTICIDES,DETERRENTS, AND REPELLENTS IN MODERN AGRICULTURE AND AN INCREASINGLY REGULATED WORLD, Murray B. Isman 45 INVASION BIOLOGY OF THRIPS, Joseph G. Morse and Mark S. Hoddle 67 INSECT VECTORS OF PHYTOPLASMAS, Phyllis G. Weintraub and LeAnn Beanland 91 INSECT ODOR AND TASTE RECEPTORS, Elissa A. Hallem, Anupama Dahanukar, and John R. Carlson 113 INSECT BIODIVERSITY OF BOREAL PEAT BOGS, Karel Spitzer and Hugh V. Danks 137 PLANT CHEMISTRY AND NATURAL ENEMY FITNESS:EFFECTS ON HERBIVORE AND NATURAL ENEMY INTERACTIONS, Paul J. Ode 163 APPARENT COMPETITION,QUANTITATIVE FOOD WEBS, AND THE STRUCTURE OF PHYTOPHAGOUS INSECT COMMUNITIES, F. J. Frank van Veen, Rebecca J. Morris, and H. Charles J. Godfray 187 STRUCTURE OF THE MUSHROOM BODIES OF THE INSECT BRAIN, Susan E. Fahrbach 209 EVOLUTION OF DEVELOPMENTAL STRATEGIES IN PARASITIC by UNIVERSITY OF ILLINOIS - CHICAGO on 08/23/07. For personal use only.

Annu. Rev. Entomol. 2006.51:441-465. Downloaded from arjournals.annualreviews.org HYMENOPTERA, Francesco Pennacchio and Michael R. Strand 233 DOPA DECARBOXYLASE:AMODEL GENE-ENZYME SYSTEM FOR STUDYING DEVELOPMENT,BEHAVIOR, AND SYSTEMATICS, Ross B. Hodgetts and Sandra L. O’Keefe 259 CONCEPTS AND APPLICATIONS OF TRAP CROPPING IN PEST MANAGEMENT, A.M. Shelton and F.R. Badenes-Perez 285 HOST PLANT SELECTION BY APHIDS:BEHAVIORAL,EVOLUTIONARY, AND APPLIED PERSPECTIVES, Glen Powell, Colin R. Tosh, and Jim Hardie 309

vii P1: JRX November 2, 2005 13:47 Annual Reviews AR263-FM

viii CONTENTS

BIZARRE INTERACTIONS AND ENDGAMES:ENTOMOPATHOGENIC FUNGI AND THEIR ARTHROPOD HOSTS, H.E. Roy, D.C. Steinkraus, J. Eilenberg, A.E. Hajek, and J.K. Pell 331 CURRENT TRENDS IN QUARANTINE ENTOMOLOGY, Peter A. Follett and Lisa G. Neven 359 THE ECOLOGICAL SIGNIFICANCE OF TALLGRASS PRAIRIE ARTHROPODS, Matt R. Whiles and Ralph E. Charlton 387 MATING SYSTEMS OF BLOOD-FEEDING FLIES, Boaz Yuval 413 CANNIBALISM,FOOD LIMITATION,INTRASPECIFIC COMPETITION, AND THE REGULATION OF SPIDER POPULATIONS, David H. Wise 441 BIOGEOGRAPHIC AREAS AND TRANSITION ZONES OF LATIN AMERICA AND THE CARIBBEAN ISLANDS BASED ON PANBIOGEOGRAPHIC AND CLADISTIC ANALYSES OF THE ENTOMOFAUNA, Juan J. Morrone 467 DEVELOPMENTS IN AQUATIC INSECT BIOMONITORING:A COMPARATIVE ANALYSIS OF RECENT APPROACHES, Nuria´ Bonada, Narc´ıs Prat, Vincent H. Resh, and Bernhard Statzner 495 TACHINIDAE:EVOLUTION,BEHAVIOR, AND ECOLOGY, John O. Stireman, III, James E. O’Hara, and D. Monty Wood 525 TICK PHEROMONES AND THEIR USE IN TICK CONTROL, Daniel E. Sonenshine 557 CONFLICT RESOLUTION IN INSECT SOCIETIES, Francis L.W. Ratnieks, Kevin R. Foster, and Tom Wenseleers 581 ASSESSING RISKS OF RELEASING EXOTIC BIOLOGICAL CONTROL AGENTS OF ARTHROPOD PESTS, J.C. van Lenteren, J. Bale, F. Bigler, H.M.T. Hokkanen, and A.J.M. Loomans 609 DEFECATION BEHAVIOR AND ECOLOGY OF INSECTS, Martha R. Weiss 635 PLANT-MEDIATED INTERACTIONS BETWEEN PATHOGENIC MICROORGANISMS AND HERBIVOROUS ARTHROPODS, Michael J. Stout, Jennifer S. Thaler, and Bart P.H.J. Thomma 663 by UNIVERSITY OF ILLINOIS - CHICAGO on 08/23/07. For personal use only.

Annu. Rev. Entomol. 2006.51:441-465. Downloaded from arjournals.annualreviews.org INDEXES Subject Index 691 Cumulative Index of Contributing Authors, Volumes 42–51 717 Cumulative Index of Chapter Titles, Volumes 42–51 722

ERRATA An online log of corrections to Annual Review of Entomology chapters may be found at http://ento.annualreviews.org/errata.shtml

View publication stats