Mexican Fruit Fly (Diptera: Tephritidae) and the Phenology of its Native Host Plant Yellow Chapote (Rutaceae) in Mexico
In northeastern Mexico the Mexican fruit fly, Anastrepha ludens (Loew), breeds on its native host, yellow chapote, Casimiroa greggii (Wats.), which typically produces fruit in the spring. Peak populations of the fly occur in late spring or early summer when adults emerge from the generation of larvae which developed in the spring crop. The parents of the spring generation are flies attracted into the groves of yellow chapote by the presence of fruit. Thus, the reproductive phenology of the fly population depends on the reproductive phenology of the host plant. The reproductive phenology of the host plant follows ambient conditions, in particular, the timing and quantity of rainfall during the preflowering and fruiting season. A prolonged fruiting period can result in production of 2 successive generations of the fruit fly. Infestation rate of the fruit is inversely proportional to the size of the crop; a predictable outcome of the synchronized mass fruiting paradigm (environmental saturation) as a strategy to escape or limit depredation. A strategy to protect a fruit crop might be more effective by targeting the early ovipositing females rather than the succeeding larger peaks of adults.Abstract
The Mexican fruit fly, Anastrepha ludens (Loew) (Diptera: Tephritidae), is a major pest of tree fruits, especially citrus, in the Republic of Mexico (Aluja and Liedo 1986). It is an invasive species that triggers quarantines and loss of markets when infestations are detected in the United States or the “fly-free zones” of Mexico (Mangan et al. 1997). Detections of this pest along the Texas border with Mexico have a distinct periodicity with adults being trapped almost exclusively in the springtime (Nilhake et al. 1991, Thomas et al. 1999). By contrast, in the mountains of northeastern Mexico where this insect is native, population peaks occur at 3 times of the year – spring, summer and fall – although usually only two peaks are found in any given year (Thomas 2003). The difference in the timing of the major peaks from one year to the next is presumably dependent on environmental factors including weather and the availability of its preferred fruit hosts. Such factors can affect the number of generations, but also account for a lag between events related to the life cycle of the pest. Although a paucity of empirical evidence exists, among local citrus growers it is widely suspected that the yellow chapote serves as a sylvatic reservoir from which the pest spreads into the commercial fruit. Moreover, some researchers (Flitters 1964, Williamson and Hart 1989) have attributed the Texas infestations to an influx from northeastern Mexico. Thus, an understanding of the timing of population peaks in those areas where the insect occurs naturally might allow a more efficient strategy of surveillance and response in the regions of the United States and Mexico where fruit crops are at risk (Mangan and Moreno 2002).
The phenology studies described herein were conducted in the canyons and premontane foothills of the Sierra Madre of Nuevo Leon, Mexico, the native habitat to which this fly is adapted and where its populations are not subject to control efforts. Yellow chapote trees (Casimiroa greggi Wats.[Rutaceae]) are dominant components of the riparian gallery forests following the canyon bottoms and outflowing streams of the eastern slope of the Sierra Madre Oriental in northeastern Mexico. The trees once occurred in south Texas in scattered locations along the Rio Grande, but these were eradicated when the fruit was discovered to be a Mexfly host (Plummer et al. 1941, Baker et al. 1944).
The objective of this study was to measure the timing and amplitude of activity levels in the fruit fly population in relation to the phenology of the host plant. For that purpose, both larval and adult populations were censused simultaneously with monitoring of the flowering and fruiting sequence of the yellow chapote. Concurrently, weather data were recorded to identify correlations among environmental factors, plant phenology, and peaks in fly abundance.
Materials and Methods
Study sites.Two sites were selected for study, 18 km apart, in the region west of the town of Linares, Nuevo Leon, where the yellow chapote trees were large (3 - 7 m high) and abundant (Fig. 1). One was a montane canyon site (24°44′N; 99°50′W) on the Santa Rosa river at an elevation of 1050 m where the surrounding vegetation is pine-oak chaparral. The second site (24°44′N; 99°40′W), was on the lower alluvial plain at elevation 500 m along a piedmont stream, the Rio Pablillo. Here, the surrounding vegetation was mainly Tamaulipan thorn scrub (Rzedowski 1978) with scattered groves of citrus, mainly Valencia oranges, in the near vicinity.
Citation: Journal of Entomological Science 47, 1; 10.18474/0749-8004-47.1.1
The study was conducted over 2 consecutive years. Prior to the beginning of the study 10 trees at each site were selected as “witness” trees. Thereafter, the sites were visited weekly and the phenology of each individual witness tree was recorded. Yellow chapotes are nondeciduous trees with leaves persistent the entire year (Standley 1926). The basic sequence of flowering, fruiting, and fruit fall was monitored with the number of witness trees in each stage recorded at each site on a weekly basis.
A weather station with recording hygrothermograph and rain gauge was maintained at each site. Yellow chapote normally produces flowers and fruit in the springtime, but with favorable rains and temperatures, can produce off-season fruit in the fall. And, although this phenomenon has been observed in other years, it did not occur during the 2 years of this study.
The weather-monitoring instruments were deployed the second week of January of the first year, so that the third week of January was the first with on-site weather records. Phenology of the host trees was recorded through the end of the fruiting season in August of both years. Therefore, for comparison between sites and years, the cumulative totals of the weather variables, including rainfall and temperatures, covered the period from wk 3 to wk 35 (wk 1 being the first wk of January of each year) of both years at both sites. Degree-days were calculated using the standard U.S. Weather Bureau method, also known as the means method (Pruess 1983), where,
The base was 10°C as established for development of A. ludens following Leyva-Vazquez (1988).
To monitor the adult population, a plastic MultiLure trap (Better World Manufacturing, Inc., Fresno, CA) baited with BioLure MFF (packets of ammonium acetate and putrescine) (Suterra LLC, Bend, OR) with dilute (10%) propylene glycol based antifreeze (Low Tox, Prestone Products Corp., Danbury CT) as the capture liquid, was hung in the lower branches of each witness tree. The traps were serviced weekly with the lures replaced monthly. All captured females were dissected to determine reproductive status.
To monitor the larval population, 100 fruits were collected weekly during the fruiting season, either from the witness trees or from the ground under the canopy, or both, depending on availability at each site. Only mature (full-sized), externally sound fruit were collected. To determine infestation rate, the fruit was placed in screen-bottom trays (Fig. 2) kept under a tarp for shade, but otherwise ambient conditions at the study site to allow any infesting larvae to egress naturally from the fruit so they could be collected and counted. Following the natural egression, all fruit was dissected to include refractory larvae in the count.
Citation: Journal of Entomological Science 47, 1; 10.18474/0749-8004-47.1.1
Statistical analysis.Mean infestation rates in fallen and tree fruit were compared by pairwise Student's t-test (Sokol and Rohlf 1973). Correlations between rainfall and adult captures were calculated with least squares linear regression (Sokal and Rohlf 1973). Probabilities were calculated with NCSS statistical software (Kaysville, UT).
Results and Discussion
There were differences in the timing and peak of the fly populations between years and between sites. Whereas the timing of the fly population peaks coincided with the host fruit phenology, the amplitude of the peaks related more to the weather pattern. In previously published studies, investigators generally failed to find close correlations between Mexfly captures and weather variables (Eskafi 1988, Celedonio-Hurtado et al. 1995, Aluja et al. 1996, Montoya et al. 2008). In part this can be explained because the availability of fruit hosts will be the overriding factor influencing recruitment into the population. In that regard the availability of hosts may actually have a negative effect on the capture rate in traps. Because fruit fly traps are baited with a food-based lure, the relative attractiveness of the trap depends on the hunger of the flies and is, thus, strongly influenced by the presence or absence of competing resources. Still, capture rate of adults is a useful index of abundance when considered in the context of other demographic parameters, such as reproductive status and development of the immature stages as detailed in the following sections.
Weather.The graphs in Figs. 3A–D illustrate the prevailing temperature patterns and rainfall events before, during, and after the spring fruiting season in the 4 replicates. The clearest difference between the 2 study sites was in the temperature regimen, e.g., the higher elevation site being decidedly cooler on average. The difference in the mean-maximum temperature between sites over the 28 wks from January to August was 5°C in the first year, and 8°C in the second. Temperatures fell briefly below freezing in the early spring of both years at the higher site but did not reach freezing in either year at the lower site. Late-spring and early-summer daytime highs were routinely in excess of 40°C at the lower site; whereas, the highest temperature recorded at the upper site on any date in either year was 38°C. Not surprisingly, peaks in the Mexfly populations tended to be earlier at the lower site.
Citation: Journal of Entomological Science 47, 1; 10.18474/0749-8004-47.1.1
The temperature difference between the sites is perhaps best reflected in the cumulative DD for each growing season (Table 1). At the higher site, the totals were 1476 and 1667 DD, respectively, for each of the 2 years of the study; whereas, at the lower site the totals were 2519 and 2320 DD over the same respective time periods. The Mexican fruit fly requires approx. 600 DD to develop from egg to adult (Leyva-Vazquez 1988). But, of course, the life cycle cannot begin until host fruits are available.
The distribution of rainfall at the 2 sites was also quite different. Springs are typically dry in northern Mexico with the summer rainy season beginning in June, and that pattern prevailed in 3 of the 4 replicates. The exception was the higher elevation site in the first year of the study. Whereas the typical dry spring occurred at the lower elevation site that year, the upper site had an unusually wet spring with ample rains in February and April followed by the normal summer rains in June. Hence, it was the distribution of rainfall, and not just the total precipitation, that was notably different in this first year. In the second year at both sites, there was ample winter rain in January, but then normal dry conditions prevailed until June. Total precipitation at the upper site was respectively 243 and 275 mm during the 2 growing seasons studied. Conversely, relative extremes in precipitation occurred at the lower site. Near drought conditions prevailed with only 129 mm falling during the first year, but an inundative 481 mm of rain fell during the second year. However, much of the total fell in August, too late to impact fruiting phenology.
Tree phenology.All of the witness trees at both sites flowered in both years of the study with the onset of flowering in early February (Figs. 4A–D). The one exception was at the lower site in the first year where possibly because of the lack of rain the first flowers did not appear until the second week of March, and then on only 2 of the witness trees. But, in spite of the dry conditions, by the fourth week of March all of the trees at this site were in flower. Despite the difference in onset, flowering was completed at both sites in both years by midApril.
Citation: Journal of Entomological Science 47, 1; 10.18474/0749-8004-47.1.1
The induction of flowering has not been studied in yellow chapote but in the other members of the Rutaceae, particularly Citrus, competence to flower depends on a complex of influences, among the most important being a prior period of chill and long nights (Nebauer et al. 2006, Moss 1969). The synchronization of flowering phenology in 3 of the 4 replicates in this study in spite of strong differences in the prevailing temperature conditions between sites suggests that night length may be an important influence on flowering phenology in the yellow chapote as well. The fact that yellow chapote can flower in the offseason following heavy rains, and the delay in flowering in the first study year associated with the lack of significant precipitation that spring, indicates that rainfall is an overriding determinant.
Fruiting phenology was also similar between sites and years as fruiting is necessarily tightly correlated with flowering. Even though flowering was delayed at the lower site in the first year by nearly a month, fruiting was delayed by only 1 - 2 weeks compared with the higher site, likely because of the warmer weather conditions. However, whereas timing of fruiting was relatively uniform, the quantity of fruit produced was much more variable. Some witness trees never set fruit, particularly at the high elevation site where only half of the trees set fruit in either year. In contrast, all of the trees set fruit at the lower site in the first year in spite of the dry conditions, and 80% of the trees set fruit in the second year (Fig. 4C). A mitigating circumstance is that the trees at the lower site, where the slope is gentle, grow in close proximity to the river and are presumably less dependent on rainfall for water. At the upper location, the slope is more severe and the river bed, dry most of the year, is denuded of vegetation. The trees at this site are rooted in the rocky outcrops set back nearly 100 m from the river bed.
Fruit persisted on the trees in all 4 study replicates through the months of April to July with a few fruits still on the trees at the beginning of August. Fruit had ripened and begun falling from the trees in numbers in the third week of May at both sites in both years except for the high elevation site in the first year of the study when ripe fruit began falling the first week of May. Overall, despite the notable differences in weather patterns, fruiting phenology was very similar in all 4 replicates, though the amount of fruit produced was much lower at the upper site in both years (Figs. 4B, 4D).
Larval infestation of chapote fruits.Infestation rates in chapote are shown for the consecutive years in Table 2 and 3. Infestation rates were much higher at the upper elevation site than at the lower elevation site in both years. Sixteen percent of the fruit was infested with A. ludens larvae in the first year of the study at the higher site, increasing to 36.3% in the second year. In contrast, the infestation rate at the lower site in those same years was only 4.3 and 3.4%, respectively. The difference between those combined means, 24.7% and 3.8% respectively, was statistically significant (t = 4.00, df = 39, P = 0.0001) and were for fallen fruit. Infestation rates in tree fruit was, in all cases, less than in fallen fruit, although the difference was only statistically significant at the more heavily infested high elevation site.
Infestation rate as a parameter should be interpreted with caution. Unlike the situation with fleshy commercial fruits, the Mexican fruit fly infests the seed and not the pulp of the chapote (Plummer et al. 1941). Thus, the fruit is only susceptible to infestation when the seed coat is soft, which is at the stage when the fruit is small and green. Secondly, because infested fruit tends to dehisce and fall, infestation rates in tree fruit is inevitably underestimated. Thirdly, although sampling was uniform among years and sites, as noted in the previous section, the total amount of available fruit differed. The lower rate of infestation at the lower site is attributable to the much higher production of fruit compared with the upper site. Similarly, production was notably greater at the upper elevation site in the first year of the study than in the second due to the heavy rains. Consequently, the lower infestation rate that year was probably due to the much larger crop. That is, when fewer fruits are available, infestation pressure is higher. Because predation is often density dependent, synchronized saturation of fruits within a habitat is a classical mechanism to promote seed survival (May 1976).
At both sites, fruits infested with 3rd instars were found from early May until the end of July with only about a 1 wk difference in initiation and termination between years except at the lower site in the first year where the drought conditions delayed fruiting with a concomitant delay in detection of larval infestation by about 2 wks. The peak in infestation occurred in early July of both years at the lower site with a level at about 20%. At the upper site, however, there appeared to be 2 peaks or pulses of infestation, the first in May and the second in July of the first year and then in May and June of the second year. At peak levels infestations reached 70% of the fruit sampled at this site.
Egression of larvae from collected fruit began almost immediately upon collection and continued over about 2 wks. Resulting puparia were held for adult eclosion which occurred in as little as 3 wks following egression to as much as 6 wks later. Survival was also variable, but because holding conditions were artificial, they are not related to the parameters measured in this study. As an example, from the peak of 72 larvae egressing and pupariating from 100 fruit collected the second week of June, 37 adult flies and 23 parasitoid wasps emerged. In addition to parasitism, which occurs mainly to the larval stage (Hernandez-Ortiz et al. 1994), puparia in nature are subject to predation by rodents, ants and other natural enemies (Thomas 1995, Aluja et al. 2005).
Adult activity.More than 7,000 flies were captured in the witness tree traps during the 2-yr study. Whereas there was no substantial difference in the numbers trapped at the lower site between the 2 yrs, at the upper site the numbers differed sharply (Table 4). Nearly 5-fold more flies were captured at the upper site during the first year of the study than during the subsequent year. Notably, there was substantial rainfall at the upper site in April and May of the first year (139 mm) but contrastingly little (13 mm) in the second. Perhaps, more importantly, the numbers of trees bearing fruit during the respective seasons was about double in the first year compared with the succeeding. It is not clear from these data whether rainfall itself or the fruit production was the determining factor, although both are likely to influence respectively survival and reproduction in the fly population. Keeping in mind that trap success depends on activity levels as well as abundance, the correlation between spring rainfall and numbers of flies captured was high (r2 = 0.9264) (Table 5).
Demographics.A notable difference between the 2 sites was in the timing in the adult peaks. At the lower site in both years, there was a marked peak in captures for a brief period (2 - 3 wks) in early June followed by a lesser peak in mid to late-July (Figs. 5A, 5C). Dissections of females revealed that both peaks consisted almost entirely of nulliparous flies. Therefore, these peaks were due to emergence of new flies. In contrast, at the upper site, the numbers of nulliparous flies was distributed over a wider time frame covering most of June and July (Figs. 5B, 5D). Even more notable is that preceding these surges in emergence at both sites and both years, the populations consisted primarily of gravid flies. Moreover, the appearance of these reproductively mature flies in the traps coincided with the availability of fruit at each study site. These data indicate that gravid females are attracted into the groves of chapote trees by the presence of fruit as potential oviposition sites. Thus, at both sites in both years there were detectable numbers of mature flies present (as indicated by dissections) throughout the period when fruit was present; but that immature flies, which account for the larger peaks in abundance, appear about 5 - 6 wks later.
Citation: Journal of Entomological Science 47, 1; 10.18474/0749-8004-47.1.1
Generations.In April of the first year, significant numbers of gravid flies were detected in the traps at the upper site (Fig. 5B). This coincided with the end of flowering and peak in fruiting at this site (Fig. 4B). Throughout the month of May high larval infestations of the fruit both on the trees and on the ground were detected (Table 2). This was followed by very large numbers of adults captured in the traps in June (Table 4). The accumulation of thermal units from the third week in April to the second week of June, calculated at 631 DD (see Table 1), was sufficient to support development of a generation, showing that the peak of flies in June can be associated with the peak of mature adults that moved into the site in April. Leyva-Vazquez (1988) calculated that the completion of development from egg to adult in A. ludens requires approx 600 DD.
At the lower site in the first year of study, there was a pulse of adult captures at the end of January. These were nulliparous flies and this pulse in captures is presumably attributable to the emergence of adults that over-wintered in the immature stages. It is interesting that this pulse occurred only at this site and was detected only in this year. The next pulse of flies appeared in the traps at midMarch (Table 4). The accumulation of thermal units during this interim was only 512 DD, somewhat less than the required 600, and thus it is not clear that the March pulse could be related to the January pulse, especially because there was no fruit for breeding at this site during this time. Rather, one must suspect that the flies had migrated into the area from another source. A third small pulse in adults at the end of April occurred after an accumulation of 648 DD suggesting that this pulse was related to the pulse at midMarch, coinciding with the predicted generation time. The pulse at the end of April was particularly interesting because it consisted largely of reproductively-mature flies. Significantly, this small peak coincided with the beginning of fruiting at this site (Fig. 4A). Between this and the next peak at the beginning of June there was detection of infestation of fruit on the trees (Table 4), and the accumulation of thermal units between these peaks, 833 DD, was more than enough to account for the development of a generation. Hence the peak in early June can be related to the peak of mature flies at the end of April at this site suggesting that the local population had completed 2 life cycles.
In all 4 replicates there was a peak in nulliparous flies in June that could be related to the pulse in gravid flies that moved into the sites at the end of flowering and beginning of fruiting in April. The June peaks were then followed by subsequent peaks in adults in early August. During all 4 replicates, the thermal unit accumulations between the second week of June and the end of July was between 600 - 800 DD (Table 1), more than enough for a life cycle. The persistence of fruit on the trees through the early summer months had been sufficient to allow 2 generations to complete development.
Implications for control.Perhaps the most notable result of this study is that the initial impulse of adult flies captured in traps is due not to an in situ population increase but rather an influx of reproductively mature females seeking oviposition sites. This relationship is reflected in the phenology data showing that the prepeak influx of flies coincides with fruiting and the dissection data showing that the females in this influx are mostly gravid. If one extrapolates to the commercial situation, the implication for pest management is that measures to protect the fruit from infestation should be implemented at this time rather than waiting until a large peak in fly abundance occurs. The large peaks in abundance consist mainly of reproductively immature adults which are recruits to the population from breeding in the host fruit after much of the damage has been done. Control programs against the Mexican fruit fly in both the US and Mexico emphasize mass releases of sterile flies. The application of chemical controls is not generally compatible with a sterile release program. In this case, however, whereas sterile releases would be useful against the new recruits to the population, they are ineffective against mature gravid flies. Hence, the application of toxic baits when the fruit is first susceptible to oviposition might be the more efficacious strategy, with release of sterile flies deployed later against the emergent populations. In this regard, another consideration is that the overflooding ratios that are necessary for SIT to be effective are more easily achieved by the application of a prior knockdown of the population as opposed to massive increases in the numbers of sterile insects released (Steiner et al. 1970, Thomas 2007).
The studies described herein involved a native host under natural, that is, uncontrolled conditions. In light of these results the hypothesis that Texas is infested by an influx from uncontrolled populations south of the border is problematic, in that the large native populations which peak in the late spring to early summer in northern Mexico occur later than the early spring infestations that afflict the citrus growing areas in Texas. The Mexican citrus crop is a different matter. Two crops are harvested in Nuevo Leon: Marrs oranges that mature in the winter and Valencia oranges that are harvested in late spring to early summer. Thus, growers of the latter may be correct in attributing their fruit fly problems to movement from the native host into their groves.
In the US there is no sylvatic reservoir. However, increasingly the commercial groves intercalate with residential areas where dooryard citrus, especially sour oranges (Citrus aurantium L.), ensure that infestation is a threat at all seasons of the year even when, and perhaps especially, when the numbers in traps are not high. Evidence that mature flies will migrate into a grove of maturing fruit from nearby sources suggests that there is an opportunity to apply control measures, such as toxic baits, which could augment and even enhance the efficacy of sterile releases.
Yellow chapote tree at Rio Pablillo, Nuevo Leon, Mexico.
Holding tray with yellow chapote fruit.
Weekly weather data at 2 study sites in Neuvo Leon, Mexico over two consecutive years: rainfall, mean-max, and mean-min temperatures.
Flowering and fruiting phenology of yellow chapote trees at 2 study sites in Nuevo Leon, Mexico over 2 consecutive seasons as percent of trees in each status, by week.
Captures of female Mexflies by week at 2 study sites in Nuevo Leon, Mexico over 2 consecutive seasons indicating proportion of gravid versus nongravid.
Contributor Notes