Seasonal Distribution of Fall Armyworm (Lepidoptera: Noctuidae) Host Strains in Agricultural and Turf Grass Habitats

Exhibit 8b, #57

POPULATION ECOLOGY Seasonal Distribution of Fall Armyworm (Lepidoptera: Noctuidae) Host Strains in Agricultural and Turf Grass Habitats

1 ROD N. NAGOSHI AND ROBERT L. MEAGHER

Center for Medical, Agricultural and Veterinary Entomology, Agricultural Research Service, U.S. Department of Agriculture, Gainesville, FL 32608

Environ. Entomol. 33(4): 881Ð889 (2004) ABSTRACT Male fall armyworm moths[ Spodoptera frugiperda (J. E. Smith)] were captured in pheromone traps over a 16- to 24-mo period in selected sites in southern Florida. Molecular markers were used to determine whether individuals were of one of two host strains (historically designated “rice-strain” and “corn-strain”). Trapsplaced in agricultural areasshoweda population peak in the spring (MarchÐMay) and fall (OctoberÐDecember), with a prolonged decline in numbers in summer (JulyÐOctober) and a smaller reduction in mid-winter (January). The host strain distribution during these periods varied signiÞcantly, suggesting strain-speciÞc and seasonal population patterns. Both strains were captured in substantial numbers during the spring peak, but surprisingly, only the rice-strain showed an increase in capture rates during the fall, despite the presence of sweet corn throughout this period. Trap captures in a sod (turfgrass) farm were composed almost entirely of the rice-strain and showed a bimodal seasonal distribution similar to that seen in the agricultural areas, with peaks in the spring and fall. These results represent the Þrst indication that the two host strains might have substantially different population dynamics in the overwintering agricultural areas of Florida and suggest that the rice-strain is the predominant fall armyworm pest during the fall and winter growing periods. It further indicates that the two strains can display a markedly different response to seasonal environmental cues. The implications of these Þndings on our understanding of fall armyworm migration are discussed.

KEY WORDS fall armyworm, Spodoptera frugiperda, host strains, population dynamics

FALL ARMYWORM, OR Spodoptera frugiperda (J. E. different plant cultivars, including transgenic Bacillus Smith), is periodically a signiÞcant pest of maize, sor- thuringiensis Berliner plants(Pashleyet al. 1987b, Jam- ghum, forage grasses for livestock, turf grasses, rice, janya et al. 1990, Quisenberry and Whitford 1988, cotton, and peanuts, and whose range extends from Adamczyk et al. 1997). Clearly strain-speciÞc differ- southern Canada to much of South America (Luginbill enceshave important ramiÞcationstoward the devel- 1928, Sparks1979). Complicating our understanding opment of effective control strategies, yet virtually of fall armyworm Þeld behavior isthe existenceof two nothing isknown about the behavior of the two strains sympatric and morphologically identical strains in the Þeld. (Pashley et al. 1985). The strains were identiÞed by Fall armyworms are not known to diapause and so molecular polymorphisms and deÞned by their host cannot survive the winters in temperate climes. This plant preferences, hence their designation as “host meansthat the infestationof much of the United States strains” (Pashley 1986). The rice-strain (R-strain) is originatesentirely from populationsoverwintering in typically found on rice and Bermuda grass, whereas southern Florida, southern Texas, and northern Mex- the corn-strain (C-strain) prefers Þeld and sweet corn. ico (Luginbill 1928, Mitchell 1979, Sparks1979, Pair et The two strains can be distinguished by strain-speciÞc al. 1986, Raulston et al. 1986). Population surveys in allozyme variantsand genetic markers(Pashley1989, southern Florida indicate a rise in the fall armyworm Lu et al. 1992, 1994, Lu and Adang 1996, McMichael population in the spring, followed by a rapid and and Prowell 1999). Host strains differ in larval devel- prolonged decline during the summer months (Pair et opment on host plants (Pashley 1988, Pashley et al. al. 1986, Mitchell et al. 1991). Thisdecline corresponds 1995), mating behaviors(Pashleyand Martin 1987, to, and presumably reßects, the northward migration Whitford et al. 1988, Pashley et al. 1992), use of food of thispopulation into northern Florida and southern resources (Veenstra et al. 1995), resistance to certain Georgia in April and May (Snow and Copeland 1969, pesticides (Pashley et al. 1987a), and susceptibility to Greene et al. 1971), which continuesinto the northern states by July and August (Mitchell 1979). It is spec- 1 E-mail: [email protected]ß.edu. ulated that thismigration allowsfall armyworm to

0046-225X/04/0881Ð0889$04.00/0 ᭧ 2004 Entomological Society of America 882 ENVIRONMENTAL ENTOMOLOGY Vol. 33, no. 4 avoid adverseweather conditionsdeleteriousto host Trap-S was place on the west side of a large sod farm plant development and to escape density-dependent (235 ha) in northcentral Collier County (26Њ21.051Ј N, natural enemies(Knipling 1980). However, the envi- 81Њ35.733Ј W). Year-round sod production was com- ronmental factorsthat trigger migratory behavior are posed of primarily St. Augustine grass [Stenotaphrum unknown. secundatum (Walt.) Kuntze], although there were If habitat-speciÞc factors initiate migration, we sections of bahiagrass grown (Paspalum notatum might expect the two host strains to differ in the timing Flugge.). These large turfgrass Þelds were bordered of their northward movement. In particular, the be- by areasof natural vegetation and a nurserycontaining havior of the corn-strain should be inßuenced by the trees and shrubs. planting and harvesting cycle for corn, whereas the Trap-C wasplaced adjacent to a large sweetcorn rice-stain would more likely reßect seasonal changes (Zea mays L.) Þeld in southern Miami-Dade County occurring in grasslands. Some evidence of strain-spe- (25Њ239.524Ј N, 80Њ28.854Ј W). Sweet corn production ciÞc differencesin population dynamicswere ob- in Miami-Dade County averagescloseto 6,000 ha (Li served in Louisiana, with the two strains showing peak et al. 2002a), with a planting season between early pheromone capture ratesat different timesof the year October and late January. Harvest is generally be- (Pashley 1988, Pashley et al. 1992). This could indicate tween January and May, and some growers produce differencesin the timing of the migration of the two two crops per season (OctoberÐDecember and Janu- strains into the tested area. aryÐMay). Planting in thisÞeld occurred in October We are currently studying the distribution of fall and December 2001Ð2003, with harvest in December armyworm strain populations in southern Florida us- and March 2002Ð2004. Many vegetable growersplant ing pheromone traps. Trap captures at three sites were a cover crop of a sorghum ϫ sudangrass hybrid (Sor- examined in particular detail because their respective ghum bicolor L. Moench) from June to September, but plant populationsallowed clear predictionsasto the Þeld adjacent to Trap-C wasallowed to go fallow. which strain should predominate. We found evidence For collectionsin November and December 2003, we for substantial and consistent differences in the sea- wanted to extend the test area and obtain a high sonal population dynamics of the two strains even in number of samples over the sampling period. In ad- collectionsmade from the samelocation. The impli- dition to Trap-C, which waskept in the original loca- cations of these results on our understanding of fall tion, two other trapswere used.Trap-C2 wasplaced armyworm migration and population behavior are dis- within 150 m of Trap-C and adjacent to a sweet corn cussed. Þeld. Trap-C3 (25Њ27.832Ј N, 80Њ29.488Ј W) waslocated Ϸ20 km east of Trap-C and adjacent to a sweet corn Þeld. Trap-T waslocated adjacent to a tomato Þeld and Materials and Methods across the road from a sweet corn Þeld in southern Trap Collection and Sites. We used pheromone Miami-Dade County (25Њ35.223Ј N, 80Њ29.488Ј W). trapsin three southernFlorida locationsto estimate Sweet corn was harvested soon after trap placement seasonal changes in population density during an (February 2002), and the land wascleared for a res- 18-mo period from January 2002 to July 2003. Male idential subdivision. Tomato production in Miami- adult mothswere captured in pheromone trapsplaced Dade County rangesbetween 1,500 and 5,500 ha, de- adjacent to a sweet corn Þeld, a mixed agricultural area pending on season and markets, and fruit is harvested with nearby tomato and sweet corn plantings, and a between December and May (Li et al. 2002b). Data sod farm growing turfgrass. We anticipated that the for thistrap wasnot obtained from 1 June 2002 to 1 July corn-strain would predominate during the growing 2002 because of a defective trap and heavy rains. seasons for the Þrst two locations, whereas the rice- Trapping was discontinued after 8 July 2003 because strain would be the primary population in the sod of continued vandalism. farm. Adult maleswere collected usingpheromone DNA Preparation. Individual adult mothswere ho- trapsaspreviouslydescribed(Meagher and Gallo- mogenized in 1 ml of homogenization buffer (0.03 M Meagher 2003). Standard plastic Unitraps were baited Tris-HCl at pH 8.0, 0.1 M NaCl, 0.2 M sucrose, 0.01 M with a commercially available fall armyworm phero- EDTA at pH 8.0, and 0.5% Triton X-100) in a 5-ml mone (3-component lure; Scenturion, Clinton, WA) Dounce homogenizer using either a hand pestle or a and contained insecticide strips (Hercon Environ- motorized mixer (Nagoshi and Meagher 2003). To mental Co., Emigsville, PA). Traps were placed 22 remove large debris, the homogenate was Þltered January 2002, and collectionswere made at Ϸ2-wk through a 5-ml plastic syringe plugged with cheese intervalsuntil 22 July 2003. This18-mo testperiod cloth (prewet with distilled water) into a 1.5-ml mi- allowed observations over two consecutive spring sea- crofuge tube. The Dounce homogenizer waswashed sonswhenthe northward migration isknown to occur with 800 ␮l of buffer, which wasÞltered and added to and should result in a decline in population numbers the homogenate. Cellsand nuclei were pelleted by in the overwintering areas. Trapped moths were sep- centrifugation at 12,000 ϫ g for 10 min at 4ЊC, and the arated from other species and counted, and a subset supernatant was removed by aspiration. The pellet (10Ð15) was randomly selected for polymerase chain was resuspended in 600 ␮l nuclei buffer (0.01 M Tris- reaction (PCR) analysis. Specimens were stored at HCl at pH 8.0, 0.35 M NaCl, 0.1 M EDTA, and 1% Ϫ20ЊC after collection. N-lauryl sarcosine), and extracted with 400 ␮l phenol- August 2004 NAGOSHI AND MEAGHER:FALL ARMYWORM HOST STRAIN DYNAMICS 883 chloroform (1:1). The supernatant was transferred to 2002, regaining the spring highs in November and a new 1.5-ml tube, precipitated with 400 ␮l isopropa- December (49 mothsper night). There wasa sharp nolfor1hatroom temperature, and centrifuged at but brief decline in January 2003 to an average of 18 12,000 ϫ g for 10 min. The DNA pellet waswashed mothsper night, followed by a rapid increasein Feb- with 70% ethanol and dried. The pellet was resus- ruary 2003 continuing to June. The average during the pended in 50 ␮l of distilled water, followed by puri- period from February to June 2003 was55 mothsper Þcation using DNA Clean and Concentrator-5 col- night. Thisbimodal annual pattern with peaksin the umns(Zymo Research,Orange, CA) according to the spring and fall is consistent with previous studies ex- manufacturerÕs instructions. Each PCR reaction used amining pheromone trapsplaced in Dade and nearby 1 ␮l of the DNA preparation (between 0.1 and 0.5 ␮g). Palm Beach counties(Pair et al. 1986). PCR Analysis. To determine strain identity from Trap-S waslocated at a sodfarm and wassur- individual moths, genomic DNA from a randomly se- rounded by extensive Þelds of turfgrass throughout lected subset (n Ն 10) of the captured samples was the year. Despite the substantial differences in plant tested by PCR for the mitochondrial COI gene re- type, we observed a similar bimodal capture pattern as striction fragment-length polymorphism as described that recorded for Traps-C and -T, with March and previously (Levy et al. 2002, Nagoshi and Meagher April 2002 capturesaveraging 21 mothsper night (Fig. 2003). PCR ampliÞcation of genomic DNA wasper- 1). Thiswasfollowed by a rapid decline during May formed in a 50-␮l reaction mix containing 5 ␮l10ϫ and June and minimal capturesfrom June to October reaction buffer with MgCl2 (Promega, Madison, WI), (an average of two mothsper night). A secondpeak 1 ␮l 10 mM dNTP (New England Biolabs, Beverly, wasobservedin November and December (15 moths MA), 0.5 ␮l20␮M primer mix, 1 ␮l DNA template, and per night average), followed by a mid-January decline 0.5 ␮l TaqDNA polymerase (Promega). AmpliÞcation (2 mothsper night). Capture numbersincreasedagain of the COI gene used primers JM76 (5Ј-GAGCT- in the spring of 2003, with an average of 18 moths per GAATTAGG(G/A)ACTCCAGG-3Ј) and JM77 (5Ј- night observed from February to April. ATCACCTCC(A/T)CCTGCAGGATC-3Ј) and be- Host Strain Population Dynamics. We next deter- gan with an initial incubation at 94ЊC (1 min), mined the proportion of each host strain captured followed by 38 cyclesof 94 ЊC (1 min), 56ЊC (1 min), during the test period by PCR analysis of a strain- 72ЊC (1 min), and a Þnal segment of 72ЊC for 5 min. On speciÞc molecular marker (Fig. 2). The proportion of completion of PCR, 0.5 ␮lofMspI wasadded to each C-strain versus R-strain observed was used to estimate reaction and incubated at 37ЊC for 1 h. Five microliters the numbers of each strain captured in the traps, of gel loading buffer wasadded to each sample,and 20 displayedgraphically asmothsper night (Fig. 3). In ␮l was loaded on a 1.5% agarose gel. The R-strain the sod farm habitat surveyed by Trap-S, the R-strain (mtR) pattern isa 569-bp PCR band, whereasthe predominated throughout the year (Fig. 3A). At no C-strain fragment (mtC) iscut by MspI to produce two time did C-strainextrapolationsexceed Þve mothsper fragmentsof 497 and 72 bp. Primerswere synthesized night, even during the spring and fall periods that had by DNAgency (Malvern, PA). the highest fall armyworm density. From February Typically, Ϸ70Ð90% of the DNA prepsgave useable 2002 to July 2003, only 11% (36/277) of the moths PCR information. In collectionswhere the strainiden- tested were of the C-strain, although at least one titiesof 10 or more samplesweredetermined, the C-strain individual was observed in 14/28 collection number of each strain present per night was extrap- periodstested. olated. Thiswasobtained by Þrstdetermining (by the A different strain distribution was observed with PCR analysis of the representative subset) the per- Traps-C and -T associated with vegetable crops. In the centage of each collection that wasof one strainand period from February to June 2002, 53% (130/243) of multiplying that by the total number of malescol- captured males tested were of the C-strain (Fig. 3B lected during the collection period. The product was and C). A similar result of 44% (103/236) was obtained divided by the number of nightsin that period. during the same period in 2003. This proportion did not change much during the summer depression from July to October, with 35% (48/139) of those tested Results showing the C-strain marker. This indicates that what- Seasonal Capture Pattern. Trap capturesfrom ag- ever wasreducing capture numbersduring thisperiod ricultural areasgrowing sweetcorn (Trap-C) or a was affecting both strains. In comparison, during the mixed crop of sweet corn and tomato (Trap-T) during fall/winter growing season from October to January, the winter and spring seasons indicated that popula- the proportion of C-strain found in the traps declined tion peaks during the spring and fall, separated by an to 13% (32/238). This reßected a substantial increase ϳ3- to 5-mo decline during the summer and a shorter in the R-strain population, whereas C-strain captures depression in late winter (Fig. 1). In the period from remained low (Fig. 3A and B). Therefore, the spring February to the end of June 2002, an average of 58 capture peak represented an approximately equal mix- mothsper night wascaptured from the two traps.This ture of both host strains, whereas the fall peak was wasfollowed by a sharpdecline from June to the end limited to the R-strain. of July. Between 1 July and 15 October, trap captures The absence of an increase in the fall C-strain pop- had declined to an average of seven moths per night. ulation was unexpected. To test whether the same Populationsbegan to increaseby the end of October pattern repeated itself in 2003, more extensive trap- 884 ENVIRONMENTAL ENTOMOLOGY Vol. 33, no. 4

Fig. 1. Seasonal pheromone trap captures at selected locations in southern Florida, 2002Ð2003. Each data point represents total malescaptured divided by the number of eveningsin the capture period. The initial samplingpoint isFebruary 2002 and sampling concluded in July 2003. ping wasperformed in agricultural areasduring No- trap. In each sample period and for every trap, the vember and December 2003, a period in the previous C-strain was a minor component, never comprising year associated with high levels of fall armyworm Ͼ28% of the test group. Overall, a total of 319 males infestation. In November and December 2003, the wasanalyzed, and 283 (89%) were of the R-strain.This areasimmediately adjacent to Trap-C contained veg- wascomparable with the resultsofcontemporaneous etable crops (beans), with a large sweet corn Þeld capturesat Trap-S (93% R-strain),a location where Ͻ100 m away. A second trap (C2) was located within the C-strain was not expected to accumulate (Table 800 m of Trap-C, immediately adjacent to a sweet corn 1). Therefore, asin 2002, the fall armyworm reinfes- Þeld with plantsat variousstagesofdevelopment. A tation of south Florida agricultural areas in fall/winter third trap (C3) wasplaced in another sweetcorn Þeld 2003 was not associated with substantial increases in Ϸ20 km to the east, with most of the plants in the whorl the C-strain population, at least as determined by stage. Consistent with the 2002 pattern, capture num- pheromone trapping. bersin all trapswere high during thisperiod, indicat- Temperature and Rainfall. The pattern of adult ing the presence of large and extensive fall armyworm captureswascompared with weekly averagesof max- populations(Table 1). In fact, Trap-C and -C2 showed imum and minimum daily temperaturesand precipi- the highest capture rates that we recorded over the tation (Fig. 3D and E). The general decline in trap past 2 yr. A minimum of 12 adult males from each capturesbeginning in June occurred coincident with collection was randomly selected and tested for strain the onset of minimum daily temperatures Ͼ20ЊC and identity, resulting in over 70 samples tested from each the onset of the summer rainy season. Trap captures August 2004 NAGOSHI AND MEAGHER:FALL ARMYWORM HOST STRAIN DYNAMICS 885

Fig. 2. Agarose gel displaying the diagnostic strain-speciÞc DNA polymorphism used to distinguish between rice- and corn-strains. DNA from 10 moths collected from Trap-C was individually ampliÞed by PCR using a primer combination speciÞc for a portion of the mitochondrial COI gene that contains a strain-speciÞc MspI site. After digestion with MspI, a single band (569 bp) indicatesthe absenceof the MspI site, a characteristic of the rice-strain (R). Two smaller bands (497 and 72 bp) are produced if the MspI site is present, indicating a corn-strain (C) identity. The gel is stained with ethidium bromide and photographed under UV illumination. of the R-strain did not recover until October, corre- and 2003. Despite the presence of whorl-stage sweet sponding with the decline in average minimum daily corn through most of the fall, the number of trapped temperature to below 20ЊC. The mid-January decline C-strain males did not increase until February, shortly in numberscoincided with a sharpdrop in minimum after the lowest temperatures of the season, with peaks daily temperaturesto 5 ЊC and below. Thiswasfol- in March and April (Fig. 3A). Relatively high numbers lowed by an increase in the number of both strains as of C-strain males continued to be captured until June, the average daily temperature rose in February, al- suggesting that the postharvest cover crop of sor- though the recovery of the C-strain may have oc- ghum/sudangrass might be able to provide a satisfac- curred 1Ð2 wk after that of the R-strain (Fig. 3A and tory habitat to support one to two additional gener- B). ations. Both strains showed minimal capture levels by We anticipated that the density of the C-strain pop- the end of July. Most tomatoes near Trap-T were ulation would largely reßect the agricultural activities planted from September to early January, with harvest in the neighboring Þelds. Sweet corn was planted in being generally completed by May. Thiswasfollowed southern Miami-Dade County (Homestead) near by a sharp decline in trap captures. As with Trap-C, Trap-C from October to mid-December in 2001 and capture numbersfor both strainswereat minimal 2002. Cropswere harvestedbeginning late December, levels by July. These results indicate that trap captures and harvesting was completed by May in both 2002 for both strains declined from their spring peak within 886 ENVIRONMENTAL ENTOMOLOGY Vol. 33, no. 4

Fig. 3. Distribution of host strains collected from pheromone traps compared with average weekly temperatures and weekly rainfall. (AÐC) The strain composition of each collection was determined by PCR and used to calculate the average number of each strain present per night during each collection period. Gray lines with open circles denote R-strain collections. Black lineswith black diamondsdenote C-strainnumbers.(D) Average air temperaturesrecorded for the Homestead,FL area during the testingperiod. Upper curve reßectsthe maximum daily temperature; lower curve reßectsthe minimum daily temperature. The squares denote weekly averages. (E) Weekly rainfall (in millimeters) recorded for the Homestead, FL region during the testing period. The weather data were obtained from the University of Florida Institute of Food and Agricultural Sciences. The initial sampling point is February 2002, and sampling was concluded in July 2003. a month of each other, providing no clear evidence for both the local residentpopulation aswell astransients strain differences in the timing of the northward mi- migrating through the region. Assuch,they can pro- gration after the spring harvest. vide a dynamic picture of short-term changes in adult population in a proscribed area. In addition, trap cap- Discussion turesdirectly measurethe number of mating-recep- Pheromone trapsare a convenient meansof obtain- tive maleswithin range of a pheromone source, ing a “snapshot” of the adult fall armyworm population thereby indicating periodsof high mating potential density at a given time and place (Pair et al. 1986, that could portend subsequent increases in larval in- Mitchell et al. 1991). Captured male adultswill reßect festation. However, there is indirect evidence that August 2004 NAGOSHI AND MEAGHER:FALL ARMYWORM HOST STRAIN DYNAMICS 887

Table 1. Fall armyworm males collected during November and December 2003 in pheromone traps placed in agricultural areas

Trap Collection period Moths/night/trap No. tested No. rice-strain Percent rice-strain C 11/19Ð11/20 165.0 16 15 94 11/21Ð11/25 142.7 27 27 100 11/26Ð12/2 75.7 15 14 93 12/3Ð12/9 36.6 16 15 94 12/10Ð12/14 111.0 18 17 94 C2 11/19Ð11/20 93.0 34 32 94 11/21Ð11/25 154.4 16 15 94 11/26Ð12/2 121.3 37 32 86 12/3Ð12/9 81.9 18 13 72 12/10Ð12/14 138.2 16 14 88 C3 11/19Ð11/20 7.0 14 11 79 11/21Ð11/25 25.3 33 30 91 11/26Ð12/2 17.1 16 13 81 12/3Ð12/9 13.6 14 11 79 12/10Ð12/14 12.3 29 24 83 C1Ð3 11/19Ð12/14 80 (average) 319 283 89% S 11/21Ð11/24 16.6 17 15 88 11/25Ð12/1 9.6 14 13 93 12/2Ð12/8 17.7 32 32 100 12/9Ð12/15 13.8 12 10 83 Total 11/21Ð12/15 14 (average) 75 70 93%

commercial pheromonesmay not be equally attractive after rainfall peaks, whereas intervals of least capture to the two strains, potentially biasing estimates of most frequently occurred 60Ð90 d after periods of least strain proportions (Pashley et al. 1992). We note that rainfall (Raulston et al. 1986). Such a mechanism such a bias, if it exists, should not affect the signiÞcance could explain our observation of a peak in R-strain of our observed geographical or seasonal differences capture in the fall and winter monthsthat followed the in strain capture ratios, because all comparisons were summer rainy season, but is not consistent with the made from samples obtained using the same trapping spring peak in captures that followed the winter dry system. season (Fig. 3). Even more problematic is the behavior The bimodal seasonal distribution of moth captures of the C-strain for which there was no evidence of a in southern Florida has been previously described population increase in the fall resulting from the sum- (Pair et al. 1986, Mitchell et al. 1991). Our data indicate mer rainy season. that for at least 2002 and 2003, this bimodal pattern was Another possibility is that the decline in capture the result of two subpopulations undergoing very dif- numbersduring the year may be related to extremes ferent seasonal behaviors. SpeciÞcally, C-strain males in the daily minimum temperature. Fall armyworm were captured in signiÞcant numbers only during the capturesfor both strainsarelowestwhen the daily Њ spring season, with the fall peak due almost entirely to minimum temperature rises above 20 C (JulyÐOcto- Њ the R-strain. While a multiyear and geographically ber) or below 5 C (January). There isprecedence for broader survey is required to establish the generality correlationsbetween daily temperature and Þeld cap- of these strain-speciÞc patterns, our results indicate ture numbersfor other insects(Butleret al. 1999, that the host strains can differ signiÞcantly in their Cammell and Knight 1992, Scott et al. 2000, Souza and temporal distribution in Florida overwintering habi- Carvalho 2002). Of particular interest are Þeld studies tats, suggesting signiÞcant strain-speciÞc differences indicating that the population of Chrysoperla externa in response to seasonal conditions. (Hagen) in citrusorchardsisnegatively affected by The observation that the sod farm area (Trap-S) increases in the minimum daily temperature (Souza displayed the same bimodal population distribution as and Carvalho 2002). The signiÞcance of this coinci- the agricultural traps (Trap-C and -T) suggests that dence with fall armyworm isunknown but could be this seasonal pattern was not caused by competition explained if an important behavior, such as mating or with the C-strain or to the cycle of sweet corn planting ßight activity, is deleteriously affected by seasonal and harvesting, but rather is a response to some en- temperature extremes. vironmental factor common to the different habitats. When temperaturesand rainfall moderated in the The July-to-October decline in R-strain capture num- fall, the incidence of R-strain captures increased dra- berswasobservedin all three localestested,even matically, comprising the majority of the moths though they differed substantially in plant composi- trapped during thisperiod. Why the C-straindid not tion. This bimodal seasonality had been previously behave similarly is unknown. The most obvious ex- observedand wasattributed to the effectsof the pro- pected inßuence on the C-strain population is the nounced wet-dry seasons characteristic of subtropical pattern of sweet corn planting and harvesting. How- and tropical areason plant quantity and quality (Pair ever, although sweet corn was grown in the agricul- et al. 1986). For example, in studies performed in tural trap areasfrom October to May, C-strainnum- Mexico, high capture ratestended to occur 60Ð90 d bersdid not increaseuntil February. Apparently, the 888 ENVIRONMENTAL ENTOMOLOGY Vol. 33, no. 4 presence of its preferred host plant and environmental from different plant hosts to selected insecticides and conditions conducive to expansion of the R-strain transgenic Bt cotton. J. Cotton Sci. 1: 21Ð28. population were not sufÞcient to stimulate C-strain Butler, L., V. Kondo, E. M. Barrows, and E. C. Townsend. increases during the fall. 1999. Effectsof weather conditionsand trap typeson The increase in moth capture numbers during the sampling for richness and abundance of forest macrolep- fall monthsistypical for southernFlorida and Texas idoptera. Environ. Entomol. 28: 795Ð811. Cammell, M. E., and J. D. Knight. 1992. Effectsof climatic and wasshownto correlate with weather and wind change on the population dynamics of crop pests. Adv. conditionsconducive to southwardmigration (Pair et Ecol. Res. 22: 117Ð162. al. 1986, Pair et al. 1987, Mitchell et al. 1991). Thisled Greene, G. L., M. J. Janes, and F. W. Mead. 1971. Fall ar- to the suggestion of a north-to-south return migration myworm, Spodoptera frugiperda, malescaptured at three to overwintering sites before the winter freeze. How- Florida locationsin trapsbaited with virgin females.Fla. ever, our observations suggest that only the R-strain Entomol. 54: 165Ð166. population increases in the fall, implying strain-spe- Jamjanya, T., S. S. Quisenberry, S. S. Croughan, and R. N. ciÞc migration behavior. If thisisthe case,it leaves Story. 1990. Comparison of bermudagrass lines grown in open the question of what is the source of the C-strain different cultural conditionsand the effect on screening population that suddenly appears in the spring. Some for fall armyworm (Lepidoptera: Noctuidae) resistance. J. Econ. Entomol. 83: 585Ð590. possible explanations include the following: (1) the Knipling, E. F. 1980. Regional management of the fall ar- depleted local population ispoisedfor a rapid expan- myworm-a realistic approach? Fla. Entomol. 63: 468Ð480. sion when favorable conditions (speciÞc to the spring Levy, H. C., A. Garcia-Maruniak, and J. E. Maruniak. 2002. season) arise, (2) trap captures during the fall (for Strain identiÞcation of Spodoptera frugiperda (Lepidop- unknown reasons) signiÞcantly underestimate the lo- tera: Noctuidae) insects and cell line: PCR-RFLP of cy- cal C-strain population, or (3) there is an unidentiÞed tochrome oxidase subunit I gene. Fla. Entomol. 85: 186Ð habitat in southern Florida that can serve as a winter 190. reservoir for the C-strain. Studies to test these possi- Li, Y. C., H. H. Bryan, W. Klassen, M. Lamberts, and T. bilitiesare currently underway. Olczyk. 2002a. Sweet corn production in Miami-Dade In summary, we showed that combining phero- County, Florida. Univ. of Florida, Florida Coop Ext. Serv., IFAS, HS-862, 2 pp. mone-trapping techniqueswith molecular methods Li, Y. C., H. H. Bryan, W. Klassen, M. Lamberts, and T. for determining host strain identity can provide un- Olczyk. 2002b. Tomato production in Miami-Dade expected and important advancesto our understand- County, Florida. Univ. of Florida, Florida Coop Ext. Serv., ing of fall armyworm population behavior. Evidence is IFAS, HS-858, 3 pp. provided that indicatesthat, while the two strainscan Lu, Y. J., and M. J. Adang. 1996. Distinguishing fall army- be sympatric, they also can differ substantially in their worm (Lepidoptera: Noctuidae) strains using a diagnos- response to seasonal changes, varying in both popu- tic mitochondrial DNA marker. Fla. Entomol. 79: 48Ð55. lation numbersand geographic distribution.We are in Lu, Y., M. J. Adang, D. J. Eisenhour, and G. D. Kochert. 1992. the process of additional long-term Þeld studies to Restriction fragment length polymorphism analysis of deÞne other strain-speciÞc differences in population genetic variation in North American populationsof the fall armyworm Spodoptera frugiperda (Lepidoptera: Noc- behavior in the wild and to identify habitatsthat could tuidae). Mol. Ecol. 1: 199Ð208. serve as reservoirs for maintaining populations, par- Lu, Y. J., G. D. Kochert, D. J. Isenhour, and M. J. Adang. 1994. ticularly the C-strain, during the less optimal seasons. Molecular characterization of a strain-speciÞc repeated Such information is essential for the development of DNA sequence in the fall armyworm Spodoptera frugi- areawide strategies to mitigate the overwintering and perda (Lepidoptera: Noctuidae). Insect Mol. Biol. 3: 123Ð migrating capacities of one or both strains. 30. Luginbill, P. 1928. The fall armyworm. U.S. Dep. Agric. Tech. Bull. 34: 1Ð91. Acknowledgments McMichael, M., and D. P. Prowell. 1999. Differencesin am- pliÞed fragment-length polymorphisms in fall armyworm We thank J. Sharp for excellent technical support. We (Lepidoptera: Noctuidae) host strains. Ann. Entomol. thank J. Shapiro (USDA) and G. Nuessly (University of Soc. Am. 92: 175Ð181. Florida) for helpful commentson the manuscript.Voucher Meagher, R. L., and M. Gallo-Meagher. 2003. Identifying specimens (FSCAÐEEÕ04, V01Ð10) were placed in the Mu- host strains of fall armyworm (Lepidoptera: Noctuidae) seum of Entomology, Florida State Collection of Arthropods, in Florida using mitochondrial markers. Fla. Entomol. 86: Florida Department of Agriculture and Consumer Services, 450Ð455. Division of Plant Industry, Gainesville, FL. The use of trade, Mitchell, E. R. 1979. Migration by Spodoptera exigua and S. Þrm, or corporation namesin thispublication isfor the frugiperda, North American style, pp. 386Ð393. In Move- information and convenience of the reader. Such use does ment of highly mobile insects: concepts and methodology not constitute an ofÞcial endorsement or approval by the in research, R. L. Rabb, and G. G. Kennedy, (eds.). Uni- United StatesDepartment of Agriculture or the Agricultural versity Graphics, North Carolina State University, Ra- Research Service of any product or service to the exclusion leigh, NC. of othersthat may be suitable. Mitchell, E. R., J. N. McNeil, J. K. Westbrook, J. F. Silvain, B. Lalanne-Cassou, R. B. Chalfant, S. D. Pair, V. H. Waddill, References Cited A. Sotomayor-Rios, and F. I. Proshold. 1991. Seasonal periodicity of fall armyworm, (Lepidoptera: Noctuidae) Adamczyk, J. J., Jr., J. W. Holloway, B. R. Leonard, and J. B. in the Caribbean Basin and northward to Canada. J. En- Graves. 1997. Susceptibility of fall armyworm collected tomol. Sci. 26: 39Ð50. August 2004 NAGOSHI AND MEAGHER:FALL ARMYWORM HOST STRAIN DYNAMICS 889

Nagoshi, R. N., and R. Meagher. 2003. FR tandem-repeat Pashley, D. P., T. C. Sparks, S. S. Quisenberry, T. Jamjanya, sequence in fall armyworm (Lepidoptera: Noctuidae) and P. F. Dowd. 1987. Two fall armyworm strains feed host strains. Ann. Entomol. Soc. Am. 96: 329Ð335. on corn, rice, and bermudagrass. Louisiana Ag. 30: 8Ð9. Pair, S. D., J. R. Raulston, A. N. Sparks, J. K. Westbrook, and Quisenberry, S. S., and F. Whitford. 1988. Evaluation of G. K. Douce. 1986. Fall armyworm distribution and pop- bermudagrass resistance to fall armyworm (Lepidoptera: ulation dynamics in the southeastern states. Fla. Entomol. Noctuidae): inßuence of host strain and dietary condi- 69: 468Ð487. tioning. J. Econ. Entomol. 81: 1463Ð1468. Pair, S. D., J. R. Raulston, D. R. Rummel, J. K. Westbrook, Raulston, J. R., S. D. Pair, A. N. Sparks, J. G. Loera, F.A.M. W. W. Wolf, A. N. Sparks, and M. F. Schuster. 1987. Pedraza, A. T. Palamon, A. Ortega, J. M. Sanchez Ruiz, Development and production of corn earworm and fall P. C. Marquez, H. A. Rueles, J. M. Perez, R. R. Rodriguez, armyworm in the Texas high plains: evidence for reverse H. R. Carrillo, R. R. Archundia, and R. F. Herrera. 1986. fall migration. Southwest Entomol. 12: 89Ð99. Fall armyworm distribution and population dynamics in Pashley, D. P. 1986. Host-associated genetic differentiation the Texas-Mexico Gulf Coast area. Fla. Entomol. 68: 686Ð in fall armyworm (Lepidoptera: Noctuidae): a sibling 691. species complex? Ann. Entomol. Soc. Am. 79: 898Ð904. Scott, T. W., A. C. Morrison, L. H. Lorenz, G. G. Clark, D. Pashley, D. P. 1988. The current status of fall armyworm Strickman, P. Kittayapong, H. Zhou, and J. D. Edman. host strains. Fla. Entomol. 71: 227Ð234. 2000. Longitudinal studies of Aedes aegypti (Diptera: Cu- Pashley, D. P. 1989. Host-associated differentiation in ar- licidae) in Thailand and Puerto Rico: population dynam- myworms(Lepidoptera: Noctuidae): an allozymic and ics. J. Med. Entomol. 37: 77Ð88. mitochondrial DNA perspective, pp. 103Ð114. In H. D. Snow, J. W., and W. W. Copeland. 1969. Fall armyworm: use Loxdale and J. der Hollander (eds.), Electrophoretic of virgin female trapsto detect malesand to determine studies on agricultural pests. Clarendon Press, Oxford, seasonal distribution. USDA Prod. Res. Rep. 110, p. 9. UK. Souza, B., and C. F. Carvalho. 2002. Population dynamics Pashley, D. P., and J. A. Martin. 1987a. Reproductive in- and seasonal occurrence of adults of Chrysoperla externa compatibility between host strains of the fall armyworm (Hagen, 1861) (Neuroptera: Chrysopidae) in a citrus (Lepidoptera: Noctuidae). Ann. Entomol. Soc. Am. 80: orchard in Southern Brazil. Acta Zool. Acad. Sci. Hun- 731Ð733. garicae. 48: 301Ð310. Pashley, D. P., S. J. Johnson, and A. N. Sparks. 1985. Genetic Sparks, A. N. 1979. A review of the biology of the fall armyworm. Fla. Entomol. 62: 82Ð86. population structure of migratory moths: the fall army- Veenstra, K. H., D. P. Pashley, and J. A. Ottea. 1995. Host- worm (Lepidoptera: Noctuidae). Ann. Entomol. Soc. Am. plant adaptation in fall armyworm host strains: compar- 78: 756Ð762. ison of food consumption, utilization, and detoxication Pashley, D. P., A. M. Hammond, and T. N. Hardy. 1992. enzyme activities. Ann. Entomol. Soc. Am. 88: 80Ð91. Reproductive isolating mechanisms in fall armyworm Whitford, F., S. S. Quisenberry, T. J. Riley, and J. W. Lee. host strains (Lepidoptera: Noctuidae). Ann. Entomol. 1988. Oviposition preference, mating compatibility, and Soc. Am. 85: 400Ð405. development of two fall armyworm strains. Florida En- Pashley, D. P., S. S. Quisenberry, and T. Jamjanya. 1987b. tomol. 71: 234Ð243. Impact of fall armyworm (Lepidoptera: Noctuidae) host strains on the evaluation of bermudagrass resistance. J. Econ. Entomol. 80: 1127Ð1130. Received 14 January 2004; accepted 6 May 2004.