Introduction

The olive fruit fly Bactrocera oleae (Rossi) likely originated in sub-Saharan Africa, where the wild olive Olea europaea cuspidata L. (Wall. ex G. Don) is found and from which the domesticated olive O. europaea europaea L. was derived. Following the path of olive cultivation, B. oleae has invaded central and northern Africa, the Mediterranean basin, south-central Asia, and recently California, USA and northwestern Mexico (Nardi et al. 2010). In California, B. oleae has spread to all commercial olive growing regions since first being detected in 1998. There are regional differences, with larger B. oleae populations in coastal regions with cooler summer temperatures than in the interior valleys where there are high summer temperatures that may limit population growth (Burrack et al. 2008; Johnson et al. 2011; Wang et al. 2009a) and adult longevity and dispersal (Wang et al. 2009b). Current management strategies in California target adult fly populations and rely primarily on frequent sprays of spinosad-based insecticidal baits and, as a result, the fly has developed resistance to spinosad in some regions (Kakani et al. 2010).

Researchers have long sought more sustainable management programs for this pest, often by using indigenous natural enemies. In the Mediterranean basin, most indigenous parasitoids found attacking B. oleae are generalist ectoparasitoids, such as Eupelmus urozonus Dalm. (Eupelmidae), Pnigalio mediterraneus Walk. (Eulophidae) and Eurytoma martellii Dom. (Eurytomidae; El-Heneidy et al. 2001; Neuenschwander et al. 1983). In California, B. oleae is attacked by a generalist ectoparasitoid Pteromalus kapaunae Heydon (Pteromalidae; Kapaun et al. 2010) and by ants (Orsini et al. 2007). However, in both the Mediterranean basin and California, these generalist natural enemies do not suppress fly populations to the economically needed levels. The absence of specialized parasitoids also argues for an origin of B. oleae outside of the Mediterranean region (Hoelmer et al. 2011) and the need for the introduction of co-adapted parasitoids that may be more effective for long-term management (Daane and Johnson 2010).

The lack of effective biological control agents attacking B. oleae in California led to the initiation of a classical biological control program in 2003. Parasitoids that were imported and evaluated in the University of California, Berkeley, USA quarantine included Bracon celer Szépligeti, Psyttalia humilis Silvestri, Psyttalia lounsburyi (Silvestri), Psyttalia ponerophaga (Silvestri), and Utetes africanus (Silvestri) (Daane et al. 2011). These parasitoids were reared from B. oleae collected from wild olives in Kenya, South Africa, Pakistan, or Namibia (Daane et al. 2008; Nadel et al. 2009; Sime et al. 2006a, b, 2007). Also evaluated were the fruit fly parasitoids Fopius arisanus (Sonan), Diachasmimorpha kraussii Viereck, and Diachasmimorpha longicaudata (Ashmead), each obtained from colonies in Hawaii (Sime et al. 2006c, 2008). At present, P. humilis and P. lounsburyi have been approved for field release in California (Daane et al. 2008; Yokoyama et al. 2008). P. ponerophaga is still under quarantine review. Although we note here that both D. longicaudata and D. kraussii were found to be effective against B. oleae (Sime et al. 2006c) because these species were considered to be host-generalists it was decided to begin the California releases with the more specialized species (P. humilis and P. lounsburyi). Here, we report on the field release and recovery efforts for P. lounsburyi and P. humilis (mainly a Namibian strain) in California that were conducted from 2006 to 2013.

Materials and methods

Insect sources and culture

Psyttalia lounsburyi and P. humilis were supplied by the USDA-ARS European Biological Control Laboratory in Montferrier, France (2008, 2009, and 2013), the Israel Cohen Institute of Biological Control in Bet Dagan, Israel (2009–2012), and the USDA-APHIS-PPQ, MOSCAMED Parasitoid Rearing Facility at San Miguel Petapa, Guatemala (P. humilis in 2010 only). At all facilities, the parasitoids were reared on the Mediterranean fruit fly, Ceratitis capitata Wiedemann, cultured on artificial diet.

The first P. lounsburyi colony was established with parasitized B. oleae collected from wild olives in Kenya’s Burguret Forest in 2002, 2003, and 2005. A second P. lounsburyi colony was established with parasitized B. oleae collected from olives in South Africa in 2005. A third colony was established with parasitized B. oleae collected from wild olives in Kenya’s Marmanet Forest in 2007. The P. humilis colony was initially established with material reared from B. oleae collected in wild olives in Grootfontein and Meteorite, Namibia in 2007 and 2008. P. humilis ‘Namibia origin’ was previously referred to as P. concolor (Szépligeti) (Rehman et al. 2009). A second P. humilis population was established with adult parasitoids collected from tephritids infesting coffee in Kenya and was mass-reared only at MOSCAMED. P. humilis ‘Kenya origin’ was previously referred to as P. cf. concolor (Wharton et al. 2000; Yokoyama et al. 2008) and P. concolor (Wang et al. 2009c). P. humilis and P. concolor are morphologically indistinguishable (Wharton et al. 2000). However, genetic analysis showed separation of these populations and researchers assigned the name P. humilis to the sub-Saharan populations (Rugman-Jones et al. 2009) and P. concolor to northern Africa. Molecular analysis was therefore used on all imported P. humilis populations to confirm their identification and origins before they were released from quarantine.

Field sites

Parasitoid releases were conducted in five California coastal counties where the summer and winter temperatures are relatively mild. In order of release efforts these were San Luis Obispo, San Mateo, Sonoma, San Diego, and Marin Counties (Fig. 1). Releases were also made in three inland counties where the summer temperatures are relatively warm (Napa) or hot (Butte and Yolo Counties; Fig. 1). The release sites were either clusters of ornamental trees, organic commercial olive groves, or abandoned olive groves. Typically the trees were Manzanillo or Mission cultivars, but some sites had a mixture of cultivars. None of the release sites received insecticides, and the coastal sites were often heavily infested by B. oleae, making them ideal habitats for field colonization and establishment of introduced parasitoids.

Fig. 1
figure 1

Map of California (USA) showing the coastal (Marin, Sonoma, San Mateo, San Luis Obispo, and San Diego Counties), intermediate (Napa County) and inland (Butte and Yolo Counties) regions where P. humilis and P. lounsburyi were released from 2006 to 2013

Full size image

In San Luis Obispo County, four sites were selected, each 5–16 km apart: (1) the Broad Street site consisted of 12 mature olive trees landscaping a parking lot. (2) The Cal Poly site consisted of three patches of olives on California Polytechnic State University campus, with parasitoid releases made at Cal Poly-1 (about 100 trees), and secondary collections made at Cal Poly-2 (1,500 m from Cal Poly-1 and consisting of only two trees) and Cal Poly-3 (1,700 m from Cal Poly-1 and consisting of only three trees). (3) The Avila Beach site consisted of three private yards that combined included 30 trees. (4) The Righetti Road site was a residential street lined with olive trees, with parasitoid releases made at Righetti Road-1 (23 ornamental olive trees), and secondary collections made at Righetti Road-2, -3, and -4 that were located about 250, 500, and 1,000 m from the release site, respectively, and each consisted of about 25 ornamental trees.

In San Mateo County, sampled sites were located on the Cañada College campus (Redwood City, California) in multiple patches of ornamental trees distributed along roadsides, parking lots, and buildings, as well as in an unmanaged olive grove and nearby patches of olive trees mixed with oak-grassland landscape. Together, the sampled areas provided about 400 trees.

There were limited releases in Sonoma, Marin, and San Diego Counties. The Sonoma County sites were in a 4 ha organic olive orchard (Stone Edge, Glen Ellen, California) and in 20 ornamental trees located along a road between two vineyards (Hanzell Vineyards, Sonoma, California). The Marin County site was located in the town of San Anselmo at a religious seminary that included about 50 olive trees. The Presidio of San Diego is a city park with 17 ha of landscaping that included 50 widely dispersed Mission olive trees.

Inland releases were concentrated in Yolo County on the University of California, Davis campus (Davis, California, USA) in an organic orchard of 30 trees, with the lower half of the trees harvested for fruit. A second Yolo County site was the University of California, Wolfskill Experimental Orchard Field Station (Winters, California), where releases were made into a USDA-ARS repository of olive varietals that consisted of about 240 trees representing over 100 varietals. The most northern release site was located in Butte County at a private residence (Leuders) with about 100 abandoned trees that were not irrigated or harvested. The Napa County site could be considered transitional between the cooler coastal and the hotter inland regions. The site consisted of about 20 ornamental trees at the Spring Mountain Vineyard (Napa, California), which were once part of an olive orchard.

Field release and recovery of parasitoids

After being cleared for release from the University of California, Berkeley quarantine, P. lounsburyi and P. humilis adults were held in organdy-screened cages (Bug Dorm 2, BioQuip, Rancho Dominquez, California, USA) with water and honey for 1–2 days prior to their field release. Adults were aspirated into small vials (40 drams) until each vial contained about 40 females and ten males. A piece of moist tissue paper was placed on the bottom of the vial to provide water, and honey was streaked on the vial lid. At the release sites, vials were typically hung on tree branches such that the parasitoids could walk or fly onto the tree. On some occasions, particularly when the numbers of olives or B. oleae were low, infested fruit were isolated with an organdy cage and the parasitoids were released into the cage, which was removed 2–3 weeks later. The number of parasitoids released on any date varied, ranging from 19 to 3,950 females per release date and site. The variation resulted from insectary production and parasitoid survival during shipment.

Pre- and post-release samples were made at all sites. Post-release fruit samples were primarily taken in the spring and fall, when olive fruit fly densities were highest, and began 1–4 weeks after a release, depending on the availability of olives (when fruit density was low collections were delayed so as to not oversample fruit that might be needed to support parasitoid establishment). On each sample date, olive fruit were randomly picked from trees within the release vicinity, depending on the number of available trees and fruit at each site, resulting in fruit collections that ranged from 102 to 2,020 fruit per site per sample date.

The collected fruit were placed in plastic containers (11 × 11 cm) that were each covered with organdy cloth and fitted with a raised metal grid (2 cm high) on the bottom to promote air circulation and facilitate efforts of pre-pupal flies to drop to the bottom of each container where they could be easily found and collected. The pupae were collected before B. oleae, P. lounsburyi, or P. humilis could emerge and form a second generation inside the collection containers (based on temperature development, Wang et al. 2012), thus representing only the field host density and field parasitism rate. Collected fruit were often held for an additional two-week period for P. kapaunae to develop from egg to pupa.

Additional samples were taken at the Cañada College sites to monitor levels of fruit fly infestation and parasitism weekly (30 September–4 November 2010) and monthly (3 December 2010–18 March 2014). On most sample dates, 50 fruit from each release tree and from adjacent trees were collected. However, in late spring when fruit were sparse less than 50 olives per site could be collected. Collected olives were placed in rearing containers as previously described.

At the Righetti Road and Cal Poly-1 sites (San Luis Obispo County) the dispersal of P. humilis and P. lounsburyi was monitored after August 2011 by sampling patches of olives located at distances of 250, 500, and 1,000 m (Righetti Roads 2–4, respectively) away from the original release point and at distances of 1,500 and 1,700 m (Cal Poly-2 and Cal Poly-3, respectively) from the Cal Poly-1 release site.

Data analysis

Results are presented as mean (±SE) for B. oleae infestation levels (estimated as the number of emerged pupae per fruit per sample) and percentage parasitism (estimated from the emergence data of adult B. oleae and parasitoid per sample). Because there were differences in release site size (number of trees) and microclimate, tree fruit load, fruit host density, the numbers of parasitoids released and number of release dates, we did not make statistical comparisons of parasitoid performance among species, species origins, or release locations. Where appropriate, we did make statistical comparisons of emergence data and parasitism percentage using a one- or two-way ANOVA, with data arcsine transformed to satisfy the assumptions of ANOVA.

Results

San Luis Obispo release and recovery

A total of 20,960 female P. humilis and 10,506 female P. lounsburyi were released from 2008 to 2013 at the four San Luis Obispo sites (Table 1). Because of insectary production, releases from 2008 to 2010 were primarily P. humilis, whereas releases from 2011 to 2013 consisted primarily of P. lounsburyi (Table 1).

Table 1 Annual and cumulative release records for adult female releases of Psyttalia humilis and Psyttalia lounsburyi at release sites in five California coastal counties, USA from 2006 to 2013
Full size table

Across all sample dates and sites in San Luis Obispo, parasitism by P. humilis ranged from 0 to 23.9 %. Recoveries of P. humilis were made on eight of ten sample dates immediately following a release date. However, P. humilis did not appear to successfully overwinter, and the longest period between a release and recovery date was 193 days at the Righetti Road-1 site (i.e., from an April 2011 release to an October 2011 post-release sample; Table 2).

Table 2 The sample locations, sample periods, numbers of olives collected, mean (±SE) olive fruit fly, Bactrocera oleae per fruit, and field recovery (as percentage parasitism ± SE) of the released Psyttalia humilis and Psyttalia lounsburyi and days after the last release to show how long after a release the recovery was made at that site, and recovery of the resident Pteromalus kapaunae near the sampled site for San Luis Obispo County, California, USA
Full size table

Psyttalia lounsburyi were first released in September 2009 with only 200 females at the Broad Street site and then November 2010 at the Cal Poly-1 site (Table 1). No recoveries were made at either site following these initial releases (Table 2). In 2011, after an April release of 4,000 P. lounsburyi at Righetti Road-1 (Table 1), 22 P. lounsburyi were recovered in August (126 days after the release, Table 2). Similarly, following a September 2011 release of 2,843 P. lounsburyi at the Cal Poly-1 site, P. lounsburyi were recovered 31 days later the same season. More importantly, at the Righetti Road and Cal Poly-1 sites, we continued to recover P. lounsburyi, up to 572 and 746 days, respectively, after the last release (Table 2).

Two resident parasitoids P. kapaunae and Eupelmus sp. were also recovered from the San Luis Obispo sites. P. kapaunae was common, although parasitism was highly variable among sites and seasons, ranging from 0 to 40.8 % (Table 2). Percentage parasitism by P. kapaunae was commonly highest in September of each year. However, those sample dates with higher percentage parasitism commonly coincided with low B. oleae infestation rates (Table 2).

San Mateo release and recovery

A total of 9,100 female P. humilis and 3,669 female P. lounsburyi were released from September 2010 to August 2013 within the Cañada College campus (Table 1). There were no releases in 2012, but this provided a 659-day period without releases to document parasitoid establishment. Similar to recoveries at the San Luis Obispo release sites, P. humilis was reared from collected fruit following releases made during the same fruit season in 2010 and 2011, with parasitism ranging from 0.9 ± 0.9 to 15.7 ± 4.7 % in 2010 and from 0.8 ± 0.6 to 4.2 ± 1.6 % in 2011 (Fig. 2a). However, no recoveries of P. humilis were made during the second half of either fruiting season (i.e., from February through May) or in pre-release samples the following year.

Fig. 2
figure 2

Mean (±SE) percentage parasitism of olive fruit fly by a Psyttalia humilis, b Psyttalia lounsburyi and c Pteromalus kapaunae from September 2010 to May 2014 at Cañada College, Redwood City, San Mateo, CA, USA. Mean values were determined by pooling samplings from all sampled trees at the different release trees. Arrows indicate field-releases of relevant parasitoid species

Full size image

Psyttalia lounsburyi was first recovered in November 2010, less than one month after 200 adults were first released on campus. Pre-release August recoveries of P. lounsburyi were then made during the beginning of each subsequent fruiting season (Fig. 2b). The recovery of P. lounsburyi in August of 2013 was 656 days after the last release at Cañada College had occurred. Percentage parasitism ranged from 0.2 ± 0.2 to 24.5 ± 6.1 % in 2011, 0.5 ± 0.2 to 3.2 ± 2.2 % in 2012, and 0.8 ± 0.2 to 42.4 ± 12.6 % in 2013. Unlike P. humilis, P. lounsburyi was collected throughout the fruiting season, with recoveries extending through April.

The resident generalist parasitoid P. kapaunae was also collected from Cañada College (Fig. 2c). Percentage parasitism averaged 4.8 ± 1.8 % in 2010, 1.4 ± 0.7 % in 2011, 0.1 ± 0.1 % in 2012 and 1.5 ± 0.6 % in 2013. Parasitism by P. kapaunae was commonly highest from September to February, although parasitism levels largely remained below 10 %.

Sonoma, Marin and San Diego County release and recovery

A total of 1,100 female P. humilis and 1,373 female P. lounsburyi were released from December 2006 to September 2009 at two sites in Sonoma County (Table 1). Pre and post-release olive collections found few B. oleae infested fruit and no parasitoids were recovered. A total of 2,050 female P. humilis and 1,037 female P. lounsburyi were released from August to September 2010 at the Presidio site in San Diego County (Table 1), and although the trees were infested with B. oleae, there were no parasitoid recoveries made. A total of 487 female P. lounsburyi were released at the San Anselmo site in Marin County in September 2013 (Table 1) and in an August 2014 post-release collection eight P. lounsburyi were recovered, nearly a year after the release.

Butte, Napa and Yolo County release and recovery

A total of 7,381 female P. humilis and 5,678 female P. lounsburyi were released from August 2006 to November 2010 at the University of California, Davis and Wolfskill sites in Yolo County; 1,840 female P. humilis and 2,220 female P. lounsburyi were released from November 2006 to November 2008 at the Spring Mountain site in Napa County; and 250 female P. humilis and 256 female P. lounsburyi were released from August 2006 to August 2008 at the abandoned orchard (Leuders) in Butte County (Table 3). Post-release fruit collections found few B. oleae infested fruit. P. lounsburyi were recovered, albeit in low numbers, at all release sites in the same season as the release, but never in pre-release collections made in the following season.

Table 3 Monthly and cumulative release records for Psyttalia humilis and Psyttalia lounsburyi for sites in five California inland counties, USA from 2006 to 2013
Full size table

Field dispersal and development of released parasitoids

The additional sample sites near the Righetti Road-1 and Cal Poly-1 allowed for a description of P. humilis and P. lounsburyi dispersal. Following a release on 15 April 2011 at the Righetti Road-1 site, both P. humilis and P. lounsburyi were recovered in fall 2011 at the original release site and two other sites located approximately 250 and 500 m away from the original release site and up to 224 days since the last release at Righetti Road-1. Additionally, two P. lounsburyi were also recovered on 25 November 2011 from Righetti Road-4, located about 1,000 m away from the original release site. At each site, parasitism by both species increased over two (Righetti Road-2) or three (Righetti Road-1 and -3) successive sampling dates (Righetti Road-1: P. humilis, F 2,10 = 62.7, P < 0.001; P. lounsburyi, F 2,10 = 5.3, P = 0.027; Righetti Road-2: P. humilis, F 1,7 = 42.1, P < 0.001; P. lounsburyi, F 1,7 = 44.3; Righetti Road-3: P. humilis, F 2,11 = 49.6, P < 0.001; P. lounsburyi, F 2,11 = 109.2, P < 0.001) and reached peak percentage parasitism levels of 20.4 ± 2.4 for P. humilis (Righetti Road-2) and 33.1 ± 1.0 and 33.0 ± 4.4 for P. lounsburyi (Righetti Road-2 and -3, respectively). Similarly, P. lounsburyi was recovered not only in the original release spot (Cal Poly-1), but also in two locations approximately 1,500 m (Cal Poly-2) and 1,700 m (Cal Poly-3) away from the original release point and 746 days since any release within 5 km. Additional samples taken near Cañada College in October 2013 also documented parasitoid dispersal.

By pooling data from 2011 collections at the Righetti Road site, more than 80 % of the parasitoids emerged from the collected fruit and host pupae that pupated during the first week following the field collection when held under laboratory conditions. Therefore, the parasitoid emergence accurately represented field parasitism rather than an artificially higher rate resulting from secondary parasitism in the emergence containers. Developmental times of the emerging adult parasitoids were collected from the pooled samples and showed that P. humilis emerged slightly earlier than P. lounsburyi (F 1,595 = 81.8, P < 0.001) and that males emerged slightly earlier than females for each species (F 1,595 = 7.2, P < 0.01, species × sex: F 1,595 = 0.2, P = 0.622).

Discussion

The field-establishment of imported biological control agents is a major step in a classical biological control program. Releases of North African populations of P. concolor have been numerous in Europe. However, these efforts led to P. concolor establishment in only southern Italy (Raspi and Loni 1994) and southern Spain (Miranda et al. 2008). There have been fewer attempts to release sub-Saharan African natural enemies of B. oleae (but see Neuenschwander et al. 1982; Silvestri 1914; Yokoyama et al. 2012). Here, we provide results from the release of sub-Saharan African populations of P. lounsburyi and P. humilis. We showed recoveries and field dispersal of both P. humilis and P. lounsburyi within the same fruit season following their releases in multiple locations. More importantly, P. lounsburyi was also recovered during fruit seasons following the last release, whereas, to date, there is no evidence that P. humilis has permanently established in California after either our current release efforts or previous releases (Yokoyama et al. 2010, 2011, 2012).

Many factors could have affected the California establishment of P. humilis and P. lounsburyi. Foremost was the limited number of parasitoids available to release, and the rearing conditions used to produce the parasitoids. An optimal release strategy would utilize large releases at sites with high target host incidence. However, insectary production and logistics associated with P. lounsburyi and P. humilis necessitated a mixed strategy of small to moderate releases at several locations. In California, maintaining large B. oleae colonies throughout the season has been difficult because the fly maggots are most easily reared in ripe olive fruit, which are unavailable throughout the year. For this reason, the parasitoids were reared on C. capitata in artificial diet, which precluded mass-rearing these parasitoid species in California, where C. capitata is a quarantined pest. Low temperature storage has been investigated for insectary-reared Psyttalia species (Daane et al. 2012), but this strategy is more conducive for colony maintenance (when ripe olives for B. oleae are sparse) than for mass rearing. The additional logistic complications of rearing parasitoids in France, Israel, and Guatemala, shipping adult parasitoids to California, and processing the material in quarantine before field release may have reduced parasitoid viability. Moreover, parasitoids that were laboratory reared on C. capitata could have lowered effectiveness against B. oleae, as studied in Trichogramma (e.g., Hoffmann et al. 2001) and other mass-reared natural enemies.

Tolerance to extreme climatic conditions could be a key attribute influencing the establishment of introduced olive fruit fly parasitoids in California. While both introduced parasitoids were recovered within the same season as the field release, only P. lounsburyi appears to have survived the winter. Previous laboratory studies suggest that P. lounsburyi is a better match with B. oleae (relative to P. humilis) regarding thermal performance and appears to be more cold tolerant than P. humilis (Daane et al. 2012; Wang et al. 2012). Field overwintering survival of both parasitoids was low in California’s interior valley, where the summer temperatures are higher and winter temperatures are colder than in coastal olive growing regions, and P. lounsburyi survival was higher than P. humilis survival at the coastal regions (Wang et al. 2013). In fact, a reason for the failed establishment of P. concolor, a species closely related to P. humilis, in northern Mediterranean regions is thought to be poor overwintering survival (Loni 1997). Therefore, climatic conditions may have affected the establishment of both parasitoids in California’s interior valley.

The olive–olive fly-parasitoid system is unique, in that the host B. oleae is also a specialist. Lack of available host material during off-season (i.e., non-fruiting) intervals may also impede parasitoid establishment. Adult parasitoids usually emerge during the early spring after overwintering as immature parasitoids within hosts. As both P. humilis and P. lounsburyi appear not to enter a winter diapause, poor host availability during the off-fruit season (late winter–late summer, depending on the California region) presents a major challenge for the survival of parasitoids that emerge in the early spring (Wang et al. 2013). Moreover, because B. oleae larvae are only found in fruit, and harvest from October to December typically remove all olives in commercial fields, only olives in ornamental trees remain available for B. oleae. These conditions would require the parasitoid to be capable of survival at low host densities. Specialization is thought to confer better host-location efficiency (e.g., Wang and Keller 2002). Indeed, most successes in classical biocontrol programs have been achieved by the introduction of specialist natural enemies, especially for parasitoids (e.g. DeBach and Rosen 1991). P. lounsburyi has been reported as a specialist on B. oleae (Daane et al. 2008), although we have reared this parasitoid on C. capitata in the insectary. While P. humilis does specialize on fruit flies, it has been commonly reared from C. capitata and other species (Wharton et al. 2000). As a specialist, P. lounsburyi may possess as yet unknown biological characteristics that facilitate survival when host densities are low, whereas P. humilis may rely on alternate hosts, with no known alternate hosts found in California. Also, in the parasitoids’ native sub-Saharan Africa range, wild olives may be found fruiting at any time of the year, which would expand the natural host reservoir. A possible solution is the planting of multiple olive tree cultivars, including some capable of carrying fruit late into spring and others having different alternative-year bearing cycles might bridge the seasonal fruit gap and improve establishment of introduced parasitoids.

Classical biological control is an attempt to reconstruct a pest–natural enemy relationship in the pest’s new environment, with success dependent upon many ecological factors. The classical biological control program for olive fruit fly in California has identified parasitoids that appeared to be highly efficient in the laboratory, such as P. humilis (Sime et al. 2006a; Wang et al. 2011, 2012), but also elucidated some inherent difficulties of establishing parasitoids in the field. Like many other agricultural pests, B. oleae originated from natural ecosystems, and the inherent tri-trophic relationships that were decoupled in the disturbed agricultural ecosystems and in newly invaded regions with different climatic conditions may disrupt this balance. Moreover, the domestication of the olive, from a small to a large fruit, changed the co-evolved parasitoid-host dynamics whereby parasitoid species with shorter ovipositors may have limited access to B. oleae larvae inside enlarged domesticated olives (Wang et al. 2009c, d). Alternative frugivorous tephritid hosts may be suitable in natural habitats where wild olives are more mixed in their maturity, providing available host fruit for the host fly or different hosts for the parasitoids (Copeland et al. 2004). Here, we have documented the California release of P. humilis and P. lounsburyi, for which most laboratory studies would suggest that P. humilis would be the likely candidate to release in California as it seems to be more effective with relatively a longer ovipositor than P. lounsburyi and outcompete resident and other imported B. oleae natural enemies, but our field results suggest that the more specialized P. lounsburyi has as yet unknown biological characteristics that enabled it to successfully overwinter and survive, even at low B. oleae densities. Continued biological control efforts, therefore, must consider not only parasitoid efficacy based on laboratory trials with an abundance of host material, but on the parasitoid species inherent abilities to survive both climatic extremes as well as periods with low host densities.