POPULATION ECOLOGY Texas Is the Overwintering Source of Fall Armyworm in Central Pennsylvania: Implications for Migration Into the Northeastern United States 1,2 3 1 RODNEY N. NAGOSHI, SHELBY FLEISCHER, AND ROBERT L. MEAGHER Environ. Entomol. 38(6): 1546Ð1554 (2009) ABSTRACT Fall armyworm, Spodoptera frugiperda (J. E. Smith) (Lepidoptera: Noctuidae), infes- tations in most of North America arise from annual migrations of populations that overwinter in southern Texas and Florida. Cytochrome oxidase I haplotype proÞles within the fall armyworm corn strain, the subgroup that preferentially infests corn (Zea mays L.), can differentiate the Texas and Florida populations. We use this molecular metric to show that fall armyworms in central Pennsylvania originate from Texas, indicating the existence of a migratory pathway from Texas to the northeastern United States. These results were compared with historical trapping data for fall armyworm and another migratory noctuid, corn earworm Helicoverpa zea (Boddie) (Lepidoptera: Noctuidae), in the Pennsylvania and Maryland corn-producing areas to better deÞne lepidopteran migratory pathways. KEY WORDS Spodoptera frugiperda, cytochrome oxidase, Noctuidae, haplotype The capacity for long distance migration often seen in relationships that might have predictive value. An the family Noctuidae has critical economic implica- example of this is the positive correlation observed tions because it is a major mechanism for the seasonal between corn acreage (the natal host resource) and expansion of pest populations. These species annually the extent of subsequent fall armyworm infestation in reinvade their northern geographic range, presumably cotton in adjacent states (Nagoshi 2009). The infor- to expand host resources, in spatial and temporal pat- mation will also facilitate the testing of aerobiological terns that use dynamic but deÞnable atmospheric models designed to predict migratory movements pathways (Westbrook and Sparks 1986, Isard et al. (Isard et al. 2009). 2005, Westbrook 2008). This is particularly true for Several attempts have been made to describe the species that do not diapause and therefore have a annual migrations of fall armyworm in North America limited temperature range for overwintering. A nota- (reviewed in Nagoshi and Meagher 2008). The most ble example in North America is fall armyworm, Spo- detailed descriptions were inferred from the timing of doptera frugiperda (J. E. Smith), a major pest of corn fall armyworm appearances through trapping and and other crops (Sparks 1979, Hall 1988, Pashley other forms of monitoring (Luginbill 1928, Pair et al. 1988a, Foster 1989). Populations that overwinter in 1986). These indicated both a northward movement southern Florida and Texas are responsible for annual from Texas into Oklahoma and a northeasterly ßow infestations that extend throughout the central and from southern Texas that follows the Gulf of Mexico eastern portions of the United States and southern coastal plain into the Mississippi River Valley, with Canada (Luginbill 1928). Understanding the pattern of population move- appearances in central Tennessee by mid-July, in ments is critical to efforts to control and predict in- southeastern Kansas in late July, and in the Ohio Val- festations by migratory pests. Such knowledge makes ley and Maryland by August and September. Luginbill possible a direct assessment of how conditions at the (1928) also suggested that populations in southern source location can inßuence the timing and severity Florida migrate into north-central Georgia by June, of subsequent infestations and expedite identifying continuing east of the Appalachians into South Caro- lina by July, and perhaps continuing northward along The use of trade, Þrm, or corporation names in this publication is the Atlantic coastal plain. There may also be a north- for the information and convenience of the reader. Such use does not westerly movement from southern Florida that con- constitute an ofÞcial endorsement or approval by the U.S. Department of Agriculture or the Agricultural Research Service of any product or tributes to the infestation in the Mississippi Valley service to the exclusion of others that may be suitable. (Pair et al. 1986). These descriptions are generally 1 Center for Medical, Agricultural and Veterinary Entomology, consistent with movements expected from average USDAÐARS, Gainesville, FL 32604. synoptic meteorological conditions (Rose et al. 1975, 2 Corresponding author, e-mail: [email protected]. 3 Department of Entomology, The Pennsylvania State University, Westbrook and Sparks 1986, Mitchell et al. 1991) and University Park, PA 16802. the geographical distribution of subpopulations that 0046-225X/09/1546Ð1554$04.00/0 ᭧ 2009 Entomological Society of America December 2009 NAGOSHI ET AL.: FALL ARMYWORM MIGRATION 1547 differed with respect to disease or pesticide resistance Table 1. Source locality information of fall armyworm col- (Young 1979, Fuxa 1987, Pitre 1988). lected from pheromone traps placed near corn fields These approaches provide a broad picture of pop- Label Location Date Reference/collector ulation movements but are limited in their resolution. In particular, inferences on the origin of immigrant TX 04-6 Hidalgo Co, TX 2004Ð6 Nagoshi et al. 2008a, b TX 06 San Angelo, TX 2006 Nagoshi et al. 2008b populations become increasingly less accurate as mi- TX 04 College Station, TX 2004 Nagoshi et al. 2008b gratory distance increases. Ultimately, the validation FL 02-3 Miami-Dade Co, FL 2002Ð3 Nagoshi et al. 2007 of projected pathways will require a direct method for FL 04-5 Miami-Dade Co, FL 2004Ð5 Nagoshi et al. 2007 identifying the overwintering origin, for example, by FL 06 FL (various sites) 2006 R.M. PA 01 College Station, PA 2001 Fleischer et al. 2005 artiÞcially tagging and recapturing immigrants or anal- PA 06 Rock Springs, PA 2006 S.F. ysis of genetic or other physical markers that can PA 07 Rock Springs, PA 2007 S.F. distinguish between overwintering populations (see, for example, Showers et al. 1989, Hendrix and Showers 1992). most of the northeast section of the United States. The Fall armyworm can be subdivided into two behav- second objective was to determine whether the data iorally distinct, but morphologically identical, strains from statewide trapping surveys can provide addi- that were initially identiÞed by differences in plant tional information about the direction of migration host distribution, hence their designation as rice strain and the entry point into the state. The PestWatch and corn strain (Pashley et al. 1985, 1987; Pashley 1986, program based at Pennsylvania State University makes 1988b). Several studies have shown that polymor- available multiyear, statewide trapping data for fall phisms in the mitochondrial cytochrome oxidase I armyworm (http://www.pestwatch.psu.edu/). This (COI) gene provides a convenient and accurate resource was used to compare the trapping patterns of marker for strain identity based on correlations with fall armyworm with corn earworm, Helicoverpa zea behavioral differences (Lu and Adang 1996, Levy et al. (Boddie), another noctuid with an overlapping host 2002, Meagher and Gallo-Meagher 2003, Prowell et al. range that also exhibits migratory ßight behavior and 2004). The corn strain population as deÞned by the nocturnal transport (Hendrix et al. 1987, Lingren et al. COI marker can be further subdivided into four hap- 1993, Westbrook et al. 1995, Westbrook et al. 1998). lotype subgroups designated as CS-h1, CS-h2, CS-h3, The implications of these observations on our under- and CS-h4 (Nagoshi et al. 2007). Surveys of popula- standing of fall armyworm migration and the useful- tions from Brazil, Texas, and Florida show that all four ness of extensive historical pest survey data to inves- subgroups are present in each area, but there are tigate migration are discussed. consistent differences in their relative proportions. This is most evident when observing the ratio of the Materials and Methods CS-h4 to CS-h2 haplotype proportions. Analysis of Florida corn strain populations over a 4-yr period and Specimen Collections and Sites. Fall armyworm over several locations spanning the state showed a specimens were obtained at several locations in the CS-h4/CS-h2 ratio that was consistently Ͼ1.0 (Na- United States (Table 1). In the Florida collections goshi et al. 2007). In contrast, populations from Brazil from 2006 and 2007, adult males were collected using and Texas, again sampled over multiple years and pheromone traps as previously described (Meagher locations, consistently showed a ratio Ͻ0.5 (Nagoshi 2001). Standard plastic Universal moth traps (Uni- et al. 2007, 2008b). Because these haplotypes are de- traps) were baited with a commercially available fall Þned by synonymous single nucleotide polymor- armyworm pheromone (Suterra, Bend, OR) and con- phisms, it is reasonable to assume that they can serve tained insecticide strips (Hercon Environmental, as neutral markers to identify a Florida or Texas origin Emigsville, PA). After collection, specimens were typ- without the complications of selective pressure. As a ically stored at Ϫ20ЊC. In 2006 and 2007, the Pennsyl- proof of concept, corn strain populations in Georgia, vania samples were collected as above, using the “Fall Alabama, Louisiana, and Mississippi were analyzed for Armyworm-PSU” lure (Scentry Biologicals, Billings, their CS-h4/CS-h2 ratios (Nagoshi et al. 2008b). These MT), which is the two-component lure described