# Environmental Entomology

### Editor-in-Chief

E. Alan Cameron

1.315
1.597
36 out of 94

Published on behalf of

## Intraspecific Larval Competition in the Olive Fruit Fly (Diptera: Tephritidae)

Hannah Joy Burrack, Angela M. Fornell, Joseph H. Connell, Neil V. O'Connell, Phil A. Phillips, Paul M. Vossen, Frank G. Zalom
1400-1410 First published online: 1 October 2009

## Abstract

Olive fruit flies [Bactrocera oleae (Gmelin)] occur at densities in California that can result in intraspecific larval competition within infested fruit. Larval B. oleae densities tracked in the field at six location were found to be highly variable and related to the proportion of fruit infested and adult densities. Egg and larval distribution within the field was generally aggregated early in the season and trended toward random and uniform as the season progressed. To determine whether B. oleae experienced fitness consequences at a range of larval densities observed in the field, olive fruits were infested with one, two, four, and six eggs, and larval and pupal developmental time, pupal weight, and pupal yield were compared. At the highest egg density, all measures of performance were negatively impacted, resulting in fewer and lighter pupae that took longer to pupate and emerge as adults, and even when only two larvae was present per olive, resulting pupae were significantly smaller. Density did not impact the sex ratio of the resulting flies or survive to adults. As field surveys showed, larval densities ranged from 1 to 11 B. oleae per fruit at some sites, and our results suggest that, at high densities, B. oleae do experience competition for larval resources. The impact of intraspecific larval competition North American in field populations of B. oleae is unknown, but the potential for competition is present.

• invasive species
• Olea europea
• larval behavior
• population dispersion

Intraspecific competition can impact insect population dynamics and structure (Nicholson 1955,1957; Klomp 1964). Competition between phytophagous insects can be divided into two broad categories: exploitative and interference. Exploitative competition denies resources or their full benefit to competitors, whereas interference competition harms one or all competitors through physical conflict or exclusion (Denno et al. 1995). It has been suggested that the olive fruit fly, Bactrocera oleae (Gmelin), larvae engage in contest competition, where only one larva is capable of completing development inside of a host (Fletcher 1987). Other tephritids, notably all other members of the genus Dacus, to which the olive fruit fly was once assigned, engage in scramble competition, where scarce resources result in reduced larval fitness but not direct morality (Fletcher 1987).

Olive trees (Olea europea L.) are primarily present in western North America and are grown in California as both a commercial crop and in ornamental plantings. The olive tree has relatively few associated insect herbivores either in North America or in its native range and even fewer of these are frugivores. Several scale insects [Parlatoria oleae (Colvée), Hemiberlesia lataniae (Signoret), H. rapax (Comstock), and Aonidiella aurantii (Maskell)], western flower thrips [Frankliniella occidentalis (Pergande)], and citrus peelminer (Marmara gulosa Guillén and Davis) occasionally feed externally on olive fruit (Katsoyannos 1992, Daane et al. 2005). Larvae of the second generation of the olive moth, Prays oleae (Bernard), which is not present in North America, feed on the olive pit and some surrounding fruit, but the olive fruit fly, Bactrocera oleae (Gmelin), is the only true olive frugivore in North America. Female B. oleae lay eggs singly in olives, and larvae feed exclusively on olive fruit tissue. B. oleae is not native to North America and was first detected in 1998 in the Los Angeles Basin (Rice 2000). B. oleae was subsequently detected in all olive growing regions of California by 2002 (Zalom et al. 2008). B. oleae is multivoltine and completes three to four overlapping generations per olive growing season in California, with the greatest adult densities typically occurring in the fall (Burrack 2007). At least three native hymenopteran parasitoids have been observed attacking B. oleae in California, but these have only been observed when B. oleae larval populations are high (Kapaun 2007). One introduced parasitoid, Psyttalia cf. concolor, has been released for B. oleae control in California, but its establishment and range are still being assessed (Yokoyama et al. 2008). Several other non-native parasitoids are at various stages of laboratory assessment to determine their suitability for release as biological control agents (Sime et al. 2006,2007,2008;Daane et al. 2008). Because their feeding niche is free of other frugivores and relatively free of parasitoids in California, larval B. oleae are only subject to competition from conspecifics.

Mechanisms that minimize intraspecific competition among larvae have been found in several tephritids. Host-marking pheromones inhibitory to oviposition by conspecifics have been found in several fruit fly species (Averill and Prokopy 1988, Aluja and Boller 1992, Nufio and Papaj 2004, Arredondo and Diaz-Fleischer 2006). Voltiles and chemotactile chemicals repulsive to ovipositing B. oleae are present in olive juice (Girolami et al. 1981, Scarpati et al. 1993, Lo Scalzo et al. 1994) and are spread near the oviposition scar by the female fly using their mouth parts and ovipositor. Host marking by B. oleae through these repulsive chemicals has been used to explain why egg dispersion in European populations seems to be nonaggregated, with few eggs and resulting larvae present per individual fruit. Laboratory populations of B. oleae, including that used for this study, readily exhibit this marking behavior but also regularly multiply infest olives (Genç and Nation 2008, unpublished data). The lack of oviposition inhibition in laboratory colonies is not surprising; however, we have also observed very high densities of olive fruit fly larvae in olives under field conditions in California (unpublished data). Given this, we set out to determine whether there are negative effects of competition at these observed densities. Competition in the North American population of this relatively recent invader has not been studied. Reduced intraspecific competition in some invasive insects, notably in the Argentine ant [Linepithema humile (Mayr)], has resulted in greater invasion success (Holway et al. 1998, Tsutsui et al. 2000, Thomas et al. 2005). We hypothesized that larval B. oleae populations in California are aggregated under field conditions and that the densities resulting from this aggregation will result in reduced fitness because of intraspecific competition. We further hypothesized that B. oleae engages in scramble rather than contest competition. We tested these hypotheses by conducting field observations of larval density and laboratory assays of competition. The results of these studies support our hypothesis that B. oleae partition resources such that more than one larva can and will mature within a single olive, although with reduced performance as density increases, and that densities that could result in reduced fitness regularly occur under field conditions.

## Materials and Methods

### Field Collections.

Olive fruit fly infestation was observed in Manzanillo and Mission olive cultivars at six California locations from May through December 2005 (Table 1). The sites were as follows: (1) Hilltop Ranch, a commercial olive grove in Oroville, CA; (2) USDA National Clonal Germplasm Repository for olives (USDA-NCGR) in Winters, CA; (3) Harry Laidlaw Bee Biology Research Center in Davis, CA; (4) an abandoned olive grove in Ojai, CA; (5) an ornamental olive planting in Sonoma, CA; and (6) Lindcove Research and Extension Center near Tulare, CA. Sites 1, 2, and 3 had both Manzanillo and Mission olive varieties. Sites 4 and 5 had only Mission olives, and site 6 had only Manzanillo olives. These locations were selected because they represent all olive growing regions of the state and had a diversity of climatic and cultural conditions.

View this table:

Eggs, indicated by oviposition scars (stings) and including hatched eggs, larvae (B. oleae has three larval instars that were combined into a total per olive), pupae, and exit holes were recorded weekly for each of 100 olives collected from four trees of the varieties present at each location, except for site 2, where there were only two trees of each variety. When fruit infestation levels reached 50%, sample size was decreased to 52 olives per tree. Oviposition results in a permanent scar on the fruit surface, and counts of these stings (eggs) are additive over the course of a growing season, during which multiple B. oleae generations occur. Larval presence, however, represents a single generation and is more accurate measure of potential competition at a given time. The weekly mean, SE, and minimum, maximum, and most frequent number of stings (eggs) and larvae were calculated. Egg and larval dispersion among olives for each sample date was analyzed using the index of dispersion as described by Averill and Prokopy (1989). Egg or larval distributions for each date were compared with a normal distribution (JMP 7 for Windows; SAS Institute, Cary, NC). Dates with egg or larval distributions differing from normal (α = 0.05) were classified as aggregated (variance [s2] to mean ($$mathtex$$\bar x$$mathtex$$) ratios >1) or uniform ($$mathtex$$s^2/\bar x\lt 1$$mathtex$$). Distributions not differing from normal were considered random ($$mathtex$$s^2/\bar x = 1$$mathtex$$). The index of dispersion was not calculated for Mission olives at site 2 and site 3 nor for Manzanillo olives at site 6 because of the high number of uninfested olives in these varieties at these locations. Season-long aggregation was evaluated using Taylor's power law (y = amb), expressed in logarithms (log y = log a + b log m), where y = variance, m = mean, a = intercept, and b = slope. The aggregation indices (b) were calculated for each site and the complete data set (Zimmerman and Garris 1987).

Adult B. oleae were monitored at each location with four plastic McPhail traps baited with torula yeast and borax food lure. Traps were checked, flies were counted and sexed, and lures were changed weekly.

### Colony Maintenance.

The B. oleae laboratory colony, which served as a source of gravid females for competition experiments, was maintained for >40 generations following methods adapted from Tzanakakis (1989) and further described by Genç and Nation (2008). The colony was kept at a constant temperature of 22 ± 2°C and at a light regimen of 14:10 h of light:dark. Adult flies were held in wire mesh cages at a density of 50-150 flies per 30 by 30 by 30-cm cage and 200-400 flies per 60 by 30 by 30-cm cage. Adult flies were fed a liquid diet of yeast hydrolysate, sucrose, and water in a ratio of 2:4:5 by weight. Each adult colony cage contained one piece of filter paper treated with diet, two 30-ml water containers, and 50-100 olives for oviposition. Larvae were reared to pupation on field-collected olives and exited the olives to pupate between paper towel on the bottom of rearing containers. Olives were collected regularly, washed, and stored at 4°C for up to 4 wk before use or in a modified atmosphere (98% N, 2% O2, 40% RH) for up to 12 wk before use.

### Laboratory Competition.

Olives were infested at four different egg densities to induce intraspecific larval competition and its impacts on offspring mortality, development rate, and pupal mass. Assays included four replicates of 50 olives each (200 olives per treatment), using Manzanillo olives collected on 9 August 2006 from a single tree at the University of California, Davis Department of Plant Sciences field station. Manzanillo olives were selected because previous experiments showed that larvae developing in Manzanillo olives performed better in the measures used in this assay than those in Mission olives and other smaller varieties (Burrack and Zalom 2008). Olives were visually inspected for infestation and size, and only uninfested olives of the same size were used for the competition assay. Olives were stored at 4°C until needed. Olives were infested over 4 d in order of replicate by the same group of ≈100 gravid female flies from a laboratory colony. Olives were infested at the following densities, one, two, four, and six eggs per olive, by placing ≈200 olives on the floor the colony cage, checking for stings at 20-min intervals, and confirming their presence under a stereomicroscope. Under laboratory conditions, virtually every sting contains a single egg. Once an olive reached the desired number of stings, it was removed from the cage and sorted into the appropriate treatment.

Two-liter paper cages containing the infested olives were randomly placed on a single shelf in a growth chamber and held at the same rearing conditions used for the general colony. Olives were observed daily for pupal emergence, water misted, and rotated to avoid position effects. Collected pupae were placed in individual 1.5-ml microcentrifuge tubes to permit tracking throughout the experiment. Pupae were individually weighed 12 h after collection, and held until adult emergence. After 20 d, pupal collection was terminated, and all olives were dissected to determine whether any flies had pupated inside of the fruit. No flies pupated inside the olives.

Proportion of eggs pupating (an indicator of offspring morality), larval and pupal development time, and pupal mass. Pupae were held until emergence, and the proportion of pupae emerging and the sex of the adult were observed. Data were subjected to analysis of variance using SAS Proc Mixed, with density, replicate, and their interaction as fixed effects, and means were separated using the Tukey-Kramer adjustment.

## Results

### Field Collections.

Olives with multiple stings were observed at all sites (Table 2), and olives with at least two live larvae present were observed at each site (except site 6; Table 3). The proportion of infested fruit and infestation intensity as indicated by the number of stings and larvae per olive was highly variable among locations. Sites 1 and 4 had the highest proportion of olives with stings and larvae (Tables 2 and 3), as well as the highest number of stings and larvae present per olive. These two locations also had the highest number of flies caught from May through December (Table 4). The maximum number of live larvae observed in a single olive was 11 in a Manzanillo olive from site 1 during the week of 25 October, and the mean number of larvae observed at the same site during this week was 1.70 ± 0.09. Although at least one multiple-infested olive was observed at each site except site 6, this density was never the most common (Table 3). Most olives were uninfested or singly infested. Sites 1, 2, and 3 had both olive varieties present, and within these locations, variety Manzanillo had higher proportions of infested olives and more stings and larvae present in individual olives (Tables 2 and 3). Mission olives at sites 2 and 3 had about half as many stings and larvae than did Manzanillo olives at the same locations, and Manzanillo olives at site 6 had the fewest number of stings and larvae present, as well as the lowest proportion of infested olives. This low level and intensity of infestation corresponded to the fewest number of flies trapped over the course of the season (Table 4).

View this table:
View this table:
View this table:

Both egg and larval distribution for the six variety and site combinations for which the index of dispersion was calculated were aggregated at the beginning of the season (Figs. 1 and 2). Egg densities remained aggregated, whereas larval densities became random during September and either trended or became uniform for the remainder of the season. Because egg counts are additive over the growing season, this aggregation suggests that flies are reinfesting fruit colonized by previous generations.

Fig. 1.

Egg dispersion pattern in (a) site 1 Manzanillo olives, (b) site 1 Mission olives, (c) site 2 Manzanillo olives, (d) site 4 Mission olives, (e) site 3 Manzanillo olives, and (f) site 5 Mission olives. * Significantlydifferentfromrandomdistribution (α = 0.05). Aggregated dispersions are represented by $$mathtex$$s^2/\bar x \lt 1$$mathtex$$, and uniform dispersions are indicated by $$mathtex$$s^2/\bar x \lt 1$$mathtex$$. Note scale differences between y-axes.

Fig. 2.

Larval dispersion pattern in (a) site 1 Manzanillo olives, (b) site 1 Mission olives, (c) site 2 Manzanillo olives, (d) site4 Mission olives, (e) site3 Manzanillo olives, and (f) site 5 Mission olives. *Significantly different from random distribution (α = 0.05). Aggregated dispersions are represented by $$mathtex$$s^2/\bar x \lt 1$$mathtex$$, and uniform dispersions are indicated by $$mathtex$$s^2/\bar x \lt 1$$mathtex$$. Note scale differences between y-axes.

Egg aggregation as assessed by Taylor's power law was similar for all sites, but larvae appeared more aggregated in Manzanillo olives at site 1 and Mission olives at site 4 (Table 5).

View this table:

### Laboratory Competition.

There were significant differences between densities of immature flies developing in olives with respect to the proportion of eggs pupating (F3,9 = 4.29, P = 0.0386), larval developmental time (F3,2025 = 23.29, P < 0.0001), pupal weight (F3,2025 = 25.58, P < 0.0001), and pupal developmental time (F3,1214 = 5.96, P = 0.0005; Table 6). The interaction between competition and replicate was nonsignificant for all measures of performance observed except for pupal weight (Comp × Repweight: F9,2025 = 6.24, P < 0.0001). This interaction might have been caused by greater variability between replicates of the highest density (six eggs/olive) but did not affect the interpretation of the results. Therefore, the simple effects were presented. Larval survival, as determined by the proportion of eggs completing larval development, was decreased only at the highest larval density: six eggs per olive. This pattern was the same for larval developmental time. Pupal mass was negatively impacted at all larval densities >1 but most severely decreased at densities of four and six eggs per olive. Pupal developmental time was also affected at densities of four and six eggs per olive. Pupal emergence was not related to larval density in this assay (F3,15 = 0.56, P = 0.6529). Although female flies developed to pupae significantly slower (F1,1197 = 20.70, P < 0.0001), emerged as adults slightly faster (F1,1197 = 5.21, P < 0.0226), and weighed on the order of 1 mg more than male flies (F1,1197 = 80.05, P < 0.0001), there was no significant effect of competition on the sex ratio of the resulting flies (F3,15 = 0.04, P = 0.9890) nor significant interaction effects between larval density and sex for any measure of performance.

View this table:

## Discussion

We sampled olive fruit for stings and larval infestation at six field sites, which included all olive growing regions of the state, and found that larval B. oleae densities were highly variable throughout California, but that at the majority of the locations, B. oleae larvae occur in densities that could result in intraspecific competition. We observed reduced fitness and greater mortality at densities more than one larva per olive in laboratory assays, but olives were able to support more than one larva through to pupation. These results, along with field observations of multiple infestations, support our assertion that B. oleae engage in scramble competition rather than contest competition.

Bactrocera oleae does not diapause during the winter, and fertile female flies can be found virtually any time of the year (Burrack 2007). Very young olives are not suitable for larval development, so fruit availability is a key limiting factor for spring population growth. Early season egg and larval distribution at the locations observed was aggregated (Figs. 1 and 2). This pattern many indicate an initially scarce food source as developing olives become attractive to female flies and suitable to larval development at different rates during the beginning of the growing season. Larval distributions became random and either trended toward or became uniform as the season progressed, indicating that all the fruit present were able to support larval development. This pattern is similar to that observed for the apple maggot fly, Rhagoletis pomonella Walsh, infesting hawthorn (Averill and Prokopy 1989). Egg distributions, however, remain aggregated. This is likely because of the use of stings, a permanent physical character of the olive, meaning that distributions at later dates represent additive values rather than point observations. Eggs observations at the end of the season likely include oviposition activity from three to four generations. The notable exception to this pattern is site 5. Olive fruit fly activity at this location was delayed because of its cooler weather, and the egg distribution pattern resembles larval distribution at the other locations.

At the three locations where both olive varieties observed were present, larval b trended higher, indicating more aggregation, in Manzanillo olives (Table 5). This effect was not observed at site 2, which had a lower number of flies captured in 2005 and is most pronounced at site 1, which had the highest trap captures. These observations support previous work by our group, indicating that, even in the presence of suitable Mission olives, Manzanillo olives are preferred by B. oleae (Burrack and Zalom 2008).

Effects of intraspecific larval competition were more pronounced for all performance measures at the highest larval density tested. These results support those reported by three European studies (Manoukas and Tsiropoulos 1977, Manoukas 1980, Cirio and Gherardini 1981). Two of these studies examined larval competition in artificial diet and found decreased pupal weights, reduced survival to pupa, extended pupal development time, and reduced adult emergence at densities of 20 eggs/g diet and above (Manoukas and Tsiropoulos 1977) and that larval weight did not decrease because of competition until the fifth day of development at densities of up to 60 eggs/g diet (Manoukas 1980). The third study was similar to ours in that it reported both field infestations and larval performance in olives in the laboratory (Cirio and Gherardini 1981). The field portion of the study equated number of stings to number of eggs per olive and assumed exit holes were evidence of larval survival. These assumptions did not account for stings and exit holes being permanent physical characters remaining on infested fruit and may have resulted from several generations that had been exposed to different infestation densities. Larval densities in the field were not reported, so competition intensity could not be determined. The laboratory portion of this study found that larval mortality increased with egg density and that pupal weight decreased with larval density, but they did not report at which larval densities these reductions in fitness occurred. They also did not observe developmental time, which may have implications for population dynamics in the field. More rapid development time may result in a greater number of generations per year when accumulated over the course of the growing season and therefore higher populations and greater associated economic injury.

The larval densities assayed by Cirio and Gherardini (1981), one, two, three, and four eggs per olive, were far lower than densities that occur in the field in California. Our laboratory study, which included densities of one, two, four, and six eggs per olive, documented detrimental effects on the proportion of eggs surviving to pupae, larval, and pupal developmental time and pupal weight at the greatest density (six per olive). Pupal weight was reduced with the addition of only one additional larva per olive (two per olive). Pupal weight in other tephritids species is well correlated to adult size, and small adult females have been documented to be significantly less fecund than larger conspecifics (Averill and Prokopy 1987, Nufio and Papaj 2004b). A size reduction occurring with as few as two larvae per olive may have significant implications for population dynamics, although the relationships between size and fecundity in B. oleae have not been studied. The flies used in our assay were from a laboratory colony reared from olives often with considerably greater egg densities than six per fruit. In similarly maintained B. oleae colonies, densities as great as 28 surviving larvae have been reported (Genç and Nation 2008). It is possible that our assay may underestimate the potential for larval competition in wild populations, because our laboratory flies may be adapted to relatively high larval densities. The densities at which we saw negative impacts of competition are still far lower than the highest density of live larvae observed per olive in the field: 11 in one Manzanillo olive from site 1.

Intraspecific competition in R. pomonella has been well studied (Averill and Prokopy 1987, Feder et al. 1995) and offers further possibilities for B. oleae. Averill and Prokopy (1987) found that larval age was important with respect to competition. Older larvae, distinguished by radiolabeling, performed significantly better in competition with younger larvae. Our laboratory assays created competition between larvae of the same age, but the larvae collected from multiple-infested fruit in the field included all three instars. It is reasonable to assume that competition may be differ between older and younger B. oleae larvae, as suggested by the work of Averill and Prokopy (1987), potentially resulting in the complete competition suggested by Fletcher (1987). Temporal aspects of larval competition, the effects of competition under field conditions, and competition in small olive varieties all warrant further study.

In summary, detrimental impacts of intraspecfic larval competition in the olive fruit fly, B. oleae, occur in California and European populations, but the significance of this intraspecific larval competition on B. oleae population regulation in the field is unknown. Some resulting performance impacts (increased development time) will only occur at locations with relatively high densities of flies and may be less important overall than the effects of temperature and host plant phenology on B. oleae populations. A reduction in size may occur at lower densities and also potentially impact field populations. Recognizing that population regulating factors such as intraspecific larval competition do indeed occur helps define an upper infestation limit for an invasive pest such as B. oleae.

## Acknowledgments

We thank M. Brown, A. Devarenne, D. Chan, J. Tsai, and E. Jovel for field and laboratory assistance during the course of this project, M. Johnson, J. Carey, and three reviewers for comments on the manuscript, and the University of California ANR Core Issues Research Grants Program for funding in support of this study.