Larval competition between three endemic fruit flies (Diptera: Tephritidae) of differing phylogenetic relatedness

Within a fruit, fruit fly larvae can be subject to scramble competition, in which density‐dependent effects can influence the fitness of subsequent adults. While there is significant research on tephritid interspecific larval competition, it has been conducted in invasive situations where the species are evolutionarily novel to each other. There has been no published research investigating larval competitive interactions between naturally coexisting, endemic species. We ran laboratory‐based, intraspecific and interspecific larval competition trials involving three co‐occurring Bactrocera species of differing genetic relatedness and also measured aspects of juvenile development rate to test possible mechanisms of competitive difference. Larval density had an influence on intraspecific competition in Bactrocera tryoni, Bactrocera neohumeralis and Bactrocera jarvisi, with a decreasing percentage of pupation with increasing larval density. Interspecific competition between B. tryoni and B. neohumeralis, and between B. tryoni and B. jarvisi was influenced by the interaction between species and density. The intensity of competition between B. tryoni and B. neohumeralis was minimal but high between B. tryoni and B. jarvisi. B. jarvisi produced larger eggs and had faster initial larval growth rates than the other two species, but it took the longest time for pupation to occur. Our results conflict with theory, as the greatest competition was observed between the two more distantly related species (B. tryoni and B. jarvisi) rather than between the two most closely related species (B. tryoni and B. neohumeralis). Further, and contrary to other studies, egg size, hatch rate and larval growth rate did not provide B. jarvisi with a competitive advantage; thus, larval size does not appear to be a mechanism of larval competition between B. tryoni and B. jarvisi.


INTRODUCTION
Competition is one of the key biotic factors influencing organismal distribution, abundance and diversity in ecological communities (Kaplan & Denno 2007).A common hypothesis associated with the role of competition in community assembly is that closely related species compete more intensely than distantly related species, reducing the ability of those species to coexist (Webb 2000).This is known as the 'competition-relatedness hypothesis' and is generally credited to Darwin, who stated that species in the same genus tend to be more ecologically analogous and should consequently compete more intensely than species of another genus (Cahill et al. 2008).Theoretically similar are the competitive displacement and coexistence principals that state that if two species share the same resources, then either one species should outcompete and displace the other or they should coexist by using the same resources in a different manner (Amarasekare 2003;Debach 1966).These principals have been supported by substantial theoretical (Chesson 2000) and empirical research (Silvertown 2004), demonstrating that stable coexistence is only possible if there is a sufficient ecological difference between species (Ruokolainen & Hanski 2016).
One group of organisms for which there has been extensive competition research are insects of the family Tephritidae-the true fruit flies (Benelli 2014).Frugivorous tephritid females lay their eggs into fruit, where the resultant larvae feed on the flesh of the fruit (Ansari et al. 2012).Many tephritid species have overlapping host fruit usage, so competition can occur between adults over access to fruit (= extrinsic competition) (Wang et al. 2008) or between larvae within a fruit piece (= intrinsic competition) (Liendo et al. 2016).Tephritid competition is asymmetrical and hierarchical: A given tephritid genus will be more or less competitive than another genera, while within a genus, a given species may be consistently more or less competitive than other species (Duyck et al. 2004;Wang et al. 2008).Competition between fruit flies can lead to the displacement of the less competitive species in time (Debach 1966), space (Reitz & Trumble 2002) or host (Ekesi et al. 2009).
Tephritid larval competition is considered to be of the scramble type as the system is closed (i.e., occurring within a single fruit piece) and moderated by the relative number of larvae and their individual size and age (older larvae are bigger) (Fletcher 1987).Under scramble competition, resources are equally distributed amongst individuals, leading to a homogeneous reduction in the resources available for each individual (Burrack et al. 2009).As larval density increases, the number of surviving larvae does not initially decline; however, individuals undergo a reduction in body size and a prolongation in developmental time (Peters & Barbosa 1977).This occurs until a density threshold has been reached, at which point the number of surviving larvae abruptly decreases (Fitt 1989).As many of the resources acquired by a larval fruit fly transfer to the subsequent pupa and adult, density-dependent larval competition effects can be particularly important for adult fly fitness (Krainacker et al. 1987).Competition between larvae can have a negative effect on the likelihood of an individual larva reaching the pupal stage or the time taken to reach the pupal or adult stage (Fitt 1989).Additionally, a resource shortage during the larval stage can negatively affect the size or weight of the subsequent pupa and then adults, which is directly correlated with adult reproductive success (Liendo et al. 2016).
The vast bulk of tephritid competition research has been done in invasive situations, studying competition between a newly invasive species and endemic species or previously established exotics (Biasazin et al. 2021;Duyck et al. 2004Duyck et al. , 2006Duyck et al. , 2022;;Ekesi et al. 2009;Fitt 1989;Moquet et al. 2021).In stark contrast, very little attention has been made to the study of competitive interactions between fruit fly species in established, endemic communities.In their endemic ranges, frugivorous tephritid fruit fly communities are highly speciose with substantial overlap in their host plant use (Clarke 2019): what, if any competition occurs in such systems, is almost entirely unknown.Only Fitt (1984Fitt ( , 1986aFitt ( , 1986bFitt ( , 1990aFitt ( , 1990b) ) has previously studied competitive interactions in a native fruit fly system, and his series of papers focused on differential adult reproductive capacity and host usage to explain competitive interactions, or their lack, between two to five co-occurring endemic Australian fruit flies of the genera Bactrocera and Zeugodacus (then all within the genus Dacus).He concluded that the evidence for competition occurring between these species was weak (Fitt 1989).An unpublished chapter from his PhD thesis (Fitt 1983) studied interspecific larval competitive interactions, and again, he found no evidence for competition effects.However, we believe this study to have a design weakness in that he inoculated larvae of his different study species (Bactrocera tryoni and Bactrocera jarvisi) on opposite sides of fruit without testing if that was the pattern observed in nature.If eggs were laid into the same oviposition hole, as Bactrocera will do (Silva & Clarke 2021), then the potential for larval competition is much more likely.
To better understand if larval competition may occur between members of an endemic fruit fly community, we carried out larval competition experiments with B. tryoni (Froggatt), B. jarvisi (Tryon) and Bactrocera neohumeralis (Hardy).All three are native to Australia and occur sympatrically over much of their geographic ranges (Osborne et al. 1997); there is also a very large overlap in their recorded host ranges (Hancock et al. 2000).B. tryoni and B. neohumeralis are a very closely related sibling species pair (Clarke et al. 2011;Gilchrist et al. 2014), while both are phylogenetically removed from B. jarvisi, with their most recent common ancestor occurring approximately 25-28 million years ago (Starkie et al. 2022).Based on the competitionrelatedness hypothesis, we hypothesise that evidence for intrinsic competition should be greatest between the closely related B. tryoni and B. neohumeralis and less between those two species and the more distantly related B. jarvisi.To ensure our experiments were relevant in the context of natural fruit infestation, we carried out two preliminary experiments.The first determined larval densities within naturally infested fruit so that the larval densities we subsequently used in the laboratory were not unnaturally high, while the second tested for female oviposition preference for existing oviposition sites versus an unexploited section of fruit.Based on these results, larval competition trials were run.Finally, with the results of laboratory trials showing that larval competition does occur between two of our study species, we carried out additional trials to help identify the mechanism(s) of competitive advantage or disadvantage.

General approach
This study utilised data obtained from natural infestations in the field, in addition to manipulative laboratory data, to determine the level of intraspecific and interspecific larval competition between B. tryoni and its phylogenetically related species B. neohumeralis and B. jarvisi.The study comprised four components.Component 1 undertook field collections of fruit from an orchard with no fruit fly management to determine naturally occurring infestation levels.Component 2 carried out behavioural observations to evaluate the oviposition site preference of the three fruit fly species with respect to existing oviposition wounds.These first two components were carried out as background to determine the experimental parameters of Component 3, the primary competition study.Component 3 was composed of two parts: The first investigated intraspecific competition between the three species at differing larval densities, while the second part examined interspecific competition between B. tryoni and B. neohumeralis, and between B. tryoni and B. jarvisi.As larval competition was observed, Component 4 investigated differences in egg-hatch rate and larval development rate as explainers for the observed patterns of intraspecific competition.We did not test for the competitive advantage of older larvae over younger larvae in either intraspecific or interspecific trials, as the existing literature evidence for older and larger larvae having a competitive advantage over younger and smaller larvae has previously been highly consistent (Averill & Prokopy 1987;Burrack et al. 2009;Dukas et al. 2001).However, the results of interspecific trials for B. jarvisi suggest that future studies should not automatically assume that larger larvae have a competitive advantage (see Discussion).

Insects
B. tryoni, B. neohumeralis and B. jarvisi pupae were obtained from laboratory colonies maintained by the Queensland Department of Agriculture and Fisheries (QDAF), Ecosciences Precinct, Brisbane, Australia.The colonies used were all less than 12 months from the wild and had been maintained on the same standard carrot-based diet (Heather & Corcoran 1985).Each species was held separately in nylon mesh cages (30 cm Â 30 cm Â 30 cm), with emergent adults supplied with water, sugar and yeast hydrolysate.The cages were held in an insectary at 26 C, 65% relative humidity (RH), and a photoperiod of 14L:10D.When used in experiments, flies were 14-18 days old and sexually mature.

Fruit
The fruits used in the study were selected based on seasonal availability, and a variety of fruit types were used to avoid results being driven by a single host effect.Fruit used were nectarines [Prunus persica (L.) Batsch (Rosaceae)], guavas [Psidium guajava L. (Myrtaceae)] and apple [Malus sylvestris (L.) Mill.var.Red Delicious (Rosaceae)], which are all known hosts of all three fruit fly species (Hancock et al. 2000).Only ripe, undamaged fruits were used, selected based on their colour and firmness.

Component 1: Field fruit collections
Fruit collections were carried out at the Redlands Research Facility, Department of Agriculture, Fisheries and Forestry, Brisbane, from established but unmanaged experimental orchards.Collection 1 (6 November 2020) consisted of 37 ripe nectarines, and Collection 2 (21 April 2021) consisted of 100 ripe guavas.Following fruit weighing, each fruit was placed into a gauze-covered container partially filled with moist vermiculite.The containers were stored under ambient temperature conditions, and the vermiculite was sieved for pupal recovery 7 and 14 days later.For both collections, the number of pupae, mean pupal weight per sample, percentage of adult emergence and adult sex were recorded.

Component 2: Oviposition site preference
Observations were carried out to determine if female flies showed a preference for ovipositing in an existing oviposition wound made by a previously ovipositing B. tryoni female, or if they preferred to lay into unexploited fruit.Flies were retained in a 25-cm Â 25-cm Â 25-cm Perspex box containing an undamaged organic apple.A single mature B. tryoni female was released into the arena and allowed to oviposit.The location of the oviposition site was noted, and the female was removed.A second B. tryoni female was released into the arena, and observations were made as to whether that female oviposited at the existing oviposition site or at another location on the fruit.New flies and fruit were introduced for each of the 15 replicates.Fifteen replicates were also completed for B. neohumeralis and B. jarvisi, with the first female (i.e., the female making the initial oviposition wound) always being B. tryoni.

Component 3: Larval competition studies
Based on the outcomes of Components 1 and 2 (see Results), densities of 5, 10 or 15 larvae per fruit per species were used for trials, while egg inoculation within a fruit occurred at the same site.
Experiment 3a: Intraspecific competition of B. tryoni, B. neohumeralis and B. jarvisi at increasing densities Artificial egging devices (plastic diet cups punctured with multiple pinholes and lined with filter paper saturated with apple juice) were placed into a cage containing a laboratory colony of B. tryoni and were left for 2 h to obtain eggs.Once the egging cups were removed, the filter paper was laid out in a petri dish and kept moistened with deionised water until eggs were used for fruit inoculation.Individual organic nectarines were weighed, and a sterile scalpel was used to cut a flap just below the skin.Using a fine-tipped paintbrush, eggs were placed under the flap, which was gently sealed with a square of elastic adhesive bandage.The nectarine was then placed into a container with vermiculite, covered with gauze cloth and placed in an incubator (26 C, RH 65%).After 14 days, the vermiculite was sieved to obtain the pupae.The number of pupae, individual pupal weight and number of emerging adults were recorded.Experiments were conducted at densities of 5, 10 and 15 eggs for each of B. tryoni, B. neohumeralis and B. jarvisi.Thirty replicates were completed for each treatment.Each day that trials were set up, a cohort of eggs (an average of 75-100 per day) on moistened filter paper were placed in a covered petri dish in an incubator and observed under a stereomicroscope 3 days later to assess egg viability.

Experiment 3b: Interspecific competition between B. tryoni and B. neohumeralis or B. jarvisi at increasing densities
The experimental details are as described above, except a treatment replicate consisted of a given number (5, 10 or 15) of B. tryoni eggs plus the same number of B. jarvisi or B. neohumeralis eggs.Because pupae of different species could not be discriminated, the pupation rate per species was not calculable.Rather, the total pupal number and total pupal weight for each replicate treatment were calculated.After adults emerged, the percentage of adults emerging for each species from the total pupal number was calculated.Partially emerged flies, deformed flies and unemerged flies were all counted as unemerged.

Component 4: Immature development rates
In related studies, competitively superior fruit fly larvae have had their success attributed to their larger size (Averill & Prokopy 1987).B. jarvisi eggs are significantly larger than B. tryoni and B. neohumeralis eggs, which are not different from each other (Fitt 1990b), and thus, at first hatching, B. jarvisi larvae may have a size advantage over the other species.But this apparent advantage could be lost if the eggs of the other species hatch sooner or if the larval growth rate varies between the species.To help explain the results obtained from the larval competition experiments, off-spring development attributes (egghatch rate, daily larval size and rate of pupation) for each species were recorded.

Experiment 4a: Egg-hatch rate
One hundred fifty B. tryoni, B. neohumeralis and B. jarvisi eggs, obtained as described above over a 2-h time period, were placed on moistened filter paper in covered petri dishes and then placed in a 26 C, RH 65% incubator for 37 h (time of earliest egg hatch based on preliminary trials).After 37 h, the petri dishes were observed under a stereomicroscope in a controlled environment (26 C) to minimise egg exposure to non-standard temperatures, and the number of newly emerging larvae was recorded.Following larval counts, the petri dishes containing unhatched eggs were immediately returned to the incubator.Larval counts were then conducted hourly until Hour 51, when new hatchings had largely ceased.A final count was carried out at Hour 72 to determine the total number of viable eggs.

Experiment 4b: Larval development and time to pupation
The experimental set-up follows that of the egg-hatch rate experiments.Once newly emerged larvae were observed under the microscope, a fine-tipped paintbrush was used to transfer 10 larvae to a petri dish containing carrot-based fruit fly larval-rearing medium (Heather & Corcoran 1985).Artificial rearing medium was used instead of fruit to minimise variation in the larval diet, while the low density of larvae per petri dish was to negate or minimise any intraspecific competition effects.A total of 240 petri dishes were plated out for each of the three species and placed in the incubator.Each day from Day 1 to Day 16, 150 larvae (= 15 petri dishes) of each species were removed from the incubator, and the larvae were killed using boiling water to prevent larval curling.Micrographs of each larva were taken, and ImageJ software was used to take exact measurements of larval length.Larval lengths were only recorded up to Day 6 due to pupation from Day 7, as pre-pupating larvae stop feeding and become smaller.The number of pupae was also recorded daily for each species, and the percentage of pupation was calculated as number of pupae=number of larvae ð Þ Ã 100.

Analysis of larval density effects
For intraspecific experiments, differences in the mean percentage of pupation and mean pupal weights for each species at each density were compared using one-way analysis of variance (ANOVA) with a Tukey's post hoc test and alpha set at 0.05.For interspecific experiments, a two-way ANOVA was conducted to test the differences in the mean percentage of adult emergence and mean pupal weights for each species at each density.A χ 2 analysis was carried out to compare oviposition observations of females ovipositing in pre-existing oviposition wounds and those ovipositing in undamaged sections of the fruit.Before all analyses, the data were evaluated for normality of variance using Levene's test and transformed as required.All analysis was done in Minitab Version 18.

Field fruit collections
The nectarines from Collection 1 had a mean weight of 26.46 ± 1.10 g.A mean of 5.30 ± 0.31 pupae were obtained from each nectarine, with a mean pupal weight of 0.0063 ± 0.0044 g.The guavas from Collection 2 had a mean weight of 72.14 ± 0.59 g.A mean of 16.2 ± 0.47 pupae were obtained from each guava, with a mean pupal weight of 0.0101 ± 0.044 g.Only one species, B. tryoni, were reared from fruit in these collections.Based on these results, subsequent trials used egg inoculations into fruit at densities of 5, 10 and 15 per fly species, which falls into the range of pupae obtained from the wildinfested fruit collections.

Intraspecific larval experiments
A Pearson correlation was run between the number of pupae and the number of subsequent emerging flies across all species and densities, and the pupation and emergence data were essentially identical (r = 0.958), and, as such, only the percentage of pupation data is presented.Larval density had a significant effect on the percentage of pupating B. tryoni females (df = 2, F = 35.1648,p < 0.001), with significantly lower pupation with each larval density increase (Figure 1a).There was a significant effect of density on the pupation rate of B. neohumeralis (df = 2, F = 6.2143, p < 0.005), with a significantly lower pupation at 10 and 15 larvae per fruit versus 5 larvae per fruit, with pupation rates at 10 and 15 larvae per fruit not different from each other (Figure 1b).For B. jarvisi, there was again a significant effect of density on the pupation rate (df = 2, F = 6.8939, p < 0.005).There was no significant difference in pupation rate between densities of 5 or 10 larvae per fruit, which were both significantly higher than the pupation rate at 15 larvae per fruit (Figure 1c).Changing larval density did not have a significant effect on the mean pupal weight of any of the three species (B.tryoni, df = 2, F = 0.56, p = 0.569; B. neohumeralis, df = 2, F = 0.08, p = 0.924; B. jarvisi, df = 2, F = 2.07, p = 0.127).When all pupae for a fly species were combined, there was a significant effect of species on pupal weight (df = 2, F = 65.03,p < 0.001: B. tryoni 0.012 ± 7.45E-5 g, B. neohumeralis 0.011 ± 8.38E-5 g and B. jarvisi 0.010 ± 0.001 g).

Interspecific larval experiments
For the species pair of B. tryoni and B. neohumeralis, there was a significant effect of increasing larval number on the combined pupation rate of the two species, with pupation decreasing as total larval density increased (df = 1, F = 50.24,p < 0.001) (Figure 2).The pupation rate at 10 larvae per fruit was significantly greater than at 20 and 30 larvae per fruit.The mean pupal weight decreased as density increased (df = 1, F = 26.73,p < 0.001) (Figure 3).Pupal weight at 30 larvae per fruit was significantly lower than pupal weights at 10 and 20 larvae per fruit, which were not different from each other.Once successful pupation had been reached, there was no significant effect of species, larval density or the interaction of the two on the number of adults of each species emerging (df = 1, 2, F = 0.50, p = 0.606) (Table 1).At very close to 50:50, the mean percentage of emergence was very similar for both B. tryoni and B. neohumeralis across the three densities (Figure 4a).
For the species pair of B. tryoni and B. jarvisi, there was a non-significant effect of increasing larval number on the combined pupation rate of the two species, with pupation remaining relatively constant at around 50%-55% as total larval density increased (df = 1, F = 0.46, p = 0.633) (Figure 2).Mean pupal weight decreased slightly, but not significantly, as density increased (df = 1, F = 0.60, p = 0.547) (Figure 3).Once pupation had been reached, there was a significant effect of species, an insignificant effect of larval density and a significant Species Â Density interaction effect on the number of adults of each species emerging (df = 1, 2, F = 3.96, p < 0.05) (Table 1).B. tryoni dominated samples with between 70% and 80% of emergent flies (Figure 4b).The percentage of B. tryoni adults from samples decreased slightly with increasing larval density, while the percentage of B. jarvisi adults did the reverse, hence the significant interaction effect.However, as the B. tryoni and B. jarvisi data sets are effectively mirrors of each other, any biological meaning of the interaction effect should be interpreted cautiously.

Egg-hatch rate
Egg hatching occurred earlier and more rapidly for B. jarvisi than for B. neohumeralis and B. tryoni.Egg hatching began at 37 h for B. jarvisi, and 50% of its eggs had hatched between 39 and 40 h.In contrast, egg hatching in B. neohumeralis and B. tryoni did not begin until 38 and 41 h, respectively, while 50% egg hatching for both species did not occur until between 45 and 46 h (Figure 5).

Larval development and rate of pupation
B. jarvisi mean larval lengths rapidly increased over the first 3 days of development, during which time the lengths of B. tryoni and B. neohumeralis larvae showed little or no size increase (Figure 6).For the latter species, a rapid increase in larval size occurred between Days 3 and 6 for B. neohumeralis, and between Days F I G U R E 1 Mean (±SE) percentage of pupation of (a) Bactrocera tryoni, (b) Bactrocera neohumeralis and (c) Bactrocera jarvisi when reared at three larval densities per nectarine fruit.N = 30 fruit replicates per species per density.For a given species, means surmounted by different letters are significantly different at p = 0.05. 4 and 6 for B. tryoni.Despite the initial fast larval growth rates, the time to pupation in B. jarvisi was slower than the other two species.B. neohumeralis began pupating on Day 7, 1 day before B. tryoni, with the majority (>70%) of pupation achieved by both species by Day 8 and 100% pupation reached on Day 12.The first B. jarvisi larvae similarly began pupating between Days 7 and 8, but greater than 70% pupation did not occur until Day 13 and 100% pupation by Day 15 (Figure 7).
F I G U R E 2 Mean (±SE) percentage of pupation of Bactrocera tryoni and Bactrocera neohumeralis, and B. tryoni and Bactrocera jarvisi at three larval densities (10, 20 or 30 larvae per nectarine).For each density, the starting larval number for both species is the same (i.e., 5, 10 or 15).N = 30 fruit per species combination per density.
F I G U R E 3 Mean (±SE) pupal weights of Bactrocera tryoni and Bactrocera neohumeralis, and B. tryoni and Bactrocera jarvisi at three larval densities (10, 20 or 30 larvae per nectarine).For each density, the starting larval number for both species is the same (i.e., 5, 10 or 15).N = 30 fruit per species combination per density.

DISCUSSION
We ran larval competition trials with a starting range of 5-15 eggs per species per fruit, loaded at a single location in the fruit.The larval load and inoculation method used vary significantly from other studies (e.g., Ekesi et al., 2009, used 40, 60 and 80 larvae per fruit, with neonate larvae each individually inoculated through separate fruit punctures), but based on our preliminary trials, we believe the methodology used accurately reflects the context of natural fruit infestation in our endemic fruit fly area.As theoretically expected based on scramble competition theory (Frouz et al. 2009;Ishii & Shimada 2008) and experimentally demonstrated in other studies (Dukas et al. 2001), the percentage of pupating larvae decreased with increasing larval densities across all three species and within each species combination.When reared together, the data strongly suggest that B. tryoni and B. neohumeralis larvae interacted with each other as if they were conspecifics rather than different species.Total pupation for this species pair declined with increasing larval load, but the proportion of one species to the other was almost exactly 50:50 and did not change with density.In stark contrast, when B. tryoni and B. jarvisi were paired, disproportionally high levels of B. jarvisi mortality occurred at even the lowest larval density, despite faster T A B L E 1 Results of a two-way analysis of variance on the mean percentage of adult emergence of Bactrocera tryoni and Bactrocera neohumeralis, or B. tryoni and Bactrocera jarvisi when larvae of each pair had been reared together at three larval densities (10, 20 or 30 larvae per nectarine).For each density, the starting larval number for both species was the same (i.e., 5, 10 or 15).N = 30 fruit per species combination per density.F I G U R E 4 Mean (±SE) percentage of adult emergence of (a) Bactrocera tryoni and Bactrocera neohumeralis, and (b) B. tryoni and Bactrocera jarvisi when larvae of each pair had been reared together at three larval densities (10, 20 or 30 larvae per nectarine).N = 30 per species combination per density.

Source
egg hatching and more rapid initial larval development.
While supporting the general observation that competition occurs between tephritid larvae, our results are contrary to theoretical predictions that more closely related species should compete more (Webb 2000) and also contrary to others' experimental data that in tephritid larvae, competitive advantage is gained by having older and bigger larvae (Duyck et al. 2004).These topics are addressed below in three sections: The first addresses pupation success versus pupal size; the second section addresses the strength of intraspecific versus interspecific competition and competition between more closely related and more distantly related species; and the final section discusses immature development as a potential mechanism of competition, comparing invasive versus endemic systems.

Pupation success versus pupal size
In other tephritid larval studies, an increase in larval density does not initially lead to a decrease in pupation success; rather, individuals undergo a reduction in body size (Diamantidis et al. 2020;Peters & Barbosa 1977).Resource shortage during the larval stage is generally considered to have a negative impact on the size or weight of pupae (Liendo et al. 2016), with larval mortality only occurring once some resource limitation threshold has been passed (Fitt 1989).However, in our system, this pattern did not hold for the intraspecific trials, although it did for the interspecific trials.Across all three species, at the intraspecific level, the mean pupal weight remained constant as larval density increased, but there was a decrease in pupation with increasing density; this was most strongly seen in B. tryoni.Larval densities used in our intraspecific studies reflected those seen in the field, and personal observation showed that fruit resources were not exhausted at any of larval densities used.We therefore conclude that the declining pupation success seen with increasing larval density was due to a declining quality of resource rather than a declining quantity of resource.Other tephritid studies have attributed poor larval survival to changing host quality (Oi & Mau 1989;Roohigohar et al. 2020), and this may be what is occurring in our study; that is, increasing larval load is increasingly changing host quality for the worse.Whether this effect would be as apparent if fruit were still on the plant as opposed to being in an incubator is worth further study (see Roohigohar et al. 2020 for a discussion of larval trials and laboratory incubator effects).Decreased pupal weight and decreased pupal survival were both recorded in the interspecific trials where higher larval densities were used, suggesting that resources may have become limiting in both quantity and quality in those trials.

Interspecific versus intraspecific competition and phylogenetic relatedness
In scramble competition, all potential competitors initially have the same access to resources, allowing better survival at the population level (Etienne et al. 2000(Etienne et al. , 2002)).Based on the 'competitive-relatedness hypothesis', which states that closely related species should compete more intensely than distantly related species, we hypothesised that interspecific larval competition would be more intense between the closely related B. tryoni and B. neohumeralis than between the more distantly related B. tryoni and B. jarvisi.Theory also predicts that intraspecific competition should be stronger than interspecific competition for any pair of stably coexisting species (Adler et al. 2018), and as B. tryoni is hierarchically dominant (Duyck et al. 2004), mechanisms of intraspecific competition could possibly contribute to the coexistence of the three species.Within the endemic, coexisting system of the B. tryoni, B. neohumeralis and B. jarvisi, we found evidence to both support and deny these theories, with the conflict possibly arising from the very close relatedness of the B. tryoni/B.neohumeralis sibling pair (Clarke et al. 2011).
While the B. tryoni/B.neohumeralis sibling species pair can now be reliably separated based on subtle genetic differences (Gilchrist et al. 2014), the time of male mating and some slight adult colour variation remain the only phenotypic traits known to separate them (Gilchrist & Ling 2006;Pike & Meats 2002).Our data strongly suggest that, at the level of larval competition, there may be no phenotypic differences between these species.Scramble competition was clearly occurring between B. tryoni and B. neohumeralis larvae, with decreasing pupal weight and decreasing percentage of pupation with increasing larval density, but the pupation success of both species was nearly identical at all densities.Functionally, it appears that each larva is competing with other larvae in exactly the same way, regardless of whether those larvae are conspecifics or heterospecifics.Thus, in the field, these two species may stably interact (at the within-fruit level) simply because they have no interspecific competition; that is, the larvae cannot tell each other apart, and all competition between them is effectively 'intraspecific'.
In stark contrast, competition between B. tryoni and the more distantly related B. jarvisi resulted in B. tryoni larvae killing B. jarvisi larvae, while B. tryoni pupation success and pupal weight remained constant across larval densities.The mechanism of competitive advantage in B. tryoni over B. jarvisi is not clear (see next section), but the fact that interspecific competition is occurring is indisputable in our data set.If these species have a stable interaction in the field, then it appears that the effect of interspecific competition is stronger than any intraspecific effect.
We note that our experiments used fruit exotic to Australia, which may have impacted our interpretation of the effects of phylogenetic relatedness and larval competition if differential adult host preference within native hosts means that larvae do not naturally co-occur within the same fruit type and so cannot compete.However, this appears not to be the case for the closely related B. tryoni/B.neohumeralis pair, as from field samples, they are known to emerge from the same fruit pieces in a linearly correlated manner, and that divergence in adult host usage to minimise competition does not appear to have happened (Irvine et al. 2023).However, B. jarvisi does show an oviposition preference for its primary native host, Planchonia careya, over exotic fruits (Fitt 1986a), which may lessen field competition with B. tryoni and B. neohumeralis in areas where P. careya is abundant relative to introduced fruits.The implications of differential adult host use and competition in naturally co-occurring Bactrocera species are researched and discussed by Fitt (1986b).

Immature development as a mechanism of competitive advantage or disadvantage
Despite the faster initial development rates that were observed for B. jarvisi, pupation took the longest to occur, in which case the majority of larvae only pupated a few days after B. neohumeralis and B. tryoni pupation.Within invasive situations, a mechanism by which a fruit fly species gains a competitive advantage over the other is through the production of larger eggs, which hatch more rapidly (Duyck et al. 2004;Liendo et al. 2018).However, within our three study species, the largest and most rapidly hatching eggs were laid by the competitively inferior B. jarvisi, indicating that in our endemic system, these egg attributes do not provide the demographic advantage documented in invasive systems.Furthermore, in competitive interactions between native and invasive species, larger larvae have been shown to have greater success in scramble competition (Duyck et al. 2004).Larvae may be initially larger (i.e., if emerging from a larger egg), or a size asymmetry may arise within equally aged and initially equally sized individuals if the development rate varies such that individuals of one species become larger than others and so have greater access to resources (Begon 1984;Duyck et al. 2006).However, as with egg size, B. jarvisi larvae were larger than those of B. tryoni and B. neohumeralis throughout development, and size cannot be a contributing factor to the competitive superiority of B. tryoni.For these reasons, the very low emergence of B. jarvisi when reared in the same fruit as B. tryoni is not easily explained through scramble competition.Rather, we suggest that B. tryoni larvae have an intrinsic competitive ability over B. jarvisi larvae, which act independently of scramble competition.What the mechanism(s) might be, we do not know.Drosophila melanogaster Meigen larvae can directly kill competitor larvae (Vijendravarma et al. 2013), while the fruit fly parasitoid Fopius arisanus can kill larvae of competing parasitoid species by the release of a chemical that physiologically suppresses competitor development by the release of toxic chemicals (Mahat & Clarke 2021).Either mechanism may be occurring in our system, but if so, it is entirely unstudied.
If we had run trials where the ages of competing cohorts were deliberately staggered at the outset, then we would predict, based on literature and our results, that the older cohort in the B. tryoni/B.neohumeralis pairing would have a competitive advantage regardless of species.But whether older B. jarvisi larvae would have had an advantage over younger B. tryoni larvae is unknown, given the intrinsic advantage shown by B. tryoni larvae in even-aged cohorts.Prior host usage by a female Bactrocera can deter subsequent host usage by later females, and this would lessen the potential for larval competition in the field (Fitt 1984).However, the deterrence effect is inconsistent and dependent on fruit type and prior female experience (Fitt 1984;Silva & Clarke 2021).Thus, larval competition may occur between different aged cohorts in the field and needs to be better studied now that the potential for intrinsic competitive ability independent of larval size is known to occur.

F I G U E 5
Cumulative hourly percentage of egg hatch of Bactrocera tryoni, Bactrocera neohumeralis and Bactrocera jarvisi over a 15-h period.N = 150 eggs per species.F I G U R E 6 Mean (±SE) daily larval measurements (mm) of Bactrocera tryoni, Bactrocera neohumeralis and Bactrocera jarvisi over the first 6 days of development (Day 0 = egg).N = 150 per species per day.

F
I G U R E 7 Total daily pupation (%) of Bactrocera tryoni, Bactrocera neohumeralis and Bactrocera jarvisi over a 16-day period.N = 150 per species.