Taxonomy, attack strategies and host relations in flowerhead Tephritidae: a review

Abstract. 1. Host relations of Palaearctic flowerhead-inhabiting tephritids (Diptera: Tephritidae) are reviewed in terms of the interaction between type of attack strategy and flowerhead size.
2. Three types of attack strategy are recognized, each associated with a particular phase of Compositae flowerhead development. These strategies are characteristic of a tribe or subfamily, viz the Tephritini (Tephritinae), Terelliini (Tephritinae) and Myopitinae.
3. Tephritini species show the most generalized feeding method, and this enables them to attack both large and very small flowerheads. In contrast, the more specialized requirements of members of the Terelliini and Myopitinae apparently restricts these groups to plant taxa with large flowerheads. As a result, over the whole range of Compositae flowerhead sizes, there exists a positive relationship between flowerhead size and the diversity of associated tephritids.
4. The occurrence of stem-feeding tephritids, and differences between a coevolutionary and functional (colonization) explanation for tephritid host-relations are discussed.     

Resource limitation of tephritid flies on lesser burdock,Arctium minus (Hill) Bernh. (Compositae)

Summary The intensity of resource exploitation by phytophagous insects is usually considered to reflect population size. For populations of two flowerhead-attacking tephritid flies, however, the resources utilised were not related to the numbers of searching adults. Tephritis bardanae Schrank attacked 11–13% of the total flowerheads each year, and Cerajocera tussilaginis (Fab) 17–65%, despite much wider and uncorrelated variation in adult numbers. Analysis of field data showed that the proportion of flowerheads used was not limited by poor flowerhead quality, but was restricted by (1) the synchronisation of adult activity with the appearance of flowerheads at the correct age for oviposition, and (2) by factors which influenced the ability of female flies to locate available heads. These restrictions were more severe in T. bardanae and explained its relatively low rate of infestation. Both tephritids tended to avoid flowerheads in open areas. The processes governing resource exploitation in each tephritid operated independently of the other, and a partial separation of the two species between flowerhead types and habitats arose simply because of their different timing of attack.     

Oviposition choice by the Onopordum capitulum weevil Larinus latus (Coleoptera: Curculionidae) and its effect on the survival of immature stages

Female Larinus latus lay eggs into the capitula of their Onopordum spp. host plants from the onset of capitulum development until the completion of flowering. This tight linkage between insect life-history and plant development optimised larval survival, apart from some initial asynchrony between the readiness of the females to oviposit and the suitability of the very small capitula to physically support eggs. Eggs were laid on bracts and stems of capitula or directly in the florets, when these became available. Both the location of the egg and time of oviposition influenced mortality factors such as egg and larval parasitism, egg desiccation and larval establishment. Overall, survival of eggs laid later in the season into florets was higher, although the adults that emerged were smaller. The change in choice of oviposition site that occurs at the onset of flowering, not only favours survival, but ensures more efficient resource use by the larvae. This occurs because sequential flowering of Onopordum spp. drives a sequential oviposition pattern and spreads the egg load more evenly over the available capitula, reversing a trend earlier in the season to clump eggs that were laid directly on the bracts. Other potential constraints, such as capitulum size, absolute egg density and previous egg-laying do not play a major role in determining the oviposition pattern of L. latus. Larval survival was positively correlated with capitulum size and not strongly influenced by egg density. Competition for resources only appeared to play a role in smaller capitula, and manifested itself in a reduction in the size of emerging adults rather than the death of immatures. In the absence of strong interspecific competition, the oviposition behavior of L. latus has become geared to maximising resource use for larvae (i.e. its own potential competitiveness) rather than minimising interactions with other members of the capitulum endophage guild.     

Egg–larval opiine parasitoids (Hym., Braconidae) of tephritid fruit fly pests do not attack the flowerhead-feeder Trupanea dubautiae (Dipt., Tephritidae)

Abstract:  We investigated the potential impact of three opiine tephritid fruit fly parasitoids: Fopius arisanus (Sonan), Fopius caudatus (Szépligeti) and Fopius ceratitivorus Wharton, on the non-target native Hawaiian tephritid, Trupanea dubautiae (Bryan), infesting flowerheads of the endemic Asteraceae shrub Dubautia raillardioides Hillebrand. The three species are the only known opiine fruit fly parasitoids that attack host eggs (but occasionally attack first instars). F. arisanus , which originated in southeast Asia, is now widely established in Hawaii and elsewhere in the world, while the other two are African species currently in quarantine in Hawaii. In the laboratory, field-collected flowerheads of D. raillardioides containing T. dubautiae eggs and first instar larvae were exposed to naïve female wasps of each of the three Fopius species in the absence (no-choice test) or presence (choice test) of papaya fruit infested with eggs of the Mediterranean fruit fly Ceratitis capitata (Wiedemann), the parasitoids' normal host. All three Fopius species visited the papaya fruit much more often than the flowerheads, and displayed normal ovipositional responses (probing and stinging) on the fruit in the choice test. None of the three parasitoid species showed ovipositional responses to flowerheads in either choice or no-choice tests. As a result, not a single T. dubautiae egg or larva was attacked by any of the three parasitoids, indicating that these parasitoids of pest tephritids are unable to recognize the microhabitats of flowerhead-feeding tephritids. The results suggest minimal risk of non-target impact in these biological control agents.     

Ability of ovipositing seed beetles to discriminate between seeds with differing egg loads

Abstract. 1. Previous work has shown that ovipositing females of Callosobruchus maculatus (F.) avoid seeds already bearing eggs, and thereby reduce competition among their larval progeny within seeds.
2. This study demonstrates that females also detect small differences in egg density, and prefer to oviposit on seeds with a lower-than-average number of eggs. A nearly uniform dispersion of eggs is thus maintained even after all seeds bear several eggs.
3. In addition, variation in egg load influences oviposition rate. Transfer of females from seeds with few eggs to seeds with many eggs inhibits oviposition; the reverse transfer stimulates it.
4. The upper surface of the egg chorion or egg 'cover' remains intact on the seed surface after the larva has entered the seed and continues to deter egg-laying for at least as long as the period required for larval development. 14-day-old egg covers provide as much deterrence as freshly laid eggs.     

The role of larval food resources and adult movement in the population dynamics of the orange-tip butterfly (Anthocharis cardamines)

Numbers of the orange-tip butterfly were recorded on a permanent transect in Monks Wood National Nature Reserve between 1982 and 1993, together with numbers of its only larval food plant in the wood, Cardamine pratensis. Females of the butterfly lay their eggs on the newly opened flower heads of Cardamine and the larvae feed on the developing seed pods. They are extremely selective in their choice of plant for egg laying, choosing mainly large flower heads growing in open sunny locations. Larval survival is greatly reduced if the flower head is more than 8 days old at the time of egg laying, because the seed pods become too tough for feeding. Earlier studies showed that only one larva can survive on a flower head, because larvae are cannibalistic, but females tend to avoid plants carrying more than one egg, in response to an oviposition-deterring pheromone laid down at the time of egg laying. The numbers of flower heads of Cardamine fluctuate enormously between years, and the numbers of eggs follow these closely. There is a strong correlation between number of eggs laid and the availability of suitable flower heads, and this correlation was shown to be real and causal by the experimental provision of extra flower heads, which resulted in more eggs, laid over an extended period of oviposition. The main causes of mortality of the butterfly's young stages were egg infertility, cannibalism, predation, starvation, the grazing of flower heads by deer, and possibly the failure of the newly hatched larvae to feed adequately. Adult numbers were only weakly correlated with the numbers of young stages in the wood, and there appears to be a considerable amount of movement by the butterfly through the wood and surrounding farmland. Received: 8 October 1996 / Accepted: 3 April 1997     

Life-history, genotypic, and environmental correlates of clutch size in the Glanville fritillary butterfly

Abstract.  1. Glanville fritillary butterfly ( Melitaea cinxia ) females lay up to 10 clutches of 50–300 eggs in their lifetime. Clutch size is an important life-history trait as larval group size affects survival throughout larval development.
2. Two experiments were carried out in a large population cage in the field to investigate the life-history and environmental correlates of clutch size.
3. Clutch size decreased with the cumulative number of eggs laid previously, increased with both female body weight and the number of days between consecutive clutches.
4. Genotypic differences among females in the glycolytic enzyme phosphoglucose isomerase had a significant influence on clutch size, partly because females of particular genotypes were able to initiate oviposition earlier in the day and thereby take advantage of the most favourable environmental conditions for oviposition.
5. Factors influencing clutch size were partly different in two summers, indicating the modulating effect of prevailing environmental conditions on reproductive performance.     

Mortality of Apatania fimbriata (Insecta: Trichoptera) during embryonic, larval and adult life stages

1. Egg masses, oviposition site preferences, and abiotic and biotic factors causing mortality during embryonic, larval and adult life stages of Apatania fimbriata were studied. Laboratory investigations provided information on the temperature dependence of embryonic development, measured as an increase in egg volume.
2. A. fimbriata laid hemispherical egg masses, consisting of a transparent matrix containing a mean of 208 eggs. Egg masses were laid on stones situated just above the water surface in dark cavities in the stream bank.
3. Two hundred egg masses were mapped and individually monitored during embryonic development. There was no significant correlation between mortality during embryonic development and any of the abiotic parameters measured. First-instar larvae of Osmylus fulvicephalus consumed developing eggs, and chironomids preyed on newly hatched larvae.
4. A mean of seventy-two females emerged per metre of stream. Mortality during the 1993/94 life cycle was measured as a percentage of the potential number of eggs laid. Female mortality between emergence and oviposition was ≈ 80%. Eight per cent of individuals were lost during embryonic development. Larval mortality to emergence in 1994 was 11.3%. This indicates that the terrestrial life stage is probably decisive in the regulation of A. fimbriata populations.
5. Duration of embryogenesis at constant temperatures (4–20 °C) in the laboratory was described best by a negative exponential function. This species is cold stenothermal and there was no hatching success at 20 °C.
6. Egg volumes during embryonic development increased sigmoidally over time.     

Mortality of Apatania fimbriata (Insecta: Trichoptera) during embryonic, larval and adult life stages

1. Egg masses, oviposition site preferences, and abiotic and biotic factors causing mortality during embryonic, larval and adult life stages of Apatania fimbriata were studied. Laboratory investigations provided information on the temperature dependence of embryonic development, measured as an increase in egg volume.
2. A. fimbriata laid hemispherical egg masses, consisting of a transparent matrix containing a mean of 208 eggs. Egg masses were laid on stones situated just above the water surface in dark cavities in the stream bank.
3. Two hundred egg masses were mapped and individually monitored during embryonic development. There was no significant correlation between mortality during embryonic development and any of the abiotic parameters measured. First-instar larvae of Osmylus fulvicephalus consumed developing eggs, and chironomids preyed on newly hatched larvae.
4. A mean of seventy-two females emerged per metre of stream. Mortality during the 1993/94 life cycle was measured as a percentage of the potential number of eggs laid. Female mortality between emergence and oviposition was ≈ 80%. Eight per cent of individuals were lost during embryonic development. Larval mortality to emergence in 1994 was 11.3%. This indicates that the terrestrial life stage is probably decisive in the regulation of A. fimbriata populations.
5. Duration of embryogenesis at constant temperatures (4–20 °C) in the laboratory was described best by a negative exponential function. This species is cold stenothermal and there was no hatching success at 20 °C.
6. Egg volumes during embryonic development increased sigmoidally over time.     

Variance in number of eggs per patch: oviposition behaviour and population dispersion in a seed parasitic moth

ABSTRACT. 1. The relationship between oviposition behaviour and population dispersion was analysed for Grey a subalba Braun (Lep., Incurvariidae), a species in which females usually, but not invariably. lay eggs in only a few schizocarps (seed-pairs) on each umbellet they visit on Lomatium dissectum (Umbelliferae). During each of four years females exhibited a constant probability of leaving an umbellet after each egg was laid. Hence, some umbellets received more eggs than others and the behaviour matched a geometric distribution of eggs among umbellets.
2. The variance in number of schizocarps attacked per umbellet by individual females did not result from females choosing schizocarps of particular sizes or avoiding schizocarps that already contained eggs or larvae. A potential, but speculative, explanation of the variance is that females distribute their eggs among umbellets in a way that maximizes unpredictability on larval dispersion to a searching parasitoid, The pattern of oviposition by Greya females fits the truncated geometric distribution that is predicted if information, as indexed by the Shannon-Wiener entropy measure, is minimized. None the less, other measures of unpredictability or minimum information are possible, and the results highlight the problem of how to identify patterns of movement that may result from selection for unpredictability to enemies.
3. In systematic surveys of the L.dissectum population during three consecutive years, virtually all plants had between 10% and 65% of their seeds attacked. Hence, the separation of oviposition bouts between umbellets and the constant probability of leaving an umbellet after each egg was laid resulted in a broad distribution of Greya attack among L.dissectum plants.     

The influence of leaf age on the oviposition preference of Chrysophtharta bimaculata (Olivier) and the establishment of neonates

1 The degree of discrimination shown by a herbivore when selecting oviposition sites has been suggested as a key factor to understanding herbivore population dynamics. Chrysophtharta bimaculata (Coleoptera: Chrysomelidae) is a primary pest of Tasmanian eucalypt forests and can cause severe defoliation. Previous work suggests that females show discrimination when selecting oviposition sites. Our aim was to test the degree of oviposition discrimination exhibited by C. bimaculata with regards to leaf toughness, a character that is critical to neonate survival. 2 We conducted an experiment examining the leaf toughness critical for neonate survival and found that significant larval mortality occurs above a toughness of 46.9 g. We also determined that the maximum toughness of leaves upon which larvae established in the field was 48.2 g, supporting the laboratory result. 3 Field surveys showed that although the majority of eggs were laid on leaves suitable for larval establishment, many eggs were laid on unsuitable, tougher leaves. However, all eggs were normally placed within 20 cm of suitable leaves and glasshouse trials demonstrated the neonates could move this distance without mortality occurring. 4 We conclude that egg batch distribution and larval performance of C. bimaculata will influence the population dynamics of C. bimaculata in two ways. Firstly, the availability of expanding/newly expanding leaves of eucalypt hosts will determine larval carrying capacity. Secondly, at a more localized level, the deposition of large numbers of egg batches on both suitable and unsuitable leaves followed by successful neonate migration increases the risk of resource depletion and poor larval development.     

On oviposition in weevils of the genus <Emphasis Type="Italic">Larinus</Emphasis> Dej. (Coleoptera,Curculionidae)

The following actions performed by females of several Larinus Dej. species during egg laying are described: search of an appropriate place on the plant, making the hole for the egg, oviposition proper, and sealing the hole. The hole preparation takes the longest time and the greatest effort. Only one individual usually completes development in one flower head. Females of Larinus vulpes Ol. prefer larger flower heads for oviposition and occasionally lay eggs into stems. The ability of females to distinguish the flower heads with already laid eggs is discussed. Species of Larinus may be divided into two groups with “early” and “late” oviposition. The evolution of egg laying in the genus Larinus is discussed.     

Polyphagy and primary host plants: oviposition preference versus larval performance in the lepidopteran pest Helicoverpa armigera

Oviposition preference and several measures of offspring performance of Helicoverpa armigera (Hübner) were investigated on a subset of its host plants that were selected for their reputed importance in the field in Australia. They included cotton, pigeon pea, sweet corn, mungbean, bean and common sowthistle. Plants were at their flowering stage when presented to gravid female moths. Flowering pigeon pea evoked far more oviposition than did the other plant species and was the most preferred plant for neonate larval feeding. It also supported development of the most robust larvae and pupae, and these produced the most fecund moths. Common sowthistle and cotton were equally suitable to pigeon pea for larval development, but these two species received far fewer H. armigera eggs than did pigeon pea. Mungbean also received relatively few eggs, but it did support intermediate measures of larval growth and survival. Fewest eggs were laid on bean and it was also the least beneficial in terms of larval growth. Among the host plant species tested, only flowering pigeon pea supported a good relationship between oviposition preference of H. armigera and its subsequent offspring performance. Australian H. armigera moths are thus consistent with Indian H. armigera moths in their ovipositional behaviour and larval performance relative to pigeon pea. The results suggest that the host recognition and acceptance behaviour of this species is fixed across its geographical distribution and they support the theory that pigeon pea might be one of the primary host plants of this insect. These insights, together with published results on the sensory responses of the females to volatiles derived from the different host plant species tested here, help to explain why some plant species are primary targets for the ovipositing moths whereas others are only secondary targets of this polyphagous pest, which has a notoriously broad host range. Handling Editor: Joseph Dickens     

Limiting the success of stem borers (Prodoxus quinquepunctellus) in yuccas: indirect effects of ants, aphids, and fruit position

Abstract.  1. Although the moth–yucca mutualism is often studied as a pairwise interaction, yucca plants are also the sole host for a variety of other visitors. One of these additional visitors is a stem-boring moth, Prodoxus quinquepunctellus.
2. In this study, it is shown how the reproductive success of Prodoxus indirectly depends on the interactions between yuccas and their pollinators ( Tegeticula , Prodoxidae) as well as the indirect effects of ants and aphids.
3. Aggressive wood ants foraging on yuccas will attack adult Prodoxus moths while attempting to oviposit. This reduces the number of eggs laid in yucca stalks, leading to fewer larvae feeding in the stalks.
4. Once in the stalk, the survival of Prodoxus eggs/larvae depends upon the rate at which the flowering stalks dry out during fruit maturation. Portions of the stalk above the highest fruit dry out quickly and survivorship approaches zero in these dry sections, while larvae in green sections of the flowering stalk have significantly higher survival rates. The presence of aphids feeding on the stalk slows down the rate of stalk drying and could lead to increased survival of Prodoxus larvae.
5. Overall, ants have strong indirect effects on P. quinquepunctellus by controlling how many eggs are laid in the stalk and by influencing the distribution of aphids. However, it is primarily the presence and position of the fruit that can affect larval survivorship, and fruit position is a function of pollinator visits and resource limitation. These complex interactions illustrate the importance of studying the yucca–moth mutualism in a community context.     

The relationships between host selection and subsequent larval performance in three free-living graminivorous sawflies

1. Selection of host plants for oviposition by females of three graminivorous sawfly species, Dolerus puncticollis C. G. Thomson, Dolerus picipes Klug. and Pachynematus clitellatus Lepeletier, was investigated experimentally using three grass species, Lolium perenne L., Festuca rubra L. and wheat, Triticum aestivum L. cv. Hunter.
2. Both species of Dolerus showed significant preferences for one grass over the other two, D. puncticollis preferring L. perenne and D. picipes selecting F. rubra. P. clitellatus females used L. perenne and F. rubra equally for oviposition, but avoided wheat.
3. The results correlated closely with larval performance trials, with D. puncticollis larvae having maximal survival and growth rates on L. perenne , and D. picipes larval performance being maximized on F. rubra. P. clitellatus larvae survived well on both grasses, but had a high level of mortality on wheat.
4. P. clitellatus larval growth rates were significantly higher on L. perenne than on F . rubra . Individual P. clitellatus adult females developing from larvae reared on L. perenne reached a significantly greater dry weight than those reared on F. rubra , but there was no significant difference in weight between males reared on the two hosts. There was some evidence that this disproportionate effect of host plant quality on the weight gain of the two sexes was translated into a bias in sex ratio of the eggs laid towards a greater proportion of females on L. perenne than on F. rubra . This manipulation of sex ratio would have a potential benefit for ovipositing females through disproportionate fecundity gains for female offspring.     

Diamondback moth performance and preference for leaves of Brassica oleracea of different ages and strata

The preference–performance hypothesis predicts that moth behaviour links plant variations with caterpillar attack and distribution, and the plant‐age hypothesis states that specialist herbivores are more successful in exploring younger plant tissue. We integrated these predictions to investigate underlying mechanisms by which moths and caterpillars of Plutella xylostella L. (Lepidoptera: Plutellidae) track and exploit within‐plant variability of leaf age and stratification. We measured leaf proteins, glucosinolates and fibre, as well as larval choice, developmental performance, and moth oviposition preference with regard to leaf age classes (young, mature and senescent) of three varieties (collard, cauliflower and cabbage) of the main host plant Brassica oleracea L. Larvae consistently fit the prediction that specialist herbivores prefer and perform better on young, upper leaves that have the highest protein level, despite the highest content of defence compounds. Conversely, moths laid more eggs on fibrous and less nutritious leaves from the lower and senescent stratum. We argue that the leaf stratification of host plants imposes conflicting selective pressures concerning offspring feeding and protection on adult females. If egg mortality is catastrophic on the upper nutritious leaves in a particular microclimatic context (e.g. sun, heat, winds, drought or rain‐washing), then oviposition preference will remain for the suboptimal lower and senescent leaves. The ability of larvae to spread upwards over the plant to access the more nutritious leaf stratum is critical when eggs are preferentially laid on the protective low‐quality leaves.     

Genetic variation in the reduction of attractive floral traits of an annual tarweed in response to drought and apical damage

Aims Foliar herbivory and water stress may affect floral traits attractive to pollinators. Plant genotypes may differ in their responses to the interplay between these factors, and evolution of phenotypic plasticity could be expected, particularly in heterogeneous environments. We aimed at evaluating the effects of simulated herbivory and experimental drought on floral traits attractive to pollinators in genetic families of the annual tarweed Madia sativa, which inhabits heterogeneous environments in terms of water availability, herbivore abundance and pollinator abundance.Methods In a greenhouse experiment with 15 inbred lines from a M. sativa population located in central Chile (Mediterranean-type climate), we measured the effects of apical bud damage and reduced water availability on: number of ray florets per flower head, length of ray florets, flower head diameter, number of open flower heads per plant, flowering plant height and flowering time.Important findings Apical damage and water shortage reduced phenotypic expression of floral traits attractive to pollinators via additive and non-additive effects. Plants in low water showed decreased height and had fewer and shorter ray florets, and fewer and smaller flower heads. Damaged plants showed delayed flowering, were less tall, and showed shorter ray florets and smaller flower heads. The number of ray florets was reduced by damage only in the low water treatment. Plant height, flowering time and number of flower heads showed among-family variation. These traits also showed genetic variation for plasticity to water availability. Ray floret length, flower head size and time to flowering showed genetic variation for plastic responses to apical damage. Plasticity in flowering time may allow M. sativa to adjust to the increased aridity foreseen for its habitat. Because genetic variation for plastic responses was detected, conditions are given for evolutionary responses to selective forces acting on plastic traits. We suggest that the evolution of adaptive floral plasticity in M. sativa in this ecological scenario (heterogeneous environments) would result from selective forces that include not only pollinators but also resource availability and herbivore damage.     

Limitation of superparasitism by Aphidius rhopalosiphi: a consequence of aphid defensive behaviour

Abstract. 1. The behavioural mechanisms limiting superparasitism by Aphidius rhopalosiphi (DeStefani-Perez) are described.
2. A.rhopalosiphi did not discriminate between parasitized and unparasitized aphids, or avoid attacking and ovipositing in previously parasitized aphids.
3. Aphid defensive behaviour was increased for a short period after a parasite stabbing attack. In response to this behaviour the number of parasite stabbing attacks subsequently undertaken declined.
4. Once a stabbing attack had been initiated, parasite oviposition success was unaffected by aphid defensive behaviour.
5. It is concluded that applied defensive behaviour, by limiting the number of stabbing attacks undertaken by A.rhopalusiphi , effectively limits the number of parasite eggs laid and the extent to which super parasitism occurs.     

Nutrient addition extends flowering display, which gets tracked by seed predators, but not by their parasitoids

Although phenological matching between two and three trophic interactions has received some attention, it has largely been disregarded in explaining the lack of strong cascade dynamics in terrestrial systems. We studied the response of the specialist seed predator, Paroxyna plantaginis (Tephritidae) and associated generalist parasitoids (Chalcidoidea) to controlled fertilisation of individuals of naturally growing Tripolium vulgare (Asteraceae) on four island populations (Skeppsvik Archipelago, Sweden). We consistently found evidence of nutrient limitation: fertilised plants increased their biomass, produced more capitula (the oviposition units for tephritid flies), were more at risk of attack by the tephritids, and puparia were heavier in fertilised plants. During some parts of the season tephritids became more heavily parasitized, supporting the presence of cascade dynamics, however net parasitism over season decreased in response to nutrient addition. We found no evidence that capitulum size complicated parasitoid access to the tephritids, however the extended bud production prolonged the flowering season. Thus, tephritids utilized the surplus production of capitula throughout the entire season, while parasitoids did not expand their oviposition time window accordingly. Implications for top down regulation and cascade dynamics in the system are discussed.     

Imperfect oviposition decisions by the pitcher plant mosquito (Wyeomyia smithii)

Summary Wyeomyia smithii mosquitoes distribute their eggs across available oviposition sites (water-holding pitcher plant leaves) of varying quality. I experimentally examined responses to three components of site quality: conspecific larval density, larval density of the pitcher plant midge,Metriocnemus knabi, and pitcher size. Responses to larval treatments were complex and apparently suboptimal. Although mosquito larval performance is better in leaves with fewer conspecific and more midge larvae, females did not lay more eggs in such pitchers. Instead, more eggs were laid in experimental pitchers containing either midge or mosquito larvae, but fewer eggs in pitchers with neither or both. More eggs were laid in larger pitchers, which tend to accumulate more resources and dry out less often. Therefore, although the oviposition decisions made were suboptimal, they were better than random.