The effects of winter length on the genetics of apple and hawthorne races of Rhagoletis pomonella (Diptera: tephritidae).
Many explanations have been offered to account for the paucity of trade-offs (Rausher 1988, 1992; Fry 1993). One compelling argument centers on the relationship between host phenology (seasonality) and insect development (Rausher 1992). To maximize fitness, an insect's life cycle must be coordinated with host-plant availability. Asynchrony in any stage is disastrous, leading either to starvation because the insect is active when host resources are not present or to immediate death because the insect is in the wrong developmental state for prevailing environmental conditions (e.g., freezing to death as an adult during a spring frost or as a larva at the onset of winter). Since many plants differ in their seasonalities, an insect that has become adapted to the phenology of one plant can become mismatched to others. Such developmental specialization could incur a serious cost if the insect where to attack a novel plant with a different phenology (Rausher 1992). But these costs are seldom incorporated into studies of fitness trade-offs. Instead, researchers have tended to concentrate on larval feeding stages and plant secondary compounds, thereby accounting for the paucity of reported trade-offs in the literature.
The apple maggot, Rhagoletis pomonella (Diptera: Tephritidae) provides an excellent opportunity to test the role of host phenology in population differentiation. Rhagoletis pomonella has been at the center of controversy ever since Walsh (1867) cited the fly's shift from its native host hawthorn (Crataegus L. spp.) to domestic apple (Malus pumila L.) as an example of an incipient sympatric speciation event. Subsequent studies have verified that apple and hawthornflies are partially reproductively isolated and genetically differentiated host races (Feder et al. 1988; McPheron et al. 1988). Significant allele frequency differences exist between apple and hawthorn fly populations for six allozymes: malic enzyme (Me), aconitase-2 (Acon-2), mannose phosphate isomerase (Mpi), NADH-diaphorase-2 (Dia-2), aspartate amino transferase-2 (Aat-2), and hydroxyacid dehydrogenase (Had) (Feder et al. 1988, 1990a; McPheron et al. 1988; Feder and Bush 1989). These six allozymes map to just three different regions of the R. pomonella genome (Berlocher and Smith 1983; Feder et al. 1989, Roethele et al. 1997). Aat-2 and Die-2 are located on linkage group I; Me, Acon-2, and Mpi are tightly linked on group II; and Had maps to linkage group III. Significant levels of linkage disequilibrium exist between loci within, but not among, these three linkage groups in natural fly populations (Feder et al. 1990a).
One key element is still missing from the Rhagoletis story, however, which is the lack of demonstrable fitness trade-offs. Such trade-offs are necessary because mark-release-recapture studies indicate that host-specific mating is not absolute; genetic exchange has been estimated to occur at 6% per generation between the host races at a field site near Grant, Michigan (Feder et al. 1994). Some form of host-associated selection is therefore needed to counteract this gene flow or the races would quickly fuse, and long-term allozyme studies indicate that this is not happening at the Grant site (Feder et al. 1993, 1997). However, reciprocal transplant experiments have given no evidence for larval feeding specialization in the races related to chemical or nutritional differences between apple and hawthorn fruits (Prokopy et al. 1988).
Past failures to document trade-offs in R. pomonella could be due to overlooking the interaction between fly development and host plant phenology. Apples and hawthorns represent different seasonal resources. Fruit on apple varieties favored by R. pomonella generally peaks about three weeks earlier than it does on hawthorns (Feder et al. 1993). An important consequence of this difference is that fly larvae emerge from apples, pupate in the soil, and enter diapause an average of 16 days earlier in the season than they do from hawthorns (Feder 1995). But apple-fly adults eclose the following year an average of only 10 days earlier than hawthorn adults (Feder 1995). Apple flies therefore overwinter for almost a week longer than hawthorn flies. Furthermore, applefly larvae and pupae develop at times in the summer when they are exposed to hotter temperatures and longer photoperiods than hawthorn flies, factors known to hasten diapause termination (Prokopy 1968). Because R. pomonella has a facultative pupal diapause, rapidly developing apple-pupae run the risk of bypassing diapause and immediately developing into adults; second generations of apple flies have been reported in the field (Caesar and Ross 1919; Porter 1928; Phipps and Dirks 1933). Such "nondiapausing" flies are inevitably doomed; either they eclose at times when suitable host fruit is no longer available or they commit to, but do not complete, adult development before the onset of winter and freeze to death. Selective pressures are likely to be different for hawthorn flies. The relatively late phenology of hawthorns means that slow developing flies may not enter a proper pupal diapause before winter and die. The pupal diapause may therefore be a key life stage under host-associated selection in Rhagoletis, with differential selection for fast development (shorter diapause) in the hawthorn race and slower development (longer diapause) in the apple race. We shall refer to this idea as the "diapause trade-off hypothesis."
Evidence suggests that the six allozymes displaying hostrelated differentiation are associated with fly development. Earlier eclosing adults in both host races possess high frequencies of the alleles Me 100, Acon-2 95, Mpi 37, Aat-2 +75, Dia-2 100, and Had 100 (Feder et al. 1993). These six alleles are generally more common in the hawthorn than the apple race in and around Grant, Michigan (Feder et al. 1990a,b; Feder et al. 1993). This suggests that, all things being equal, hawthorn flies are genetically programmed to develop faster than apple flies. Since the timing of adult eclosion is directly tied to diapause termination in most insects (Morris and Fulton 1970; Tauber and Tauber 1976), these findings concur with the diapause trade-off hypothesis. They are also interesting in that they suggest that the reason that apple flies eclose earlier in the field season than hawthorn flies in Michigan is environmental; apple pupae have, on average, a 16-day head start over hawthorn pupae. If flies from Grant, Michigan, could be reared under standardized conditions, then apple flies should eclose later than hawthorn flies.
Before delving further into the Rhagoletis system, it would be useful to review some background information about diapause. As discussed by Tauber and Tauber (1976)
diapause is a dynamic state, as the season progresses, diapause depth or intensity decreases, even when insects are held under constant conditions (McLeod and Beck 1963; Beck 1968; Tauber and Tauber 1970). . . . Sometime during winter, a population ceases to respond to diapause-maintaining factors and the incidence of diapause decreases, often over a broad span of time. . . . Therefore, in most species with an overwintering diapause, diapause ends by mid-winter rather than in the spring. . . . Subsequently, postdiapause development is prevented, sometimes for a considerable period, until temperatures rise above the lower thermal threshold for development. This type of diepause termination has been demonstrated for numerous species, e.g., the sweet clover weevil, Sitone cylindricollis (Hans 1961); the ladybug, Coccinella septempunctata (Hodek 1962); Pyrrhocoris apterus (Hodek 1971, 1974); the flesh fly, Sarcophaga bullata (Denlinger 1972); the thrips, Anaphothrips obscurus (Kamm 1972); and the lacewings Chrysopa carnea and Chrysopa harrisii (Tauber and Tauber 1973, 1974).
We will concentrate on the question of diapause termination and its relationship to winter length in this manuscript. But we will also explore how summer/fall conditions interact with winter duration to affect the genetics of the host races. One reason for concentrating on diapause termination is that this trait is often genetically controlled in insects and subject to rapid rates of natural selection (Danilevsky 1965; Tauber and Tauber 1976). Danilevskii (1965) has written that "experimental investigations have not disclosed any considerable intraspecific geographic variation in temperature requirements during the period of active growth and development, or of frost-resistance in the diapausing stages." Rather, adaptation to latitude is accomplished via genetic differences governing diapause duration (Masaki 1965; Tauber and Tauber 1976). We hypothesize that the difference in apple and hawthorn phenology affects R. pomonella in a similar manner that latitude does for other insects; the earlier phenology of apples selects the apple race to be more recalcitrant to diapause termination. Indeed, later we will argue that both latitude and host phenology likely shape the genetics of R. pomonella through their effects on diapause.
Some of the most elegant work on diapause was conducted by Morris and Fulton (1970) on the fall webworm, Hyphantria cunea. These researchers manipulated winter length for diapausing pupae and found that survivorship peaked at four months (Morris and Fulton  reported that the optimal chilling period was six months, but inspection of their fig. 5.6 shows that it was actually closer to four months). They hypothesized that survivorship was low when the overwintering period was shorter than four months because many pupae were unable to break diapause without adequate chilling. Survivorship decreased rapidly after eight months because fat reserves of pupae became depleted. Morris and Fulton (1970) confirmed this latter claim by measuring pupal weight loss. Nearly all pupae could tolerate a weight loss of up to 8% during chilling and survive. Beyond this point, however, mortality increased progressively and the pupae that died were those with the highest rates of weight loss. The authors also manipulated the duration of the prechilling period. They found that exposure of pupae to warm weather for even relatively short periods prior to chilling influenced survival rates. Pupae could tolerate a weight loss of up to 2% during the prechilling period, but after this point, they started to die. The longer they made the prechilling period, the greater the extent of weight loss in pupae and the higher the rate of overwintering mortality. In addition, weight loss was an increasing linear function of prechilling temperature. These results suggest that similar processes affect pupal survival before and during winter in Hyphantria. Morris and Fulton (1970) also showed that differences in pupal heat requirements for development arose largely through heritable differences associated with diapause termination. In other words, the time it took for an adult moth to eclose after being removed from the cold as a pupa was primarily a function of when it broke diapause and became developmentally competent, rather than how fast it underwent morphogenesis. Morris and Fulton (1970) related their laboratory findings to nature by showing that longer growing seasons in coastal areas of Nova Scotia and New Brunswick, Canada, permitted the survival of a larger proportion of moths with high pupal heat requirements (i.e., insects that took longer to terminate diapause and eclose) than in colder, inland areas with shorter seasons. Cold years at sites were also shown to favor moths with low pupal heat requirements.
We have discovered many parallels between the fall webworm and the apple maggot fly with respect to the prewintering period. In an earlier study (Feder et al. 1997), we manipulated the prechilling period for hawthorn pupae in an effort to induce a genetic response at the six allozymes displaying host-associated differentiation (we shall refer to this study as the "prewinter experiment"). Similar to Hyphantria, extending the prewintering period for more than one week at 26 [degrees] C, 15:9 L:D drastically reduced overwintering survivorship in the hawthorn race, presumably due to pupae exhausting their fat reserves. We also found that populations of eclosing haw adults, when challenged as pupae with the earlier apple phenology, evolved toward the allozyme frequencies characteristic of the apple race. Finally, nondiapausing flies that developed rapidly and eclosed prior to chilling had the highest frequencies of "hawthorn-race alleles." These findings are consistent with hawthorn flies generally having shallower pupal diapauses and/or higher metabolic/development rates than apple flies. However, there was one anomalous result in the prewinter experiment. The diapause trade-off hypothesis predicts that short prechilling periods should favor hawthorn race alleles. However, allele frequencies for adults eclosing after brief prechilling treatments (e.g., 2 days) in the prewinter experiment were not appreciably different from those of untreated controls (Feder et al. 1997).
Two lines of evidence tie the results from the prewinter experiment to the ecology and genetics of R. pomonella. First, the allozyme loci Me, Acon-2, Mpi, Dia-2, Aat-2, and Had all display latitudinal allele frequency clines among both apple and hawthorn populations across eastern North America (Feder and Bush 1989; Feder et al. 1990a; Berlocher and McPheron 1996). Other polymorphic allozymes show little geographic variation in the races. Fly populations from northern (i.e., colder) latitudes possess higher frequencies of the alleles Me 100, Acon-2 95, Mpi 37, Aat-2 +75, Dia-2 100, and Had 100, which are more common to the hawthorn than the apple race at the Grant site and correlate with earlier eclosion (Feder et al. 1993). In contrast, populations from southern latitudes possess higher frequencies of alleles more common to the apple race. The clines also show several perturbations that coincide with differences in ambient temperature conditions among local collecting sites (Feder and Bush 1989, 1991). Second, frequencies for Me 100, Acon-2 95, Mpi 37, Aat-2 +75, and Dia-2 100 in the hawthorn race at the Grant site were significantly related to ambient temperature in the spring of the preceding year over an 11-year period beginning in 1984 (Feder et al. 1993, 1997). Because a large majority of R. pomonella pupae break diapause after their first winter (92-98%; Phipps and Dirks 1933; Dirks 1935), flies sampled in year N are representative of pupae that survived the previous season (year N - 1). Hot spring temperatures in year N - 1 meant an early season and correlated with increased frequencies in year N of alleles common to the apple race (or southern populations). Conversely, cold springs favored alleles more common in northern populations.
Here, we continue our studies of pupal diapause in R. pomonella by investigating the effects of winter duration on the genetics of the host races. We shall refer to this current study as the "overwinter experiment." As in Hyphantria, if the period preceding winter is important for R. pomonella, then so too should be the length of winter. Our rationale for this is that pupae that break diapause too early in the winter and/or have high baseline metabolic rates will use up their energy reserves too quickly and die during prolonged cold storage. Conversely, short winters may not provide adequate chilling for diapause termination, thereby favoring pupae in shallower states of diapause. To test these possibilities, we overwintered apple- and hawthorn-origin pupae at 4 [degrees] C for time periods ranging from one week to two years. Our prediction was that longer periods of cold storage would select against the allozymes markers associated with earlier eclosion in our previous studies (Me 100, Acon-2 95, Mpi 37, Aat-2 +75, Dia2 100, and Had 100). Conversely, short overwintering periods should favor these alleles.
MATERIALS AND METHODS
Infested fruits were collected from beneath apple and hawthorn trees at a field site on Ferris road near the town of Grant, Michigan, on August 30 and September 16, 1990, respectively. The Ferris site is approximately 3 km southeast of the 112th street study site where flies used in the prewinter experiment were collected in 1989. Fruits were placed on wire screens above plastic trays and the trays stored at 26 [degrees] C in a constant temperature room with a 15:9 h L:D cycle. Puparia were collected from the plastic trays on a daily basis and divided into 22 subsamples. One daily subsample for each race was immediately frozen at - 70 [degrees] C to serve as an untreated genetic control. Pupae from the remaining subsamples were put into petri dishes that contained moist vermiculite and the petri dishes were held in the constant temperature room for seven days before being placed in cold storage at 4 [degrees] C to simulate winter. (We used a seven-day prechilling period because this resulted in maximal overwintering survivorship in the prewinter experiment.) The positions of petri dishes were randomly shuffled every week to account for possible temperature variation in the refrigerator. Subsamples of pupae were removed from the cold after one, three, five, seven, eight, nine, 10, 11, 13, 15, 17, 19, 22, 26, 30, 35, 40, 45, 52, 78, and 108 weeks and placed in an incubator at 21 [degrees] C with a 14:10 h L:D cycle. The petri dishes were monitored for eclosing flies over a five-month period before the dishes were returned to the refrigerator to be exposed to a second yearly cycle of chilling and heating. We did this to ensure that flies in deep diapauses or with two-year life cycles had adequate time to eclose. However, no adults eclosed during the second year of the study. Flies were genetically scored for Me, Acon-2, Mpi, Aat-2, Dia-2, Had, and Isocitrate dehydrogenase (Idh) using standard starch gel electrophoretic techniques (Feder et al. 1989). Idh was included as a genetic control because it displays limited host-associated and geographic allele frequency variation in R. pomonella (Feder et al. 1990a).
The diapause trade-off hypothesis makes four specific predictions about the overwinter experiment: (1) there should be an optimal overwintering period that is long enough to ensure diapause termination but short enough to prevent excessive energy loss in pupae; (2) the apple race should be more recalcitrant to the effects of prolonged chilling than the hawthorn race because it overwinters for a longer period of time in nature - two expected manifestations of this are that apple flies should survive long winters better and eclose later than hawthorn flies when the two races are reared under standardized conditions; (3) as in previous studies (Feder et al. 1993), earlier eclosing adults in both host races should have high frequencies of the alleles Me 100, Acon-2 95, Mpi 37, Aat-2 +75, Dia-2 100, and Had 100; and (4) short overwintering periods should favor the aforementioned alleles associated with earlier eclosion and by inference shallower diapause; in contrast, long chilling periods should select against these allozymes.
For the purposes of data analysis, results from different overwinter treatments were pooled to increase sample sizes. This usually involved combining 1-3, 5-8, 9-13, 15-19, 22-26, 30-35, 40-52, and 78-108 week subsamples to produce a total of eight different winter treatments. We also pooled alleles at the Aat-2 locus into two different electrophoretic classes for analysis. One class contained all alleles whose relative mobilities were [greater than or equal to] 75 compared to the most common allele in the population (the 100 allele). The other class consisted of all alleles whose relative mobilities were [less than] 75. Previous studies have indicated that the +75 and - 75 classes of Aat-2 alleles are in very strong linkage disequilibrium with the 100 and 70 alleles, respectively, at the Dia-2 locus (Feder et al. 1988, 1990a; Berlccher and McPheron 1996).
Differences in the frequencies of eclosing adults between the host races for a given overwintering treatment were tested for significance by two-tailed, randomized Fisher exact tests. Differences in the eclosion times of the races were tested for significance using two-tailed Mann-Whitney U-tests with tied ranks. The relationship between the allozymes and eclosion time was analyzed in three different ways. First, Spearman rank correlation coefficients ([r.sub.s]) corrected for tied ranks were calculated between genotypic scores of flies and their eclosion times. Genotypic scores were the number of Me 100, Acon-2 95, Mpi 37, Aat-2 +75, Dia-2 100, Had 100, or Idh 100 alleles that a fly possessed (e.g., Me 100/100 homozygotes = 2, Me 100/-heterozygotes = 1 and Me-/-homozygotes = 0, where - indicates any Me allele other than 100). Rank correlations were calculated separately for each locus for each overwintering treatment in both host races. Correlation coefficients were z-transformed to test them for significance. For Me, Acon-2, Mpi, Aat-2, Dia-2, and Had, one-tailed tests were performed under the alternate hypothesis that flies possessing the alleles Me 100, Acon-2 95, Mpi 37, Aat-2 +75, Dia-2 100, and Had 100 should eclose the earliest. Two-tailed tests were performed for the control locus Idh. Second, allozyme frequencies were calculated for flies eclosing within 10 successive time intervals in the nine to 19 week overwinter treatment. Linear regressions were then performed between these arcsine-transformed allele frequencies and the mean time to eclosion for flies within the 10 intervals. Third, allele frequencies were calculated for adults that eclosed prior to 43 days after being removed from the cold and flies that took 43 days or longer to eclose. Differences between the [less than] 43-day and [greater than or equal to] 43-day populations were tested for significance using one-tailed randomized Fisher exact tests under the alternate hypothesis that the [less than]43-day sample should have higher frequencies of the alleles Me 100, Acon-2 95, Mpi 37, Aat-2 +75, Dia-2 100, and Had 100. The effect of winter length was examined by performing linear regressions between arcsine-transformed allozyme frequencies in eclosing adults and the duration of chilling in weeks. Two-tailed randomized Fisher exact tests were calculated to test for significant allozyme frequency differences between eclosing adults and the appropriate control sample.
We must stress that the overwinter experiment is not designed to test whether the allozymes are directly under selection. Rather, its purpose is to examine the relationship between winter length and pupal development through an analysis of allele frequency changes induced at allozyme loci. As such, the allozymes represent genetic markers tagging three regions of the R. pomonella genome known to differ between the races and which have previously been associated with a diapause related trait, the timing of eclosion (Feder et al. 1993). We assume that any genetic response observed for the allozymes will, at the very least, be indicative of selection at linked loci. Of course, if the allozymes are indirectly responding to selection, then the overwinter experiment will underestimate the true magnitude of the genetic response.
The overwinter experiment is also not designed to elucidate the exact physiological basis for any observed genetic response to selection, whether it be due to a difference in diapause depth, metabolic level, or development rate (all three traits are probably interrelated and are refered to interchangeably in this manuscript). Estimating metabolic rates and times of diapause termination for pupae are notoriously difficult tasks (Tauber and Tauber 1976). Even measuring pupal weight loss inflicts a high degree of mortality (Morris and Fulton 1970). It is therefore important to first establish that these traits are the likely foci of selection and to identify the environmental factors responsible for selection. These are the goals of the overwinter experiment. Future studies can then be performed to pinpoint the physiological basis for selection.
Eclosion Patterns for Adult Flies
Pupal survival was related to the duration of the chilling period [ILLUSTRATION FOR FIGURE 1A OMITTED]. Percentages of eclosing adults rose sharply until the 15-19-week overwintering treatment and then declined afterward in both host races [ILLUSTRATION FOR FIGURE 1A OMITTED]. Fifteen to 19 weeks of chilling at 4 [degrees] C therefore appears to be the optimal overwintering period for Ferris site flies, a result paralleling the finding of Morris and Fulton (1970) for Hyphantria. The decline in survivorship after 22 weeks of cold storage was more precipitous for the hawthorn-than the apple-fly race [ILLUSTRATION FOR FIGURE 1 OMITTED]. This suggests that the apple race was more recalcitrant to the effects of long winter than the hawthorn race, as predicted by the diapause trade-off hypothesis.
Mean eclosion times (i.e., the time it took for adults to emerge following removal from the cold) also varied significantly across winter treatments [ILLUSTRATION FOR FIGURE 1B OMITTED]. Similar to the survivorship curves, flies took progressively longer to eclose up until the 15-19-week treatment, but then took less time thereafter [ILLUSTRATION FOR FIGURE 1B OMITTED]. This suggests that survivorship and mean eclosion times were lower in the shorter overwintering treatments because pupae in deep states of diapause did not receive adequate chilling to initiate development. By 22 weeks, however, it is likely that most pupae had terminated diapause. Extending the overwintering period for these flies subsequently increased their mortality and decreased their eclosion times [ILLUSTRATION FOR FIGURE 1A, B OMITTED]. As in Hyphantria, the increase in mortality was probably due to pupae exhausting their fat reserves, but this hypothesis requires further testing in Rhagoletis.
Unlike the survivorship curves, mean eclosion times decreased more rapidly for the apple than the hawthorn race following the 19-week treatment [ILLUSTRATION FOR FIGURE 1B OMITTED]. Up to and including the 15-19-week winter treatment, apple flies had greater mean eclosion times than hawthorn flies [ILLUSTRATION FOR FIGURE 1B OMITTED]. If the 15-19-week treatment is the optimal chilling period for pupae in nature, then this suggests that apple flies generally have deeper diapauses or slower metabolic rates than hawthorn flies, consistent with the diapause trade-off hypothesis. But in the 30-35- and 40-52-week treatments hawthorn flies had greater mean eclosion times than apple flies, a finding in accord with previous studies of adult eclosion in the host races (Smith 1988; Feder et al. 1993). We suspect that the change in the relative order of eclosion for the races was related to the pronounced drop in survivorship for hawthorn flies following the 22-26-week treatment [ILLUSTRATION FOR FIGURE 1A OMITTED]. It is possible that a greater proportion of fast developing pupae were culled from the hawthorn-than the apple-fly population at this time. But regardless of whether this is true, our results send a cautionary note concerning the interpretation of diapause experiments in R. pomonella, as mean eclosion times and the relative order of eclosion for the races are affected by winter rearing conditions.
The shapes of adult eclosion curves were also dependent upon winter rearing conditions [ILLUSTRATION FOR FIGURE 2 OMITTED]. Eclosion curves for adult flies surviving the [less than or equal to] 19-week winter treatments were bi-modal, while curves for treatments [greater than or equal to] 22 weeks were unimodal [ILLUSTRATION FOR FIGURE 2 OMITTED]. This difference was due to the elimination of the first mode of earliest eclosing flies following 19 weeks of cold storage (see [ILLUSTRATION FOR FIGURES 2I-1 OMITTED] for detailed comparisons of the 15-22-week eclosion curves for the apple race that illustrate the culling of the first mode of flies). The second mode of flies was itself not immune to the effects of long winters, however. Survivorship percentages and mean eclosion times also decreased for these flies after the 22-26-week treatment [ILLUSTRATION FOR FIGURES 1A, B OMITTED]. This suggests that a similar culling process occurred in both the first and second mode of eclosing flies, but it happened later, after a longer chilling period, for the second mode of flies.
Genetic Analysis of the Overwinter Experiment
The Genetics of Eclosion Time
The allozymes Me, Acon-2, Mpi, and Had showed consistently significant relationships with the timing of adult eclosion in the overwinter experiment. We document these relationships in three different ways. First, Table 1 gives nonparametric Spearman rank correlation coefficients ([r.sub.s]) between Me 100, Acon-2 95, Mpi 37, and Had 100 genotypes of flies and their eclosion times. Second, Figure 3 plots allozyme frequencies for flies that eclosed within 10 successive time intervals in the 9-19-week overwinter treatments. Finally, Figure 4 compares allele frequencies between flies that eclosed prior to 43 days after being removed from the cold and flies that took 43 days or longer to eclose. The 43-day cutoff was used because this was the dividing time between first and second modes of flies in the [less than or equal to] 19-week treatments [ILLUSTRATION FOR FIGURE 2 OMITTED]. The take home message was the same in all three analyses; flies possessing the alleles Me 100, Acon-2 95, Mpi 37, and Had 100 eclosed significantly earlier than flies with alternate alleles at these loci. The alleles Me 100, Acon-2 95, Mpi 37, and Had 100 are generally found in higher frequencies in the hawthorn than the apple race near Grant, Michigan (Feder and Bush 1989; Feder et al. 1990a, b, 1993), although the frequency of HAD 100 was slightly higher in the apple race at the Ferris site in the current study [ILLUSTRATION FOR FIGURE 5 OMITTED]. These findings therefore support our contention that there is a genetic basis for the hawthorn race developing faster (eclosing earlier) than the apple race, at least with respect to the allozymes Me, Acon-2, Mpi, and Had.
The results for Aat-2 and Dia-2 were more equivocal than those for the other four allozymes. Aat-2 +75 and Dia-2 100 were both significantly related to earlier adult eclosion in the hawthorn race in the 9-19-week treatment (Table 1, [ILLUSTRATION FOR FIGURES 3, 4 OMITTED]), as was Dia-2 100 in the 22-35-week treatment (Table 1). For the apple race, Dia-2 100 frequencies were significantly different between [less than] 43- and [greater than or equal to] 43-day eclosing adults in the 1-8-week treatment [ILLUSTRATION FOR FIGURE 4 OMITTED]. However, only the results for DIA-2 for hawthorn flies were significant when tested on a tablewide basis within each host race after applying a sequential Bonferroni correction (Holm 1979; Rice 1989).
The control locus Idh displayed one significant Spearman rank regression with adult eclosion out of eight total tests (Table 1). This result was not significant on a tablewide basis.
The Effects of Winter on Allozyme Frequencies
Genetic analysis of the overwinter experiment supported the major premise of the diapause trade-off hypothesis; extended chilling periods selected against the allozymes associated with earlier eclosion in R. pomonella. Allele frequencies for Me 100, Acon-2 95, Mpi 37, and Had 100 all significantly declined in flies surviving longer winter treatments ([ILLUSTRATION FOR FIGURES 5A-D OMITTED]; [r.sup.2]-value for linear regression between arcsine-transformed Me 100 frequencies in the hawthorn race and the length of the chilling period = 0.91, P = 0.0002; [r.sup.2] Me 100 apple race = 0.59, P [less than] 0.027; [r.sup.2] Acon-2 95 hawthorn race = 0.57, P = 0.031; [r.sup.2] Acon-2 95 apple race = 0.75, P [less than] 0.006; [r.sup.2] Mpi 37 hawthorn race = 0.51, P = 0.047; [r.sup.2] Mpi 37 apple race = 0.12, P = 0.485; [r.sup.2] Had 100 hawthorn race = 0.89, P = 0.0004; [r.sup.2] Had 100 apple race = 0.26, P = 0.198; 7 df for all tests). The only exception was Had 100 in the apple race, which displayed a V-shaped response pattern rather than a consistent linear decline [ILLUSTRATION FOR FIGURE 5D OMITTED]. We do not know the reason for this. One possibility is the existence of hidden electrophoretic variation at, or linked to, Had. More detailed molecular genetic analysis in and around the Had locus is required to resolve this issue.
The results for Aat-2 and Dia-2 were again more equivocal than those for the other four allozymes. Neither locus displayed a significant regression with winter length in either host race ([ILLUSTRATION FOR FIGURE 5E, F OMITTED]; [r.sup.2] Aat-2 +75 hawthorn race = 0.08, P = 0.487; [r.sup.2] Dia-2 100 hawthorn race = 0.38, P = 0.10; Aat-2 +75 apple race = 0.04, P = 0.642; [r.sup.2] Dia-2 100 apple race = 0.01, P = 0.82; 7 df in these and subsequent tests involving Aat-2 and Dia-2 in the overwinter experiment). But a significant sex x chilling period effect was found for Aat-2 +75 and Dia-2 100 in the hawthorn race, as frequencies for these two allozymes significantly fell in males ([r.sup.2] Aat-2 +75 = 0.51, P = 0.044; [r.sup.2] Dia-2 100 = 0.58, P = 0.027) but not in females ([r.sup.2] Aat-2 +75 = 0.15, P = 0.347; [r.sup.2] Dia-2 100 = 0.001, P = 0.938). No significant sex effect was found for either Aat-2 or Dia-2 100 in the apple race, but males did have higher re-values ([r.sup.2] Aat-2 +75 males = 0.28, P = 0.178; [r.sup.2] Aat-2 +75 females = 0.001, P = 0.940; [r.sup.2] Dia-2 100 males = 0.18, P = 0.299; [r.sup.2] Dia-2 100 females = 0.02, P = 0.774). Aat-2 and Dia-2 were also unusual in that frequencies for Aat-2 +75 and Dia-2 100 in the hawthorn race were always higher in eclosing flies than in the untreated control sample, while in the apple race the opposite was true [ILLUSTRATION FOR FIGURE 5E, F OMITTED]. We may therefore have either failed to identify the source of [TABULAR DATA FOR TABLE 1 OMITTED] selection acting on Aat-2 and Dia-2 in the host races or correctly identified the source but it displays a sex x host interaction.
If the sex-specific responce observed for Aat-2 and Dia-2 to winter length is real, then a similar effect would be expected for these two loci with respect to the prechilling period. A reanalysis of the prewinter experiment revealed this to be true. Extending the prewintering period resulted in significant linear declines in the frequencies of Aat-2 +75 and Dia-2 100 in surviving hawthorn-fly males ([r.sup.2] Aat-2 +75 = 0.45, P = 0.049, 8 df; [r.sup.2] Dia-2 100 = 0.45, P = 0.049, 8 df; note: we included new data for 35-, 42-, and 65-day prewinter treatments in this analysis that were not published in Feder et al. 1997). But allele frequencies did not significantly decline in females ([r.sup.2] Aat-2 +75 = 0.003, P = 0.896, 8 df; [r.sup.2] Dia-2 100 = 0.03, P = 0.660, 8 df).
Allele frequencies for the control locus Idh were not affected by the length of winter ([r.sup.2] hawthorn race = 1.5 x [10.sup.-6], P = 0.998, 7 df; [r.sup.2] apple race = 0.192, P = 0.278, 7 df).
The goal of the overwinter experiment was to clarify the nature of host-associated selection acting on apple- and hawthorn-infesting races of R. pomonella. To this end, we put forth the diapause trade-off hypothesis: the earlier seasonality of apples selects for slower development rates in apple flies in the form of lower baseline metabolic levels or more entrenched pupal diapauses than in hawthorn flies. If this hypothesis is correct, then one environmental factor that should have an important bearing on this selection is the length of winter. We investigated this possibility by exposing fly pupae to varying periods of cold storage at 4 [degrees] C to test for a genetic response at six allozyme loci displaying host-associated differentiation. Four predictions follow from the diapause tradeoff hypothesis: (1) there should be an optimal overwintering period that maximizes pupal survivorship; (2) the apple race should be more recalcitrant to the effects of long winters than the hawthorn race; as a consequence, apple pupae should survive longer periods of chilling better than hawthorn pupae; also, apple-fly adults should eclose later, and by inference have deeper pupal diapauses, than hawthorn flies when the races are reared under controlled conditions; (3) earlier eclosing adults in both host races should have higher frequencies of the alleles Me 100, Acon-2 95, Mpi 37, Aat-2 +75, Dia-2 100, and Had 100; (4) long chilling periods should select against the aforementioned allozymes.
The results from the overwinter experiment supported all four predictions of the diapause trade-off hypothesis. First, adult survivorship was highest when pupae were chilled for 15-19 weeks [ILLUSTRATION FOR FIGURE 1A OMITTED]. This suggests that chilling periods shorter than 15 weeks do not provide adequate time for all pupae to terminate diapause. Chilling periods longer than 19 weeks, however, probably deplete the energy reserves of pupae below levels needed for morphogenesis.
Second, pupal survivorship was higher in the apple than the hawthorn race following extended chilling periods. Apple flies also eclosed significantly later than hawthorn flies in the [less than or equal to] 19 week treatments [ILLUSTRATION FOR FIGURE 1B OMITTED]. The reason for these findings was that a greater proportion of apple adults eclosed in the second than the first mode of flies compared to the hawthorn race [ILLUSTRATION FOR FIGURE 6 OMITTED]. Apple and hawthorn flies experienced similar environmental conditions in the overwinter experiment. The main difference was that apple-fly larvae were exposed to higher temperatures and longer photoperiods during the period before infested fruits were collected and transported back to the laboratory. These conditions are known to hasten diapause termination in R. pomonella (Prokopy 1968). Therefore, if 15-19 weeks of chilling is reflective of nature, then hawthorn pupae at the Ferris Road site developed earlier (broke diapause sooner) than apple pupae under standardized rearing conditions.
Third, the allozymes displaying host-associated differentiation were significantly related to adult eclosion. In both races, earlier eclosing adults tended to have higher frequencies of alleles more common to the hawthorn race at the latitude of the Grant, Michigan, site (Table 1; [ILLUSTRATION FOR FIGURES 3, 4 OMITTED]; Feder and Bush 1989). This concurs with our working hypothesis that there is a genetic basis for the hawthorn race developing earlier than the apple race, at least for the allozymes we studied.
Fourth, and finally, Me 100, Acon-2 95, Mpi 37, and Had 100 frequencies significantly declined in apple and hawthorn flies surviving successively longer chilling periods [ILLUSTRATION FOR FIGURE 5 OMITTED], while Aat-2 +75 and Dia-2 100 showed a sex-specific effect in the hawthorn race. This again implies that there is a genetic basis for apple pupae developing slower than hawthorn pupae, as predicted by the diapause trade-off hypothesis.
At least three factors were responsible for the declines in allozyme frequencies observed in the overwinter experiment. (1) Up until the 15-19-week treatment, an increasing percentage of second relative to first mode flies eclosed as the chilling period was increased [ILLUSTRATION FOR FIGURE 6 OMITTED]. The second mode of eclosing flies had significantly lower Me 100, Acon-2 95, Mpi 37, and Had 100 frequencies than the first mode [ILLUSTRATION FOR FIGURE 4 OMITTED]. Consequently, as greater proportions of second mode flies eclosed, the frequencies of Me 100, Acon-2 95, Mpi 37, and Had 100 fell. (2) The first mode of flies was eliminated from the population after 19 weeks of cold storage [ILLUSTRATION FOR FIGURE 2 OMITTED]. Since these flies had high frequencies of Me 100, Acon-2 95, Mpi 37, and Had 100 [ILLUSTRATION FOR FIGURE 4 OMITTED], these allozymes decreased in flies after the 19-week treatment. (3) Survivorship percentages and mean eclosion times decreased for the second mode of flies after 19 weeks [ILLUSTRATION FOR FIGURES 1A, B OMITTED]. Allele frequencies for Me 100, Acon-2 95, Mpi 37, and Had 100 were also correlated with adult eclosion times in the [greater than or equal to] 22-week treatments (Table 1, [ILLUSTRATION FOR FIGURE 4 OMITTED]). Consequently, the differential culling of earlier eclosing flies in the second mode after 19 weeks of chilling caused allozyme frequencies to continue to decline throughout the course of the experiment.
The bimodal emergence curves are interesting because they suggest that overwintering pupae exist in one of two different developmental states, one class being comprised of pupae in a true diapause and the other class being "nondiapausing" (i.e., pupae capable of immediately undergoing morphogenesis given permissive temperatures; see Fig. 7 for a summary diagram of this model). We assert that the first mode of eclosing flies were nondiapausing pupae. We base this claim on the observation that they took an average of 32 days [+ or -] 0.1 SE to eclose following removal from the cold (n = 1009 for both host races combined). Nondiapausing laboratory strains of R. pomonella, when exposed to continuous heating at 21 [degrees] C with a 14:10 h L:D cycle, also take around 30 days to develop from pupae into adults (range 25-40 days; Feder, unpubl. data). This suggests that 30 days is close to the minimum time required for flies to complete development. We hypothesize that the first mode of nondiapausing flies died in the overwintering experiment when they exhausted their energy reserves after 19 weeks in the cold [ILLUSTRATION FOR FIGURE 7 OMITTED]. The second mode of flies were diapausing pupae [ILLUSTRATION FOR FIGURE 7 OMITTED]. They survived much better than the nondiapausing flies following prolonged chilling because, being in a true state of diapause, they did not expend as much metabolic energy during cold storage. Nevertheless, many of these pupae eventually did break diapause during the course of the overwinter experiment. Once these flies ended diapause, they too had to experience permissive temperatures within a set amount of time to complete adult development or they would die [ILLUSTRATION FOR FIGURE 7 OMITTED]. Flies possessing the alleles Me 100, Acon-2 95, Mpi 37, and Had 100 had a greater probability of entering winter in a nondiapausing state and, if they were diapausing, tended to break diapause sooner than others [ILLUSTRATION FOR FIGURE 7 OMITTED]. These flies were favored during short-winter treatments because they required little or no chilling to complete development. However, they were selected against during long winters, resulting in the declines in the frequencies of Me 100, Acon-2 95, Mpi 37, and Had 100 in the overwinter experiment [ILLUSTRATION FOR FIGURES 5, 7 OMITTED]. Of course, final verification of this bimodal diapause model requires further physiological testing and genetic crosses to confirm that a significant proportion of overwintering pupae are nondiapausing and to establish a direct link between the allozyme markers and diapauserelated traits in the fly.
Another noteworthy finding from the overwinter experiment is that nondiapausing flies are not immediately killed by cold weather (or at least by winter as simulated in a refrigerator). Rather, they can survive for periods of up to 19 weeks. This is important for two reasons. First, it shows how pre- and overwintering conditions interact to affect pupal survivorship. Why, one may ask, should winter be important since the absolute period of cold weather experienced by apple and hawthorn pupae at a given site in nature is the same? The answer is that winter poses different developmental challenges to apple and hawthorn flies due to the earlier phenology of apples. Let us suppose that 17 weeks is the optimal overwintering period for hawthorn flies. If these flies were to infest apples, then their development times would be shifted two to three weeks earlier in the season to correspond to the apple phenology. Fast-developing hawthorn flies would lose appreciable amounts of energy before the onset of cold weather. Consequently, a 17-week winter would now be much too long for these pupae and would kill a large proportion of them. The length of the prewinter period therefore dictates the developmental state and energy reserves that pupae carry into winter, but it does not necessarily kill them. It is the nature of the ensuing winter that determines which pupae will survive and which will die.
The second reason is that the onset of mortality after 19 weeks of chilling also helps to explain why we failed to detect a genetic response to short prewinter treatments in the prewinter experiment. Recall that one prediction of the diapause trade-off hypothesis that was apparently not realized in the prewinter experiment was that brief prewinter treatments did not select for the allozymes Me 100, Acon-2 95, Mpi 37, Aat2 +75, Dia-2 100, and Had 100. The basis for this prediction was that slow-developing flies would not enter pupal diapause quickly enough before the onset of winter and die. But in the prewinter experiment we chilled pupae for 30 weeks at 4 [degrees] C. The overwinter experiment suggests that a 30-week chilling period is excessively long and kills most fast-developing flies [ILLUSTRATION FOR FIGURE 1A OMITTED]. Such an overwintering period would therefore counteract any selection imposed by short prechilling periods. Consequently, allozyme frequencies in flies surviving brief prechilling treatments in the prewinter experiment were not significantly different from those of untreated control flies (Feder et al. 1997). If, however, we had used a shorter overwintering period in the study, such as 13 weeks, then we would probably have observed a significant genetic response in the predicted direction.
One paradox remains, however, which has to do with the allele frequency clines in R. pomonella. Allozyme frequencies for Me 100, Acon-2 95, Mpi 37, Aat-2 +75, Dia-2 100, and Had 100 are higher in fly populations from northern than from southern latitudes (Feder and Bush 1989; Feder et al. 1990a; Berlocher and McPheron 1996). This relationship makes sense with respect to the prewinter period, where colder summer and fall temperatures in the North should favor faster developing pupae. But winters are also longer in the North. This should select against the allozymes Me 100, Acon-2 95, Mpi 37, Aat-2 +75, Dia-2 100, and Had 100.
One possible explanation for the clines centers on the 4 [degrees] C temperature used in the overwinter experiment. This temperature is fairly close to the 6.4 [degrees] C lower threshold for adult development in R. pomonella (Reissig et al. 1979). Pupae from the Ferris site, which is in the northern portion of R. pomonella's distribution, may therefore have had to endure higher winter temperatures in the overwinter experiment than they would normally face in nature. In effect, we challenged flies from a northern population with over-wintering temperatures they would face in the South. Consequently, longer chilling periods selected against pupae possessing the alleles Me 100, Acon-2 95, Mpi 37, and Had 100, "chilling" being something of a misnomer in the overwinter experiment. If we had used 0 [degrees] C instead, then the results may have been different.
A second possibility is that winter temperatures have little to do with the clines; It does not matter to a pupa whether it is 4 [degrees] C or 0 [degrees] C, as long as it is colder than the developmental threshold. If this is true then our results should be taken at face value; it is simply the length of time below 6.4 [degrees] C that is selecting on pupae. The implication of this is that the allozyme clines exist despite, not because of, longer winters in the North. It is the shortness of the growing season that is primarily responsible for the higher frequencies of Me 100, Acon-2 95, Mpi 37, Aat-2 +75, Dia-2 100, and Had 100 in northern latitudes. Long winters act as a counterbalance to this selection and help explain why several "slow-development" alleles can still be found in moderate frequencies even in the northernmost fly populations (Feder and Bush 1989).
A third possibility is that winter may not be significantly longer in the North. Although this seems counterintuitive, few data exist on winter soil temperatures. It is therefore conceivable that at one to two inches beneath the soil surface (the depth at which most pupae overwinter), temperatures remain below 6.4 [degrees] C for roughly the same amount of time across much of the range where apple and hawthorn flies coexist in the northeastern United States and Canada. Consequently, winter length may not be a key variable affecting the allozyme clines.
It remains to be seen which view (or combination of views) for the clines is correct. Additional studies are needed in which we measure soil temperatures in nature and expose pupae to a range of chilling temperatures before we have a complete picture of how winter affects geographic variation in R. pomonella.
One aspect of the diapause trade-off hypothesis that we have yet to fully explore is the timing of adult eclosion. Our discussion of phenology has concentrated on the metabolic dynamics of diapause termination during winter and, in particular, on the depletion of energy reserves needed for morphogenesis. But another important aspect of diapause termination is the synchrony between adult eclosion and host fruit availability If the diapause trade-off hypothesis is correct and apple flies develop more slowly than hawthorn flies, then [F.sub.1] "hybrids" between the races should display intermediate eclosion phenotypes under controlled laboratory conditions, but eclose at the extremes of the apple- and hawthorn-fly distributions in nature The latter is true because hybrids that infest apples will have genotypes causing them to terminate diapause earlier than apple flies, while hybrids that infest hawthorns will have genotypes that cause them to develop later than hawthorn flies. The repercussions of this are potentially disastrous for hybrids, as they would eclose too early in the season to optimally utilize apples and mate with a majority of apple flies (provided that they survive winter) and too late to attack hawthorns. Further work on this component of host-associated isolation is still needed for R. pomonella.
In conclusion, the pre- and overwinter experiments show that it is unwise to focus exclusively on larval feeding performance and plant secondary compounds if one is interested in detecting fitness trade-offs for phytophagous insects Yes, it can be critical for an insect to feed efficiently while it is on its host plant. But the life-history of the insect must also be coordinated with the phenology of its host plant, especially if the insect is univoltine like R. pomonella. Host plants are temporal resource islands. Developing too rapidly or slowly in one life-history stage may upset the timing of other stages and make the island uninhabitable. Failure to properly consider the interplay of development with host phenology and climatic conditions across the entirety of an insect's life cycle and geographic range can help explain why there are so few documented examples of fitness trade-offs (Rausher 1988, 1992). The current study is particularly relevant because it not only highlights the intricacies of host-associated tradeoffs, but also shows how postmating reproductive isolation can evolve in sympatry as a pleiotropic byproduct of insect populations adapting to phenologically diverged plants. We therefore feel that it is too early to draw any firm conclusions about nonallopatric (sympatric) speciation based on the current paucity of reported fitness trade-offs in phytophagous insects
We would like to thank G. Williams, E Wang, A. Wang, and E. Wang, M. Kreitman, N. Pierce, A. Berry, D. Atkins, E. Wong, C. Chen, K. Ardlie, H. Akashi, M. Taylor, J. MacDonald, K. Filchak, S. Berlocher, G. Bush, J. Smith, the Hansens, the Nelsons, the Riders, and two anonymous reviewers for their help on this project and/or comments on this manuscript. Also special thanks to J. Fry for alerting us to the work of Morris and Fulton on Hyphantria. This work was supported, in part, by grants from the Alfred P. Sloan Foundation, the USDA (Agri-94-37302-0540) and National Science Foundation (DEB 95-08559) to JLF.
BECK, S. D. 1968. Insect photoperiodism. Academic Press, New York.
BERLOCHER, S. H., AND B. A. McPHERON. 1996. Population structure of Rhagoletis pomonella, the apple maggot fly. Heredity 77: 83-99.
BERLOCHER, S. H. AND D.C. SMITH. 1983. Segregation and mapping of allozymes of the apple maggot fly. J. Hered. 74:337-340.
BUSH, G. L. 1969a. Sympatric host race formation and speciation in frugivorous flies of the genus Rhagoletis (Diptera: Tephritidae). Evolution 23:237-251.
----- 1969b. Mating behavior, host specificity, and the ecological significance of sibling species in frugivorous flies of the genus Rhagoletis (Diptera: Tephritidae). Am. Nat. 103:669-672.
----- 1975. Sympatric speciation in phytophagous parasitic insects. Pp. 187-206 in P. W. Price, ed. Evolutionary strategies of parasitic insects and mites. Plenum, New York.
----- 1992. Host race formation and sympatric speciation in Rhagoletis fruit flies (Diptera: Tephritidae). Psyche 99:335-358.
BUTLIN, R. K. 1987. A new approach to sympatric speciation. Trends Ecol. Evol. 2:310-311.
CAESAR, L., AND W. A. Ross. 1919. The apple maggot. Ontario Dept. Agric. Bull. 275.
DANILEVSKY, A. S. 1965. Photoperiodism and seasonal development of insects (English transl.). Oliver and Boyd Publishers, London.
DENLINGER, D. L. 1972. Seasonal phenology of diapause in the flesh fly Sarcophaga bullata. Ann. Entomol. Soc. Am. 65:410-414.
DETHIER, V. G. 1954. Evolution of feeding preference in phytophagous insects. Evolution 8:33-54.
DIRKS, C. O. 1935. Larval production and adult emergence of the apple fruit fly in relation to apple varieties. J. Econ. Entomol. 28:198-203.
FEDER, J. L. 1995. The effects of parasitoids on sympatric host races of the apple maggot fly, Rhagoletis pomonella (Diptera: Tephritidae). Ecology 76:801-813.
FEDER, J. L., AND G. L. BUSH. 1989. Gene Frequency Clines for Host Races of Rhagoletis pomonella (Diptera: Tephritidae) in the Midwestern Unites States. Heredity 63:245-266.
-----. 1991. Genetic variation among apple and hawthorn host races of Rhagoletis pomonella (Diptera: Tephritidae) across an ecological transition zone in the mid-Western United States. Entomol. Exp. Appl. 59:249-265.
FEDER, J. L., C. A. CHILCOTE, AND G. L. BUSH. 1988. Genetic differentiation between sympatric host races of Rhagoletis pomonella. Nature 336:61-64.
-----. 1989. Inheritance and linkage relationships of allozymes in the apple maggot fly. J. Hered. 80:277-283
-----. 1990a. The geographic pattern of genetic differentiation between host associated populations of Rhagoletis pomonella (Diptera: Tephritidae) in the eastern United States and Canada. Evolution 44:570-594
-----. 1990b. Regional, local and micro-geographic allele frequency variation between apple and hawthorn populations of Rhagoletis pomonella in western Michigan Evolution 44:595608.
FEDER, J. L., T. A. HUNT, AND G. L. BUSH. 1993. The effects of climate, host plant phenology and host fidelity on the genetics of apple and hawthorn infesting races of Rhagoletis pomonella. Entomol. Exp. Appl. 69:117-135.
FEDER, J. L., S. OPP, B. WLAZLO, K. REYNOLDS, W. GO, AND S. SPISAK. 1994. Host fidelity is an effective pre-mating barrier between sympatric races of the apple maggot fly. Proc. Nat. Acad. Sci. USA 91:7990-7994
FEDER, J. L., J. B. ROETHELE, B. WLAZLO, AND S. H. BERLOCHER. 1997. Selective maintenance of allozyme differences between sympatric host races of the apple maggot fly. Proc. Nat. Acad. Sci. USA 94:11417-11421.
FRY, J. D. 1990. Trade-offs in fitness on different hosts: evidence from a selection experiment with a phytophagous mite. Am. Nat. 136:569-580.
----- 1993. The "general vigor" problem: can antagonistic pleiotropy be detected when genetic covariances are positive? Evolution 47:327-333.
FUTUYMA, D. J., AND G. MORENO. 1988. The evolution of ecological specialization. Annu. Rev. Ecol. Syst. 19:207-233.
GOULD, F. 1979. Rapid host range evolution in a population of the phytophagous mite Tetranycchus urticae Koch. Evolution 33: 791-802.
HANS, H. 1961. Termination of diapause and continuous laboratory rearing of the sweet clover weevil, Sitona cylindricollis. Entomol. Exp. Appl. 4:41-46.
HODEK, I. 1962. Experimental influencing of the imaginal diapause in Cocinella septempunctata Acta Soc. Entomol. Cesk. 59:297-313.
----- 1971. Termination of adult diapause in Pyrrhocoris apterus in the field Entomol. Exp. Appl. 14:212-222.
----- 1974. Development of diapause in Pyrrhocoris apterus females in the middle period of winter dormancy. Vestn. Cesk. Spol. Zool. 3:161-169.
HOLM, S. 1979. A simple sequentially rejective multiple test procedure Scand. J. Stat. 6:65-70.
JAENIKE, J. 1990. Host specialization in phytophagous insects. Annu. Rev. Ecol. Syst. 21:243-273.
KAMM, J. A. 1972. Environmental influence on reproduction, diapause and morph determination of Anaphothrips obscurus. Environ. Entomol. 1:16-19.
KAROWE, D. N. 1990. Predicting host range evolution: colonization of Coronnilla varia by Collos philodice (Lepidoptera: Pieridae). Evolution 44:1637-1647.
MACKENZIE, A. 1996. A trade-off for host plant utilization in the black bean aphid, Aphis fabae. Evolution 50:155-162.
MASAKI, S. 1965. Geographic variation in the intrinsic incubation period: a physiological cline in the Emma field cricket (Teleogryllus) Bull. Fac. Agric. Hirosaki Univ. 11:59-90.
MAYNARD SMITH, J. 1966. Sympatric speciation. Am. Nat. 100: 637-650.
McLEOD, D. G. R., AND S. D. BECK. 1963. Photoperiodic termination of diapause in an insect. Biol. Bull. 124:84-96.
MCPHERON, B. A., D.C. SMITH AND S. H. BERLOCHER. 1988 Genetic differences between Rhagoletis pomonella host races. Nature 336:64-66
MITTER, C., D. J. FUTUYMA, J. C. SCHNEIDER, AND J. D. HARE. 1979. Genetic variation and host plant relations in a parthenogenic moth. Evolution 33:770-790
MORRIS, R. F., AND W. C. FULTON. 1970. Models for the development and survival of Hypantria cunea in relation to temperature and humidity. Mem. Entomol. Soc. Can. 70:1-60.
PHIPPS, C. R., AND C. O. DIRKS. 1933. Notes on the biology of the apple maggot fly. J. Econ. Entomol. 26:349-358
PORTER, B. A. 1928. The apple maggot U.S. Dept. Agicr. Tech. Bull. 66.
PROKOPY, R. J. 1968. The influence of photoperiod, temperature and food on the initiation of diapause in the apple maggot. Can. Entomol. 100:318-329
PROKOPY, R. J., S. R. DIEHL, AND S.S. COOLEY. 1988. Behavioral evidence for host races in Rhagoletis pomonella flies. Oecologia 76:138-147.
RAUSHER, M.D. 1988. Is coevolution dead? Ecology 69:898-901.
-----. 1992. Natural selection and the evolution of plant-insect interactions. Pp. 20-88 in B. D. Roitberg and M.B. Isman, eds. Insect chemical ecology: an evolutionary approach. Chapman and Hall Publ., New York.
REISSIG, W. H., J. BARNARD, R. W. WEIRES, E. H. GLASS, AND R. W. DEAN. 1979. Prediction of apple maggot fly emergence from thermal unit accumulation. Environ. Entomol. 8:51-54.
RICE, W. R. 1989. Analyzing tables of statistical tests. Evolution 43:223-225.
ROETHELE, J. B., J. L. FEDER, S. H. BERLOCHER, M. E. KREITMAN, AND D. LASHKARI. 1997. Towards a molecular genetic linkage map for the apple maggot fly, Rhagoletis pomonella (Diptera: Tephritidae): a comparison of alternative strategies. Ann. Entomol. Soc. Am. 90:470-479
SMITH, D.C. 1988. Heritable divergence of Rhagoletis pomonella host races by seasonal asynchrony. Nature 336:66-67.
TAUBER, M. J., AND C. A. TAUBER. 1970. Adult diapause in Chrysopa carnea: stages sensitive to photoperiodic induction J. Insect Physiol. 16:2075-2080.
-----. 1973. Seasonal regulation of dormancy in Chrysopa carnea. J. Insect Physiol. 19:1455-1463.
-----. 1974. Thermal accumulations diapause, and oviposition in a conifer-inhabiting predator, Chrysopa harrisii. Can. Entomol. 106:969-978.
-----. 1976. Insect seasonality: diapause maintenance, termination, and postdiapause development. Annu. Rev. Entomol. 21: 81-107.
VIA, S. 1991. The genetic structure of host plant adaptation in a spatial patchwork: demographic variability among reciprocally transplanted pea aphid clones. Evolution 45:827-852.
WALSH, B. D. 1867. The apple-worm and the apple maggot. J. Horticulture 2:338-343.
|Printer friendly Cite/link Email Feedback|
|Title Annotation:||part 2|
|Author:||Feder, Jeffrey L.; Stolz, Uwe; Lewis, Kristin M.; Perry, William J.; Roethele, Joseph B.; Rogers, Al|
|Date:||Dec 1, 1997|
|Previous Article:||Molecular phylogenetic analysis of life-history evolution in asterinid starfish.|
|Next Article:||The genetics of adaptation: the genetic basis of resistance to wasp parasitism in Drosophila melanogaster.|