ECOLOGY AND BEHAVIOR Evidence for Competitive Displacement of Ceratitis cosyra by the Invasive Fruit invadens (Diptera: ) on Mango and Mechanisms Contributing to the Displacement

SUNDAY EKESI,1 MAXWELL K. BILLAH, PETERSON W. NDERITU, SLAWOMIR A. LUX, AND IVAN RWOMUSHANA

International Centre of Physiology and Ecology (icipe), P.O. Box 30772-00100 GPO, Nairobi, Kenya

J. Econ. Entomol. 102(3): 981Ð991 (2009) ABSTRACT Bactrocera invadens Drew, Tsuruta & White (Diptera: Tephritidae) invaded Kenya in 2003. Before the arrival of B. invadens, the indigenous fruit ßy Ceratitis cosyra (Walker) was the predominant fruit ßy pest of mango (Mangifera indica L.). Within 4 yr of invasion, B. invadens has displaced C. cosyra and has become the predominant fruit ßy pest of mango, constituting 98 and 88% of the total population in traps and mango fruit at Nguruman, respectively. We tested two possible mechanisms responsible for the displacement namely; resource competition by larvae within mango fruit and aggression between adult ßies. Under interspeciÞc competition, larval duration in B. invadens was signiÞcantly shorter (6.2 Ϯ 0.6Ð7.3 Ϯ 0.3 d) compared with C. cosyra (8.0 Ϯ 1.2Ð9.4 Ϯ 0.4 d). Pupal mass in C. cosyra was affected by competition and was signiÞcantly reduced (7.4 Ϯ 0.3Ð9.6 Ϯ 0.6 mg) under competitive interaction compared with the controls (12.1 Ϯ 1.5Ð12.8 Ϯ 1.1 mg). InterspeciÞc competition also had a signiÞcant adverse effect on C. cosyra eclosion, with fewer adults emerging under co-infestation compared with the controls. Interference competition through aggressive be- havior showed that fewer C. cosyra (3.1 Ϯ 0.8) landed on mango dome compared with the controls (14.2 Ϯ 1.5) when adults were mixed with B. invadens adults in Plexiglas cages. Similarly the number of times C. cosyra was observed ovipositing was signiÞcantly lower (0.2 Ϯ 0.2) under competitive interaction compared with the controls (6.1 Ϯ 1.8). Aggressive encounters in the form of lunging/ head-butting and chasing off other species from the mango dome was higher for B. invadens compared with C. cosyra. Our results suggest that exploitative competition through larval scrambling for re- sources and interference competition through aggressive behaviors of the invader are important mechanisms contributing to the displacement of C. cosyra by B. invadens in mango agroecosystems.

KEY WORDS invasive species, Bactrocera invadens, native species, Ceratitis cosyra, competitive displacement

The fruit ßy Ceratitis cosyra (Walker) has long been (Lux et al. 2003a; S.E. et al., unpublished data). In 2004, recognized as the most damaging tephritid fruit ßy a shift in dominance between C. cosyra and an invasive pest of mango (Mangifera indica L.) in Africa, includ- species Bactrocera invadens Drew, Tsuruta & White ing Kenya (Lux et al. 2003a). Few other tephritid fruit was observed in mango orchards at Nguruman in the ßies, such as Ceratitis fasciventris (Bezzi), Ceratitis Rift Valley Province of Kenya, just 1 yr after its de- rosa (Karsch), Ceratitis anonae (Graham), and to a tection in the country (Lux et al. 2003b). Ekesi et al. limited extent Ceratitis capitata (Weidemann), also (2006) in assessing the level of damage of the new coexist with C. cosyra on mango in different parts of invasive species on mango speculated that competitive Africa, but C. cosyra has generally been regarded as displacement appeared to be in progress. the primary pest of mango. Losses are because of B. invadens is believed to have invaded Africa from direct feeding damage and loss of export market op- the Indian subcontinent and was discovered in Sri portunities through quarantine restrictions imposed Lanka after it was Þrst reported from Africa, where it by importing countries to avoid entry and establish- has become a signiÞcant pest of quarantine and eco- ment of unwanted fruit ßies. For example, because of nomic importance (Mwatawala et al. 2004, Drew et al. the threat posed by invasive fruit ßies, Mauritius and 2005). The insect is rapidly spreading across tropical South Africa have recently banned the importation of Africa and in addition to Kenya, it is now reported mango and avocado from Kenya. Direct damage from 28 other African countries, including the Como- caused by the fruit ßies usually ranged from 20 to 80% ros Island (Drew et al. 2005; French 2005; Ekesi and Billah 2006; S.E. et al., unpublished data). Although it 1 Corresponding author, e-mail:[email protected]. has now been reported from Ͼ30 plant species, its

0022-0493/09/0981Ð0991$04.00/0 ᭧ 2009 Entomological Society of America 982 JOURNAL OF ECONOMIC ENTOMOLOGY Vol. 102, no. 3 primary host is currently mango (Mwatawala et al. surrounded by a dry savannah belt and thus could be 2004, Drew et al. 2005, Ekesi et al. 2006, Ekesi and described as an “ecological island.” The average max- Billah 2006, Rwomushana et al. 2008). The introduc- imum and minimum temperature during the experi- tion of species into a new area can alter successional mental period was 21 and 35ЊC, whereas minimum and patterns, mutualistic relationships, community dy- maximum relative humidity was 33 and 76%, respec- namics, ecosystem function, and resource distribution tively. (Mooney and Cleland 2001). Invasive species can also Trapping System and Collection of . A small- negatively impact resident populations through eco- holder mango farm (Mololo farm) measuring Ϸ 0.5 ha logical interactions such as competition leading to was used for the experiments. Interviews with the displacement. Reitz and Trumble (2002) deÞned com- farmer before monitoring and fruit sampling revealed petitive displacement as “the removal of a formerly that no pesticide treatments had previously been em- established species from a habitat through superior ployed against fruit ßies or any other insect pests in the use, acquisition or defense of resources by another farm. Early harvesting as a means to evade fruit ßy species.” This can occur through many different facets infestation was the only fruit ßy management strategy that are often broadly categorized as exploitative or used by the farmer. interference in nature. Competitive displacement is McPhail traps were the only traps used to monitor frequently difÞcult to document under natural con- ßy densities during the experiments. Traps were ditions, but it is occasionally conspicuous when or- baited with liquid protein bait consisting of 300 ml of ganisms invade a new continent or islands (Mooney aqueous solution of 9% NuLure and 3% borax. Five and Drake 1986). In general, opportunities to study traps were distributed randomly in the orchard at a the invasion process have usually been missed because distance of 30 m between traps. Traps were rotated such community revolutions are generally very swift sequentially after each sampling. The traps were and most research efforts are concentrated on con- placed on trees at between 1.5 and 2.0 m above ground trolling the invading pest (Brown et al. 1995). Several and were checked once a week. NuLure was renewed studies have however tried to identify attributes of every 7 d. At each check, the number of fruit ßies invasive species, determine factors that govern their captured was taken to the laboratory where they were establishment and subsequent rate of spread (Wil- counted and their identity determined. liamson 1996, Reitz and Trumble 2002; Bruno et al. Fruit Sampling and Handling of Insects. Ripe and, 2005). In spite of the progress in these areas, predict- occasionally, unripe mango fruit were sampled from ing the outcome of a particular invasion remains a the tree and the ground (as “windfalls”). Fruit were daunting task (Lodge 1993). transported to the laboratory in screened 3-liter plas- Among fruit ßies, competitive displacements of in- tic buckets, weighed, and counted. Fruit from separate digenous species by introduced species have also been collections were placed individually in 2-liter rectan- reported (Duyck et al. 2004, 2006). However, like in gular, plastic containers that had Ϸ0.5- and 2.5-cm other species of insects, they are often too rapid for ellipsoidal holes cut at the bottom (Copeland et al. systematic studies of the casual factors. Since 2001, we 2002). The container was then nested inside a 4-liter had been investigating seasonal population ßuctuation rectangular plastic container with heat sterilized of C. cosyra at Nguruman, Rift Valley Province, Kenya. moistened sand at the bottom for pupation. The holes By October 2003, we recorded B. invadens in McPhail in the smaller container allowed mature pupariating traps baited with NuLure at the start of our routine larvae to fall into the larger container below to pu- seasonal population studies of C. cosyra in the orchards. pariate after leaving the mango fruit. We continued to monitor population of both species in Sample containers were checked every 3Ð4 d for this area, and in the present article we document evi- puparia and adults ßies. Puparia were sieved from the dence of competitive displacement of the indigenous sand and held for adult emergence in transparent species C. cosyra by the invader B. invadens on mango Plexiglas cages (25 by 25 by 25 cm). Adult ßies were agroecosystem and provide some laboratory data on pos- fed on a diet consisting of 3 parts sugar and 1 part sible mechanisms of the displacement. enzymatic yeast hydrolysate ultrapure (USB Corpo- ration, Cleveland, OH). Water also was provided on pumice granules. After 5 d when adult body color had Materials and Methods fully developed, ßies were allowed to die within the cages or were killed by freezing. Fruit samples were Field Experiments discarded after Ϸ4 wk. Fruit infestation index was Description of Experimental Site. The study was calculated as the ratio of number of adults per kilo- carried out from October/November to January 2000Ð gram of fruit collected (Cowley et al. (1992). 2007 at Nguruman, Kajiado District in the Rift Valley Province of Kenya. Nguruman is an irrigation scheme Laboratory Studies on Mechanisms of Displacement located at 01Њ 54Ј 44 S, 36Њ 17Ј 15 E and altitude of 700 m. The mango growing area is supplied, in addition Resource Competition between Larvae. Adult fe- to rains, with furrow irrigation water from three rivers male B. invadens and C. cosyra were obtained from a (Oloibortoto, Entasopia, and Sampu), which cut culture maintained on mango (ÔAppleÕ) for six gener- across the area and eventually discharge into Lake ations. The ßies were exposed separately to a mango Natron of Tanzania. This rich and productive area is dome (a whole fruit in which the pulp and seed had June 2009 EKESI ET AL.: COMPETITIVE DISPLACEMENT AMONG FRUIT 983 been removed leaving just the skin) in a 50- by 50- by versus C. cosyra females, 2) B. invadens males versus 50-cm Plexiglas cage to obtain eggs of each species. C. cosyra female, 3) B. invadens female versus C. cosyra After 24 h of oviposition, eggs were collected from the male, 4) B. invadens male versus C. cosyra female, 5) dome by using a Þne brush and placed on a wet Þlter B. invadens females only, 6) B. invadens males only, 7) paper in petri dishes for hatching. After hatching, 20 C. cosyra females only, and 8) C. cosyra males only. In or 40 newly emerged larvae of each species of fruit ßy treatments 1Ð4, Þve insects of each species was used, were collected from the dishes and gently introduced whereas in treatments 5Ð8, 10 insects were used. with a Þne brush into each of 60 or 80 holes (20Ð40 Mango dome with droplets of NuLure also was pro- holes per each fruit ßy species) perforated with an vided as in the landing and egg-laying assay, and insect entomological pin on the surface of a single large ripe age was also the same. Four observers collected data mango. Controls with single species treatment (one between 10 and 12 h (2 h) and recorded the number fruit for each species) had 20 or 40 holes for the larvae. and outcome of aggressive interactions in the form of Each hole was Ϸ1 mm in diameter and 1 cm in depth. lunging, head-butting, and pushing between species The experiments therefore consisted of the following when on the dome. Departure from the mango by Þve treatments: 1) 40 larvae of B. invadens versus 20 either species was classiÞed as either unprovoked larvae of C. cosyra, 2) 20 larvae of B. invadens versus (without apparent cause) or as a result of aggression. 40 larvae of C. cosyra, 3) 40 larvae of B. invadens versus Four replications were maintained per each treatment 40 larvae of C. cosyra, 4) Control (40 larvae each), and and the experiment was repeated twice. The labora- 5) control (20 larvae each). Prospects for interspeciÞc tory experimental conditions were similar to the re- competitions depend both on the frequency of co- source competition experiments. infestation and the density of larvae within the fruit. Daily oviposition rates of B. invadens at peak period of egg laying can vary from 15 to 76 eggs per female per Statistical Analysis d, whereas C. cosyra can lay between 10 and 40 eggs A nonlinear regression based on GaussÐNewton per female per d (Ekesi et al. 2006; S.E. et al., unpub- method was applied to Þeld data by using the Proc lished data). The larval densities and combinations for NLIN procedure. In this analysis, the function y is the treatments described above were therefore cho- observed fruit ßy response, x is time, and exp is the sen on the basis of the level of fecundity of each exponential function; a, b and c are parameters esti- species and these densities are within the range used mated by the nonlinear least squares estimation by several other authors studying larval development (Draper and Smith 1981, KÞr 1997). In the resource and interspeciÞc competition of Tephritidae (Keiser competition and aggression studies, we protected et al. 1974, Fitt, 1986, Krainacker et al. 1987, Duyck et against type I error by including global comparison by al. 2006). After larval introduction, each mango was using analysis of variance (ANOVA) for a complete transferred into 2-liter rectangular, plastic containers randomized design (Scheiner and Gurevitch 1993) for as described above, and fruit handling and rearing of comparisons within a species among treatments and insects were similar to the procedures described for were appropriate post-ANOVA mean separation was the Þeld experiment. At pupation, puparia were held done using TukeyÕs honestly signiÞcant difference individually in 95- by 30-mm vials with a screen glued (HSD) test (P ϭ 0.05). Wilcoxon paired test (Siegel to one cut end of the cage until adult emergence and and Castellan 1988) was applied to test for interspe- species determination. Developmental duration, total ciÞc differences within a treatment. Data for larval pupal harvest, pupal weight, and percentage of eclo- developmental time, pupal harvest, and pupal weight sion were determined. Each fruit containing both fruit were transformed to natural logarithms, whereas the ßy species served as a replicate, and there were four proportion of adult emergence data were arcsine replications per treatment and the experiment was transformed before analyses. All analyses were per- repeated twice. The experiments were carried out in formed using the SAS (SAS Institute 1989) software. a room maintained at 28 Ϯ 1ЊC, 50 Ϯ 8% RH, and a photoperiod of 12:12 (L:D) h. Aggression. Aggressive interaction was quantiÞed Results by transferring Þve pairs of B. invadens and Þve pairs Field Data of C. cosyra (mixed together) to a 50- by 50- by 50-cm Plexiglas cage containing a mango dome. Flies were In the 2000Ð2002 season, the indigenous fruit ßy C. 10Ð14 d old. Several droplets of 2% NuLure were applied cosyra was the principal pest detected in monitoring on the surface of the dome. The dome served as a source traps and reared from mango fruit apparently because of egg laying and the NuLure as food. In the control B. invadens had not invaded the study area (Fig. 1). cages, each species (Þve pairs) was held separately with- During the October 2003 to January 2004 growing out cohabitation. One to three observers collected data season, of 2,916 fruit ßies collected from monitoring for between 10 and 12 h (2 h) and recorded the number traps, 72% of the population was C. cosyra, 27% was B. of landings and egg laying on the mango dome. invadens, and other ßies consisting mainly of C. rosa, A second independent assay assessed aggressive be- C. fasciventris, C. capitata, and Dacus spp. made up 1% havior in the form of lunging, head-butting, or pushing of the population (Fig. 1A). In the same year, fruit between the species. In this assay, the following treat- infestation data showed that 82% were C. cosyra, ment combinations were used: 1) B. invadens females whereas 18% were B. invadens (Fig. 1B). By Novem- 984 JOURNAL OF ECONOMIC ENTOMOLOGY Vol. 102, no. 3

Fig. 1. Displacement of C. cosyra by B. invadens at Nguruman, Kenya, between 2000 and 2001 and 2007 and 2008 mango seasons. (A) Percentage of ßy catches in trap (n ϭ 5 traps): C. cosyra:Yϭ axb exp (cx): a ϭ 22.6218, b ϭ 0.5430. c ϭϪ3.1417, R2 ϭ 0.81, P ϭ 0.0041. B. invadens:Yϭ axb exp (cx): a ϭ 15.2451, b ϭ 0.7622, c ϭϪ1.7845, R2 ϭ 0.81, P ϭ 0.0021. (B) Percentage of ßies in fruit samples: C. cosyra:Yϭ axb exp (cx): a ϭ 16.8734, b ϭ 0.4116. c ϭϪ1.3542, R2 ϭ 0.88, P ϭ 0.0012. B. invadens: Y ϭ axb exp (cx): a ϭ 32.1487, b ϭ 0.8028, c ϭϪ1.2402, R2 ϭ 0.72, P ϭ 0.0001. (C) Flies per trap per d (n ϭ 5 traps): C. cosyra: Y ϭ axb exp (cx): a ϭ 8.2142, b ϭ 2.6451, c ϭϪ1.3042, R2 ϭ 0.92, P ϭ 0.0027. B. invadens:Yϭ axb:aϭ 2.1125, b ϭ 1.3682, R2 ϭ 0.92, P ϭ 0.0001. (D) Percentage of ßies in fruit samples: C. cosyra:Yϭ aexp (bx): a ϭ 1.0426, 0.8642, R2 ϭ 0.91, P ϭ 0.0038. B. invadens:Yϭ axb:aϭ 0.4834, b ϭ 1.2075, R2 ϭ 0.82, P ϭ 0.0008. ber 2004, B. invadens rapidly increased its percentage January 2008, a complete reversal of the 2003 Þgure of the ßy population reaching 75% in monitoring traps was observed, with 98 and 88% of the ßy population (Fig. 1A). On mango fruit, the invader had occupied consisting of B. invadens in traps and fruit, respectively 80% of the population (Fig. 1B). By October 2007 to (Fig. 1A and B). June 2009 EKESI ET AL.: COMPETITIVE DISPLACEMENT AMONG FRUIT FLIES 985

Fig. 2. Effect of interspeciÞc competition on life history traits of B. invadens and C. cosyra. (A) Larval duration. (B) Pupal weight. (C) Adult emergence. Error bars denote SE. For each species, bars superscripted with same letter do not differ signiÞcantly by TukeyÕs HSD test (P ϭ 0.05). Bi, B. invadens; Cc, C. cosyra.

In the estimated regression parameters the high under competitive interactions between both species values of coefÞcient of determination indicate a good compared with the controls: B. invadens (F ϭ 24.43, Þt to the data generated. The mean number of C. df ϭ 4, 25; P ϭ 0.0001), C. cosyra (F ϭ 18.56; df ϭ 4, cosyra per trap per d was high from the beginning but 35; P ϭ 0.0001) (Fig. 2A). Under co-infestation, larval started to decrease rapidly after the invasion of B. duration of B. invadens was found to be shorter (6.8Ð invadens (Fig. 1C). A similar case was also observed in 7.3 d) compared with C. cosyra (8.0Ð9.4 d). Wilcoxon fruit. By the 2003Ð2004 season, the density of C. cosyra paired test revealed signiÞcant differences in larval on mango fruit was 4.5 ßies per kg fruit but decreased duration among the two species when 40 larvae of B. to 0.7 ßies per kg fruit in the 2007Ð2008 season. In invadens was co-infested with 20 larvae of C. cosyra contrast, the level of infestation by B. invadens has (Z ϭϪ3.3706, df ϭ 7, P ϭ 0.0008) and when 40 larvae showed an increase from 4.1 ßies per kg fruit in the of B. invadens was co-infested with 40 larvae of C. 2004Ð2005 season to 5.6 ßies per kg fruit in the 2007Ð cosyra (Z ϭϪ3.3731, df ϭ 7, P ϭ 0.0007) but no 2008 season (Fig. 1D). signiÞcant difference in the treatment fruit with 20 larvae of B. invadens versus 40 larvae of C. cosyra (Z ϭ Ϫ2.7674, df ϭ 7, P ϭ 0.0057). Mechanisms of Displacement Pupal weight did not differ signiÞcant among co- Resource Competition by Larvae. Larval duration infested treatments in B. invadens when compared was affected by interspeciÞc competition among the with the controls (F ϭ 8.32; df ϭ 4, 35; P ϭ 0.3215) (Fig. different infestation combinations and was shorter 2B). However, in C. cosyra pupal weight was signiÞ- 986 JOURNAL OF ECONOMIC ENTOMOLOGY Vol. 102, no. 3

Fig. 3. Behavioral responses (mean Ϯ SE) of B. invadens and C. cosyra under competitive interaction in the laboratory: (A) Number of landings. (B) Number of times observed laying eggs. (C) Number of aggressions. Bi, B, invadens; Cc, C. Cosyra; ǨǨ, females; ((, males. cantly reduced (7.4Ð9.6 mg) under competitive inter- difference in adult emergence across treatment for C. action with B. invadens compared with the controls cosyra (F ϭ 12.78; df ϭ 4, 35; P ϭ 0.0001). When mango (12.1Ð12.8 mg) (F ϭ 22.56; df ϭ 4, 35; P ϭ 0.0001) (Fig. fruit was infested at the level of 40 B. invadens larvae 2B). Wilcoxon test for the following treatments were versus 20 C. cosyra, the proportion of adult C. cosyra signiÞcant among the two species: 40 larvae of B. that emerged was 0.21 compared with 0.79 B. invadens invadens versus 20 larvae of C. cosyra (Z ϭ 3.3706, df ϭ adults (Z ϭ 3.4164, df ϭ 7, P ϭ 0.0006) (Fig. 2C). 7, P ϭ 0.0008), 20 larvae of B. invadens versus 40 larvae Infestation of 40 B. invadens larvae versus 40 C. cosyra of C. cosyra (Z ϭ 3.3681, df ϭ 7, P ϭ 0.0008), and 40 larvae resulted in 0.23 C. cosyra compared with 0.77 B. larvae of B. invadens versus 40 larvae of C. cosyra (Z ϭ invadens adults (Fig. 2C) (Z ϭ 3.3731, df ϭ 7, P ϭ 3.3756, df ϭ 7, P ϭ 0.0007). 0.0007). Competitive interaction therefore had a sig- The total puparia (B. invadens and C. cosyra) har- niÞcant adverse effect on C. cosyra survivorship. vested under co-infestation was 39.1 Ϯ 2.6, 34.2 Ϯ 1.8, Aggression. Under interspeciÞc competition, fewer and 42.2 Ϯ 3.2 in the respective treatments: 40 B. C. cosyra landed on the mango dome compared with invadens larvae versus 20 C. cosyra larvae, 20 B. inva- the controls (F ϭ 11.13; df ϭ 1, 14, P ϭ 0.0001) (Fig. dens larvae versus 40 C. cosyra larvae, and 40 B. inva- 3A). Between species, more B. invadens were re- dens larvae versus 40 C. cosyra larvae. The proportion corded on the dome compared with C. cosyra (Z ϭ of initial larvae of B. invadens that emerged as adults 3.8706, df ϭ 7, P ϭ 0.0008). Similarly, the number of was not signiÞcantly different across treatments (F ϭ times C. cosyra was observed ovipositing was signiÞ- 1.36; df ϭ 4, 35; P ϭ 0.2876), but there was a signiÞcant cantly lower under competitive interaction compared June 2009 EKESI ET AL.: COMPETITIVE DISPLACEMENT AMONG FRUIT FLIES 987 with the controls (F ϭ 21.30; df ϭ 1, 14; P ϭ 0.0001) elevation regions of the country but the magnitude of (Fig. 3B). Aggressive encounters recorded as the damage to mango is not well known (M.K.B. et al., number of times each insect was observed lunging and unpublished data). chasing off the other species from the mango dome The mechanisms that trigger competitive displace- was higher for B. invadens compared with C. cosyra ment are usually very difÞcult to establish and may be (Fig. 3C). In general, both females and males of B. speciÞc to each pair of competing species. In the invadens were highly aggressive to either sex of C. conventional niche theory, the primary determinant cosyra in the various encounters and rarely was the of competition is overlap in resources (or niche over- reverse observed: For example, B. invadens male ver- lap). B. invadens and C. cosyra are ecological homologs sus C. cosyra male (Z ϭ 3.3933, df ϭ 7, P ϭ 0.0007) and that, in spite of substantial difference in the range of B. invadens female versus C. cosyra male (Z ϭ 3.3400, host plants and the level of polyphagy, largely com- df ϭ 7, P ϭ 0.0007) (Fig. 3C). pete for the same ecological nicheÑmango fruit. Re- sults from larval competition for food resources re- vealed that co-infestation of mango by larvae of both Discussion species adversely affected C. cosyra. Most notable is Increased levels of travel and trade have led to a that few adults of C. cosyra eclosed when B. invadens heightened spread of invasive species, a leading an- larvae infest mango at a higher density (40 larvae of B. thropogenic disturbance with far-reaching implica- invadens and 20 larvae of C. cosyra). In growing pop- tions (Naeem et al. 1995). About one Þfth of invasive ulations, larval duration may be the most important species may cause extensive economic and ecological parameter, but it may be less critical compared with damage with unpredictable negative effects on native pupal mass in stable populations (Yoshimura and populations that are second only to habitat destruction Clark 1991), and we did observe that pupal weight was (DÕAntonio and Vitousek 1992; Jenkins 1996). The adversely affected in C. cosyra under competitive in- results of the current study clearly indicate rapid dis- teraction where it was signiÞcantly lower compared placement of C. cosyra by B. invadens at Nguruman, with the control. One signiÞcant and interesting ob- Kenya, 4 yr after its detection in the African continent. servation was that under competitive interaction, the Several Bactrocera species are notorious and well- duration of larval development was shortened in both documented invaders and rank high on quarantine species. This may be explained in part by the fact that lists worldwide (Clarke et al. 2005). In 2003 when B. under interspeciÞc interaction, individuals of each invadens was introduced, it caused 18% of fruit infes- species exert a higher per capita competitive intensity tation on mangoes at Nguruman. However, 4 yr later, upon each other by taking up more resources to meet the invader has become the predominant species, con- their higher metabolic needs resulting in faster devel- stituting Ͼ80% of fruit ßy population infesting mango opment and shorter duration. Another explanation fruit. could be that the larvae were responding to compet- Although fruit ßies commonly invade new zones itive stress by developing sooner because of less food (Fletcher 1986), relatively few cases of competitive availability reaching the next developmental stage in displacement are documented. Perhaps, the most no- small sizes. This was clearly observed in C. cosyra with table examples include displacement of C. capitata by lower pupal weights when exposed to competition the Queensland fruit ßy, Bactrocera tryoni (Froggatt), with B. invadens. around the Sydney area in Australia (Debach 1966) Overall, our laboratory study demonstrates that B. and displacement of the same species by Bactrocera invadens was a superior competitor to C. cosyra. In- dorsalis (Hendel) from the coastal areas in Hawaii in deed, the success of many invasive species is believed 1945 (Duyck and Quilici 2002). In the latter case, the to result primarily from their superior competitive displacement is to some extent mediated by host fruit abilities relative to native species (Williamson 1996, in that C. capitata persists in lowlands on coffee (Cof- Bruno et al. 2005). In a series of tephritid invasions on fea arabica L.), their presumed ancestral host in Africa La Re´union, Duyck et al. (2006) demonstrated that to which it is better adapted (Vargas et al. 1995). In the the invasive species B. zonata, tended to have higher Mascarene Islands, the indigenous Ceratitis catoirii ranks than the previously established invasive (C. rosa Guerin-Meneville seems to have been displaced by C. and C. capitata) and native (C. catoirii) species in the capitata and C. rosa in La Re´union occurring in small hierarchy. In their study, B. zonata, which was the numbers on the east and south coast of the island most recently established species was dominant in (Duyck et al. 2004, 2006), whereas in Mauritius it both forms of competition (scramble and interfer- seems to have disappeared (White et al. 2000). Pre- ence), which the authors attributed to its large body sumably, the invasion of Bactrocera zonata (Saunders) size and shorter developmental period. In a similar in Mauritius in 1987 and La Re´union in 1991 may have laboratory scrambled competition experiment, B. dor- compounded the displacement of the indigenous spe- salis was observed to out compete C. capitata and cies (Duyck et al. 2004, 2006). In our study, the speed inhibit its development (Keiser et al. 1974). Although with which B. invadens displaced C. cosyra and its not many studies have addressed competitive inter- efÞcient utilization of mango fruit suggests that there action between tephritids of different genera, our re- has been a prolonged adaptation by B. invadens to sults agree with those of Duyck et al. (2006) and Keiser using mango as host fruit. In its aboriginal home of Sri et al. (1974) that Bactrocera species tend to be com- Lanka, the pest has been reared from mango at low petitively superior to Ceratitis species. 988 JOURNAL OF ECONOMIC ENTOMOLOGY Vol. 102, no. 3

The precise interactions that take place inside the drive off C. cosyra from the mango dome. In his fruit is not obvious and this was not quantiÞed. All study on B. tryoni, Pritchard (1969) noted that insects were infested on the same day and initial age nearly 20% of females observed ovipositing in the differences among the cohort cannot be responsible wild interrupted egg laying to drive off conspeciÞc for the pronounced asymmetry. However, asymmetric females through threat displays that occasionally competition may also arise within a cohort of equally escalated to head-butting and pushing. Male B. in- aged individuals of the different species if there is vadens were also highly aggressive toward both variation in growth rate so that individuals become sexes of C. cosyra. In Bactrocera species, male ßies larger than the others (Begon 1984). Indeed, the larg- have been reported to respond aggressively in de- est individuals may be Þtter than the rest and, because fense of sites containing resources vital to female of large size, gain asymmetric advantage in competi- ßies, thus increasing their mating opportunities tion. Larval development of B. invadens was shorter (Fletcher 1987). compared with C. cosyra, both in the controls and Biological invasions are increasingly being viewed under competitive interaction. Duyck et al. (2006) as one of the components of global change and several noted that shorter larval developmental time of B. ecological factors interact that favor introduced spe- zonata compared with C. catoirii, C. capitata, and C. cies (Vitousek et al. 1996, Shea and Chesson 2002). In rosa conferred superior competitive ability on B. addition to the evidence provided above, there are zonata than the Ceratitis species. Because B. invadens possibly several other reasons why B. invadens has develop faster, the cohorts in the fruit were probably been able to adapt so quickly in the introduced range larger and older than C. cosyra even when they were and able to displace C. cosyra, and this warrants ap- infested on the same day. When two groups of differ- praisal. First, the high reproductive capability of B. ently sized and aged juvenile insects are reared to- invadens allows it to achieve enormous population gether, the smaller and younger cohort suffers from size. When one has greater realized fecun- increased mortality and reduced size (Averill and dity than a competitor, that competitor will be dis- Prokopy 1987, Edgerly and Livdahl 1992, Cameron et placed (Reitz and Trumble 2002), and this mechanism al. 2007). Another crucial factor in the case of fruit ßies is especially enhanced when more females are re- may be resource degradation arising from variation in cruited into the population. A female B. invadens has nutritional quality inside the mango fruit and it is likely a fecundity rate of 1056.8 eggs as against 356 eggs in C. that more of the lower quality resources are consumed cosyra (Ekesi et al. 2006; S.E. et al., unpublished data). by the inferior competitor. Over time, the numerical advantage of B. invadens may Interference competition through behavioral ag- reduce the probability of C. cosyra individuals having gression of B. invadens to C. cosyra may also have access to available resources. Differences in repro- given the invader a competitive advantage over ductive capability alone have been attributed to rapid the resident. In our experiments, both male and displacement in many insects (Brown et al. 1995, Reitz female B. invadens responded aggressively to the and Trumble 2002). Second, B. invadens is a very presence of C. cosyra by lunging at the opponent mobile insect and arriving at a resource Þrst through interrupting the process of landing on a protein food their high mobility and dispersive power (S.E., un- source or ovipositing on the mango dome, and the published data) probably confers competitive supe- level of female aggression in B. invadens was even riority over C. cosyra. In fruit ßies and other insect more pronounced than in males. In female Bactro- species, early arrival can result either from early sea- cera species, host marking pheromones are appar- sonal emergence or rapid colonization (Denno et al. ently absent (Fletcher and Prokopy 1991), as a re- 1995; Vargas et al. 1995). B. invadens has an earlier sult, female aggression especially in defense of seasonal phenology than C. cosyra based on trap oviposition sites is frequently intense even among catches in McPhail traps baited with NuLure (S.E. et conspeciÞcs. Shelly (1999) reported that females of al., unpublished data). When coupled with the rapid B. dorsalis defended oviposition sites on mango population growth and/or the preemption of re- against conspeciÞc females by lunging at opponents sources, this factor can dictate competitive outcome in and chasing them off. If such aggression can occur fruit ßies (Fitt 1984). Third, like many invading or- among conspeciÞcs, then it is perhaps not surprising ganisms in novel environments, B. invadens is released that both sexes of B. invadens launched several ag- from the harmful effects of their coevolved natural gressive behaviors against C. cosyra in our study. enemies. For example, there is evidence that host According to Shelly (1999), territoriality in female speciÞc parasitoids may inßuence competitive out- Bactrocera is presumably related to effect on larval comes between fruit ßy species (Clarke et al. 2005). So competition. By defending oviposition sites (even far, no parasitoids have been observed attacking B. temporarily), Shelly argued that females may pro- invadens in Kenya from host fruit samples collected vide their larvae with a “head start” in growth over from March 2003 to date (Rwomushana et al. 2008; S.E. related larvae and hence a competitive advantage et al., unpublished data). Because the parasitoids that for host fruit resources. Ostensibly, a similar thing beset B. invadens in their aboriginal home are absent occurs in B. invadens in defense of its offspring in Kenya, their population may have attained higher against other species of fruit ßies, but this warrants levels than they do in their native home. These fac- further studies. In few a cases, we did observe that torsÑrapid growth rate, male and female territorial- female B. invadens even interrupted egg laying to ity/aggressive interaction, greater reproductive po- June 2009 EKESI ET AL.: COMPETITIVE DISPLACEMENT AMONG FRUIT FLIES 989 tential, early arrival and lack of natural enemiesÑ may remain a permanent part of the mango agro- probably combine with one another allowing ecosystem. competitive displacement of C. cosyra at Nguruman. C. cosyra however has not been completely dis- placed in mango orchards at Nguruman, and the rea- Acknowledgments son for this also warrants assessment. There are prob- We greatly appreciate the efforts by P. Agola and E. F. ably some advantages that the insect has that allows for Malto for various contributions to the Þeld collection of data, some level of coexistence with B. invadens. One ad- and we thank M. Wanyonyi for aid with laboratory experi- vantage may be its more specialized host-searching ments. The research activity was supported by various grants abilities on mangoes having been linked more closely from the International Fund for Agricultural Development to this host plant over a long period in Africa. Second, (IFAD) and the German Ministry for Economic Cooperation C. cosyra has been recorded from just nine plants and Development (BMZ). species in Kenya (Lux et al. 2003a, Copeland et al. 2006) compared with the increasing host range of B. References Cited invadens that currently stands at Ͼ14 in Kenya (Ekesi and Billah 2006, Rwomushana et al. 2008). It is there- Averill, A. L., and R. J. Prokopy. 1987. IntraspeciÞc compe- fore likely that B. invadens can switch to other suitable tition in the tephritid fruit ßy Rhagoletis pomonella. Ecol- hosts when there is a bottleneck in carrying capacity, ogy 68: 878Ð886. Begon, M. 1984. Density and individual Þtness: asymmetric providing some niche on mango for C. cosyra to sur- competition, pp. 175Ð194. In B. Shorocks [ed.], Evolu- vive. For example, Nguruman is a lowland ecology tionary ecology. Blackwell ScientiÞc Publications, Ox- with an altitude of 700 m above sea level. Generally, ford, United Kingdom. most Bactrocera species, including B. invadens are be- Brown, M. W., H. W. Hogmire, and J. J. Schmitt. 1995. lieved to be lowland residents (Vargas et al. 1983, Competitive displacement of apple aphid by spirea aphid Wong et al. 1985, Harris et al. 1986; Ekesi et al. 2006), (Homoptera: Aphididae) on apple as mediated by human enabling B. invadens to displace C. cosyra in lowland activities. Environ. Entomol. 24: 1581Ð1591. ecologies. At the higher elevation areas of Kenya, such Bruno, J. F., J. D. Fridley, K. D. Bromberg, and M. D. Bertness. 2005. Insights into biotic interactions from as Embu in the Eastern Province, C. cosyra remains the studies of species invasions, pp. 9Ð40. In D. F. Sax, J. J. dominant species, probably because of poor toler- Stachowicz, and S. D. Gaines [eds.], Species invasions: ance of B. invadens to low temperatures (Ekesi et al. insights into ecology, evolution and biogeography. Si- 2006; I. Rwomushana, unpublished data). It is there- nauer, Sunderland, MA. fore probable that B. invadens may be restricting Cameron, T. C., H. J. Wearing, P. Rohani, and S. M. Sait. populations of C. cosyra to the highlands. Indeed 2007. Two species asymmetric competition: effects of such phenomena have been reported from Hawaii, age structure on intra- and interspeciÞc interactions. J. where B. dorsalis largely displaced C. capitata from Anim. Ecol. 76: 83Ð93. Copeland, R. S., R. A. Wharton, Q. Luke, and M. De Meyer. the low-elevation coastal zones and restricted C. 2002. Indigenous hosts of Ceratitis capitata (Diptera: Te- capitata populations to cooler climates at high alti- phritidae) in Kenya. Ann. Entomol. Soc. Am. 95: 672Ð694. tudes where B. dorsalis does not occur (Vargas et al. Copeland, R. S., R. A. Wharton, Q. Luke, M. De Meyer, S. 1995). Lux, N. Zenz, P. Machera, and M. Okumu. 2006. Geo- The results of this study have shown that compet- graphic distribution, host fruit, and parasitoids of African itive displacement of C. cosyra by B. invadens has fruit ßy pests Ceratitis anonae, Ceratitis cosyra, Ceratitis occurred at Nguruman, a low-elevation region of Ke- fasciventris and Ceratitis rosa (Diptera: Tephritidae) in Kenya. Ann. Entomol. Soc. Am. 99: 261Ð278. nya. Generally, the species composition of a habitat is Cowley, J. M., R. T. Baker, and D. S. Harte. 1992. DeÞnition dynamic with species appearing and going extinct and determination of host status for multivoltine fruit ßy over time (Southwood 1977), making competitive dis- (Diptera: Tephritidae) species. J. Econ. Entomol. 85: 312Ð placement difÞcult to prove. Changes are often very 317. rapid occurring unnoticed or if noticed, it is after the Clarke, A. R., K. F. Armstrong, A. E. Carmichael, J. R. Milne, changes have occurred making it difÞcult to examine S. Raghu, G. K. Roderick, and D. K. Yeates. 2005. Inva- the causes. Several factors may be involved in the sive phytophagous pests arising through a recent evolu- displacement of C. cosyra by B. invadens on mango. tionary radiation: the Bactrocera dorsalis complex of fruit ßies. Annu. Rev. Entomol. 50: 293Ð319. But our results suggest that the most probable mech- Debach, P. 1966. Competitive displacement and coexist- anisms of displacement include exploitative com- ence principles. Annu. Rev. Entomol. 11: 183Ð212. petition through larval scrambling for resources Denno, R. F., M. S. McClure, and J. R. Ott. 1995. Interspe- within mango fruit and interference competition ciÞc interactions in phytophagous insectsÑcompetition through aggressive behaviors of the invader. What- re-examined and resurrected. Annu. Rev. Entomol. 40: ever the mechanisms involved, these Þndings dem- 297Ð331. onstrate that the role of C. cosyra as the predomi- Draper, N. R., and H. Smith. 1981. Applied regression anal- nant pest of mango at Nguruman has been replaced ysis, 2nd ed. Wiley, New York. Drew, R.A.I., K. Tsuruta, and I. M. White. 2005. A new by the new invasive mango fruit ßy species, B. in- species of pest fruit ßy (Diptera: Tephritidae: ) vadens. However, because of the systems variability from Sri Lanka and Africa. Afr. Entomol. 13:149Ð154. (Ebenhoh 1994), it is unlikely that the insect will be Duyck, P. F., and S. Quilici. 2002. Survival and development completely replaced by the invader and C. cosyra of different life stages of three Ceratitis spp. (Diptera: 990 JOURNAL OF ECONOMIC ENTOMOLOGY Vol. 102, no. 3

Tephritidae) reared at Þve constant temperatures. Bull. ment, pp. 277Ð293. In P. Neuenschwander, C. Borgemeis- Entomol. Res. 92: 461Ð469. ter, and J. Langewald [eds.], Biological control in inte- Duyck, P-F., P. David, and S. Quilici. 2004. A review of grated pest management systems in Africa. CABI, relationships between interspeciÞc competition and in- Wallingford, United Kingdom. vasions in fruit ßies (Diptera: Tephritidae). Ecol. Ento- Lux, S. A., R. S. Copeland, I. M. White, A. Manrakhan, and mol. 29: 511Ð520. M. K. Billah. 2003b. A new invasive fruit ßy species from Duyck, P.-F., P. David, G. Junod, C. Brunel, R. Dupont, and the Bactrocera dorsalis (Hendel) group detected in East S. Quilici. 2006. Importance of competition mechanisms Africa. Insect Sci. Appl. 23: 355Ð361. in successive invasions by polyphagous tephritids in La Mooney, H. A., and E. E. Cleland. 2001. The evolutionary Re´union. Ecology 87: 1770Ð1780. impact of invasive species. Proc. Natl. Acad. Sci. U.S.A. 98: Ebenhoh, W. 1994. Competition and coexistence: model- 5446Ð5451. ling approaches. Ecol. Model. 75/. 76: 83Ð98. Mooney, H. A., and J. A. Drake. 1986. Ecology of biological Edgerly, J. S., and T. P. Livdahl. 1992. Density-dependent invasions of North America and Hawaii. Springer, New interactions with a complex life cycle: the role of cohort York. structure and mode of recruitment. J. Anim. Ecol. 61: Mwatawala, M. W., I. M. White, A. P. Maerere, F. J. Sen- 139Ð150. kondo, and M. De Meyer. 2004. A new invasive Bactro- Ekesi, S., P. W. Nderitu, and I. Rwomushana. 2006. Field cera species (Diptera: Tephritidae) in Tanzania. Afr. En- infestation, life history and demographic parameters of tomol. 12: 154Ð156. Bactrocera invadens Drew, Tsuruta & White, a new in- Naeem, S., L. J. Thompson, S. P. Lawler, J. H. Lawton, and vasive fruit ßy species in Africa. Bull. Entomol. Res. 96: R. M. Woodfin. 1995. Empirical evidence that declining 379Ð386. species diversity may alter the performance of terrestrial Ekesi, S., and M. K. Billah. 2006. A Field guide to the man- ecosystems. Philos. Trans. R. Soc. Lond. 347: 249Ð262. agement of economically important tephritid fruit ßies in Pritchard, G. 1969. The ecology of natural population of Africa. ICIPE Science Press, Nairobi, Kenya. Queensland fruit ßy Dacus tryoni II. The distribution of Fitt, G. P. 1984. Oviposition behaviour of two tephritid fruit eggs and its relation to behaviour. Aust. J. Zool. 17: 293Ð ßies as inßuenced by the presence of larvae in the host 311. fruit. Oecologia (Berl.) 62: 37Ð46. Reitz, S. R., and J. T. Trumble. 2002. Competitive displace- Fitt, G. P. 1986. The role of adult and larval specialisations ment among insects and arachnids. Annu. Rev. Entomol. in limiting the occurrence of Þve species of Dacus 47: 435Ð465. (Diptera, Tephritidae) in cultivated fruits. Oecologia Rwomushana, I., S. Ekesi, I. Gordon, and C.K.P.O. Ogol. (Berl.) 69: 101Ð109. 2008. Host plants and host plant preference studies for Fletcher, B. S. 1986. Some issues in fruit ßy ecology, pp. Bactrocera invadens (Diptera: Tephritidae) in Kenya, a 117Ð133. In M. Mangel, J. R. Carey, and R. Plant [eds.], new invasive fruit ßy species in Africa. Ann. Entomol. Soc. Pest control: operations and system analysis in fruit ßy Am. 101: 331Ð340. management. Springer, Berlin, Germany. SAS Institute. 1989. SAS/STAT userÕs guide, version 6, 4th Fletcher, B. S. 1987. The biology of dacine fruit ßies. Annu. ed., vol. 2. SAS Institute. Carey, NC. Rev. Entomol. 32: 115Ð144. Scheiner, S. M., and J. Gurevitch. 1993. The design and Fletcher, B. S., and R. J. Prokopy. 1991. Host location and analysis of ecological experiments. Chapman & Hall, New oviposition in tephritid fruit ßies, pp. 139Ð171. In W. J York. Bailey and J. Ridsdill-Smith [eds.], Reproductive behav- Shea, K., and P. Chesson. 2002. Community ecology theory iour of Insects. Chapman & Hall, London, United King- as a framework for biological invasions. Trends Ecol. Evol. dom. 17: 170Ð176. French, C. 2005. The new invasive Bactrocera species, pp. Shelly, T. E. 1999. Defense of oviposition sites by female 19Ð20. In Insect Pest Control Newsletter, No. 65. Inter- oriental fruit ßies (Diptera: Tephritidae). Fla. Entomol. national Atomic Energy Agency, Vienna, Austria. 82: 339Ð346. Harris, E. J., J. M. Takara, and T. Nishida. 1986. Distribution Siegel, S., and J. C. Castellan Jr. 1988. Nonparametric sta- of the melon ßy Dacus cucurbitae (Diptera: Tephritidae) tistics for the behavioural sciences, 2nd ed. Chapman & and host plants on Kauai, Hawaiian Islands. Environ. Hall, New York. Entomol. 15: 488Ð493. Southwood, T.R.E. 1977. Habitat, the template for ecolog- Jenkins, P. T. 1996. Free trade and exotic species introduc- ical strategies. J. Anim. Ecol. 46: 337Ð365. tions. Conserv. Biol. 10: 300Ð302. Vargas, R. I., T. Nishida, and J. W Beardsley. 1983. Distri- Keiser, I., R. M. Kobayashi, D. H. Miyashita, E. J Harris, and bution and abundance of Dacus dorsalis (Diptera: Te- E. L. Schneider. 1974. Suppression of Mediterranean phritidae) in native and exotic forest areas of Kauai. fruit ßies by oriental fruit ßies in mixed infestations in Environ. Entomol. 12: 1185Ð1189. guava. J. Econ. Entomol. 67: 355Ð360. Vargas, R. I., W. Walsh, and T. Nishida. 1995. Colonization Kfir, R. 1997. Competitive displacement of Busseola fusca of newly planted coffee Þelds: dominance of Mediterra- (Lepidoptera: Noctuidae) by Chilo partellus (Lepidop- nean fruit ßy over oriental fruit ßy. J. Econ. Entomol. 88: tera: Pyralidae). Ann. Entomol. Soc. Am. 90: 619Ð624. 620Ð627. Krainacker, D. A., J. R. Carey, and R. I. Vargas. 1987. Effect D’Antonio, C. M., and P. M. Vitousek. 1992. Biological in- of larval host on life-history traits of the Mediterranean vasions by exotic grasses, the grass/Þre cycle, and global fruit ßy Ceratitis capitata. Oecologia (Berl.) 73: 583Ð590. change. Annu. Rev. Ecol. Syst. 23: 63Ð87. Lodge, D. M. 1993. Species invasions and deletions, pp. 367Ð Vitousek, P. M., C. M. D’Antonio, L. L Loope, and R. West- 387. In P. M. Kareiva, J. G. Kingsolver, and R. B. Huey brook. 1996. Biological invasions as global environmen- [eds.], Biotic interactions and global change. Sinauer, tal change. Am. Sci. 84: 468Ð478. Sunderland, MA. White, I. M., M. De Meyer, and J. Stonehouse. 2000. A Lux, S. A., S. Ekesi, S. Dimbi, S. Mohamed, and M. K. Billah. review of native and introduced fruit ßies (Diptera: Te- 2003a. Mango infesting fruit ßies in AfricaÑperspectives phritidae) in the Indian Ocean Islands of Mauritius, Re- and limitations of biological approaches to their manage- union, Rodrigues and Seychelles, pp. 15Ð21. In Proceed- June 2009 EKESI ET AL.: COMPETITIVE DISPLACEMENT AMONG FRUIT FLIES 991

ings of the Indian Ocean Commission Regional Fruit Fly fruit ßies (Diptera: Tephritidae) in Kula, Maui, Hawaii. J. Symposium, 5Ð9 June 2000. IPC, Mauritius. Econ. Entomol. 78: 1267Ð1271. Williamson, M. 1996. Biological invasions. Chapman & Yoshimura, J., and C. W. Clark. 1991. Adaptation in Stoi- Hall, London, United Kingdom. chastic Environment. Spinger-Verlag, Berlin, Germany. Wong, T.M.Y., D. O. McInnis, and N. Mochizuki. 1985. Sea- sonal distribution and abundance of adult male oriental Received 15 August 2008; accepted 25 February 2009.