A Comparison of Behavioral Interactions in Solitary and Presocial Zetoborinae Cockroaches (Blattaria,Blaberidae)

The social behavior of five species of Zetoborinae cockroaches is compared with respect to inter-individual interactions of nymphs in the laboratory. These species belong to the same Neotropical subfamily and were characterized as gregarious (Lanxoblatta emarginata, Parasphaeria boleiriana, Phortioeca nimbata, Schultesia lampyridiformis) and solitary (Thanatophyllum akinetum) by previous field studies. Our results show that gregarious species accept closer contacts than does the solitary one. The solitary species did not display especially short, infrequent or less diverse behavioral sequences when forced to remain aggregated, but its interactions are characterized by fewer acts promoting contact and more dominance-like acts. The solitary species symmetrically interacts with conspecifics and does not show specific dispersal-promoting behaviors. This suggests that the solitary behavior observed in the field for species of Zetoborinae mainly results from a passive spacing tendency and a lack of attraction for conspecifics. One of the gregarious species, P. boleiriana, was previously described as subsocial with nymphs remaining with the female in a wood chamber after brood birth. This species does not show a peculiar behavioral repertoire but its interactions are characterized by more dominance-like behaviors than are those of the non-subsocial gregarious species. Electronic supplementary material  The online version of this article (doi:) contains supplementary material, which is available to authorized users.     

Behaviours promoting grouping or dispersal of mother and neonates in ovoviviparous cockroaches

Summary. The nature of early relationships between mother and nymphs and among siblings was compared in four cockroach species belonging to the same ovoviviparous family (Zetoborinae) in order to characterise the behavioural interactions favouring dispersal or maintenance of the group of neonates. Behavioural interactions between mothers and their new-born nymphs and between two sibling neonates were video recorded and analysed with flow charts on factorial maps. In the solitary species Thanatophyllum akinetum, nymphs dispersed a few hours after birth without aggressiveness between siblings or between mother and offspring. In contrast females of Schultesia lampyridiformis displayed a behaviour never previously observed in cockroaches: aggression towards their own neonate nymphs, which could contribute to their dispersal. In Phortioeca nimbata and Lanxoblatta emarginata, nymphs stayed with their mother for 10 days, but social interactions differed between the two sp ecies: P. nimbata mothers actively searched for their nymphs, this behaviour being favoured by an active search for mutual contact by the nymphs themselves, while L. emarginata nymphs sought more actively the proximity of their mother but less contact between themselves. Thus, different species of Zetoborinae presented two types of dispersal of the young and two types of maintenance of the birth group, both achieved by specific behavioural interactions. The relevance of behavioural interactions for the characterisation of early gregarism and parental care is discussed.     

Aggregation behaviour and interspecific relationships in Dermaptera

The rate of gregariousness and intra- and inter-specific attractivity were evaluated under laboratory conditions in five species of Dermaptera. The three species of the genus Forficula (Forficula auricularia L., F. decipiens Gené and F. pubescens Gené) showed distinctly gregarious behaviour in old larvae and adults in the pre-reproductive phase. This gregariousness was seen at the intraspecific level and also among individuals of different species. The other two species, Labidura riparia Pallas and Euborellia moesta Gené, were randomly distributed over the available shelters, with no distinct tendency for aggregation or isolation. In individual tests, insects of all five species were attracted by shelters previously visited by conspecifics in preference to control shelters. For the two non-gregarious species, this phenomenon might function as a burrow marker. The same tests showed an interattractivity for four of the species studied. Only L. riparia did not respond to the trails of the four other species, although its trails were attractive to the other species. The pheromonal nature of the substances involved in the interspecific relationships and the involvement of behaviour in the biology and speciation of Dermaptera are discussed.     

Interspecific competition between the braconid endoparasitoids Glyptapanteles porthetriae and Glyptapanteles liparidis in Lymantria dispar larvae

The effect of interspecific competition between the solitary endoparasitoid Glyptapanteles porthetriae Muesebeck (Hymenoptera: Braconidae) and the gregarious Glyptapanteles liparidis Bouché (Hymenoptera: Braconidae), was investigated in larvae of Lymantria dispar L. (Lepidoptera: Lymantriidae). Host larvae were parasitized by both wasp species simultaneously in premolt to the 2^nd or the 3^rd host instar or in an additional approach with a 4-day delay in parasitization by the second wasp species. Host acceptance experiments revealed that both wasp species do not discriminate between unparasitized host larvae and larvae parasitized previously by the same or the other species. In more than 90% female wasps parasitized the larva they encountered first. During the period of endoparasitic development, larvae of the competing parasitoid species never attacked the egg stage of the other species. When host larvae were parasitized simultaneously by both wasp species, the rate of successful development of both species depended on the age of the host larva at the time of its parasitization; G. liparidis emerged successfully from 44% of host larvae parasitized during the premolt to 2^nd instar, G. porthetriae from 28%, and in 20% of the hosts both parasitoid species were able to develop in one gypsy moth larva. However, when host larvae were parasitized simultaneously during premolt to the 3^rd instar, G. liparidis was successful in 90% of the hosts, compared to 8% from which only G. porthetriae emerged. In the experiments with delayed oviposition, generally the species that oviposited first succeeded in completing its larval development. Larvae of the species ovipositing with four days delay were frequently attacked and killed by larvae of the first parasitizing species or suffered reduced growth. As the secondary parasitoid species, G. porthetriae-larvae were never able to complete their development, whereas G. liparidis developed successfully in at least 12,5% of the multiparasitized host larvae. Thus, multiparasitism of gypsy moth larvae by both Glyptapanteles species corresponds to the contest type; however, G. porthetriae is only able to develop successfully as the primary parasitoid of young host larvae.     

Influence of host behavior and host size on the success of oviposition ofCotesia urabae andDolichogenidea eucalypti (Hymenoptera: Braconidae)

Behavioral interactions among Cotesia urabaeAustin and Allen, Dolichogenidea eucalyptiAustin and Allen (Hymenoptera: Braconidae), and their host Uraba lugensWalker, the gum leaf skeletonizer (Lepidoptera: Noctuidae), were observed at three host sizes over a 20-min period. These sizes were first instar (small, gregarious), fourth-fifth instar (mid, gregarious), sixth-seventh instar (large, solitary) larvae. Unlike C. urabae, D. eucalyptiused its legs to hold small larvae before ovipositor insertion. D. eucalyptialso visited patches of small larvae more frequently, proceeded less often through patches of mid larvae, and made significantly fewer ovipositions in mid and large larvae. Small larvae responded to both parasitoids by dispersing outward, while mid larvae responded to parasitoids by moving inward to form a denser group. Larvae reared or thrashed after each parasitoid visit, especially mid larvae, and some continued to do so for up to 2h after parasitoid departure. Mid and large larvae occasionally injured parasitoids by biting their appendages. By rearing or thrashing immediately prior to an encounter with a parasitoid, mid and large larvae decreased the likelihood of being parasitized by up to 50%.     

Interpreting dragonfly diversity to aid in conservation assessment: lessons from the Odonata assemblage at Middle Creek, north-eastern Victoria, Australia

In order to evaluate single-occasion sampling in compiling inventories for Odonata, larvae were sampled on 20 occasions from 1987–1990 at a site on Middle Creek, north-eastern Victoria, and adults sought also on each visit to more fully evaluate the diversity of the asemblage, and limitations of depending on a single life stage for this purpose. A total of 18 species (7 Zygoptera, 11 Anisoptera) included 15 species collected as larvae and 16 as adults. Few species were common as larvae, and about half the 2806 specimens identified were Austrogomphus cornutus Watson. Orthetrum caledonicum (Brauer) and Ischnura heterosticta (Burmeister) were also abundant, and these three species were also the most common as adults. The number of species obtained ranged from 2–9 on different occasions, and represent different 'habitat groups' within the local fauna. The Middle Creek odonate fauna appears to be much richer than that of a nearby site on the Kiewa River (12 species), and reasons for this are discussed. Seasonal variation in species representation and relative abundances are noted. Any single sample occasion provided insufficient knowledge of the total assemblage to interpret odonate diversity reliably.     

Susceptibility of Delia radicum to steinernematid nematodes

Isolates of different Steinernema species (S. affine, S. bicornutum, S. feltiae and Steinernema C1) were used in mortality assays with third instar larvae of Delia radicum (L.) (Diptera: Anthomyiidae). The nematode isolates had been obtained by baiting soil regularly grown with cabbage. One isolate (S. feltiae) was the result of a natural infection of a D. radicum puparium. The highest mortality (77%) was obtained with an isolate of S. feltiae (DK1). The isolate DK1 was also used in tests with all larval stages of D. radicum. Mortality around 60% was observed for second and third instar larvae, while first instar larvae showed very low or no susceptibility. Maximum mortality of second and third instar larvae was reached applying only 25 nematodes per larva. Observations of larvae that pupated revealed that some of these puparia contained nematodes. Experiments with hatching puparia showed that a high proportion was infected by nematodes if the flies were prevented from leaving nematode-containing soil. In addition to mortality, the ability of the nematodes to successfully reproduce in the insects was studied. It was found that the species S. feltiae and S. bicornutum reproduced in D. radicum larvae and adults with S. feltiae being the most successful.     

Variation in hydraulic architecture and gas-exchange in two desert sub-shrubs,<Emphasis Type="Italic"> Hymenoclea salsola</Emphasis> (T. &; G.) and<Emphasis Type="Italic"> Ambrosia dumosa</Emphasis> (Payne)

Adjustment of hydraulic architecture in response to environmental conditions was studied in two warm-desert sub-shrubs, Hymenoclea salsola and Ambrosia dumosa, both at the level of genetic adaptation along a climatic gradient and plastic response to immediate growth conditions. Individuals of both species originating from southern populations developed higher leaf-specific hydraulic conductance in the common greenhouse than individuals from northern populations. Hydraulic conductance was higher in plants grown at high temperature, but did not vary as a function of growth relative humidity. Hydraulic conductance was not correlated within species with individual variation in vessel diameter, cavitation vulnerability, or root:shoot ratio, but was strongly, negatively correlated with the fraction of total plant biomass allocated to leaves. For both species, stomatal conductance (g _s) at high leaf-to-air vapor pressure difference (ν) was tightly correlated with variability in hydraulic conductance, as was the sensitivity of stomatal closure to increasing ν. Experimentally increasing shoot water potential by soil pressurization, under conditions where high ν had already caused stomatal closure, led to substantial stomatal reopening in both species, but recovery was significantly higher in H. salsola. Hydraulic conductance was higher in H. salsola than A. dumosa. H.salsola also differed from A. dumosa by being a representative of a highly specialised group of desert shrubs which use the twigs as a major photosynthetic organ. The southern population of H. salsola produced far fewer leaves and relied much more heavily on twig photosynthesis than the northern population. At the whole-plant level, increased reliance on twig photosynthesis was associated with higher leaf-specific hydraulic conductance, but equivalent whole-plant photosynthesis on either a dry weight (μmol CO₂ g^–1) or nitrogen basis (μmol CO₂ g^–1)). This suggests that twig photosynthesis might be one way of increasing hydraulic conductance per unit photosynthetic canopy by increasing allocation to an organ which simultaneously performs photosynthetic, support, and transport functions. Received: 13 December 1999 / Accepted: 31 March 2000     

Competitive interactions between parasitoid larvae and the evolution of gregarious development

We report experiments using two closely related species of alysiine braconids directed at understanding how gregarious development evolved in one subfamily of parasitoid wasps. Theoretical models predict that once siblicide between parasitoid wasps has evolved, it can only be lost under stringent conditions, making the transition from solitary to gregarious development exiguous. Phylogenetic studies indicate, however, that gregariousness has independently arisen on numerous occasions. New theoretical models have demonstrated that if gregarious development involves reductions in larval mobility, rather than a lack of fighting ability (as in the older models), the evolution of gregariousness is much more likely. We tested the predictions of the older tolerance models (gregariousness based on non-fighting larval phenotypes) and the reduced mobility models (gregariousness based on non-searching larval phenotypes) by observing larval movement and the outcome of interspecific competition between Aphaereta genevensis (solitary) and A. pallipes (gregarious) under multiparasitism. Differences in larval mobility matched the prediction of the reduced mobility model of gregarious development, with the solitary A. genevensis having larvae that are much more mobile. The proportion of hosts producing the solitary species significantly declined after subsequent exposure to females of the gregarious species. This contradicts the prediction of the older models (fighting vs non-fighting phenotypes), under which any competitive interactions between solitary and gregarious larvae will result in a highly asymmetrical outcome, as the solitary species should be competitively superior. The observed outcome of interspecific competition offers evidence, with respect to this subfamily, in favour of the new models (searching vs non-searching phenotypes).     

Biology, rearing and field release on Guam of Euplectrus maternus, a parasitoid of the fruit-piercing moth, Eudocima fullonia

The eulophid parasitoid, Euplectrus maternus, is a gregarious ectoparasitoid of the larvae of thefruit-piercing moth, Eudocima (Othreis) fullonia. This parasitoid is indigenous to India and was released into Guam in 1998–99 to aid in the biological control of E. fullonia, an important pest of ripeguava, banana, mango, pomegranate, litchi,papaya, tomato, orange and other fruit crops.In the laboratory, biology and rearing of E. maternus on the larvae of E. fulloniawere studied. In no-choice tests, ovipositionby E. maternus was significantly greateron first and second instar larvae than on thirdinstar larvae. However, in free-choice tests,oviposition by E. maternus wassignificantly greater on second instar larvae,followed by first and third instar larvae. Thisparasitoid did not lay eggs on later instars,either in no-choice or free-choice tests.Females laid a significantly greater number ofeggs (>80%) on the dorsal surface of thelarvae than on either lateral or ventralsurfaces (1 to 8%). In general, more eggs werelaid on abdominal segments 1–3 on larvalinstars one to three. The survival of theimmature stages (from egg to adult) of E. maternus on second instar E. fullonialarvae was 61%, which was significantlygreater than the survival rates on first andthird instar larvae (32% and 26%), indicatingthat second instar host larvae are ideal formass rearing of the parasitoid. Mated femaleE. maternus continuously laid eggs on thesecond instar larvae of E. fullonia forup to 30 days, but the greatest number of eggswere laid during the first week after exposure.When E. maternus laid more than two eggson host larva, more female progeny wereproduced, indicating female-biasedreproduction. Adult parasitoids lived longerwhen fed with a honey: water (50% w/v)solution than with pure honey. Similarly, thefecundity of females increased significantlywhen fed with the honey-water solution whencompared to feeding with pure honey. To date,E. maternus has not established in Guam.     

Facilitated predation through interaction between life stages in the stinkbug predatorperillus bioculatus (Hemiptera: Pentatomidae)

Competitive interactions in arthropod predators are well-known, but positive interactions have received less attention. The two-spotted stinkbugPerillus bioculatus often feeds gregariously on leaf beetle larvae and caterpillar prey. Consequences of prey sharing amongP. bioculatus conspecifics of dissimilar size (instar) was studied using Colorado potato beetle (CPB) prey. Rearing second-instar (N2) nymphs ofP. bioculatus with an N5 conspecific facilitated early feeding on L4 CPB larvae (a difficult prey to handle by N2 nymphs but not by N5’s), thus increasing survival and accelerating development. One in every 20 cases ofP. bioculatus foraging in the field was accounted for by pairs or small groups of mostly feeding individuals. CPB egg masses and L4’s represented a disproportionate number of cases of aggregated feeding byP. bioculatus, compared to feeding singly. Small CPB larvae decreased in the diet of aggregated stinkbugs compared to L4 larvae and egg masses, suggesting that sharing these prey may be favorable or unavoidable. In a field test measuring residence/survival of N2’s limited to L4 prey, the N2’s rate of residence/survival increased significantly when large nymphs acting as food providers were also present. The function of communal feeding inP. bioculatus is discussed, as well as the potential for greater impact on prey density that may be expected from tolerance to opportunistic feeding by conspecifics in slightly gregarious predators.     

Pathogenicity of hyphomycetous fungi against <Emphasis Type="Italic">Cyclocephala signaticollis</Emphasis>

Susceptibility of the white grub Cyclocephala signaticollis Burmeister (Coleoptera: Scarabaeidae: Dynastinae) larvae to seven isolates of Beauveria bassiana (Balsamo) Vuillemin, five of Metarhizium anisopliae (Metschnikoff) Sorokin and two of Paecilomyces lilacinus (Thom) Samson (Deuteromycotina: Hyphomycetes) was investigated. Among 14 fungal isolates screened the most virulent was a B. bassiana isolate (Bb 53) that caused 70% mortality of third instar larvae in 40 days after inoculation at 1 × 10⁸ conida/ml. Strains of M. anisopliae and P. lilacinus showed low efficacy or no virulence to the target host.     

Predation by three hemipterans:Tropiconabis nigrolineatus,Oechalia schellenbergii andCermatulus nasalis,onHeliothis punctiger larvae in two types of searching arenas

Three species of hemipteran predators preyed differently upon 1^st instarHeliothis punctiger Wallengren larvae.Cermatulus nasalis consumed more larvae thanOechalia schellenbergii which consumed more larvae thanTropiconabis nigrolineatus. All the species consumed significantly less 1^st instar larvae on plants than what they consumed in Petri-dishes. Fifth instar predators showed significant differences in terms of prey consumption due to sex independent of searching conditions. Only 4^th and 5^th instars ofT. nigrolineatus attacked and captured 2^nd instars ofH. punctiger larvae. The other 2 species however readily attacked and consumed 2^nd instarH. punctiger larvae. Their prey consumption was similar in Petri-dishes and on plants. Only 5^th instars ofT. nigrolineatus could subdue and capture 3^rd instarH. punctiger larvae. Second instar pentatomids captured just one 3^rd instar larva but older instars killed and ate more. Fourth instarH. punctiger larvae were immune to attacks by allT. nigrolineatus and younger pentatomids due to their defense ploys but 5^th instar pentatomids could subdue and capture them. None of the predators captured 5^th instarH. punctiger larvae except few 5^th instar females ofC. naslis andO. schellenbergii.      

Evaluation ofHadena perplexa [Lepidoptera: Phalaenidae] as a biocontrol agent of bladder campionSilene vulgaris [Caryophyllaceae] in Canada: Rearing and host specificity

Rearing techniques and results of preliminary host range tests are reported forHadena perplexa (Denis & Schiffermuller) (Lep.: Phalaenidae) a candidate biocontrol agent against the weed bladder campion,Silene vulgaris (Moench) Garcke, in Manitoba, Canada. In the laboratory, it was necessary to pipette a 15% honey solution in water into the flowers as food for the adult moths. When reared singly to avoid cannibalism, 56% of the 1^st instar larvae developed to pupae. Larvae fed on a natural diet for 10 days can then be reared on either one of 2 artificial diets. Choice oviposition tests and no-choice larval feeding tests were conducted with plant species closely related toS. vulgaris in the generaSilene, Dianthus, Gypsophila, Lychnis, Saponaria. Species in 4 of 5 of these genera were accepted for oviposition, and species in all 5 genera supported the development of 1^st instar larvae to the pupal stage.H. perplexa should not be introduced into Canada.      

Does superparasitism improve host suitability for parasitoid development? A case study in the Microplitis rufiventris-Spodoptera littoralis system

Superparasitism refers to the oviposition behavior of parasitoid females who lay their eggs in an already parasitized host. Recent studies have shown that allocation of additional eggs to an already parasitized host may be beneficial under certain conditions. In the present work, mortality of Microplitis rufiventris wasps was significantly influenced by both host instar of Spodoptera littoralis larvae at parasitism and level of parasitism. In single parasitization, all host instars (first through sixth) were not equally suitable. Percentage of emergence success of wasp larvae was very high in parasitized first through third (highly suitable hosts), fell to 60% in the fourth instar (moderate suitable) and sharply decreased in the penultimate (5th) instars (marginally suitable). Singly parasitized sixth (last) instar hosts produced no wasp larvae (entirely unsuitable), pupated and eclosed to apparently normal adult moths. The scenario was different under superparasitism, whereas supernumerary individuals in the highly suitable hosts were almost always killed as first instars, superparasitization in unsuitable hosts (4th through 6th) had significant increase in number of emergence success of wasp larvae. Also, significantly greater number of parasitoid larvae successfully developed in unsuitable hosts containing three wasp eggs than counterparts containing two wasp eggs. Moreover, the development of surplus wasp larvae was siblicidal in earlier instars and nonsiblicidal gregarious one in the penultimate and last “sixth” instars. It is suggested that the optimal way for M. rufiventris to deal with high quality hosts (early instars) is to lay a single egg, while the optimal way to deal with low quality hosts (late instars) might be to superparasitize these hosts.     

Interference and competitive behaviour of the aphid predators,Chrysoperla carnea andCoccinella septempunctata in the laboratory

Interference and competition betweenChrysoperla carnea (Stephens) andCoccinella septempunctata L. was investigated in the absence and presence of aphid prey. When larvae of similar vigour encountered each other,C. carnea larvae were superior toC. septempunctata. Otherwise the larger of 2 individuals always killed the smaller. In the absence of preyC. carnea adults were attacked by their own 2nd and 3rd instar larvae as well as by 3rd instar larvae ofC. septempunctata. In all encounters the adults ofC. septempunctata were superior. Since the eggs ofC. carnea are protected to some extent by being on egg stalks, they were less susceptible to cannibalism and predation thanC. septempunctata eggs which are deposited in batches directly on plant leaves. In the presence of prey cannibalism and predation were reduced, especially between larvae. Only eggs and 1st instar larvae were endangered. In the present experimentsC. carnea showed a slight superiority overc. septempunctata. Paper presented at the 17th International Congress of Entomology held in August 20–26 1984 in Hamburg, F.R. Germany.     

Understanding gregariousness in a larval Lepidopteran: the roles of host plant, predation, and microclimate

Abstract.  1. Many moth and butterfly larvae are gregarious early in development, but become solitary in late instars. This ontogenetic variation in behaviour is probably the result of temporal changes in the costs and benefits associated with gregariousness. This study provides observational and experimental evidence that, in one particular moth species, a series of different ecological factors influence larval behaviour at different times during development.
2. Field observations show that young caterpillars of the limocodid Doratifera casta form large aggregations while foraging, but that mature larvae are largely solitary.
3. A field experiment revealed that individual first to third instar larvae in larger groups develop more rapidly, but that group size had no detectable influence on survival. The developmental advantage associated with gregariousness is affected by host plant species, but not by predator exclusion, suggesting that group living in these cryptic early instar larvae promotes feeding facilitation, but does not provide individuals with protection from natural enemies.
4. Laboratory experiments revealed that aposematic fourth instar caterpillars in large groups were less likely to be attacked by a generalist insect predator than those in small groups.
5. Field observations provided no evidence that group living affects body temperature, suggesting that microclimatic factors do not favour gregariousness in this species.
6. It is concluded that gregariousness in D. casta confers at least two different advantages on larvae at different stages early in development, but that these advantages disappear, or are outweighed by costs associated with intraspecific competition, in final instars.     

The life-history,distribution, and production of Chironomus riparius and Glyptotendipes paripes in a prairie pond

C. riparius and G. paripes exhibited univoltine life-cycles in Stephenson Pond; pupation, emergence and oviposition occurred mainly during May, and both species overwintered as mature fourth instar larvae. The marjority of larval growth for both species took place during the fourth instar stage (August–October), and growth and production were very low during late May to mid July when only young instars were present. Low production occurred during an interval when sestonic chlorophyll a concentration was very low, and the high production period corresponded to the Aphanizomenon bloom (August) and the autumn diatom pulse. None of the growth and production parameters investigated were correlated with temperature at the mud-water interface. Tube structure and behavior of the larvae indicate that G. paripes larvae are filter-feeders, whereas, C. riparius larvae are deposit-feeders.     

Interspecific competition between the polyembryonic wasp Copidosoma floridanum and the gregarious endoparasitoid Glyptapanteles pallipes

We examined interspecific competition between the egg-larval polyembryonic parasitoid Copidosoma floridanum (Hymenoptera: Encyrtidae) and the gregarious larval endoparasitoid Glyptapanteles pallipes (Hymenoptera: Braconidae). Approximately 72% of multiparasitized Acanthoplusia agnata (Lepidoptera: Noctuidae) produced C. floridanum adults, 14% produced G. pallipes adults, and 14% died without producing any parasitoids. Development was delayed and the rate of weight gain was reduced in multiparasitized hosts that produced C. floridanum compared with singly-parasitized hosts. In contrast, only weight was reduced in multiparasitized hosts that produced G. pallipes adults. Compared with single parasitism, the brood size of wasps emerging from multiparasitized hosts was reduced in both species. The percentage of hosts containing precocious larvae of C. floridanum, which are considered to be soldiers, did not increase in response to parasitization by G. pallipes. However, developmental cessation and death of G. pallipes eggs and larvae may be closely related to the number of coexisting precocious larvae.     

Specificity of an ant-lycaenid interaction

Summary Many lycaenid butterflies are believed to be mutualists of ants — the butterfly larvae secrete sugars and amino acids as rewards for the ants, and the ants protect the larvae from predation or parasitism. We examined the specificity of the relationship between the lycaenid Plebejus argus and ants in the genus Lasius. Eggs were not attractive to Lasius ants until the emerging larvae had broken through the chorion. First instar larvae were palpated and picked up by Lasius workers and taken to the nest. First instars were mostly ignored by Myrmica sabuleti ants and they were rarely detected by Formica fusca. Older larvae were more attractive to Lasius than to the other ant genera. Pupae were very attractive to Lasius, moderately so to Myrmica, and were ignored by Formica fusca. Teneral adults were palpated by Lasius, but were attacked by Myrmica and Formica workers. We conclude that P. argus is a specialist associate of Lasius ants. Two populations of Plebejus argus were compared: one is naturally associated with Lasius niger, and the other with Lasius alienus. In reciprocal trials, larvae were slightly more attractive to their natural host ant species. Since test larvae were reared on a single host plant species in captivity, this differentiation probably has a genetic basis.