Sterile Insect Technique and Control of Tephritid Fruit Flies: Do Species With Complex Courtship Require Higher Overflooding Ratios?

Forum Annals of the Entomological Society of America, 109(1), 2016, 1 11 doi: 10.1093/aesa/sav101 Advance Access Publication Date: 28 October 2015 Forum Sterile Insect Technique and Control of Tephritid Fruit Flies: Do Species With Complex Courtship Require Higher Overflooding Ratios? Todd Shelly 1,2 and Donald McInnis 3 1 USDA-APHIS, 41-650 Ahiki St., Waimanalo, HI 96795 (todd.e.shelly@aphis.usda.gov), 2 Corresponding author, e-mail: todd.e.shel ly@aphis.usda.gov, 3 USDA-ARS (retired), 1216 Manu Aloha Street, Kailua, HI 96734 (moscamed@aol.com) Received 9 August 2015; Accepted 7 October 2015 Abstract The sterile insect technique (SIT) is widely used to suppress or eradicate tephritid fruit fly pests that threaten agricultural crops. The SIT entails the production, sterilization, and release of large numbers of the target pest species, with the goal of achieving sterile male by wild female matings, which result in infertile eggs and the subsequent reduction of the wild pest population. The success of this control strategy depends, to a large extent, on the ability of released, sterile males to compete successfully against wild males to achieve copulations with wild females. Lance and McInnis (2005) proposed that species with greater male involvement in courtship and mating are less amenable to the SIT than species with simple courtship, because strong artificial selection characteristic of mass-rearing environments is more likely to generate greater behavioral modification to complex than simple courtship and thus to result in increased rejection by wild females. Consistent with Lance and McInnis (2005), the mating competitiveness of mass-reared males of the Mediterranean fruit fly, Ceratitis capitata (Wiedemann), a species with complex male courtship, is substantially lower than that of mass-reared males of the melon fly, Bactrocera cucurbitae (Coquillett), or the oriental fruit fly, Bactrocera dorsalis (Hendel), two species with very simple male courtship. In light of this difference in mating ability, we tested the prediction that higher overflooding ratios (sterile:wild males; OFR) would be required to control populations of C. capitata than of the Bactrocera spp. Levels of induced egg sterility achieved under different OFRs were compared to the net sterility target, computed as the proportional decrease in a species intrinsic birth rate (realized via SIT) required to balance the intrinsic rate of death and found to be 80% for the three species considered. Although the data are scant, they generally support the prediction. For B. cucurbitae, OFRs < 50:1 resulted in egg sterility levels near 80%, whereas in C. capitata similarly high sterility values were achieved only with OFRs >300:1. Likewise, relative to populations in control (untreated) sites, populations in release areas were reduced >99% at OFRs of 50:1 100:1 for B. cucurbitae in Japan compared to reductions for C. capitata of 50 93% at OFRs between 700:1 3,160:1 in Nicaragua and about 80% for OFRs between 100:1 400:1 in Hawaii. The implications of these findings for fruit fly SIT are discussed. Key words: sterile insect technique, overflooding ratio, melon fly, Mediterranean fruit fly Courtship and the Effectiveness of the Sterile Insect Technique True fruit flies (Diptera: Tephritidae) are agricultural pests that attack many commercially important fruits and vegetables, particularly in tropical and subtropical regions (White and Elson-Harris 1992). A variety of management tools have been developed to suppress or eradicate these pests, including male annihilation, protein bait sprays, mass trapping, and the sterile insect technique (SIT; Mau et al. 2007, Jessup et al. 2007). The latter strategy entails production of large numbers of the target pest species, sterilization via irradiation, and release of the sterile insects into the environment. The objective is to achieve sterile male by wild female matings, which result in the deposition of infertile eggs and the eventual reduction of the wild pest population. Because its autocidal action avoids insecticide use, the SIT is considered an environmentally benign approach, which has led to its increased implementation worldwide (Enkerlin 2005). The success of the SIT depends, to a large extent, on the ability of released, sterile males to compete successfully against wild males to achieve copulations with wild females (Calkins 1984). Recognizing this basic requirement, Lance and McInnis (2005) proposed that species with greater male involvement in courtship and Published by Oxford University Press on behalf of Entomological Society of America 2015. This work is written by US Government employees and is in the public domain in the US. 1

2 Annals of the Entomological Society of America, 2016, Vol. 109, No. 1 mating would be less amenable to the SIT than species with no, or very simple, courtship, where matings follow more directly from the mere incidence of intersexual encounters. This argument follows directly from the notion that artificial selection occurring in massrearing environments, and deriving largely from the high adult holding densities, is more likely to generate greater behavioral modification to complex than simple courtship and thus to result in increased rejection of mass-reared sterile males by wild females. Mating Competitiveness of Mass-Reared Males Among economically important tephritid species, the SIT has been used most widely against the Mediterranean fruit fly (medfly) Ceratitis capitata (Wiedemann) (Enkerlin 2005), a species with very complex male courtship behavior (Féron 1962; Arita and Kaneshiro 1986, 1989). In the medfly, receptive females seek perching males, who signal their location via emission of pheromones, intense wing-fanning, which may serve to disperse the pheromone, and production of audible sounds (a consequence of the rapid wing movements). Upon detection of an approaching female, the male produces a suite of close-range signals, including rapid, forwardbackward wing flicks, continued pheromone emission but with the abdomen curved downward and forward, and oscillatory head movements, all of which are performed while directly facing the female. Consistent with Lance and McInnis (2005), mating studies of the medfly conducted worldwide generally show that mass-reared, sterile males are inferior to wild males in procuring copulations with wild females (Table 1). In these studies, most of which were conducted in field cages (typically 3 m in diameter and 2.5 m in height) placed over host plants using equal numbers of mass-reared males, wild males, and wild females, the proportion of total matings obtained by mass-reared males (or the Relative Sterility Index, RSI; McInnis et al. 1996) was usually below 40% and averaged 31% across all studies. Although this pattern is widely recognized (Hendrichs et al. 2002), its exact cause remains unknown. However, several studies (Liimatainen et al. 1997; Briceño and Eberhard 1998, 2000; Lux et al. 2002) have documented differences in male courtship between wild and mass-reared male medflies. Importantly, owing to frequent interruptions by other flies under crowded holding conditions, mass-reared male medflies have evolved abbreviated premount courtship, which results in frequent rejection by wild females (Briceño and Eberhard 2000, Shelly 2012). The SIT is also used, albeit to a lesser extent, against agriculturally important species of Bactrocera, particularly the oriental fruit fly Bactrocera dorsalis (Hendel) and the melon fly Bactrocera cucurbitae (Coquillett). As in the medfly, receptive females of these species seek perching males, who signal their location via emission of pheromones, intense wing-fanning, and accompanying sound production (Fletcher 1987). However, unlike the medfly, when an approaching female is detected, males of these three species cease wing-fanning and attempt to mount the female. Mounting is, more or less, immediate in B. dorsalis (Kobayashi et al. 1978, Arakaki et al. 1984) and B. cucurbitae (Suzuki and Koyama 1980, Kuba et al. 1984), although wing vibrations produced by just-mounted males may provide the female with information on male quality (Arakaki et al. 1984). Although fewer data are available, mass-rearing may alter calling and courtship signaling of Bactrocera males. Kanmiya et al. (1987) identified subtle differences in the temporal patterns of wing-fanning between wild and mass-reared B. cucurbitae males, with wing-fanning bouts being generally of shorter duration but greater periodicity in wild than mass-reared flies. Although fewer data are available, existing studies indicate that mass-reared males of the oriental fruit fly and melon fly exhibit mating frequencies equivalent to their wild counterparts in Table 1. Results of mating competitiveness trials for the Mediterranean fruit fly, C. capitata, in which mass-reared males competed with wild males (or wild-like males from recently established laboratory colony) for copulations with wild (or wild-like) females Mass-reared strain Source of wild flies RSI (%) Reference Name Type a Metapa Bisexual Guatemala 38 Zapién et al. 1983 T101 GSS-pc Italy 23 Robinson et al. 1986 Vienna-42 GSS-tsl Greece 39 Hendrichs et al. 1996 Hi-Lab Bisexual Hawaii 38 b McInnis et al. 1996 SEIB 6-96 GSS-pc Argentina 26 Cayol et al. 1999 Hi-Lab Bisexual Hawaii 32 Shelly 2001a Maui-Med Bisexual Hawaii 26 Shelly and McInnis 2001 Vienna-4/Tol-94 GSS-tsl Israel 48 Taylor et al. 2001 T (Y;5) 1-61 GSS-pc Greece 34 Economopoulos and Mavrikakis 2002 Vienna-7/mix-2000 GSS-tsl Guatemala 17 Shelly et al. 2002 Vienna-7/mix-2000 GSS-tsl Portugal 17 Shelly et al. 2002 Vienna-7/Tol-99 GSS-tsl Hawaii 25 Shelly et al. 2004 Vienna-7-97 GSS-tsl South Africa 30 Barnes et al. 2004 Vienna-7/mix-2000 GSS-tsl Portugal 23 Pereira et al. 2007 Vienna-7/mix-99 GSS-tsl Western Australia 27 Steiner et al. 2013 Vienna-8 GSS-tsl Spain 40 Juan-Blasco et al. 2013 Vienna-8 GSS-tsl Brazil 39 Paranhos et al. 2013 RSI, the relative sterility index, is the percentage of the total matings with wild females achieved by males of the mass-reared strain. In some cases, RSIs represent averages computed over several experiments within a study. In all studies, mass-reared flies were sterilized, except Shelly (2001a); mass-reared males were not exposed to aromatic mating stimulants; mating tests were conducted in field cages, except Juan-Blasco et al. (2013) and Paranhos et al. (2013), which were performed in laboratory cages; and equal numbers of wild males, wild females, and mass-reared males were used, except Zapien et al. (1983), Robinson et al. (1986), and McInnis et al. (1996), which in addition released an equal number of mass-reared females. a Bisexual standard; GSS-pc genetic sexing strain, pupal color; GSS-tsl genetic sexing strain, temperature sensitive lethal. b Data for nonresistant females.

Annals of the Entomological Society of America, 2016, Vol. 109, No. 1 3 competitiveness trials (Table 2). Average RSI values computed for lab and field cages/field cage studies only were: 56/47% for B. cucurbitae, 53/51% for B. dorsalis (including B. philippinensis), and 53/ 48% over all Bactrocera species included in Table 1. Thus, while mass-rearing may modify sexual signaling by Bactrocera males to some degree (as noted above), the changes appear to have only a small effect on their mating competitiveness. Although a different experimental design was used, Pérez-Staples et al. (2009) reported a similar finding for the Queensland fruit fly, Bactrocera tryoni (Froggatt): sterile males, when fed a full diet (with protein), obtained approximately the same number of matings with wild females as did wild males. Overflooding Ratios and the Effectiveness of SIT With medfly SIT in particular, management response to the competitive inferiority of mass-reared males has been to increase the production and release numbers of sterile males (Calkins and Parker 2005). Simply increasing the ratio of sterile:wild males (the so-called overflooding ratio, hereafter OFR) will, of course, not be effective if released males, regardless of their abundance, fail to meet some minimum threshold of acceptable courtship performance required by most or all wild females (Itô and Yamamura 2005). However, if choosiness varies inversely with the encounter rate with acceptable mates as shown for other insects (Cratsley and Lewis 2005, Lehmann 2007), then wild medfly females that encounter a series of released males may eventually mate with a released male to insure some reproduction prior to death. If this scenario is true, increasing the OFR in an SIT program may lead to more effective population suppression. If this, in turn, is true, then the OFR required to attain effective control would likely be much higher for C. capitata than for B. cucurbitae or B. dorsalis, as, as noted above, mass-reared medfly males appear far less competitive relative to wild males than is the case for these Bactrocera species. The objective of the present report was to assess the relationship between OFR and the effectiveness of SIT in suppressing populations of C. capitata, B. cucurbitae, and B. dorsalis. We focused on these species, because field data were available that related OFR to 1) induced egg sterility in wild females and, in a small subset of the studies, 2) trap captures of wild flies in SIT-treated and untreated (control) areas. Pilot or demonstration studies of the SIT conducted in small, nonisolated areas (e.g., McInnis et al. 1994) were excluded, because immigration likely had a nontrivial influence on the results. Likewise, studies that estimated OFRs and the size of the target population in the release area but not simultaneously in an untreated area (e.g., Cheikh et al. 1975) were also excluded. However, we included several studies that experimentally assessed the impact of OFR on egg sterility using large field enclosures. The present paper includes no data on Anastrepha spp., because 1) relevant data from the open field are not available, and 2) recent data (Flores et al. 2014) for the Mexican fruit fly A. ludens (Loew) that related OFR to induced egg sterility in large field cages were obtained using flies from a mass-reared strain exclusively. The Data The methods used in the relevant studies are summarized briefly below, and the original papers should be consulted for more thorough accounts. Mediterranean Fruit Fly, C. capitata Data were available from three field studies and two studies conducted using large field cages. Rhode et al. (1971) aerially released sterile medflies from a mass-reared, bisexual strain over a 48 km 2 area of coffee (Coffea arabica L.) in Nicaragua from 1967 1969. However, the data were most complete for September 1968 May 1969, and our analysis is limited to this period. In addition to the release area, a control area (termed Check 1) that received no sterile flies was established. Before and after start of the releases, aerial spraying of a protein bait-malathion mixture was performed over both release and nonrelease areas. Concurrent spraying confounds assessment of the SIT s Table 2. Results of mating competitiveness trials for species of Bactrocera in which mass-reared males competed with wild males (or wildlike males from recently established laboratory colony) for copulations with wild females Species Mass-reared strain Source of wild flies Testing cage Type/sex ratio a RSI (%) Reference B. correcta Pathumthani Thailand Field 1:1:1:0 43 b Orankanok et al. (2013) B. cucurbitae USDA-ARS Hawaii Lab 1:1:1:0 86 Shelly and Villalobos (1995) Okinawa Taiwan Lab 1:1:1:1 44 Matsuyama and Kuba (2004) USDA-ARS Hawaii Field 1:1:1:1 37 McInnis et al. (2004) T1 Hawaii Field 1:1:1:0 56 McInnis et al. (2004) B. dorsalis USDA-ARS Hawaii Lab 1:1:1:0 46 Shelly (1995) DTWP c Hawaii Lab 1:1:1:0 64 Shelly et al. (2000) Pathumthani Thailand Field 1:1:1:0 38 d Orankanok et al. (2013) B. oleae Marathon Greece Lab 1:1:1:1 52 Economopoulos (1972) B. philippinensis e PNRI f Philippines Field 1:1:1:0 54 Shelly et al. (1996) PNRI Philippines Field 1:1:1:1 62 Obra and Resilva (2013) RSI, the relative sterility index, is the percentage of the total matings with wild females achieved by males of the mass-reared strain. In some cases, RSIs represent averages computed over several experiments within a study. In all studies, mass-reared flies were sterilized, except Matsuyama and Kuba (2004); mass-reared strains were bisexual (standard), except McInnis et al. (2004; T1 data) and Shelly et al. (2000), which used pupal color genetic sexing strains. a Wild #: Mass-reared #: Wild $: Mass-reared $. b Data for mass-reared males aged 10 19 d reared on protein diet without methyl eugenol. c Dorsalis Translocation White Pupae. d Data for mass-reared males aged 6 15 d reared on protein diet without methyl eugenol. e Recently synonymized with B. dorsalis (Schutze et al. 2015). f Philippines Nuclear Research Institute.

4 Annals of the Entomological Society of America, 2016, Vol. 109, No. 1 effectiveness, and here we assume that 1) its impact on wild fly populations was similar in the release and control areas and 2) its impact on sterile fly numbers was negligible in the release area. Hatch rate was determined using eggs dissected from coffee berries collected in the release and control areas. However, sample sizes were quite small for several months, and we included egg hatch data from the release area for only 3 mo (when n ¼ 17, 43, and 205 eggs, respectively) and used the overall average hatch rate of 93% for wild X wild matings (in Check 1) to compute net induced egg sterility (here and throughout, egg sterility in wild X wild crosses was subtracted from the egg sterility observed in release areas to obtain net percent egg sterility, i.e., that induced by sterile males over and above naturally occurring egg sterility). Rhode et al. (1971) also estimated the populations of wild males in the release and control areas using traps baited with trimedlure (a male attractant routinely used for monitoring medfly populations, Jang and Light 1996). Comparison of wild male captures between the release and control areas on a given date yielded an estimate of population reduction expressed as 100% X (1 wild flies per trap per day in treated site/ wild flies per trap per day in control site). Wong et al. (1986) compared egg sterility and the abundance of wild medflies in release (Kula) and control (nonrelease, Keokea) areas on Maui, Hawaii, from December 1981 June 1982. Sterile flies from a mass-reared, bisexual strain were released weekly from the air and ground over a 6-mo period. Male captures in traps baited with trimedlure were used to estimate OFRs in the release area and numbers of wild flies in both areas. Eggs were dissected from peaches (Prunus persica (L.)) collected in both Kula and Keokea, and hatch was measured in the laboratory. Establishment of release and control areas allowed estimation of net percent egg sterility (as described above) and degree of population reduction (both calculated as described above). In a large-scale, multiyear project, Rendon et al. (2004) compared the effectiveness of aerial releases of bisexual versus malesonly strains in inducing egg sterility in wild medfly populations in coffee fields in Guatemala. OFRs were measured using trimedlurebaited traps, and eggs were dissected from field-collected coffee berries; natural hatch rates were determined via egg collections from a nonrelease area. Releases spanned 1995 1997, but our analysis includes only results from 1997, which afforded the most complete data set. No trapping of wild males was conducted in nonrelease areas in any year, and consequently no estimates of SIT-mediated population suppression were available. In addition to these field studies, Shelly et al. (2005) examined the relationship between OFR and egg sterility using large field enclosures (16 by 6 by 2.5 m 3, l:w:h) placed over nonfruiting guava trees (Psidium guajava L.) on Oahu. Irradiated males from a temperature-sensitive lethal (tsl) genetic sexing strain were released at OFRs ranging from 5:1 to 60:1 (seven replicates were performed per OFR). Replicates involved sterile males that were exposed to ginger root oil (a substance known to enhance male mating success, Shelly 2001b) prior to release or were nonexposed, but only data for the nonexposed males were included in this report. Several days after the flies were released, apples were introduced into the enclosures as oviposition sites. Eggs were dissected from the apples, and hatching rates were then obtained. To calculate net induced sterility, we determined natural hatch rate by releasing wild flies in tents covering single guava trees, allowing mating, and then collecting eggs deposited in apples as in the large enclosures. In another study employing large field enclosures (7.5 by 6.4 by 2.4 m 3, l:w:h), Rendon et al. (2006) measured the effects of sterile male medflies, parasitoids, or these two in combination on reproduction of fertile females. Data presented here derive exclusively from the treatment involving sterile males only. Cages were placed over coffee plants at four different plantations of varying altitude in Guatemala. At each location, fertile flies (110 individuals per sex) were released in one set of (control) cages, and later all coffee berries within the cages were harvested. Berries were held in the laboratory to allow larval development, and resultant pupae and adults were counted (only data for pupae are included in our analysis). Concurrently, in another set of (SIT) cages at each location, sterile males were released along with the fertile flies (again 110 individuals per sex) at an OFR of 100:1 (i.e., 11,000 sterile males were released per cage), and coffee fruits were harvested and pupae counted as above. Four control cages and four SIT cages were established per location. SIT-mediated population reduction was estimated following the above method. Melon Fly, B. cucurbitae Two field studies included data relating OFRs to egg sterility and reduction of wild melon fly populations. To our knowledge, no studies involving large field cages have been conducted for this species. Iwahashi (1977) presented comprehensive data from the eradication project on Kume Island, Okinawa, which included extensive ground releases of sterile, mass-reared flies from a bisexual strain over a period of 2 yr. Prior to the releases, male annihilation (involving aerial distribution of strings soaked in the male attractant cuelure [Jang and Light 1996] and naled solution) and protein bait sprays were used to reduce the wild melon fly population to 5% of peak levels, but these methods were suspended with the start of the SIT. Iwahashi et al. (1977) provided data on male captures (males per trap per day in traps baited with cue-lure and naled) and percent egg hatch (based on laboratory oviposition of field-caught wild females) for both Kume Island (release area) and Naha Island, which received no sterile flies and thus served as a control area. Male capture data were used to compute OFRs on Kume Island, and the comparison of wild male captures between the release and control areas on a given date yielded an estimate of population reduction (as described above). Regarding egg sterility, for a given sampling date the egg hatch estimate for Naha was subtracted from that for Kume Island to yield net percent egg sterility. The second data set derived from McInnis et al. (2007), which was completed as part of the USDA/ARS Area-Wide IPM program against tephritid fruit flies in Hawaii (Vargas et al. 2008). McInnis et al. (2007) released sterile males from a pupal color sexing strain (McInnis et al. 2004) for 7 mo over 10 km 2 on the island of Maui. No other control measures were applied at this site during the period of sterile male releases. Monthly measurements were taken of OFRs (based on male captures per trap per day in cue-lure baited traps) and percent egg sterility (based on hatch rate of eggs dissected from field-collected host fruit). This study did not include a control area and thus did not describe the level of population suppression achieved via the SIT. However, an average egg hatch rate was obtained for concurrent wild by wild matings from a nonrelease area on the island of Hawaii, thus allowing computation of net percent egg sterility as described above. Oriental Fruit Fly, B. dorsalis Analysis included one study conducted in the field and one using large field tents. McInnis et al. (2011) conducted weekly ground releases of sterile flies from a males-only, pupal color genetic sexing strain (McCombs and Saul 1995) in two fruit orchards on Oahu, Hawaii, over a

Annals of the Entomological Society of America, 2016, Vol. 109, No. 1 5 period of 8 mo in 2006. Prior to release, males designated for one orchard were exposed to methyl eugenol (a male attractant known to enhance male mating success in this species, Tan and Nishida 1996), while males in the other area were nonexposed. To be consistent with field studies on the other species, we here included only data for the area that received nonexposed males. Captures of sterile and wild males in methyl eugenol-baited traps were monitored at weekly intervals in (n ¼ 31 sampling dates). Measurements of egg sterility were made every 4 7 wk (n ¼ 12 sampling dates) and were based on eggs dissected from a single host (pomelo, Citrus grandis (L.) Osbeck). Control egg hatch rates were measured for wild X wild crosses by collecting eggs from pomelos immediately prior to the sterile male releases. A designated control (nonrelease) area was ultimately considered unsuitable, owing to differing availability of host plants from the two release sites, consequently no data on population suppression were presented in this study. Also, for B. dorsalis, Shelly et al. (2010) measured egg sterility for wild oriental fruit flies using the same field enclosures and the same basic protocol as described above for the medfly. Irradiated males were derived from the same genetic sexing strain used by McInnis et al. (2011), and OFRs of 5:1, 10:1, 30:1, and 60:1 were tested in the field enclosures (seven replicates were performed per OFR). Replicates involved sterile males that were exposed to methyl eugenol prior to release or were nonexposed, but as above we here included only data for the nonexposed males. Several days after the flies were released, apples were introduced into the enclosures as oviposition sites. Eggs were dissected from the apples, and hatching rates were then obtained. To calculate net induced sterility, we determined natural hatch rate by releasing wild flies in tents covering single guava trees, allowing mating, and then collecting eggs deposited in apples as in the large enclosures. Induced Sterility and Effective OFRs As evident from the above summary of available data, studies assessing the impact of OFRs in SIT programs have more frequently provided measurements of induced egg sterility than population reduction. To gauge whether the elevated net percent egg sterility achieved via the SIT might actually reduce wild populations, we computed, for C. capitata, B. cucurbitae, and B. dorsalis, respectively, the relative decrease in the intrinsic birth rate (b) required to exactly balance the intrinsic rate of death (d), such that the overall intrinsic rate of increase (r) would equal zero. For brevity, and acknowledging SIT s goal of reducing birth rate, we label the proportional decrease in b required to achieve b ¼ d as the net sterility target of the SIT. We then compared the net egg sterility levels achieved with varying OFRs with this target to estimate the minimum effective OFR for the species. We fully acknowledge that this procedure, which uses a single, static measure (percent egg sterility) to assess SIT s impact on a complex, time-dependent process (per capita birth rate per unit time), provides only a rough estimate of the net sterility target but nonetheless consider this value sufficiently robust to make broad comparisons between different tephritid species. Estimates of population growth parameters are, by necessity, gathered by demographic studies conducted in the laboratory, and these vary in the environmental regimes (temperature, humidity, photoperiod, holding densities, etc.), strains (source location, degree of laboratory adaptation, etc.), and larval diets (artificial media, natural hosts, amount, etc.) used. Given this variability, our approach was to calculate average b and d values using all data available for a given species, and the following values were obtained: for C. capitata, b¼ 0.145 and d ¼ 0.0295 (Carey 1982, Vargas et al. 1984, Krainacker et al. 1987, Diamantidis et al. 2011); for B. cucurbitae, b¼ 0.145, d ¼ 0.026 (Vargas et al. 1984, Yang et al. 1994, Vayssières et al. 2008); and for B. dorsalis, b¼ 0.133, d ¼ 0.023 (Vargas et al. 1984, Foote and Carey 1987). Thus, the net sterility target (percent reduction of b to equal d) was estimated as 80% for C. capitata, 82% for B. cucurbitae, and 83% for B. dorsalis, respectively. Results For C. capitata, the relationship between OFR and induced egg sterility varied greatly among studies (Fig. 1A). Egg sterility levels between 65 83% were observed for OFRs varying from 102:1 to 390:1 for medfly on Maui, Hawaii. However, data collected in Nicaragua and Guatemala indicated that much higher OFRs were required to induce similarly high egg sterility. In Nicaragua, egg sterilities of 75 92% were observed but only at extremely high OFRs (1,120:1 3,280:1). In Guatemala, male-only releases resulted in OFRs as high as 500:1 and 775:1 with associated egg sterility levels of 68 73%. Still, these releases appeared more effective than bisexual releases, where OFRs of 750:1 and 1,350:1 yielded egg sterility levels <15%. The Nicaragua and Maui studies permitted characterization of the association between OFR and reduction of the wild fly population (Fig. 1B). In Nicaragua, OFRs between 700:1 3,160:1 resulted in declines in the wild fly population varying from 50 93%; over the 4 mo with available data, the average OFR achieved was 1,364:1, and the wild population was reduced an average of 75%. In Maui, with the exception of one estimate, the data showed a uniform level of population suppression of 70 80% across a wide range of OFRs. The large field cage study conducted in Hawaii yielded markedly different results from the aforementioned field data. Whereas field measurements of egg sterility typically exceeded 60% only at very high OFRs (Fig. 1A), induced sterility recorded in field cages in Hawaii was high even at low OFRs (Fig. 2). Sterility values >80% were observed for all OFRs 10:1, and even at the 5:1 OFR, egg sterility averaged 62%. Results from the large enclosures established in Guatemala were more consistent with the data from the open field. In these enclosures, the average reduction in pupal yield from host fruits (coffee berries) was only 44% at an OFR of 100:1. For B. cucurbitae, field studies conducted on both Kume and Maui Islands reported low to moderate OFRs but relatively high levels of egg sterility (Fig. 3A; this and following references to sterility refer to net sterility). On Kume Island, OFRs <1:1 resulted in 9 40% egg sterility, an OFR of 7:1 yielded 58% egg sterility, and OFRs of 28:1 and 45:1 were associated with egg sterilities >70%. On Maui, low OFRs (<5:1) did not induce any detectable egg sterility, but when the highest OFR (22:1) was realized, egg sterility was 64%, an estimate comparable to that observed for a similar OFR (28:1) on Kume. Data from Okinawa on the abundance of wild populations indicated that, when OFRs >45:1, the wild population in the release area was reduced >99% relative to wild population in the control area (Fig. 3B). As noted above, the computed net sterility target for a species is best considered a rough estimate, and the finding that population reduction occurred even though egg sterility levels did not exceed the computed target indicates that this value was an overestimate in this particular case. For B. dorsalis, no clear relationship between OFR and egg sterility was detected in the field study. OFRs ranged from 2:1 to 64:1 with an average of 29:1, and egg sterility measurements ranged from

6 Annals of the Entomological Society of America, 2016, Vol. 109, No. 1 A 100 80 Egg sterility (%) 60 40 20 Maui Nicaragua Guatemala - bisexual Guatemala - males-only B Wild population reduction (%) 0 100 80 60 40 20 0 0 200 400 600 800 2000 3000 4000 Sterile:wild males 0 200 400 600 800 2000 3000 4000 Sterile:wild males Maui Nicaragua Fig. 1. For C. capitata, the relationship between overflooding ratio and (A) net egg sterility reported for Maui, Hawaii (Wong et al. 1986), Nicaragua (Rhode et al. 1971), and Guatemala (Rendon et al. 2004) and (B) reduction of wild population in Nicaragua (Rhode et al. 1971) and on Maui, Hawaii (Wong et al. 1986). Data from Guatemala are presented separately for bisexual and male-only releases, respectively. Horizontal dotted line in A represents net sterility target of 77%. 18 80% with an average of 50%. In the large field enclosures, egg sterility levels increased noticeably with increasing OFR (Fig. 2). Average egg sterilities were 55 and 77% for OFRs of 5:1 and 60:1, respectively. Discussion Authors frequently mention an adequate OFR as a requisite to successful application of the SIT. To cite but a few examples, Klassen (2005) lists substantial overflooding ratios and Lance and McInnis (2005) consider a sufficiently high overflooding (sterile:wild insect) ratio critical for the success of SIT. Likewise, Steiner (1969) stated that eradication based on SIT will succeed only when there is a sustained and adequate overflooding of the native population. Beyond this general recognition, however, there is often little attempt to more precisely link OFR with SIT efficacy. For example, Villaseñor et al. (2000) listed 80:1 as the minimum OFR for medfly control, and Steiner (1969) proposed 20:1 as the critical OFR for eradication of populations of B. dorsalis, B. cucurbitae, and C. capitata. However, in neither case was evidence provided that justified the critical OFR values proposed. On the other hand, mathematical models have been developed that explicitly define the minimum OFR needed to suppress reproduction of the target population. These models, which originated with Knipling (1955, 1959), essentially modify geometric population growth through inclusion of terms that include the number of sterile and wild males in the population (as an index of infertile matings) and a coefficient of the competitive ability of sterile males that serves to weight their numerical abundance (Berryman 1967, Barclay 2005). While offering precision, these models contain parameters, such as male and female population size and the population s intrinsic rate of increase, that are difficult to measure in the field. As a result, to our knowledge, a modeling approach has never been used to determine a target OFR for a field SIT program against tephritid pests.

Annals of the Entomological Society of America, 2016, Vol. 109, No. 1 7 100 B. dorsalis C. capitata 80 Egg sterility (%) 60 40 20 0 5:1 10:1 30:1 60:1 Sterile:wild males Fig. 2. The relationship between overflooding ratio and net egg sterility for experiments conducted in large field enclosures for C. capitata and B. dorsalis on Oahu, Hawaii. Re-drawn from Shelly et al. (2005, 2010). As stated above, our objective in this paper was to compare likely indicators of SIT performance namely, net egg sterility and population reduction at different OFRs for C. capitata and two Bactrocera species. Lance and McInnis (2005) proposed that courtship complexity influences the success of the SIT, and we sought to determine whether, in general, suppression of C. capitata, with relatively complex courtship, requires higher OFRs than does suppression of Bactrocera, with relatively simple courtship. Importantly, then, we did not aim to explicitly define the threshold OFRs needed to reduce populations of these pests, but, based on presumably reliable indices of the SIT s success, we investigated whether effective OFRs differed widely for these species. With one exception, the data offered broad support for Lance and McInnis (2005) prediction. Regarding the field data, egg sterility values of 70% were recorded for B. cucurbitae on Kume Island at OFRs ranging from roughly 20:1 to 50:1. Similarly, on Maui, where only low OFRs were realized, an OFR of 22:1 was associated with an egg sterility level of 64%. The data for B. dorsalis, which derived from a single study on Oahu, Hawaii, were not clear-cut but suggest a general consistency with B. cucurbitae in that the relatively low average OFR (29:1) realized for this study was associated with a relatively high average egg sterility (50%). In contrast, for C. capitata in Nicaragua, OFRs of 1,100:1 were associated with egg sterility levels of 75 80, and 92% egg sterility was recorded only when the OFR was nearly 4,000:1. For medfly in Guatemala, release of a male-only strain resulted in egg sterility levels that were 30% for OFRs 500 and 75% at an OFR of nearly 800:1. Bisexual releases were far less efficient: egg sterility levels were all <15% even when the OFR exceeded 1,000:1. Even the sterile male releases data in Hawaii, which appeared to yield relatively high induced sterility, did not result in sterility values > 60% until OFR exceeded 60:1. Although field data relating OFR to population suppression are scant, they too indicate higher OFRs are required to dramatically reduce or eradicate populations of C. capitata than is the case for B. cucurbitae. The wild population of B. cucurbitae on the release island of Kume was >97% smaller than the corresponding population on the control island at OFRs as low as 30:1 45:1. By comparison, data from Nicaragua indicated that, over a 4-mo period, an average OFR > 1,000:1 reduced the wild medfly population only by 75%, and an extremely high OFR (3,160:1) was required to achieve a population reduction >90%. On Maui, although OFRs <100:1 appeared to result in unexpectedly high suppression (wild population in the treated area was reduced 60 70% relative to that in the control area), population reduction never exceeded 80% even when the OFR approached 400:1 Three studies presented data from large field enclosures, and two of these were consistent with Lance and McInnis (2005), while one was not. The single cage study on B. dorsalis found high levels of induced egg sterility at relatively low OFRs, thus supporting the trend noted above for field data on Bactrocera. Similarly, the cage study on C. capitata conducted in Guatemalan coffee fields reported only a 44% reduction in pupal production at an OFR of 100:1, which indicates that very high OFRs are required for reduction of target medfly populations. However, results from the Hawaiian cage study on C. capitata differed substantially from the open field studies and Lance and McInnis (2005), since OFRs ranging from 10:1 60:1 produced >80% egg sterility. In fact, under the same experimental protocol and very similar environmental conditions in Hawaii, sterile C. capitata males induced higher levels of egg sterility than did sterile B. dorsalis males. We have no ready explanation for this anomalous result, although it possibly reflects intergeneric differences in dispersal tendency. Although some individuals of C. capitata may disperse over long distances (up to 9.5 km, Meats and Smallridge 2007), the species, in general, appears relatively sedentary, with most individuals moving only tens or hundreds of meters over time (Wong et al. 1982, Barry et al. 2002, Meats and Smallridge 2007). In contrast, Bactrocera are larger, more robust flies and typically appear to move much greater distances than medfly (Fletcher 1987, Froerer et al. 2010). By confining sterile males and wild females in the same, discrete space, studies performed in large field enclosures obviously remove the need for sterile males to disperse in the environment and

8 Annals of the Entomological Society of America, 2016, Vol. 109, No. 1 A 100 80 Egg sterility (%) 60 40 20 Kume Island Maui Island 0 0 10 20 30 40 50 Sterile:wild males B 100 Wild population reduction (%) 80 60 40 20 Kume Island 0 0 20 40 60 80 100 Sterile:wild males Fig. 3. For B. cucurbitae, the relationship between overflooding ratio and (A) net egg sterility reported for Kume Island, Okinawa (Iwahashi 1977) and Maui, Hawaii (McInnis et al. 2007) and (B) reduction of the wild population on Kume Island, Okinawa (Iwahashi 1977). Horizontal dotted line in A represents net sterility target of 82%. locate leks and attendant females, a fact that would presumably provide a relatively larger advantage to sterile males of C. capitata than those of Bactrocera spp. Thus, the high OFRs apparently required to gain effective control of medfly may reflect, not only the rearingmediated modification of male courtship, but also the low vagility of sterile males, which lessens their ability to locate population hotspots. In this regard, it is noteworthy that Wong et al. (1986), which employed both air and ground releases (the latter explicitly made in known hotspots), obtained higher sterility at lower OFR values than Rhode et al. (1971) or Rendon et al. (2004), which relied on aerial release exclusively and the indiscriminate distribution of sterile males. While the low vagility of sterile medfly males may, in part, explain the need for high OFRs for effective control, it remains unclear why minimizing the importance of dispersal in the large field enclosures would necessarily result in high induced sterility, since mating trials, conducted in even smaller cages, show repeatedly that sterile male medflies are inferior sexual competitors relative to their wild counterparts (as evidenced in Table 1). In other words, even though the large field enclosures may eliminate dispersal as a factor, they would not eliminate the need for male courtship displays that are suitable or acceptable to wild females. One possible explanation for the apparent success of mass-reared males in the large field cages in Hawaii was the presence of a-copaene hotspots on particular guava trees, which served to boost the mating competitiveness of the mass-reared males. This chemical is known to be highly attractive to male medflies (Flath et al. 1994), to enhance male medfly mating success (Shelly 2001b), and to occur in the bark of guava trees (Shelly and Villalobos 2004). Males exposed to a-copaene-rich spots on guava trees gain a mating advantage over nonexposed males (Shelly and Villalobos 2004), and such exposure may have increased the mating frequency of massreared males, leading to heightened levels induced egg sterility.

Annals of the Entomological Society of America, 2016, Vol. 109, No. 1 9 In conclusion, limited data obviously render our analysis and interpretation preliminary. Moreover, an explanation resting solely on a complex vs. simple courtship dichotomy is simplistic, as many other factors may affect the success of SIT. In particular, and as noted above, the dispersal ability of released, sterile males is most likely a critical influence. We are keenly aware of these limitations and so present the Ceratitis Bactrocera comparison, not only as an initial step toward characterizing potential differences between them, but also as a stimulus to collect additional field data thus permitting a more rigorous evaluation of the relationship between OFR and SIT efficacy. The knowledge thus gained would have great practical value, as release numbers could be adjusted to the minimum target level needed to start driving the population downward to eradication. The release of too few flies would only slow the rate of increase of the wild population, while releasing much more than the target minimum number may be unnecessary, and possibly wasteful of program resources, unless a faster decline of the wild population is desired. SIT projects are large and complex operations that require great attention to a myriad of logistical problems. When running, these projects are by their very nature not particularly conducive to critical analysis and evaluation: eradication is the goal, releasing sterile males is the means to this end, and the nature of the relationship between fly release numbers and the probability of eradication is given little, or no, attention. We acknowledge the political and economic pressures on an ongoing SIT program but nevertheless urge ongoing quantitative assessment of its effectiveness, which could save critical resources and provide useful information for any future tephritid fruit fly program, thereby increasing their chance for success. Acknowledgments We are grateful to Pablo Liedo and Teresa Vera for information on Anastrepha species, Dave Lance for suggesting the review of RSI values, and Pablo Liedo, Pedro Rendon, Bill Woods, and two anonymous reviewers for helpful comments on an earlier draft. References Cited Arakaki, N., H. Kuba, and H. Soemori. 1984. Mating behavior of the oriental fruit fly, Dacus dorsalis Hendel (Diptera: Tephritidae). Appl. Entomol. Zool. 19: 42 51. Arita, L., and K. Y. Kaneshiro. 1986. Structure and function of the rectal epithelium and anal glands during mating behavior in the Mediterranean fruit fly. Proc. Hawaii. Entomol. Soc. 26: 27 30. Arita, L., and K. Y. Kaneshiro. 1989. Sexual selection and lek behavior in the Mediterranean fruit fly, Ceratitis capitata (Diptera: Tephritidae). Pac. Sci. 43: 135 143. Barclay, H. J. 2005. Mathematical models for the use of sterile insects, pp. 147 174. In V. A. Dyck, J. Hendrichs, and A. S. Robinson (eds), Sterile insect technique: Principles and practice in area-wide integrated pest management. Springer, Dordrecht, The Netherlands. Barnes, B. N., D. K. Eyles, and G. Franz. 2004. South Africa s fruit fly SIT programme - the Hex River Valley pilot project and beyond, pp. 131 141. In B.N. Barnes (ed.), Proceedings of the 6th International Symposium on Fruit Flies of Economic Importance. Isteg Scientific Publications, Irene, South Africa. Barry, J. D., R. V. Dowell, and J. G. Morse. 2002. Comparison of two sterile Mediterranean fruit fly (Diptera: Tephritidae) strains released in California s preventative release program. J. Econ. Entomol. 95: 936 944. Berryman, A. 1967. Mathematical description of the sterile male principle. Can. Entomol. 99: 858 865. Briceño, R. D., and W. G. Eberhard. 1998. Medfly courtship duration: a sexually selected reaction norm changed by crowding. Ethol. Ecol. Evol. 10: 369 382. Briceño, R. D., and W. G. Eberhard. 2000. Possible Fisherian changes in female mate-choice in a mass-reared strain of Ceratitis capitata (Diptera: Tephritidae). Ann. Entomol. Soc. Am. 93: 343 345. Calkins, C. O. 1984. The importance of understanding fruit fly mating behavior in sterile male release programs (Diptera, Tephritidae). Folia Entomol. Mex. 61: 205 213. Calkins, C. O., and A. G. Parker. 2005. Sterile insect quality, pp. 269 296. In V. A. Dyck, J. Hendrichs, and A. S. Robinson (eds), Sterile insect technique: Principles and practice in area-wide integrated pest management. Springer, Dordrecht, The Netherlands. Carey, J. R. 1982. Demography and population dynamics of the Mediterranean fruit fly. Ecol. Model. 16: 125 150. Cayol, J. P., J. Vilardi, E. Rial, and M. T. Vera. 1999. New indices to measure the sexual compatibility and mating performance of medfly (Diptera: Tephritidae) laboratory reared strains under field cage conditions. J. Econ. Entomol. 92: 140 145. Cheikh, M., J. F. Howell, E. J. Harris, H. Ben Salah, and F. Soria. 1975. Suppression of the Mediterranean fruit fly in Tunisia with released sterile insects. J. Econ. Entomol. 68: 237 243. Cratsley, C. K., and S. M. Lewis. 2005. Seasonal variation in mate choice of Photinus ignitus fireflies. Ethology 111: 89 100. Diamantidis, A. D., J. R. Carey, C. T. Nakas, and N. T. Papadopoulos. 2011. Population-specific demography and invasion potential in medfly. Ecol. Evol. 4: 479 488. Economopoulos, A. P. 1972. Sexual competitiveness of c-ray sterilized males of Dacus oleae. Mating frequency of artificially reared and wild females. Environ. Entomol. 1: 490 497. Economopoulos, A. P., and P. G. Mavrikakis. 2002. Mating performance and spatial distribution of medfly (Diptera: Tephritidae) white pupa genetic sexing males under field cage conditions. Fla. Entomol. 85: 58 62. Enkerlin, E. R. 2005. Impact of fruit fly control programmes using the sterile insect technique, pp. 651 676. In V. A. Dyck, J. Hendrichs, and A. S. Robinson (eds), Sterile insect technique: Principles and practice in area-wide integrated pest management. Springer, Dordrecht, The Netherlands. Féron, M. 1962. L instinct de reproduction chez la mouche Mediterranéan des fruits Ceratitis capitata Wied. (Diptera: Trypetidae). Comportement sexuel. Comportement de ponte. Rev. Pathol. Veg. Entomol. Agric. Fr. 41: 1 129. Flath, R. A., R. T. Cunningham, T. R. Mon, and J. O. John. 1994. Additional male Mediterranean fruit fly, Ceratitis capitata Wied., attractants from angelica seed oil (Angelica archangelica L.). J. Chem. Ecol. 20: 1969 1984. Fletcher, B. S. 1987. The biology of the dacine fruit flies. Ann. Rev. Entomol. 32: 115 144. Flores, S., P. Montoya, J. Toledo, W. Enkerlin, and P. Liedo. 2014. Estimation of populations and sterility induction in Anastrepha ludens (Diptera Tephritidae) fruit flies. J. Econ. Entomol. 107: 1502 1507. Foote, D., and J. R. Carey. 1987. Comparative demography of a laboratory and a wild strain of the oriental fruit fly, Dacus dorsalis. Entomol. Exp. Appl. 44: 263 268. Froerer, K. M., S. L. Peck, G. T. McQuate, R. I, Vargas, E. B. Jang, and D. O. McInnis. 2010. Long-distance movement of Bactrocera dorsalis (Diptera: Tephritidae) in Puna, Hawaii: How far can they go? Am. Entomol. 56: 88 94. Hendrichs, J., B. Katsoyannos, K. Gaggl, and V. Wornoayporn. 1996. Competitive behavior of males of Mediterranean fruit fly, Ceratitis capitata, genetic sexing strain Vienna-42, pp. 405 414. In B. A. McPheron and G. J. Steck (eds.), Fruit fly pests: A world assessment of their biology and management. St. Lucie Press, Delray Beach, FL. Hendrichs, J., A. S. Robinson, J. P. Cayol, and W. Enkerlin. 2002. Medfly areawide sterile insect technique programmes for prevention, suppression or eradication: the importance of mating behavior studies. Fla. Entomol. 85: 1 13. Itô, Y., and K. Yamamura 2005. Role of population and behavioural ecology in the sterile insect technique, pp. 177 208. In V. A. Dyck, J. Hendrichs, and A. S. Robinson (eds), Sterile insect technique: Principles and practice in area-wide integrated pest management. Springer, Dordrecht, The Netherlands. Iwahashi, O. 1977. Eradication of the melon fly, Dacus cucurbitae from Kume Is., Okinawa with the sterile insect release method. Res. Popul. Ecol. 19: 87 98.