The recent shift of Rhagoletis pomonella (Diptera: Tephritidae) from its native host downy hawthorn, Crataegus mollis, to introduced domesticated apple, Malus domestica, in the eastern United States is a model for sympatric host race formation. However, the fly is also present in the western United States, where it may have been introduced via infested apples within the last 60 years. In addition to apple, R. pomonella also infests two hawthorns in the West, one the native black hawthorn, C. douglasii, and the other the introduced English ornamental hawthorn, C. monogyna. Here, we test for behavioral evidence of host races in the western United States. through flight tunnel assays of western R. pomonella flies to host fruit volatile blends. We report that western apple, black hawthorn, and ornamental hawthorn flies showed significantly increased levels of upwind-directed flight to their respective natal compared to nonnatal fruit volatile blends, consistent with host race status. We discuss the implications of the behavioral results for the origin(s) of western R. pomonella, including the possibility that western apple flies were not introduced, but may represent a recent shift from local hawthorn fly populations.

Behavior is often the pacemaker of evolution and population divergence (Mayr 2001). Behavioral shifts in habitat choice may be particularly important drivers of population divergence for host plant-specific phytophagous insects (Berlocher and Feder 2002; West and Cunningham 2002; Smadja and Butlin 2009). When an insect mates on a preferred host, host choice translates directly into mate choice, generating positive assortative mating. Differences in host choice can also affect female oviposition preference, selecting for offspring performance on the preferred hosts. The evolutionary feedback between selection for increased host choice and performance has been argued to allow populations to diverge into host-associated races and potentially species in the absence of complete geographic isolation and in the face of gene flow (Diehl and Bush 1989).

The Rhagoletis pomonella sibling species group is a model for divergence-with-gene-flow speciation for phytophagous insects (Bush 1966; Berlocher et al. 1993). The group contains a number of races and species that are similar morphologically yet attack nonoverlapping sets of host plant fruit. The recent shift (<150 years) of the species R. pomonella from its native host Crataegus mollis (downy hawthorn) to introduced domesticated apple (Malus domestica) in the northeastern United States is an example of incipient sympatric ecological speciation in action (Bush 1966; Feder et al. 1988; McPheron et al. 1988).

Here, we test for behavioral evidence of host race formation for R. pomonella flies in the western United States. The fly is thought to have been recently introduced to the Pacific Northwest via larval-infested apples (AliNiazee and Penrose 1981; Brunner 1987; Tracewski et al. 1987; Yee 2008). Rhagoletis pomonella was first reported attacking apple in the Portland, Oregon area in 1979 (AliNiazee and Penrose 1981), although it may have colonized the western United States earlier in the 1950s and gone undetected (AliNiazee and Wescott 1986). Flies have subsequently spread northward into Washington and British Columbia and southward through Oregon into northern California (Brunner 1987; Yee 2008; Yee and Goughnour 2008; Canadian Food Inspection Agency 2009). Although apple populations are found in feral trees west of the Cascade Mountains, the fly is a serious quarantine pest-threatening apple producing areas of central Washington and Oregon (Yee 2008; Yee and Goughnour 2008). In addition to apple, R. pomonella also infests two species of hawthorns as primary hosts in the Pacific Northwest (Tracewski et al. 1987; Yee 2008; Yee and Goughnour 2008). One species, C. douglasii (black hawthorn), is native to the area, whereas the second, C. monogyna (English ornamental hawthorn), was introduced perhaps sometime between the first and second world wars (Phipps 1998).

If R. pomonella was introduced into the Pacific Northwest via apples, then the current C. douglasii and C. monogyna infesting flies may represent recent sympatric shifts of R. pomonella back to hawthorn. It is also possible, although less likely, that R. pomonella is native to the Pacific Northwest and shifted from black hawthorn to apple and C. monogyna when these hosts were introduced. Indeed, black hawthorn-infesting populations of R. pomonella are known from Utah, where apple is not attacked (Allred and Jorgensen 1993). It is even possible that R. pomonella was independently introduced to the West via apple and ornamental hawthorn, and shifted to black hawthorn from ornamental hawthorn and not apple. However, there are no agricultural or museum records of the fly attacking hawthorns in the Pacific Northwest prior to the finding of infested apple in the region (S. Berlocher, pers. comm.), consistent with the apple introduction hypothesis. Regardless, the presence of R. pomonella on a number of different alternate hosts in the West presents a potential example of how ecologically differentiated host races may exist in sympatry.

In the eastern United States, apple and downy hawthorn flies discriminate between M. domestica versus C. mollis fruit volatiles (Linn et al., 2003, 2005; Forbes et al., 2005). This behavioral difference is an important ecological adaptation involved in host choice that reduces gene flow between the sympatric fly races (Linn et al., 2003, 2005; Forbes et al., 2005; Forbes and Feder 2006), as Rhagoletis flies mate on or near the unabscised fruit of their host plants (Prokopy et al. 1971). Mark-release-recapture experiments have indicated that differential host choice reduces movement of adult flies between apples and downy hawthorn to ∼4%–6% per generation at sympatric sites and has the potential to completely prezygotically isolate Rhagoletis sibling species (Feder and Bush 1989; Feder et al. 1994). Once flies have migrated to nonnatal host plants, they tend to readily mate with residents, so estimates of interhost movement are accurate gauges of potential gene flow prior to host-related selection on performance (Feder and Bush 1989; Feder et al. 1994). Volatile compounds emitted from the surface of ripening fruit are the first cues flies use to distinguish between host plants from distances of up to 10 m (Prokopy and Roitberg 1984). The behaviorally active compounds from apples and downy hawthorns have been characterized and synthetic fruit volatile blends developed that induced the same level of response in flies as whole fruit extracts (see Table S1) (Zhang et al. 1999; Nojima et al. 2003). Flight tunnel assays and field trapping studies have demonstrated that eastern apple and downy hawthorn flies preferentially orient to their respective natal host fruit blend, resulting in prezygotic reproductive isolation (Linn et al. 2003; Forbes et al. 2005; Forbes and Feder 2006).

Given the role that host fruit odor discrimination contributes to assortative mating and serves as an ecological barrier to gene flow between apple and downy hawthorn races in the East, if sympatric host races of R. pomonella exist in the western United States, then western flies should also display fruit odor discrimination. To test this hypothesis, we developed synthetic fruit volatile blends for the introduced English ornamental hawthorn and the native black hawthorn (Table S1) (Cha et al. 2012). Here, we report the results from flight tunnel assays comparing the behavioral responses of western and eastern R. pomonella to apple and hawthorn host fruit volatile blends.

Materials and Methods


Three different categories of adult flies were tested in the study (Table 1). Most of the flies were collected in the field as larvae in infested host fruit and reared to adulthood in the laboratory using standard Rhagoletis husbandry techniques, as discussed in Linn et al. (2003). We also tested adult flies aspirated directly off of host fruit in the field. Finally, we assayed adult apple flies from a laboratory colony established from apple-infesting flies collected from Geneva, New York in the 1970s. The Geneva lab colony has been replenished every few years with a portion of wild collected apple flies from the Geneva area. Treatment conditions for adult flies were the same as those in Linn et al. (2003). In addition to the Geneva, NY lab colony, flies were collected and tested in the flight tunnel from six other sites, five from the western United States and one from the East (Table 1; Fig. S1). Two of the locations in the West (site 2 = Burnt Bridge Creek Greenway, Vancouver, WA; site 3 = St. Cloud Park, Skamania, WA) represented “sympatric” sites where apple, black hawthorn, and ornamental hawthorn-infesting fly populations co-occurred in close proximity (<0.4 km) (Fig. S1). The same was also true for the eastern apple and downy hawthorn site at Fennville, MI (site 7). At the Puyallup, WA site 1, apple and ornamental hawthorn fly populations occurred within 2 km. We note that the apple flies used in the study came from a number of different apple varieties. In previous and well as the current study, we have yet to detect a behavioral difference related to apple variety (Linn et al. 2003, 2004). Similarly, we have also not found any sex-related differences in behavior between males and females (Linn et al. 2003, 2004).

Table 1.  Host plant origin, locations (latitude [N] and longitude [W] in degrees and minutes), year sampled, and whether flies tested in the flight tunnel were reared to adulthood from larval-infested field fruit (Reared lab), were captured as adults directly off of trees in the field (Field collect), or were assayed from a laboratory colony of flies (Lab colony). WA = Washington state, MI = Michigan, NY = New York. Site # refers to the designation of locations that appear in the map in Fig. S1 of the Washington sites and Table S2.
Host plant originSite #LocationLatitudeLongitudeYearTested
Western apple flies1Puyallup, WA47.11122.182008Reared lab
  2 Vancouver, WA (Burnt Bridge Creek Greenway) 45.18 122.36 2007 Reared lab
 2Vancouver, WA (Burnt Bridge Creek Greenway)45.18122.362009Reared lab
  2 Vancouver, WA (Burnt Bridge Creek Greenway) 45.18 122.36 2010 Reared lab
 3Skamania, WA (St. Cloud Park, Hyw. 14)45.35122.062008Reared lab
  3 Skamania, WA (St. Cloud Park, Hyw. 14) 45.35 122.06 2010 Reared lab
 4Woodland, WA (Donaldson Farm)45.52122.452010Field collect
Eastern apple flies 6 Geneva, NY 42.52 77.00 2001 Lab colony
 6Geneva, NY42.5277.002008Lab colony
  6 Geneva, NY 42.52 77.00 2010 Lab colony
 7Fennville, MI (125th St.)42.3686.092001Reared lab
  7 Fennville, MI (125th St.) 42.36 86.09 2010 Field collect
 7Fennville, MI (125th St.)42.3686.092010Reared lab
Western black haw flies 2 Vancouver, WA (Burnt Bridge Creek Greenway) 45.18 122.36 2009 Reared lab
 2Vancouver, WA (Burnt Bridge Creek Greenway)45.18122.362010Reared lab
  3 Skamania, WA (St. Cloud Park, Hyw. 14) 45.35 122.06 2008 Reared lab
 3Skamania, WA (St. Cloud Park, Hyw. 14)45.35122.062010Reared lab
Western ornamental haw flies 1 Puyallup, WA 47.11 122.18 2009 Reared lab
 2Vancouver, WA (Burnt Bridge Creek Greenway)45.18122.362009Reared lab
  5 Vancouver, WA (WSU Exp. Station, 78th St.) 45.52 122.39 2009 Reared lab
Eastern downy haw flies7Fennville, MI (62nd St.)42.3686.092001Reared lab
  7 Fennville, MI (62nd St.) 42.36 86.09 2008 Reared lab
 7Fennville, MI (62nd St.)42.3686.092010Reared lab


The five synthetic host fruit blends used in this study (western apple, eastern apple, downy hawthorn, black hawthorn, and ornamental hawthorn) were the same as those reported in Zhang et al. (1999), Nojima et al. (2003), and Cha et al. (2012) (Table S1). In brief, synthetic blends are developed by first isolating and characterizing whole fruit extracts made separately from fruit immediately after it is collected from host plants at several different sites. The extracts are then assessed by coupled gas chromatography/mass spectrometry and electroantennographic detection (GC/MS-EAD) analysis to identify the critical subset of active volatiles that a fly is responding to with its antennae. The behavior of flies is next reiteratively tested to different blends of these neuronally active compounds in concentrations dictated by their relative concentrations in the GC/MS analysis to arrive at a synthetic blend that induces the same degree of response for flies in the flight tunnel as the original whole fruit extract.

The development of synthetic blends is important for behavioral analysis because it allows for standardized testing and comparison of results across experiments performed at different times and on different fly populations, eliminating uncontrolled noise related to fruit quality and condition. Indeed, Rhagoletis flies are most interested in fruit that is just reaching peak ripeness. Unripe green or over ripe fruit is of reduced interest and can act to actively deter flies. Moreover, once a fruit is picked from a tree, its volatile chemistry starts to change. Thus, the standardized synthetic blends are critical because they represent as near as possible a mix of volatiles a fly would identify as being associated with ripe fruit of a particular host species.


The response of flies to host fruit volatile blends was measured in a sustained-flight tunnel as described in Nojima et al. (2003), Linn et al. (2003), and Cha et al. (2011). From our previous studies, we have determined that an accurate measure of behavioral acceptance of a blend is whether a fly initiates “upwind directed flight” of 20 cm or more toward an odor source applied to a rubber septum attached to a 7.5-cm diameter red plastic sphere positioned 1 m upwind of the fly within a 1-min test period. Flies do not respond in the flight tunnel unless fruit volatiles are present. Indeed, we have yet to observe a fly show upwind-directed flight to the red sphere in control experiments in the absence of a fruit blend (Linn et al. 2003). The baseline control level for statistical testing of fly response to fruit volatiles is therefore 0% upwind-directed flight. Differences in the frequencies of upwind-directed flight between trials were compared for statistical significance using Fisher's exact or G-heterogeneity tests (R development core team, Vienna, Austria).

We tested flies for their responses to only a single synthetic blend at a time in the flight tunnel (no-choice design) due to our previous work showing that R. pomonella is behaviorally antagonized by nonnatal fruit volatiles (Forbes et al. 2005; Linn et al. 2005). Thus, the mixing of volatile plumes that will inevitably occur when more than one blend is tested in a choice design will lead to avoidance behavior for flies and will not represent an accurate gauge of host preference. Moreover, in nature, flies are usually faced with a situation in which they have to accept or reject a specific host plant, not a choice between two different plants. Thus, no choice experiments are the most natural and realistic behavioral experiments for R. pomonella and we note that this may also be the case for many, if not most, phytophagous insects, in general.



Flies reared and collected from apples in the western United States displayed significantly reduced behavioral response in the flight tunnel to the eastern apple volatile blend compared to eastern apple flies (western flies = 34.8% upwind-directed flight [n = 442]; eastern flies = 72.4%[n = 308]; P≤ 1 × 10−10 Fisher's exact test) (Figs. 1A, B). There was no significant variation in upwind-directed flight to the eastern apple blend either among the four western apple fly sites tested (G = 3.53, P = 0.318, df = 3) or across years at the Burnt Bridge Creek Greenway or St. Cloud Park sites (G = 0.19, P = 0.662, df = 1; G = 0.43, P = 0.512, respectively). These results imply that either western apple flies have evolved to be less sensitive to or prefer a different blend of compounds than eastern apple flies. The addition of hexyl acetate and hexyl propionate to the eastern apple blend to constitute the western apple blend (Table S1) significantly increased the response of western apple flies (63.4%) to levels comparable to those for eastern apple flies to the eastern apple blend (72.4%), indicating a behavioral difference between western and eastern apple flies. However, western and eastern apple flies did not differ in their response to the western apple blend (63.4%[n = 418] versus 64.0%[n = 175] upwind-directed flight; P = 0.999) (Figs. 1A, B).

Figure 1.

Percentages of (A) western apple-origin flies, (B) eastern apple-origin flies, (C) western black hawthorn-origin flies, (D) western English ornamental hawthorn-origin flies, and (E) eastern downy hawthorn-origin flies displaying upwind-directed flight when tested to the western apple blend (WA), eastern apple blend (EA), black hawthorn blend (BH), ornamental hawthorn blend (OH), and downy hawthorn (DH) blend in the flight tunnel. n = sample sizes. Results shown are for pooled data across study sites. For individual site results see Table S2. Blend tests that do not share a letter in common differ significantly from each other, as determined by Fisher's exact tests.

Western apple flies showed significantly lower responses in the flight tunnel to the nonnatal western black hawthorn (17.5% upwind-directed flight, n = 371), western ornamental hawthorn (13.7%, n = 233), and eastern downy hawthorn (12.8%, n = 321) blends compared to both the western apple (63.4%, n = 418) and eastern apple blends (34.8%, n = 442; Fig. 1A, Table S2). There were slight, but nonsignificant, trends for western apple flies to orient more to the western hawthorn blends compared to eastern apple flies (black hawthorn: 17.5% vs. 11.4%, P = 0.077; ornamental hawthorn: 13.7% vs. 11.8%, P = 0.733) (Figs. 1A, B).

Eastern apple flies showed a similar pattern of behavior as western apple flies to hawthorn fruit volatiles in the flight tunnel (Fig. 1B). Like western apple flies, eastern apple flies displayed significantly lower percentages of upwind-directed flight to the nonnatal black hawthorn (11.4% upwind-directed flight, n = 175), ornamental hawthorn (11.8%, n = 110), and downy hawthorn blends (21.1%, n = 308) compared to either the western apple (64.0%, n = 175) or eastern apple (72.4%, n = 308) blends (Fig. 1B; Table S2). There were two differences, however. First, eastern apple flies responded at a significantly higher level to the eastern downy hawthorn blend than they did to either of the two western hawthorn blends (Fig. 1B). Second, eastern apple flies responded at a significantly higher level to the eastern downy hawthorn blend than western apple flies (21.1% vs. 12.8%, P = 0.0056). These results suggest possible migration of eastern downy hawthorn flies into the eastern apple fly population.


Except for the downy hawthorn blend, there was no significant variation between sites in the responses of black hawthorn flies to a given fruit volatile blend (Table S2). We, therefore, present the pooled results for blends except for the downy hawthorn blend, which we discuss separately below. Black hawthorn flies showed greater orientation to their natal black hawthorn blend (61.8% upwind-directed flight, n = 238) than to the nonnatal western apple (21.6%, n = 259, P≤ 1 × 10−10), eastern apple (11.9%, n = 227, P≤ 1 × 10−10), and ornamental hawthorn blends (18.8%, n = 122, P≤ 1 × 10−10) (Fig. 1C). Black hawthorn flies from the St. Cloud Park site displayed a higher level of upwind-directed flight to the downy hawthorn blend (26%, n = 50) than flies from the Burnt Bridge Creek Greenway site (7.1%, n = 170, P = 0.0016). However, at both sites, black hawthorn flies responded significantly more often to their natal black hawthorn blend than to the downy hawthorn blend (Fig. 1C). In addition, a higher percentage of black hawthorn flies from both collecting sites responded to the western apple (21.6%) compared to eastern apple blend (11.9%, P = 0.0053). These results again imply interhost migration, in this instance of western apple flies moving into the black hawthorn population. There was also a trend in the eastern United States for a higher percentage of downy hawthorn flies to respond to the eastern apple compared to western apple blend, but the difference was not significant (P = 0.0851) (Fig. 1E).


There was no significant difference among collecting sites in the percentages of upwind-directed flight displayed by ornamental hawthorn-origin flies to any of the five volatile blends tested (Table S2). We therefore discuss the pooled results for each of the five blends. Ornamental hawthorn flies showed significantly higher orientation to their natal ornamental hawthorn blend than to the nonnatal western apple, eastern apple, black hawthorn, and eastern downy hawthorn blends (Fig. 1D). In contrast to the results for apple, black hawthorn, and downy hawthorn flies, ornamental hawthorn flies displayed no significant increase in response to the western apple or black hawthorn blends compared to the eastern apple or downy hawthorn blends (Fig. 1D).


At both the Burnt Bridge Creek Greenway and St. Cloud sympatric sites, flies displayed significantly higher percentages of upwind-directed flight to their natal compared to nonnatal blends (Fig. 2). Therefore, despite evidence suggesting a degree of interhost migration, sympatric populations of R. pomonella infesting different hosts in the West nonetheless showed substantial differences in their behavioral discrimination for different fruit volatile blends.

Figure 2.

Percentages of (A) apple, black hawthorn, and ornamental hawthorn-origin flies from the sympatric Burnt Bridge Creek Greenway site in Vancouver, WA, and (B) apple and black hawthorn-origin flies from the St. Could Park sympatric site near Skamania, WA displaying upwind-directed flight when tested to the western apple blend (WA), black hawthorn blend (BH), and in the case of Burnt Bridge Creek Greenway flies, ornamental hawthorn blend (OH). Blend tests that do not share a letter in common differ significantly from each other, as determined by Fisher's exact tests.



The results from the flight tunnel assays provide evidence for behaviorally differentiated host races of R. pomonella flies infesting apple, black hawthorn, and ornamental hawthorn fruit in the western United States. All western fly populations displayed substantially higher percentages of upwind flight to their natal than to any of the other alternative nonnatal fruit blends (Fig. 1). This was true even at the Burnt Bridge Creek Greenway and St. Cloud Park sites where host trees were sympatric (Fig. 2). Moreover, a Neighbor joining tree based on Cavalli-Sforza and Edward's (1967) chord distances between fly populations calculated from their responses to the western apple, eastern apple, black hawthorn, ornamental hawthorn, and downy hawthorn fruit blends, as implemented in Powermarker (Liu and Muse 2005), clustered populations from the same host together and distinct from flies attacking alternative hosts (Fig. 3). As fruit odor discrimination in R. pomonella is directly linked to host choice and prezygotic reproductive isolation (Forbes et al. 2005; Forbes and Feder 2006), the observed behavioral differences imply a degree of nonrandom mating and reduced levels of gene flow between flies-infesting apples and different hawthorn species in the West. Indeed, trapping studies of black hawthorn-origin, ornamental hawthorn-origin, and apple-origin flies in Washington state have confirmed the relevancy of the laboratory flight tunnel results to nature and shown that R. pomonella display host-related differences in their orientation to natal versus nonnatal fruit volatile in the field (Sim et al. 2012).

Figure 3.

Neighbor joining network of behavioral chord distances between fly populations infesting black hawthorn (gray-filled triangles), ornamental hawthorn (gray squares), western apple (gray circles), eastern apple (black circles), and eastern downy hawthorn (black diamond). Numbers denote population sites (see Table 1 for site designations and descriptions). Only population sites that were tested in the flight tunnel against all five volatile blends are included in the network.

Genetic surveys are needed to complement and extend the flight tunnel and field trapping results (Sim et al. 2012) to verify the existence of R. pomonella host races in the West. However, unless behavioral phenotypes are nongenetic and represent conditioned responses determined by prior experience with fruit volatiles (see below for why this is unlikely), populations of western flies are at the least differentiated for loci affecting fruit odor discrimination. Moreover, host trees in the western United States have different fruiting phenologies during the season. The native black hawthorn C. douglasii typically ripens a couple of weeks before favored varieties of apple (Tracewski et al. 1987; Note that this differs from the East where C. mollis fruit later than apples). In addition, many favored varieties of apples peak earlier in the season than C. monogyna in the West. Variation in fruiting phenology therefore likely creates another ecological axis for genetically differentiating western R. pomonella populations, as it does for apple and downy hawthorn host races in the East (Feder et al. 1993, 1994; Filchak et al. 2000). Further study is needed to quantify the consequences of differences in host fruiting time on pre- and postzygotic isolation and genetic differentiation among western R. pomonella flies.


It is conceivable that the preferences displayed by the western flies in the flight tunnel were not genetically based, but this is unlikely for several reasons. First, although experience can affect behavior, it is usually mediated by exposure of adult insects with plant compounds (Kester and Barbosa 1991; Fellowes et al. 2005). However, Rhagoletis larvae have no extended contact with fruit surface volatiles, as they are internal fruit feeders. There is a brief period when larvae leave fruit and burrow into the soil when they may be exposed to surface volatiles. However, at this time, fruit are past peak ripeness and are beginning to rot, which represents a different mixture of volatiles than flies will experience as adults. Rhagoletis adults are not attracted to rotting fruit and are antagonized by high concentrations of natal fruit volatiles, which may indicate fruit that are past peak condition (Linn et al. 2003). In addition, adult flies are also antagonized by nonnatal volatiles (Forbes et al. 2005; Linn et al. 2005), which they will have no prior exposure to as larvae.

Second, although it was not practical to rear all of the fly populations tested on a standardized host fruit in the laboratory, such a controlled experiment was performed for one population of eastern downy hawthorn-infesting flies from Urbana, Illinois. Downy hawthorn-origin flies reared for two generations in the laboratory on red delicious apple did not differ in their upwind-directed flight responses to the apple or the downy hawthorn fruit blend compared to flies reared from field-collected fruit (Linn et al. 2003). These results argue against environmental conditioning in larvae and maternal effects affecting behavior.

Third, prior studies have shown that individual flies are consistent in their behavioral response to fruit blends across multiple trials (Linn et al. 2003), again implying that experience, in this case in the adult stage, does not influence fruit odor discrimination. Similarly, the field-collected apple fly adults tested in the current study, which presumably had prior exposure to apple fruit volatiles, responded similarly as naïve, laboratory-raised apple flies in the flight tunnel (Table S2).

Fourth, mapping studies for eastern apple × downy hawthorn flies have implied a genetic basis for fruit odor discrimination (Dambroski et al. 2005; Feder and Forbes 2008).


An important finding bearing on the origin of western flies was that apple flies from the West and East differed in their responses to apple fruit volatiles. A previous field study similarly suggested that the eastern apple blend may not be as attractive to western flies as to eastern flies (Yee et al. 2005). Apple volatile profiles from the eastern and western United States do not appear to differ greatly from one another. Yet flies from these two regions behave differently.

One possible explanation for the difference is that during the introduction of apple flies from the East, a population bottleneck occurred that altered the behavior of western apple flies. A portion of the eastern apple fly population does preferentially orient to the western apple blend. It is therefore conceivable that these flies disproportionately contributed to the establishment of apple flies in the West. An alternative possibility is that western apple flies may not represent an introduction from the East but rather a host shift from either black or ornamental hawthorn to apple. In this case, the difference between eastern and western apple flies would result from the independent evolution of behavioral discrimination for apple fruit volatiles from different ancestral hawthorn populations. Genetic data could help resolve whether eastern and western apple fly populations formed independently in different regions of the country or represent a single host shift in the East followed by introduction to the West (McPheron 1990). For example, it may be that western and eastern fly populations infesting apples and hawthorn possess unique alleles distinguishing them from each other and supporting the independent origins hypothesis. However, it is also possible that black hawthorn flies are native to the west and apple flies introduced. In this case, local gene flow from black hawthorn flies could mask the genetic signature of the introduction.

Another important subplot of the western R. pomonella story is that black and ornamental hawthorn flies not only differ from each other and from western apple flies, but also from eastern flies in their responses to host fruit volatiles (Fig. 3). Hence, even if apple flies were introduced and western hawthorn fly populations were recently derived, the shift “back” to hawthorn would not merely represent the reassembly of the original downy hawthorn behavioral phenotype, but the evolution of new host odor discrimination behaviors. Moreover, the differences between eastern and western R. pomonella fly behavior serve to underscore that the eastern apple fly race did not originate via the introduction of a previously unrecognized, preassembled apple accepting form of the fly existing elsewhere in the United States (e.g., from the Washington state). Regardless of the eventual outcome, western apple, and hawthorn-infesting fly populations highlight the remarkable evolutionary potential of R. pomonella to rapidly diversify when the fly is presented with new host resource opportunities. Indeed, along with the classic host shift from downy hawthorn to apple (Feder et al. 1988; McPheron et al. 1988), the hybrid origin of the Lonicera fly (Schwarz et al., 2005), and the existence of differentiated populations on native hawthorns in the southern United States (Powell et al. 2012), western flies appear to represent a fourth example of sympatric host races of R. pomonella complex flies, suggesting that at least for some groups of insects, ecological adaptation associated with host shifts may be an important driver of rapid diversification.

Associate Editor: T. Craig


We thank the Clark County Washington 78th St. Heritage Farm, the Washington State Univ. Research and Extension Unit, Vancouver, T. Arcella, S. Egan, T. Glen, G. Hood, M. Mattsson, M. Nash, T. Porter, K. Rogers, G. St. Jean, D. Stienbarger, S. Tracy, and B. Wolfley. We also thank C. Musto, K. Poole, and P. Fox for maintaining the flies received from Washington State and Notre Dame, and H. Reissig, D. Combs, and C. Smith for use of the Geneva, NY apple maggot colony. We also thank A. Zhang for the synthesis of (3E)-4,8-dimethyl-1,3,7-nonatriene. This work was supported in part by grants to JLF and CEL from NSF and USDA, and to WLY and CEL by the Washington Tree Fruit Research Commission and Washington State Commission on Pesticide Registration.