Ant Predation and the Cost of Egg Carrying in the Golden Egg Bug: Experiments in the ﬁeld

OIKOS 89: 254–258. Copenhagen 2000

Ant predation and the cost of egg carrying in the golden egg bug: experiments in the ﬁeld

Arja Kaitala, Xavier Espadaler and Risto Lehtonen

Kaitala, A., Espadaler, X. and Lehtonen, R. 2000. Ant predation and the cost of egg carrying in the golden egg bug: experiments in the ﬁeld. – Oikos 89: 254–258.

Golden egg bug (Phyllomorpha laciniata) females lay eggs on the backs of both female and male conspecifics. Bugs receive eggs voluntarily and involuntarily, and even when males carry eggs, many eggs are not fertilised by the carrier. Carrying the eggs of another individual is unexpected, particularly if egg carrying bears a cost in survival. We examined the predation risk associated with egg carrying experimentally in the field. P. laciniata individuals were enclosed with workers of one of two ant species, Pheidole pallidula or Cataglyphis piliscapus, which co-occur with the bug in the wild. Pheidole pallidula workers preyed on golden egg bugs and their eggs, but Cataglyphis piliscapus workers did not. In P. pallidula enclosures, golden egg bugs carrying larger egg loads were eaten first. These results suggest that golden egg bugs experience high predation pressure and that egg carrying increases the risk of predation. Due to the direct survival costs associated with egg carrying and the lack of relatedness between the eggs and the carrier, we suggest the golden egg bug as the first known intraspecific parasite in which parasitism is not related to active parental care.

A. Kaitala, Dept of Biology, Uni6. of Oulu, Box 3000, FIN-90014 Oulu, Finland (arja.kaitala@oulu.ﬁ).–X. Espadaler, CREAF Uni6ersitat Auto`noma de Barcelona, E-08193 Bellaterra, Spain.–R. Lehtonen, Dept of Statistics, Uni6. of Jy6a¨skyla¨, Box 35, FIN-40351 Jy6a¨skyla¨, Finland.

Predation is an important selective determinant affect- female conspeciﬁcs (Jeannel 1909, Reuter 1909, Kaitala ing prey spatial and temporal distributions, host selec- 1996). Eggs do not survive in nature unless they are tion and social behaviour (for review, see Taylor 1984, carried because of predation by ants (Kaitala 1996; but Crawley 1992). The higher the risk of predation, the see Mineo 1984). Newly laid eggs are white but turn more individuals should act to avoid those risks. Be- golden and shiny after 3–6 d, making the bug highly havioural and morphological adaptations may evolve visible when carrying eggs. In most cases, eggs are not to reduce predation risk. For example, individuals may fertilised by the carrying male (M. Miettinen and A. avoid predators by being cryptic (Cott 1940, Endler Kaitala unpubl.). Individuals receive eggs both volun- 1988). However, adaptations directed against one tarily and involuntarily (Kaitala 1996, 1998, Kaitala predator may be of limited or no value in protection and Miettinen 1997), and in the laboratory males often from another predatory species whose search image accept eggs without mating with the female. All females differs (Endler 1988). Visually cryptic prey may never- carry eggs they have not fertilised. Copulating individu- theless be detected by chemical or other means. als seem especially unable to resist oviposition attempts The golden egg bug, Phyllomorpha laciniata Vill (Kaitala and Miettinen 1997). Egg carrying is theoreti- (Heteroptera, Coreidae), is a highly cryptic species cally problematic if the costs of carrying are higher (Cott 1940) which lays eggs on the backs of male and than the beneﬁts. Why should an individual help

254 OIKOS 89:2 (2000) another if it suffers from helping? To date, no field Predators used in the experiments data exist regarding the predation risk faced by the golden egg bug or the potential costs associated with Pheidole pallidula is an extremely common ant species egg carrying. of the Mediterranean region, inhabiting dry, open habi- The present study was undertaken to examine pre- tats and borders of woods (Bernard 1968).Workers are dation by ants on the golden egg bug, and to evalu- small and dimorphic (minor worker 1.6–2.6 mm, major ate the costs of egg carrying. Ants have an important worker 3.3–4.9 mm, Bernard 1983) compared to (10 effect on the distribution and abundance of terrestrial mm) P. laciniata. However, workers are efficient preda- organisms (Wilson 1990). They detect moving objects tors quickly recruiting tens of workers to any edible while foraging and commonly use chemical cues to object (Szlep-Fessel 1970, Detrain 1990, Retana et al. locate prey (Ho¨lldobler and Wilson 1990). We focus 1992). They exhibit a bimodal diurnal rhythm with on predation of P. laciniata by Pheidole pallidula foraging activity interrupted during the hottest hours (Nylander) (Hymenoptera, Formicidae) and (Delalande 1985). Cataglyphis piliscapus (Forel) (Hymenoptera, Formi- By comparison, Cataglyphis piliscapus is both ecolog- cidae). These were the most common ant species in ically and geographically more restricted (Cagniant the study area and in habitats occupied by P. lacini- 1976, Lenoir et al. 1988). Cataglyphis species are indi- ata. P. pallidula individuals have also been noted to vidual foragers with C. piliscapus scavenging primarily attack eggs and bugs in the field (A. Kaitala, R. Ma- dead arthropods and collecting nectar (Retana et al. cias-Ordonez, M. Miettinen and R. L. Smith un- 1986). Their searching activity is strictly diurnal (Dela- publ.). lande 1985), coinciding with the activity patterns shown for P. laciniata.

Materials and methods Experiments Natural history of the golden egg bug Our study was carried out in an abandoned vineyard located 3 km from Sant Climent Sescebes in Catalonia, P. laciniata is found in the Mediterranean region in north-eastern Spain. The principal ground cover is association with its primary host plants Paronychia Paronychia argentea, in which P. laciniata individuals sp. (Caryophyllaceae; Moulet 1995, Kaitala 1996). In are found in abundance (Kaitala 1996). the field individuals of both sexes carry eggs, but The experiments were carried out in open (32 cm males carry on average twice as many eggs as females diameter) circular enclosures. The inner and outer sur- (Kaitala 1996). Males receive more eggs than females face of the 4 cm high enclosures were covered by Fluon because they are often close to females due to copula- to prevent bugs from escaping and other invertebrates tion attempts and due to mate guarding after copula- from entering. The ants could easily escape via their tion (Kaitala and Miettinen 1997, Kaitala 1998). Egg tunnel system, and other insects could fly in. carrying does not increase male mating success We had three treatments: a) enclosures placed (Kaitala 1998). In populations where eggs are laid on around the entrance of Pheidole pallidula nests (or the backs they do not survive due to invertebrate tunnel entrance), b) enclosures placed on the entrance predation unless carried. In Catalonia, north-eastern of nests of Cataglyphis piliscapus and c) control enclo- Spain, the reproductive period starts in May, and sures placed randomly 2–5 m from the experimental during reproduction the bugs are most active in the enclosures. Control enclosures were placed on sites middle of the day (1100–1400 h). Although females where no ant activity was found before the experiment lay eggs for more than two months, they normally started and where no ant tunnels were found. Control lay only one or two eggs at a time (Kaitala and enclosures were treated with a small amount of insect Miettinen 1997). Eggs are glued firmly on the backs repellent (Merck Insect repellent 3535) to prevent ants of conspecifics, and they are very difficult to remove from digging an entrance. This was necessary because without causing damage to the host individual. Eggs in pre-experiments P. pallidula found bugs in control are large; one egg corresponds to about 4% of male enclosures within one or two days. body weight, and one male may carry up to 28 eggs Each enclosure contained the host plant P. argentea. on its back (Kaitala and Miettinen 1997). An egg-car- We marked bugs individually and counted the number rying individual does not provide any visible care to of eggs they carried. Ten bugs were released in each the eggs, and after hatching, the larvae leave the back enclosure (5 females, 5 males), and surviving bugs were of the carrying bug although egg shells remain on the recaptured after two days. If the wings were un- carrier’s back. Eggs removed from an individual in damaged, the tip of one flight wing was cut from P. the wild, survive and hatch well in the laboratory laciniata individuals with no or only a few eggs to (unpubl.). prevent escape by flying. This was done because in a

OIKOS 89:2 (2000) 255 pre-experiment most bugs without eggs took off soon after being released in the enclosures. Eggs were mostly attached on the wings so that wings could not be cut if an individual carried eggs. However, bugs carrying several eggs are unable to fly (pers. obs.). Wings are small and transparent, and cutting the tip of one wing caused hardly any changes on their appearance. In some bugs the wings were already damaged, and cutting was unnecessary. This damage may be due to egg scraping (Kaitala 1999) or due to the fact that ants or other enemies had taken off the eggs and a piece of wing on which eggs were attached. The controls and P. pallidula experiments were car- Fig. 1. The loss of bugs from enclosures with workers of the ried out once on 27–29 May and once during 2–4 ant Pheidole pallidula (n=2×5 enclosures) and from controls June, 1997 with five replicates each time. The C. pilisca- (n=2×5 enclosures). pus experiment (with five replicates) was carried out only in 27–29 May since the species did not show any ally ants carried a living bug directly to the nest. aggressiveness towards P. laciniata. During the experi- Sometimes ants took eggs without preying on the carry- ment P. pallidula workers excavated a tunnel in one of ing bug by cutting eggs off while the bug was still alive. the C. piliscapus enclosures. Subsequently, P. pallidula P. pallidula workers leave unpalatable parts of inverte- workers were found to be very active and eating one of brates near nest entrances. Several times we found body the bugs. This replicate was excluded from the analysis, parts of insects with paint spots from our marked P. because it was obvious that P. pallidula and not C. laciniata bugs, suggesting that P. pallidula workers were piliscapus workers caused the loss of bugs. responsible for the disappearance of the bugs. We We recaptured and identified surviving individuals, attribute the lower rate of disappearance and activity, and counted the remaining number of eggs they carried. of bugs and ants, in 2–4 June replicates to inclement Individuals without eggs at the start of an experiment weather. were excluded from the analyses examining egg loss. Only one of 40 bugs disappeared from the C. pilisca- Only 27–29 May replicates were used in determining pus enclosures suggesting that C. piliscapus workers whether bugs disappearing carried more eggs than rarely (if ever) preyed on P. laciniata. those recaptured. The number of bugs disappearing was In P. pallidula enclosures established 27–29 May, too low for statistical analysis using replicates from 2–4 bugs that disappeared carried more eggs when released

June. Bug disappearance was analysed using logistic than those that were left (ANOVA, F(1,40) =16.8, PB ANOVA since the response variable was binary, de- 0.001), and the interaction between replicates and dis- scribing whether a bug did or did not disappear. The appearance was not signiﬁcant (F(4,16) =0.98, P=0.43, maximum likelihood method was used in estimating the Fig. 2). During the experiment some individuals re- model parameters (Sokal and Rohlf 1995). ceived eggs laid by females and some lost the eggs. Altogether 59% of the bugs (n=58) in P. pallidula enclosures had lost more eggs than they received, while

Results More bugs disappeared from enclosures with P. pallidula workers than from controls (Fig. 1). The difference between the experimental enclosures and the controls was signiﬁcant (two-way logistic ANOVA, Wald statistic=14.2, d.f.=1, PB0.001). The difference between the two experiment dates was not signiﬁ- cant (Wald statistic=0.02, d.f.=1, P=0.90) nor was the interaction between dates and experimental groups (Wald statistic=1.92, d.f.=1, P=0.17). P. pallidula workers were seen feeding on or carrying P. laciniata individuals to their nest on seven occasions during the experiment. Sometimes ants were cutting the bug in pieces while others were cutting the eggs. In that Fig. 2. The number of eggs carried at the beginning of experiment by recaptured and lost bugs in the Pheidole pallidula case, tens (if not hundreds) of workers encircled the bug enclosures. The number of bugs in each category is denoted in (so that the bug was totally covered by ants). Occasion- parentheses below each column.

256 OIKOS 89:2 (2000) 24% of the bugs (n=82) in controls had lost eggs (the fly. Loss of ability to fly may be an additional factor difference between the groups was significant, x2 =16.8, increasing predation risk for an egg-carrying individual. d.f.=1, PB0.001). As noted before we several times The fact that a few bugs and eggs disappeared from saw ants taking eggs from individuals’ backs, indicating control enclosures suggests predation by other preda- that ants were responsible for egg disappearance. tors, like spiders, birds and lizards. However, these losses were minimal and ants seem to be the main enemies for the bugs. Under natural conditions, P. laciniata eggs appear Discussion not to survive unless carried by bugs. In earlier experiments, most eggs of P. laciniata disappeared quickly if P. pallidula workers prey on adult P. laciniata bugs, placed on the host plant (Kaitala 1996). Similarly, eggs and egg carrying increases predation risk because bugs of the Mediterranean flour moth (Anagasta kuehniella) carrying many eggs were captured by ants more often placed on the soil were quickly found and eaten by P. than those carrying only a few eggs (see Fig. 2). P. pallidula (Du Merle et al. 1978). Thus, it is evident that pallidula individuals have also been observed attacking the unique host ‘‘back’’ selection by P. laciniata is at and killing egg-carrying bugs in natural situations (A. least partly explained as a strategy against predation. Kaitala, R. Macias-Ordonez, M. Miettinen, and R. Mineo (1984) has found a mountain population of P. Smith unpubl.). Individuals carrying many eggs are laciniata in which most eggs are laid on plants and only either more attractive to ants (Kaitala and Axeń 2000) a small proportion on backs of conspecifics. Unfortu- or else are incapable of escaping ants. In the laboratory nately, there is no data on the presence or absence of experiments where the bugs were exposed to workers of invertebrate predators in the mountain habitat. It Formica rufa, egg-bearing bugs were attacked signifi- would be interesting to see if the lack of predatory ants cantly faster and more often than non-egg-bearing could explain why eggs are laid on plants. bugs, suggesting that eggs attract ants. These experi- To date, we have found no evidence suggesting that ments also show that mating with an egg-bearing bug egg carrying is beneficial for the carrying individual. increases risk of predation because mating pairs were Egg carrying does not increase male mating success attacked more often than two single bugs together (Kaitala 1998). Preliminary DNA analysis confirms (Kaitala and Axeń 2000). that most eggs carried in the field are not fertilised by Egg-carrying P. laciniata individuals are found in the carrier (M. Miettinen and A. Kaitala unpubl.). habitats where predation risk by ants is high. Ants are Thus paternal care is an unlikely explanation (at least efficient foragers, and in similar habitats as those occu- not the whole explanation) for egg carrying. An indirect pied by P. laciniata they are able to detect arthropod evidence of costs of egg carrying is that individuals corpses in less than five minutes (Retana et al. 1991). P. often scrape off eggs (Kaitala 1999). pallidula workers are quickly able to recruit a vast number of workers to a food source (Szlep-Fessel 1970, Individuals are not expected to help each other if Detrain 1990, Retana et al. 1992). Despite their small costs of helping are larger than benefits or if the size, they have powerful mandibles capable of cutting receiver is not a relative (Hamilton 1963). The increased prey into smaller pieces (Detrain 1990), and they cap- predation risk caused by egg carrying, and the fact that ture large prey, such as adult bugs, in a co-operative most eggs are unrelated to the carrier calls for an effort. In our experiments, P. pallidula workers occa- evolutionary explanation; why individuals sometimes sionally took only eggs without preying on the carrying accept eggs without resistance? One plausible explana- bug by cutting eggs off, sometimes they killed the bug tion may be that accepting eggs is less costly than or carried a living bug directly to the nest. High preda- rejecting them. tion pressure by P. pallidula was also indicated in the Intraspecific parasitism occurs when individuals pre-experiments. In those experiments P. pallidula avoid the costs of parental care by exploiting the time found the controls soon, and therefore insect repellent and energy of unrelated conspecifics (Field 1992). In- was used in controls to avoid P. pallidula digging an traspecific brood parasitism is well described in several entrance to the enclosures. animal groups (for reviews, e.g. Field 1992, Brockmann P. pallidula is active early in the morning and in the 1993), and has evolved in species which actively care for evening (Cammaerts et al. 1993) when P. laciniata is their young. According to current knowledge individu- inactive (pers. obs). Although P. laciniata individuals als of the golden egg bug suffer fitness costs from egg hide in the bushes when inactive (pers. obs.), it is carrying, and egg carrying can be regarded as intraspe- unlikely that they can totally avoid P. pallidula cific parasitism. However, the kind of intraspecific para- predation. sitism described here, in which parasitism occurs Individuals carrying no eggs or only a few eggs do independently of parental care, and where eggs are often fly away in the laboratory, and did so also in our attached to bugs and carried passively by conspecifics, pre-experiment but those carrying many eggs cannot has not been previously described.

OIKOS 89:2 (2000) 257 Acknowledgements – Thanks are due to Denson Kelly Mclain, Ho¨lldobler, B and Wilson E. O. 1990. The ants. – Belknap Bohdan Sklepkovych and Christer Wiklund for comments on Press of Harvard Univ. Press, Cambridge, MA. the manuscript. This study was supported by the Swedish Jeannel, R. 1909. Sur les moeurs et les me´tamorphoses de Natural Science Research Council to A.K. and by PB96-1164 Phyllomorpha laciniata Vill (Hem. Coreidae). – Bull. Soc. from Direccioń General de Ensenãnza Superior to X.E. We Entomol. France 15: 282–286. also acknowledge Ricardo de la Barrera for hospitality and for Kaitala, A. 1996. Oviposition on the back of conspecifics: an helping us to find suitable field sites. unusual reproductive tactic in a coreid bug. – Oikos 77: 381–389. Kaitala, A. 1998. Is egg carrying attractive: mate choice in the golden egg bug (Heteroptera, Coreidae). – Proc. R. Soc. Lond. B 265: 779–783. References Kaitala, A. 1999. Counterstrategy to egg dumping in a coreid bug: recipient individuals discard eggs from their backs. – Bernard, F. 1968. Faune de l’Europe et du Bassin Me´diterra- J. Insect Behav. 12: 225–232. neén 3. Les fourmis (Hymenoptera Formicidae) d’Europe Kaitala, A. and Miettinen, M. 1997. Female egg dumping and occidentale et septentrionale. – Masson et Cie, Paris. the effect of sex ratio on male egg carrying in a coreid bug. Bernard, F. 1983. Les fourmis et leur milieu en France me´d- – Behav. Ecol. 8: 429–432. iterraneénne. – Enciclope´die entomologique 45. Lecheva- Kaitala, A. and Axeń, A. H. 2000. Predation risk: egg load lier, Paris. and mating status of the golden egg bug matter. – Ecology Brockmann, J. H. 1993. Parasiting conspecifics – comparisons 81 (in press). between Hymenoptera and birds. – Trends Ecol. Evol. 8: Lenoir, A., Querard, L., Pondicq, N. and Berton, F. 1988. 2–4. Reproduction and dispersal in the ant Cataglyphis cursor Cagniant, H. 1976. Distribution, ećologie et nid de la fourmi (Hymenoptera: Formicidae). – Psyche 95: 21–44. Cataglyphis cursor Fonscolombe. Hymeńopte`res Formi- Mineo, G. 1984. Notizie biologiche su Phyllomorpha laciniata cidae. – Vie Milieu 26: 265–276. (Vill.) (Rhynchota, Het., Coreidae). – Phytophaga 117– Cammaerts, M. C., Cammaerts, R. and Dejean, A. 1993. 132. Propriete´s physiques et e´thologiques des pistes de six espe`- Moulet, P. 1995. Hemipte`res Coreoidea Euro-Me´diterraneéns. ces de Myrmicines. – Actes Colloq. Insectes Soc. 8: 95– 102. Faune de France 81 – Federation Francaise des Socie´tes Cott, H. B. 1940. Adaptive coloration in animals. – Methuen, de Sciences Naturelles, Paris. London. Retana, J., Cerda´, X., Alsina, A. and Bosch, J. 1986. Impor- Crawley, M. J. (ed.) 1992. Natural enemies: the population tancia del alimento lı´quido en el re´gimen tro´fico de la biology of predators, parasites and diseases. – Blackwell hormiga Cataglyphis cursor (Formicidae). – Sessio´ Ento- Scientific, Oxford. mol. I.C.H.N.-S.C.L. 4: 139–146. Delalande, C. 1985. Rythmes d’activite´ de quelques espe`ces de Retana, J., Cerda´, X. and Espadaler, X. 1991. Arthropod fourmis en re´gion me´diterraneénne. – Actes Colloq. In- corpses in a temperate grassland: a limited supply? – sectes Soc. 2: 303–318. Holarct. Ecol. 14: 63–67. Detrain, C 1990. Field study on foraging by the polymorphic Retana, J., Cerda´, X. and Espadaler, X. 1992. Coexistence of ant species, Pheidole pallidula. – Insectes Soc. 37: 315–332. two sympatric ant species, Pheidole pallidula and Tetramor- Du Merle, P., Jourdheuil, P., Marro, J. P. and Mazet, R. 1978. ium semilae6e (Hymenoptera: Formicidae). – Entomol. E´ volution saisonnie`re de la myrmećofaune et de son activ- Gener. 17: 29–40. ite´ pre´datrice dans un milieu forestier: les interactions Reuter, O. M. 1909. Quelques mots sur les Phyllomorphes clairie`re-lisie`re-fore`t. – Ann. Soc. Entomol. France 14: (Hem Coreidae). – Bull. Soc. Entomol. France 15: 264– 141–157. 268. Endler, J. A. 1988. Frequency-dependent predation, crypsis Sokal, R. R. and Rohlf, F. J. 1995. Biometry. – Freeman, and aposematic coloration. – Philos. Trans. R. Soc. Lond New York. B 319: 505–523. Szlep-Fessel, R. 1970. The regulatory mechanisms in mass Field, J. 1992. Intraspecific parasitism as an alternative repro- foraging and the recruitment of soldiers in Pheidole.– ductive tactic in the nest-building wasps and bees. – Biol. Insectes Soc. 17: 233–244. Rev. 67: 79–126. Taylor, R. J. 1984. Predation. – Chapman & Hall, London. Hamilton, W. D. 1963. The evolution of altruistic behavior. – Wilson, E. O. 1990. Success and dominance in ecosystems: the Am. Nat. 197: 354–356. case of the social insects. – Ecology Institute, Oldendorf.

258 OIKOS 89:2 (2000)