Assessing the climatic potential for epizootics of the gypsy moth fungal pathogen Entomophaga maimaiga in the North Central United States

Assessing the climatic potential for epizootics ofthe gypsy moth fungal pathogen Entomophagamaimaiga in the North Central United States

Nathan W. Siegert, Deborah G. McCullough, Robert C. Venette, Ann E. Hajek, andJeffrey A. Andresen

Abstract: The fungal pathogen Entomophaga maimaiga Humber, Shimazu et Soper has become an important biocontrolfor gypsy moth (Lymantria dispar (L.)) in the northeastern United States and is commonly introduced into new areas withestablished gypsy moth populations. Germination of the fungus is dependent on spring temperature and moisture, but spe-cific conditions associated with epizootics have not been determined. Whether E. maimaiga will be as effective in other re-gions that experience different weather conditions is not yet known. We examined similarity of weather conditionsassociated with 16 documented E. maimaiga epizootics with conditions at 1351 North American locations using theclimate-matching software CLIMEX. Based on CLIMEX’s overall index of climatic similarity, long-term annual climaticpatterns across much of the eastern United States were 60%–80% similar to the conditions associated with epizootics.Monthly weather records from 1971 to 2000 in nine North Central states were examined to compare precipitation and tem-perature with conditions observed during epizootics. Based on climatic averages identified with the documented epizootics,temperature and precipitation conditions in Illinois, Indiana, Iowa, Kentucky, Missouri, and Ohio were more conducive forepizootics than conditions in Minnesota, Wisconsin, and Michigan, which were likely to support E. maimaiga epizooticsin fewer than 6 of the 30 years considered.

Resume : Le champignon pathogene Entomophaga maimaiga Humber, Shimazu et Soper est devenu un important moyende lutte biologique contre la spongieuse (Lymantria dispar (L.)) dans le nord-est des Etats-Unis et est couramment intro-duit dans de nouvelles regions ou sont etablies des populations de spongieuse. La germination du champignon depend desconditions printanieres de temperature et d’humidite mais les conditions specifiquement associees aux epizooties n’ont pasete determinees. On ne sait pas encore si E. maimaiga sera efficace dans d’autres regions ou les conditions meteorologi-ques sont differentes. Nous avons etudie la similitude entre les conditions meteorologiques associees a 16 epizooties docu-mentees d’E. maimaiga et les conditions meteorologiques dans 1351 endroits en Amerique du Nord a l’aide du logiciel decomparaison du climat CLIMEX. Sur la base de l’indice general de similitude du climat de CLIMEX, les patrons annuelsa long terme du climat a travers presque tout l’est des Etats-Unis sont de 60 a 80 % similaires aux conditions associeesaux epizooties. Les donnees meteorologiques mensuelles de 1971 a 2000 dans neuf Etats du centre-nord ont ete examineesen comparant la temperature et les precipitations aux conditions observees lors des epizooties. Sur la base des moyennesobservees lors des epizooties documentees, les conditions de temperature et les precipitations en Illinois, en Indiana, enIowa, au Kentucky, au Missouri et en Ohio etaient plus propices aux epizooties que les conditions au Minnesota, au Wis-consin et au Michigan ou moins de six des 30 annees etudiees ont connu des conditions qui auraient pu entraıner une epi-zootie d’E. maimaiga.

[Traduit par la Redaction]

Introduction

Substantial reductions in gypsy moth, Lymantria dispar(L.) (Lepidoptera: Lymantriidae), defoliation in the north-eastern United States in the last 10–15 years have beenlargely attributed to epizootics of Entomophaga maimaigaHumber, Shimazu et Soper (Zygomycetes: Entomophthor-ales) (Elkinton et al. 1991; Hajek 1999). Epizootics of

E. maimaiga, a fungal pathogen native to Japan, were firstobserved in 1989 in North America (Andreadis and Weseloh1990; Hajek et al. 1990b). In addition to natural spread, E.maimaiga has been introduced into many areas with estab-lished gypsy moth populations (Hajek and Roberts 1991;Smitley et al. 1995; Hajek et al. 1996b, 2005). Entomophagamaimaiga is a desirable biological control agent because ithas few impacts on nontarget organisms (Hajek et al.

Received 21 November 2008. Accepted 10 July 2009. Published on the NRC Research Press Web site at cjfr.nrc.ca on 14 October 2009.

N.W. Siegert1 and D.G. McCullough. Departments of Entomology and Forestry, 243 Natural Sciences Building, Michigan StateUniversity, East Lansing, MI 48824-1115, USA.R.C. Venette. Department of Entomology, 1980 Folwell Avenue, 219 Hodson Hall, University of Minnesota, St. Paul, MN 55108, USA.A.E. Hajek. Department of Entomology, 6126 Comstock Hall, Cornell University, Ithaca, NY 14853-2601, USA.J.A. Andresen. Department of Geography, 116 Geography Building, Michigan State University, East Lansing, MI 48824-1117, USA.

1Corresponding author (e-mail: [email protected]).

1958

Can. J. For. Res. 39: 1958–1970 (2009) doi:10.1139/X09-117 Published by NRC Research Press

1996a, 2000, 2004) and is compatible with other natural en-emies and pathogens, including the gypsy moth nuclear pol-yhedrosis virus (NPV) (Malakar et al. 1999).

Entomophaga maimaiga produces two types of spores,both of which can lead to infection of gypsy moth larvae(Hajek 1999). Germination of soilborne E. maimaiga restingspores (i.e., azygospores) occurs in the spring if environ-mental conditions are suitable (Hajek and Humber 1997;Weseloh and Andreadis 1997). These spores, which producegerm conidia that infect gypsy moth larvae as they movefrom tree to tree, are integral in the early development oflarge-scale epizootics (Hajek et al. 1998b). Infected early-instar cadavers externally produce E. maimaiga conidio-phores that discharge conidia, which typically infect mid- tolate-instar gypsy moth. Late-instar cadavers drop to the soiland decompose, releasing resting spores that remain dormantin the soil until the following spring (Hajek et al. 1998a).

Fungal entomopathogens can be highly efficacious in con-trolling host insect populations (Carruthers and Soper 1987;McCoy et al. 1988), but are generally effective within a nar-row range of environmental conditions (Benz 1987; Hajekand St. Leger 1994; Burges 1998). Large-scale epizootics ofE. maimaiga, where more than 60% infection of gypsy mothlarvae occurs, do not necessarily develop consistently in areaswhere E. maimaiga is established though. Previous studieshave suggested that E. maimaiga is sensitive to abiotic condi-tions, particularly temperature and moisture (Shimazu andSoper 1986; Hajek et al. 1990a; Hajek and Humber 1997),and that weather plays a critical role in the development ofepizootics (Elkinton et al. 1991; Weseloh and Andreadis1992; Hajek et al. 1993). Average meteorological conditions,such as 30 year averages for temperature or precipitation, canbe useful for estimating potential effectiveness of E. mai-maiga introduced into a new area. Specifically, springweather conditions are likely to strongly influence E. mai-maiga resting spore germination, infection of young gypsymoth larvae, and initiation of epizootics (Hajek and Roberts1991). The variability of spring weather from year to year,however, is also likely to affect the long-term viabilityE. maimaiga and its success as a biocontrol of gypsy moth ina local area.

Whether E. maimaiga will play a consistent role in regu-lating gypsy moth populations is an important question, par-ticularly in states along or just beyond the leading-edge ofgypsy moth range expansion. Differences in climate betweenthe northeastern United States, where most E. maimaigastudies have been conducted, and the North Central region,which we define here as the states of Illinois, Indiana,Iowa, Kentucky, Michigan, Minnesota, Missouri, Ohio, andWisconsin, could strongly influence the effect of E. mai-maiga on gypsy moth dynamics in this region. Understand-ing more about how meteorological factors affect E.maimaiga germination and epizootics could help forest man-agers and pest specialists predict the potential effectivenessof E. maimaiga as they develop gypsy moth managementstrategies.

Here we assess the likelihood that E. maimaiga will be aconsistently effective biological control agent for regulatinggypsy moth populations in the North Central region. First,we identified precipitation and temperature conditions asso-ciated with years in which epizootics of E. maimaiga oc-

curred at 16 different sites in six US states and Ontario,Canada. Then we used the climatological software CLI-MEX, which previous researchers have used to estimate thepotential geographic distributions of many exotic pest spe-cies (e.g., Carnegie et al. 2006; Venette and Cohen 2006;Sutherst et al. 2007 and references therein), to compare theepizootic-year weather conditions that occurred at these siteswith long-term weather patterns across North America. Thisallowed us to identify areas likely to frequently experienceconditions suitable for the development of E. maimaiga epi-zootics. Next, we focused on precipitation and temperatureconditions in the nine states within the North Central regionfrom 1971 to 2000, encompassing the western leading edgeof expanding gypsy moth populations (http://www.fs.fed.us/ne/morgantown/4557/gmoth/atlas/). Localized annual andspring (i.e., May and June) weather conditions in the NorthCentral region were compared with weather at the 16 epi-zootic sites during years with documented E. maimaiga epi-zootics. We also examined the temporal variability ofprecipitation and temperature across the North Central re-gion to identify areas likely to consistently experienceweather conducive for the development of E. maimaiga epi-zootics. Finally, we estimated the number of years, from1971 to 2000, in which precipitation and temperature condi-tions in the North Central region would have been suitablefor E. maimaiga epizootics if gypsy moth and E. maimaigahad been established.

Methods

Documented epizooticsSpecific environmental conditions associated with 16

documented E. maimaiga epizootics (Table 1) were usedfor climate comparisons. These 16 epizootic locations withtheir respective year-specific weather data, referred to hereinas ‘‘epizootic-specific sites’’, were selected from the scien-tific literature, because the reported level of E. maimaiga in-fection was greater than 60%, or it could otherwise bediscerned that a large-scale epizootic had occurred. Weatherdata for the years with epizootics were retrieved from theweather stations nearest to each of the 16 locations from theMidwestern, Southeastern, and Northeastern Regional Cli-mate Centers (http://www.wrcc.dri.edu/rcc.html).

Comparisons of temperature and precipitationEnvironmental conditions at the 16 epizootic-specific sites

(Table 1) were individually compared with 1351 locations inNorth America (Fig. 1A) using the climate-matching soft-ware CLIMEX for Windows version 1.1 (CommonwealthScientific and Industrial Research Organization (CSIRO)Publishing, Victoria, Australia) (Sutherst et al. 1999). TheCLIMEX software generates indices of similarity for fiveclimatic parameters at each location compared with a targetlocation(s) and can be used to identify geographic areas withsimilar climatic conditions. Models in CLIMEX use temper-ature and moisture as primary determinants in a species’ bi-ogeography (Sutherst and Maywald 1985; Sutherst et al.1995). The CLIMEX software contains a meteorological da-tabase of approximately 3000 locations worldwide and 300locations in North America (Sutherst et al. 1999). The mete-orological database is composed of 30 year averages for

Siegert et al. 1959

Published by NRC Research Press

Table 1. Summary of Entomophaga maimaiga epizootic locations used in the CLIMEX climate-matching analyses.

Site of epizootic

Source Month

Precipitation (mm) Air temperature (8C)

Year State County Observed30 yearaverage

Departure from 30year average Observed

30 yearaverage

Departure from30 year average

1989 CT Fairfield Andreadis and Weseloh1990

April 80 101 –21 8.8 9.3 –0.5May 242 102 140 15.3 15.0 0.3June 142 90 52 20.3 20.1 0.2July 87 96 –9 22.1 23.4 –1.3

1989 MA Hampshire Hajek and Roberts 1991;Hajek et al. 1998b

April 109 97 12 6.5 8.2 –1.7May 223 105 118 15.6 14.4 1.2June 146 97 49 20.2 19.2 1.0July 97 100 –3 22.1 21.9 0.2

1989 NY Westchester Hajek and Roberts 1991 April 96 112 –16 8.2 8.9 –0.7May 280 123 157 15.0 15.0 0June 183 107 76 20.0 19.7 0.3July 57 118 –61 21.8 22.5 –0.7

1992 VA Augusta Hajek et al. 1996a April 88 72 16 10.6 10.8 –0.2May 140 96 44 13.8 15.8 –2.0June 97 91 6 18.2 20.2 –2.1July 100 97 3 22.8 22.6 0.2

1992 NY Tompkins Hajek 1997; Hajek andHumber 1997

April 84 84 0 5.9 6.7 –0.8May 81 82 –1 12.4 12.8 –0.4June 73 98 –25 16.7 18.1 –1.4July 191 90 101 18.8 20.3 –1.5

1993 MI Lake Smitley et al. 1995 April 141 74 67 4.6 6.4 –1.8May 72 75 –3 12.8 13.3 –0.5June 149 87 62 16.8 17.9 –1.1July 89 70 19 21.4 20.3 1.1

1994 VA Augusta Hajek et al. 1996b April 47 72 –25 12.8 10.7 2.1May 52 96 –44 13.4 15.7 –2.3June 42 91 –49 21.5 20.2 1.3July 96 97 –1 22.8 22.6 0.2

1995 MD Queen Anne’s Hajek et al. 1998a April 55 88 –33 12.0 11.7 0.4May 137 108 29 16.9 16.9 0June 55 102 –47 22.4 21.7 0.7July 61 105 –44 25.8 24.3 1.5

1995 VA Page Hajek et al. 1998b April 44 80 –36 12.1 11.5 0.6May 151 100 51 16.3 16.3 0June 141 101 40 21.3 20.6 0.7July 59 96 –37 24.0 22.9 1.1

1995 VA Rockbridge Webb et al. 1999 April 36 80 –44 12.9 12.9 0.1May 104 100 4 16.9 17.4 –0.5June 432 104 328 20.6 21.6 –1.0July 45 99 –54 23.6 23.8 –0.2

1960C

an.J.

For.R

es.V

ol.39,

2009

Publishedby

NR

CR

esearchPress

Table 1 (concluded).

Site of epizootic

Source Month

Precipitation (mm) Air temperature (8C)

Year State County Observed30 yearaverage

Departure from 30year average Observed

30 yearaverage

Departure from30 year average

1996 NY Tompkins Hajek 1997 April 130 84 46 5.7 6.7 –1.0May 145 82 63 11.9 12.8 –0.9June 96 98 –2 19.4 18.1 0.7July 71 90 –19 19.4 20.3 –0.9

1996 MD Worcester Hajek et al. 1998b April 103 76 27 13.8 13.4 0.5May 96 93 3 17.9 18.1 –0.3June 123 87 36 22.8 22.3 0.5July 223 114 109 24.2 24.6 –0.4

1996 VA Rockbridge Webb et al. 1999 April 65 80 –15 13.3 12.8 0.5May 125 100 25 17.5 17.5 0.1June 110 103 7 22.3 21.6 0.7July 76 99 –23 23.2 23.8 –0.6

1999 MI Clare N.W. Siegert,unpublished data

April 61 64 –3 8.1 6.7 1.4May 60 74 –14 14.5 13.4 1.2June 164 82 82 19.8 18.5 1.4July 166 72 94 21.8 20.9 0.9

2003 Ont. Algoma Villedieu andvan Frankenhuyzen 2004

April 63 65 –2 1.3 3.6 –2.3May 50 64 –14 10.6 10.6 0June 53 76 –23 14.8 15.0 –0.2July 119 80 39 18.3 17.8 0.5

2003 MD Prince George’s Webb et al. 2005 April 65 85 –20 11.2 11.9 –0.7May 193 115 78 15.5 17.2 –1.7June 274 91 183 20.8 22.2 –1.4July 150 104 46 24.6 25.0 –0.4

Average (SE) values for all sites combinedApril 79±8 82±3 –4±8 9.2±0.9 9.5±0.7 –0.3±0.3May 134±17 95±4 40±15 14.8±0.5 15.1±0.5 –0.4±0.3June 143±24 94±2 48±24 19.9±0.6 19.8±0.5 0.04±0.3July 105±13 95±3 10±14 22.3±0.5 22.3±0.5 –0.02±0.2

Note: Following each location are the observed conditions during the epizootic year, 30 year averages, and departures from 30 year averages for precipitation (mm) and air temperature (8C) recorded bythe nearest available weather station for April, May, June, and July.

Siegert

etal.

1961

Publishedby

NR

CR

esearchPress

monthly minimum and maximum air temperature, precipita-tion, and morning and afternoon relative humidity.

Additional long-term meteorological data (30 year aver-ages 1971–2000) from 1051 locations in nine North Centralstates were imported into the standard CLIMEX meteorolog-ical database to more thoroughly represent variability intemperature and precipitation within the region (Fig. 1B).Specifically, meteorological data from 116 locations in Illi-nois, 94 locations in Indiana, 123 locations in Iowa, 72 loca-

tions in Kentucky, 123 locations in Michigan, 140 locationsin Minnesota, 137 locations in Missouri, 104 locations inOhio, and 142 locations in Wisconsin were added to the da-tabase (http://mcc.sws.uiuc.edu/).

For climate-matching analyses, spring and annual weatherthat occurred during the 16 documented epizootics werecompared with 30 year climatic averages at 1351 NorthAmerican locations (Fig. 1A). Climatic comparisons werebased on minimum and maximum air temperatures, total

Fig. 1. Green points mark the North American locations used in the CLIMEX climate-matching analyses (total n = 1351 locations) in (A)North America. Additional locations were added to the 300 location CLIMEX database to better represent climatic variability in (B) theNorth Central region of the United States (n = 1051 locations). Images in this figure are presented in color.

1962 Can. J. For. Res. Vol. 39, 2009

Published by NRC Research Press

precipitation, and monthly precipitation patterns. Relativehumidity data were not available for the additional locationsimported into the CLIMEX database, so this parameter wasexcluded from the climate-matching analyses. The CLIMEXsoftware expresses similarity between two locations as anindex for each climatic parameter. Indices were scaled be-tween 0 and 100, with higher values reflecting greater simi-larity in a given climatic parameter between the twolocations. Sutherst et al. (1999) provide formulas for the cal-culation of similarity indices. The maximum overall similar-ity of a North American location to any one of the 16epizootic-specific sites was used in the climate-matchinganalyses. To accommodate for regional differences in gypsymoth phenology and associated E. maimaiga epizootics, weused April through June weather for epizootic-specific sitessouth of latitude 41.28N (n = 9) and May through Julyweather for epizootic-specific sites north of latitude 41.28N(n = 7) in climate-matching analyses assessing spring cli-matic similarity to epizootic sites. This selection of springmonths corresponded well with the average dates of 50%cumulative gypsy moth egg hatch at epizootic sites (Gray2004) and timing of gypsy moth development and E. mai-maiga epizootics in the literature (Table 1).

The geographically referenced index of climatic similarityfor a given North American location with maximum similar-ity to any one of the epizootic-specific sites was exported tothe ArcView 3.2 geographic information system (Environ-mental Systems Research Institute (ESRI), Redlands, Cali-fornia). Isoclines were generated using the ArcView SpatialAnalyst extension (ESRI).

Variability in weatherFor each of the 16 epizootic-specific sites (Table 1), we

determined departures from 30 year average values for pre-cipitation and temperature conditions. April through Julymonthly precipitation and temperature values during yearswhen epizootics occurred were tested for differences from30 year averages using two-tailed t tests (SYSTAT 2000).

To assess variability in weather across the North Centralregion, we determined the average monthly standard devia-tions from 30 year averages (1971–2000) of precipitationand temperature for each climate division in the North Cen-tral region. A climate division is defined as a climaticallysimilar region within a state (http://lwf.ncdc.noaa.gov/oa/climate/normals/usnormals.html). Divisional climate dataare used for numerous research applications, including as-sessment of large-scale climatic trends over long time peri-ods. Spatial trends in variation of precipitation andtemperature were examined using spring (May and June)and annual (January through December) weather data foreach of the 75 climate divisions in the North Central region.

Favorable years for epizootics in the North Centralregion 1971–2000

Precipitation and temperature data for May and June atthe 16 epizootic-specific sites (Table 1) were examined todetermine the approximate range of environmental condi-tions that were associated with E. maimaiga epizootics.Monthly weather records for the 75 climate divisions innine North Central states from 1971 to 2000 (http://lwf.ncdc.noaa.gov/oa/climate/normals/usnormals.html) were as-

sessed to estimate the number of years during that periodthat may have been favorable for an E. maimaiga epizooticto occur. This exercise was hypothetical because E. mai-maiga and gypsy moth were not established throughout theNorth Central region between 1971 and 2000. When esti-mating the number of years between 1971 and 2000 thatmay have been favorable for an E. maimaiga epizootic tooccur, we assumed no phenotypic variation among isolatesof E. maimaiga and that both the fungus and host werepresent and poised for an epizootic.

The number of years between 1971 and 2000 that mayhave been favorable for an E. maimaiga epizootic to occurin the North Central region were estimated based on theaverage weather conditions associated with epizootic-spe-cific sites (Table 1). A year was considered favorable ifweather conditions met or exceeded the average for a givenmeteorological parameter estimated from the 16 documentedepizootics. Specifically, our criteria were thefollowing: ‡134 mm of total precipitation inMay; ‡143 mm of total precipitation in June; ‡14.8 8C aver-age daily air temperature for May; and ‡19.9 8C averagedaily air temperature for June. Scenarios were estimatedbased on precipitation only, temperature only, and both pre-cipitation and temperature for a given month.

Results

Documented epizooticsThe 16 epizootic-specific sites used in our study occurred

in six US states and the Canadian province of Ontario, andexperienced a range of meteorological conditions. The siteswere not, however, distributed systematically across the cur-rent ranges of gypsy moth and E. maimaiga. Whether the 16sites adequately represent the range of meteorological condi-tions associated with development of E. maimaiga epi-zootics remains to be determined in future studies.Microclimatic site-level weather data combined with E. mai-maiga infection levels from additional sites across the cur-rent and future range of E. maimaiga are needed to morefully understand how interactions of precipitation and tem-perature affect this pathogen.

Precise upper and lower climatic thresholds governing thedevelopment of E. maimaiga epizootics remain unclear, butprecipitation was the meteorological parameter that mostclearly varied from the 30 year averages at the 16 epi-zootic-specific sites (Table 1). For the documented epi-zootics south of latitude 41.28N that occurred April throughJune (n = 9), total precipitation tended to be above averageat most of the sites in May (25–157 mm total precipitationabove the 30 year average) and June (36–328 mm total pre-cipitation above the 30 year average), but below average inApril (15–44 mm total precipitation below the 30 year aver-age). April precipitation was above average at three of thenine sites south of latitude 41.28N (16–27 mm total precipi-tation above the 30 year average). May precipitation wasabout average at two sites (3–4 mm total precipitation abovethe 30 year average) and below average at one site (44 mmtotal precipitation below the 30 year average). June precipi-tation was about average at two sites (6–7 mm total precip-itation above the 30 year average) and below average at twosites (47–49 mm total precipitation below the 30 year aver-

Siegert et al. 1963

Published by NRC Research Press

age). Overall, however, total precipitation in April, May,and June was not significantly different from the 30 yearaverages (P > 0.05) recorded for these sites. Temperaturesat the sites south of latitude 41.28N were not as variable asprecipitation and, overall, departures in temperature did notdiffer significantly from long-term average values in April,May, or June (P > 0.05). Average monthly temperatures atthese sites ranged from 8.2 to 13.8 8C in April, from 13.4to 17.9 8C in May, and from 18.2 to 22.8 8C in June(Table 1).

For the documented epizootics north of latitude 41.28Nthat occurred from May through July (n = 7), total precipita-tion tended to be above average at most of the sites in June(49–82 mm total precipitation above the 30 year average)and July (19–101 mm total precipitation above the 30 yearaverage), with only three of the seven sites with above aver-age values in May (63–140 mm total precipitation below the30 year average). May precipitation was about average attwo sites (1–3 mm total precipitation below the 30 yearaverage) and below average at two sites (14 mm total pre-cipitation below the 30 year average). June precipitationwas about average at one site (2 mm total precipitation be-low the 30 year average) and below average at two sites(23–25 mm total precipitation below the 30 year average).July precipitation was about average at two sites (3–9 mmtotal precipitation below the 30 year average) and belowaverage at one site (19 mm total precipitation below the30 year average). Overall, however, total precipitation inMay, June, and July was not significantly different from the30 year averages (P > 0.05) recorded for these sites. Tem-peratures at the epizootic sites north of latitude 41.28N werealso not as variable as precipitation and, overall, departuresin temperature did not differ significantly from long-termaverage values in May, June, and July (P > 0.05). Averagemonthly temperatures at these sites ranged from 10.6 to15.6 8C in May, from 14.3 to 20.3 8C in June, and from18.3 to 22.1 8C in July (Table 1).

Climate comparisonsBased on climate-matching analyses using 30 year aver-

age precipitation and temperature values, spring climaticconditions (i.e., April–June for epizootics south of latitude41.28N and May–July for epizootics north of latitude41.28N) throughout most of the United States and Canadawere ‡60% similar to at least one of the epizootic-specificsites (Fig. 2A). An area south of the Great Lakes regionthat extended from central Kansas east through West Vir-ginia, and portions of central Minnesota and Wisconsin,were ‡80% similar in overall spring climate. Individual indi-ces of climatic similarity (i.e., similarity indices based solelyon minimum air temperature, maximum air temperature, to-tal precipitation, or precipitation pattern) throughout theNorth Central region were typically ‡80% similar to anyone of the epizootic-specific sites for the spring climatecomparisons. When annual climatic conditions were as-sessed, much of the eastern half of the United States andsoutheastern Canada were 60%–80% similar to any one ofthe epizootic-specific sites (Fig. 2B).

Variability in spring and annual weatherSpring weather experienced at the 16 sites during years

with epizootics was compared with the 30 year average forprecipitation and temperature for each location (Table 1).Average monthly departures in temperature from 30 yearaverages were minimal, typically <0.5 8C for April, May,June, or July. Total precipitation in May and June, however,often exceeded the 30 year average (departures of 40 ±15 mm and 48 ± 24 mm, respectively) (Table 1). For nearly70% of the epizootic-specific sites, May or June precipita-tion exceeded the 30 year average precipitation by morethan 40%. In 13 of the 16 sites, May or June precipitationexceeded the 30 year average precipitation by more than25%.

In the North Central region, variability in spring (Mayand June) precipitation was greatest in the southwestern por-tion of the region, extending from Iowa into Kentucky(Fig. 3A). Monthly precipitation in spring was less variablein northern Minnesota, northern Wisconsin, and most ofMichigan. Variations in annual precipitation over the 30year time period were greatest through Missouri, much of Il-linois, Iowa, Kentucky, and southern Indiana (Fig. 3B). An-nual precipitation was relatively consistent across much ofMinnesota, Wisconsin, and Michigan.

Spring (May and June) temperatures were more variablethroughout most of Minnesota, Iowa, and portions of Wis-consin and Michigan compared with spring temperaturesthroughout the rest of the North Central region (Fig. 3C).Annual deviations in temperature were greatest in the north-western portion of the region, from the Upper Peninsula ofMichigan through Wisconsin and Minnesota to northernIowa (Fig. 3D), while annual temperatures varied little inthe southern portions of the North Central region.

Favorable years for epizootics in the North Centralregion 1971–2000

The estimates of the number of favorable years for epi-zootics that we generated suggest that adequate precipitationfor E. maimaiga epizootics generally occurred in less than athird of the years between 1971 and 2000 in the North Cen-tral region. There was also a gradient of fewer favorableyears from south to north, with northern areas typically re-ceiving adequate precipitation in fewer years than southernareas (Figs. 4A and 4B). This gradient was strongest inMay, with adequate precipitation occurring in 0–6 yearsthroughout Minnesota, Wisconsin, and Michigan, whilemuch of the southern half of the North Central region re-ceived adequate precipitation in 7–12 years. Based onweather observed during years with epizootics at the 16 sites(Table 1), we selected 134 and 143 mm of total precipitationfor May and June, respectively, as the minimum levels con-ducive for epizootics. Adequate precipitation for an E. mai-maiga epizootic occurred in 12 or fewer years throughoutmost of the North Central region, with large portions receiv-ing adequate precipitation in less than 6 of the 30 years ex-amined.

In analyses based on temperature alone, years from 1971to 2000 where daily temperatures averaged at least 14.8 and19.9 8C for May and June, respectively, were assumed to befavorable for the development of E. maimaiga epizootics.When scenarios were based on temperature alone, a strongergradient of favorable years from south to north emergedgeographically for both May and June (Figs. 4C and 4D)

1964 Can. J. For. Res. Vol. 39, 2009

Published by NRC Research Press

compared with estimates based solely on precipitation. Theestimates based on temperature alone suggested that, in thesouthern portion of the North Central region, temperatureswould have been adequate for E. maimaiga epizootics in atleast 19 years between 1971 and 2000 (Figs. 4C and 4D).Temperatures in much of Minnesota, Wisconsin, and Michi-gan, however, generally met or exceeded average tempera-ture estimates in fewer than 12 years during the 1971 to2000 period, with the northern portions of each state experi-

encing favorable temperature conditions in fewer than 6 ofthe 30 years.

In comparisons incorporating both precipitation and tem-perature, northern areas of the region generally had fewer fa-vorable years for epizootics than southern areas (Figs. 4E and4F). The estimates suggest that much of the North Central re-gion, except for the most southern portions in May, experi-enced adequate levels of precipitation and temperature in nomore than 6 years over the 30 year period (Figs. 4E and 4F).

Fig. 2. Maximum similarity to any one of 16 locations where a documented Entomophaga maimaiga epizootic occurred based on overallclimatic similarity (i.e., Match Index), using (A) spring (April–June for epizootics south of 41.28N; May–July for epizootics north of41.28N) and (B) annual climate data (January–December) 1971–2000. Images in this figure are presented in color.

Siegert et al. 1965

Published by NRC Research Press

DiscussionClimatic similarity to the current range of an organism is

one of several factors that have been proposed to influencethe likelihood of establishment of a nonindigenous species(Williamson 1996; Mack et al. 2002). This can be the casefor both pestiferous species and organisms that function asbeneficial biological control agents. Recruiting natural ene-mies from areas with climatic conditions similar to those ofthe target pest increases the probability that the beneficialorganism will persist and provide effective control (Hoelmerand Kirk 2005). Previous researchers have used CLIMEX’sclimate-matching capabilities, which are typically based onprecipitation, temperature, or related meteorological varia-bles averaged by calendar month over a 30 year time period,in classical biological control (e.g., Julien et al. 1995; vanKlinken et al. 2003; Goolsby et al. 2005). While CLIMEXserves as a useful tool to initially estimate the potential dis-tribution of a nonindigenous pest or beneficial organism, itrelies on relatively coarse descriptors of climate. Moreover,simple comparisons using CLIMEX do not account for tem-poral variability in precipitation and temperature, which may

be substantial within monthly periods. The stability or varia-bility of weather conditions can strongly influence dynamicsof organisms such as gypsy moth larvae and E. maimaiga.In general, organisms with broad ecological tolerances canbe successful in climatically variable environments, whilethose with narrow tolerances, such as the E. maimaiga fun-gal pathogen, are more likely to be consistently successful inclimatically stable areas that are physiologically appropriatefor that species (Leigh 1981; Crawley et al. 1986).

Results from our climate-matching analyses, which werebased on long-term monthly averages for temperature andprecipitation across North America, indicate that averagespring weather conditions in an area south of the GreatLakes region that extended from central Kansas east throughWest Virginia, and portions of central Minnesota and Wis-consin were highly similar to weather associated with E.maimaiga epizootics. Interestingly, since weather observedduring epizootics generally differed considerably from 30year average conditions to which they were compared(Table 1), spring similarity indices generated by CLIMEXwere only 60%–80% similar to epizootic conditions in areas

Fig. 3. North Central region average monthly standard deviations, 1971–2000, of (A) spring precipitation (mm, May–June), (B) annual pre-cipitation (mm, January–December), (C) spring temperature (8C, May–June), and (D) annual temperature (8C, January–December) by cli-mate division (n = 75). Areas covered by darker colors indicate greater average monthly standard deviations than areas covered by lightercolors. Images in this figure are presented in color.

1966 Can. J. For. Res. Vol. 39, 2009

Published by NRC Research Press

where epizootics have been documented to occur. Phenologyand interactions of gypsy moth larvae and E. maimaiga,however, are clearly mediated by current-year weather con-ditions. Whether the dynamics of these organisms are de-fined satisfactorily by long-term average values oftemperature or precipitation depends on the extent of thetemporal variability in weather experienced at a given lo-cale.

Our assessment of weather conditions that occurred from1971 to 2000 in the North Central region indicated that pre-

cipitation and temperature conditions were not uniformlyconsistent; some areas of the region were more climaticallyvariable on a year to year basis than others. Portions of theNorth Central region with high climatic variability may notexperience the particular conditions necessary for the devel-opment of E. maimaiga epizootics as often as areas that areboth generally suitable for the fungus and less variable inprecipitation or temperature. Alternatively, when long-termaverage conditions are relatively unfavorable for E. mai-maiga epizootics, highly variable areas are more likely than

Fig. 4. Number of years from 1971 to 2000 in the North Central region with precipitation (A and B), temperature (C and D), and bothprecipitation and temperature (E and F) similar to 16 documented Entomophaga maimaiga epizootics in May and June. A year was consid-ered favorable for an epizootic if the climatic parameter during a given time period met or exceeded the average that occurred in areas withdocumented epizootics. Images in this figure are presented in color.

Siegert et al. 1967

Published by NRC Research Press

areas with low variability to at least occasionally experiencesuitable conditions for epizootics to occur, if adequate fun-gal inoculum and suitable hosts are present. This illustratessome potential issues associated with relying simply on 30year averages for weather-related variables; such long-termaverages can provide only a coarse estimate of how condu-cive an area will be for E. maimaiga activity. Additionally,sites may vary more in one meteorological parameter thananother at a given time of year, which could affect E. mai-maiga germination and infection rates and, thus, the likeli-hood of an epizootic. In our analysis, for example, thesouthern portion of the North Central region was fairly sim-ilar in overall spring weather to one or more of the epi-zootic-specific sites (Fig. 2A). While this region was theleast variable in temperature (Fig. 3C), it was the most vari-able in terms of precipitation (Fig. 3A).

For estimating the number of favorable years for E. mai-maiga epizootics in the North Central region from 1971 to2000, we assessed climatic conditions and variability basedon precipitation and temperature on both a spring and an-nual basis. It is currently unclear how E. maimaiga is specif-ically affected by spring versus annual climate (Hajek1999). Spring weather is likely to be integral to the develop-ment of E. maimaiga epizootics because of the organism’ssensitivity to adequate precipitation for resting spore germi-nation and close phenological association with gypsy mothlarval development (Hajek and Roberts 1991). Conversely,annual weather conditions may affect survival of E. mai-maiga resting spores, their persistence in the environment,or perhaps the synchrony between resting spore germinationand activity of young gypsy moth larvae. In addition, cli-matic deviations from 30 year averages (i.e., normal condi-tions) are not uniform throughout the year, and trends differfor precipitation and temperature. For instance, in the NorthCentral region, more than 25% of the total departures of pre-cipitation from normal average conditions during the yearoccurred in the spring, particularly in May and June. Inother words, precipitation levels in spring were more varia-ble than at other times of the year. Conversely, air temper-atures in spring were less variable than at other times of theyear; less than 25% of the total departures of temperaturefrom normal conditions occurred in the spring.

Results from our analyses have direct implications forgypsy moth management in the North Central region, espe-cially for areas along or beyond the leading edge of estab-lished gypsy moth populations. States in the southernportion of the North Central region, including Indiana, Illi-nois, Kentucky, Missouri, and Ohio, appear likely to fre-quently experience weather conditions that are conducivefor the development of E. maimaiga epizootics. In the north-ern tier of states, specifically Minnesota, Wisconsin, andMichigan, precipitation or temperature conditions that ap-pear to be associated with E. maimaiga epizootics did notoccur as frequently as in areas to the south. Additionally,the northern tier of states tended to be less variable in pre-cipitation compared with the rest of the region, suggestingthat the levels of spring precipitation necessary for an epi-zootic may be relatively uncommon. Furthermore, if climatechange affects the North Central region by making itwarmer, drier, or more variable (Hayhoe et al. 2007), thelikelihood of frequent E. maimaiga epizootics would de-

crease. Unfortunately, Minnesota, Wisconsin, and Michiganhave large forested areas that are highly susceptible to gypsymoth, as measured by the total basal area of preferred treespecies (Liebhold et al. 1997). In these areas, E. maimaigamay exert less influence on gypsy moth population dynam-ics, while factors such as severe winter weather or other nat-ural enemies could increase in importance.

Observations from established gypsy moth populations inthe North Central region appear to support our predictions.For example, while E. maimaiga was widely distributed inlower Michigan by 1993 (Smitley et al. 1995; Buss 1997),gypsy moth defoliation exceeded 10 000 ha in at least 11 ofthe 15 years between 1993 and 2007, and in nearly half ofthose years, gypsy moth defoliation exceeded 34 000 ha(http://na.fs.fed.us/fhp/gm/defoliation/index.shtm). Most ofthe recorded defoliation occurred in northern counties whereour results suggest weather may be marginal for E. mai-maiga. Similarly, E. maimaiga was introduced into numer-ous gypsy moth populations in eastern Wisconsin in themid- to late-1990s, and cadavers of E. maimaiga-killed lar-vae have been recovered from about half of the release sites(http://www.uwex.edu/ces/gypsymoth/bio.cfm). Gypsy mothpopulations in Wisconsin, however, have increased in den-sity and spread substantially faster than populations in thesouthern region of the advancing front (http://www.gmsts.org/operations/). Hoffman Gray et al. (2008) reported thatwhile E. maimaiga was widely distributed in oak-dominatedstands in Wisconsin, gypsy moth mortality attributable tothis pathogen was substantially lower than mortality re-corded in the northeastern United States. While these pat-terns are not conclusive, they are consistent with predictionsfrom our analyses.

Other regions of the United States still beyond the ad-vancing front of gypsy moth include a substantial compo-nent of oaks (Quercus sp.), aspen (Populus sp.), or otherpreferred hosts, and the potential suitability of weather forE. maimaiga may similarly vary. A thorough assessment ofthe role that weather and climate play in the occurrence ofE. maimaiga epizootics will likely require long-term moni-toring and landscape-level studies but would enhance ourability to assess potential impacts of gypsy moth.

AcknowledgementsWe gratefully acknowledge the assistance of Aaron M.

Pollyea, Michigan State University (MSU), in obtaining cli-mate data. We further acknowledge Joseph S. Elkinton (Uni-versity of Massachusetts), Therese M. Poland (USDA ForestService), David E. Rothstein (MSU), Suzanne M. Thiem(MSU), and three anonymous reviewers for their helpfulcomments on an earlier draft of this manuscript. Fundingfor this research was provided by the Special TechnologyDevelopment Program, State and Private Forestry, USDAForest Service, and the Michigan Department of Agriculture.This study was supported in part by Michigan AgriculturalExperiment Station Project No. MICL01700.

ReferencesAndreadis, T.G., and Weseloh, R.M. 1990. Discovery of Entomo-

phaga maimaiga in North American gypsy moth, Lymantria dis-

1968 Can. J. For. Res. Vol. 39, 2009

Published by NRC Research Press

par. Proc. Natl. Acad. Sci. U.S.A. 87(7): 2461–2465. doi:10.1073/pnas.87.7.2461. PMID:11607071.

Benz, G. 1987. Environment. In Epizootiology of insect diseases.Edited by J.R. Fuxa and Y. Tanada. John Wiley and Sons, NewYork. pp. 177–214.

Burges, H.D. (Editor). 1998. Formulation of mycoinsecticides. InFormulation of microbial biopesticides: beneficial microorgan-isms, nematodes and seed treatments. Kluwer Academic Pub-lishers, Dordrecht, Netherlands. pp. 131–185.

Buss, L.J. 1997. Evaluation of three egg mass survey methods andtwo biological control agents for gypsy moth management. M.S.thesis, Department of Entomology, Michigan State University,East Lansing, Mich.

Carnegie, A.J., Matsuki, M., Haugen, D.A., Hurley, B.P., Ahu-mada, R., Klasmer, P., Sun, J., and Iede, E.T. 2006. Predictingthe potential distribution of Sirex noctilio (Hymenoptera: Sirici-dae), a significant exotic pest of Pinus plantations. Ann. For.Sci. 63(2): 119–128. doi:10.1051/forest:2005104.

Carruthers, R.I., and Soper, R.S. 1987. Fungal diseases. In Epizoo-tiology of insect diseases. Edited by J.R. Fuxa and Y. Tanada.John Wiley and Sons, New York. pp. 357–416.

Crawley, M.J., Kornberg, H., Lawton, J.H., Usher, M.B., South-wood, R., O’Connor, R.J., and Gibbs, A. 1986. The populationbiology of invaders. Philos. Trans. R. Soc. Lond. Ser. B Biol.Sci. 314(1167): 711–731. doi:10.1098/rstb.1986.0082.

Elkinton, J.S., Hajek, A.E., Boettner, G.H., and Simons, E.E. 1991.Distribution and apparent spread of Entomophaga maimaiga(Zygomycetes: Entomophthorales) in gypsy moth (Lepidoptera:Lymantriidae) populations in North America. Environ. Entomol.20: 1601–1605.

Goolsby, J.A., DeBarro, P.J., Kirk, A.A., Sutherst, R.W., Canas, L.,Ciomperlik, M.A., Ellsworth, P.C., Gould, J.R., Hartley, D.M.,Hoelmer, K.A., Naranjo, S.E., Rose, M., Roltsch, W.J., Ruiz,R.A., Pickett, C.H., and Vacek, D.C. 2005. Post-release evalua-tion of biological control of Bemisia tabaci biotype ‘‘B’’ in theUSA and the development of predictive tools to guide introduc-tions for other countries. Biol. Control, 32(1): 70–77. doi:10.1016/j.biocontrol.2004.07.012.

Gray, D.R. 2004. The gypsy moth life stage model: landscape-wideestimates of gypsy moth establishment using a multi-genera-tional phenology model. Ecol. Model. 176(1-2): 155–171.doi:10.1016/j.ecolmodel.2003.11.010.

Hajek, A.E. 1997. Fungal and viral epizootics in gypsy moth (Lepi-doptera: Lymantriidae) populations in central New York. Biol.Control, 10(1): 58–68. doi:10.1006/bcon.1997.0541.

Hajek, A.E. 1999. Pathology and epizootiology of Entomophagamaimaiga infections in forest Lepidoptera. Microbiol. Mol.Biol. Rev. 63(4): 814–835. PMID:10585966.

Hajek, A.E., and Humber, R.A. 1997. Formation and germinationof Entomophaga maimaiga azygospores. Can. J. Bot. 75(10):1739–1747. doi:10.1139/b97-888.

Hajek, A.E., and Roberts, D.W. 1991. Pathogen reservoirs as a bio-logical control resource: introduction of Entomophaga maimaigato North American gypsy moth, Lymantria dispar, populations.Biol. Control, 1: 29–34. doi:10.1016/1049-9644(91)90098-K.

Hajek, A.E., and St. Leger, R.J. 1994. Interactions between fungalpathogens and insect hosts. Annu. Rev. Entomol. 39: 293–322.doi:10.1146/annurev.en.39.010194.001453.

Hajek, A.E., Carruthers, R.I., and Soper, R.S. 1990a. Temperatureand moisture relations of sporulation and germination by Ento-mophaga maimaiga (Zygomycetes: Entomophthoraceae), a fun-gal pathogen of Lymantria dispar (Lepidoptera: Lymantriidae).Environ. Entomol. 19: 85–90.

Hajek, A.E., Humber, R.A., Elkinton, J.S., May, B., Walsh, S.R.A.,

and Silver, J.C. 1990b. Allozyme and restriction fragment lengthpolymorphism analyses confirm Entomophaga maimaiga re-sponsible for 1989 epizootics in North American gypsy mothpopulations. Proc. Natl. Acad. Sci. U.S.A. 87(18): 6979–6982.doi:10.1073/pnas.87.18.6979. PMID:11607100.

Hajek, A.E., Larkin, T.S., Carruthers, R.I., and Soper, R.S. 1993.Modeling the dynamics of Entomophaga maimaiga (Zygomy-cetes: Entomophthorales) epizootics in gypsy moth (Lepidop-tera: Lymantriidae) populations. Environ. Entomol. 22: 1172–1187.

Hajek, A.E., Butler, L., Walsh, S.R.A., Silver, J.C., Hain, F.P.,Hastings, F.L., O’Dell, T.M., and Smitley, D.R. 1996a. Hostrange of the gypsy moth (Lepidoptera: Lymantriidae) pathogenEntomophaga maimaiga (Zygomycetes: Entomophthorales) inthe field versus laboratory. Environ. Entomol. 25: 709–721.

Hajek, A.E., Elkinton, J.S., and Witcosky, J.J. 1996b. Introductionand spread of the fungal pathogen Entomophaga maimaiga (Zy-gomycetes: Entomophthorales) along the leading edge of gypsymoth (Lepidoptera: Lymantriidae) spread. Environ. Entomol. 25:1235–1247.

Hajek, A.E., Bauer, L., McManus, M.L., and Wheeler, M.M.1998a. Distribution of resting spores of the Lymantria disparpathogen Entomophaga maimaiga in soil and on bark. BioCon-trol, 43(2): 189–200. doi:10.1023/A:1009933525540.

Hajek, A.E., Tatman, K.M., Wanner, P.H., and Wheeler, M.M.1998b. Location and persistence of cadavers of gypsy moth, Ly-mantria dispar, containing Entomophaga maimaiga azygo-spores. Mycologia, 90(5): 754–760. doi:10.2307/3761315.

Hajek, A.E., Butler, L., Liebherr, J.K., and Wheeler, M.M. 2000.Risk of infection by the fungal pathogen Entomophaga mai-maiga among Lepidoptera on the forest floor. Environ. Entomol.29: 645–650.

Hajek, A.E., Strazanac, J.S., Wheeler, M.M., Vermeylen, F.M., andButler, L. 2004. Persistence of the fungal pathogen Entomo-phaga maimaiga and its impact on native Lymantriidae. Biol.Control, 30(2): 466–473. doi:10.1016/j.biocontrol.2004.02.005.

Hajek, A.E., McManus, M.L., and Delalibera, I., Jr. 2005. Catalo-gue of introductions of pathogens and nematodes for classicalbiological control of insects and mites. USDA For. Serv.FHTET-2005-05.

Hayhoe, K., Wake, C.P., Huntington, T.G., Luo, L., Schwartz,M.D., Sheffield, J., Wood, E., Anderson, B., Bradbury, J., De-Gaetano, A., Troy, T.J., and Wolfe, D. 2007. Past and futurechanges in climate and hydrological indicators in the US North-east. Clim. Dyn. 28(4): 381–407. doi:10.1007/s00382-006-0187-8.

Hoelmer, K.A., and Kirk, A.A. 2005. Selecting arthropod biologi-cal control agents against arthropod pests: Can the science beimproved to decrease the risk of releasing ineffective agents?Biol. Control, 34(3): 255–264. doi:10.1016/j.biocontrol.2005.05.001.

Hoffman Gray, R., Lorimer, C.G., Tobin, P.C., and Raffa, K.F.2008. Preoutbreak dynamics of a recently established invasiveherbivore: roles of natural enemies and habitat structure instage-specific performance of gypsy moth (Lepidoptera: Lyman-triidae) populations in northeastern Wisconsin. Environ. Ento-mol. 37(5): 1174–1184. doi:10.1603/0046-225X(2008)37[1174:PDOARE]2.0.CO;2. PMID:19036196.

Julien, M.H., Skarratt, B., and Maywald, G.F. 1995. Potential geo-graphical distribution of alligator weed and its biological controlAgasicles hygrophila. J. Aquat. Plant Manage. 33: 55–60.

Leigh, E.G., Jr. 1981. The average lifetime of a population in avarying environment. J. Theor. Biol. 90(2): 213–239. doi:10.1016/0022-5193(81)90044-8. PMID:7311579.

Siegert et al. 1969

Published by NRC Research Press

Liebhold, A.M., Gottschalk, K.W., Mason, D.A., and Bush, R.R.1997. Forest susceptibility to the gypsy moth. J. For. 95: 20–24.

Mack, R., Barrett, S.C.H., deFur, P.L., MacDonald, W.L., Madden,L.V., Marshall, D.S., McCullough, D.G., McEvoy, P.B., Nyrop,J.P., Reichard, S.E.H., Rice, K.J., and Tolin, S.A. 2002. Predict-ing invasions of nonindigenous plants and plant pests. NationalAcademy Press, Washington, D.C.

Malakar, R., Elkinton, J.S., Carroll, S.D., and D’Amico, V. 1999.Interactions between two gypsy moth (Lepidoptera: Lymantrii-dae) pathogens: nucleopolyhedrovirus and Entomophaga mai-maiga (Zygomycetes: Entomophthorales): field studies and asimulation model. Biol. Control, 16(2): 189–198. doi:10.1006/bcon.1999.0751.

McCoy, C.W., Samson, R.A., and Boucias, D.G. 1988. Entomogen-ous fungi. In CRC handbook of natural pesticides. Vol. V. Mi-crobial insecticides, Part A: Entomogenous protozoa and fungi.Edited by C.M. Ignoffo and N.B. Mandavas. CRC Press, BocaRaton, Fla. pp. 151–236.

Shimazu, M., and Soper, R.S. 1986. Pathogenicity and sporulationof Entomophaga maimaiga Humber, Shimazu, Soper and Hajek(Zygomycetes: Entomophthorales) on larvae of the gypsy moth,Lymantria dispar L. (Lepidoptera: Lymantriidae). Appl. Ento-mol. Zool. (Jpn.), 21: 589–596.

Smitley, D.R., Bauer, L.S., Hajek, A.E., Sapio, F.J., and Humber,R.A. 1995. Introduction and establishment of Entomophaga mai-maiga, a fungal pathogen of gypsy moth (Lepidoptera: Lyman-triidae) in Michigan. Environ. Entomol. 24: 1685–1695.

Sutherst, R.W., and Maywald, G.F. 1985. A computerised systemfor matching climates in ecology. Agric. Ecosyst. Environ.13(3-4): 281–299. doi:10.1016/0167-8809(85)90016-7.

Sutherst, R.W., Maywald, G.F., and Skarratt, D.B. 1995. Predictinginsect distributions in a changed climate. In Insects in a chan-ging environment. Edited by R. Harrington and N.E. Stork. Aca-demic Press, New York. pp. 59–91.

Sutherst, R.W., Maywald, G.F., Yonow, T., and Stevens, P.M.1999. CLIMEX: predicting the effects of climate on plants andanimals. CSIRO Publishing, Collingwood, Victoria, Australia.

Sutherst, R.W., Maywald, G.F., and Kriticos, D. 2007. CLIMEX.

Version 3. User’s guide. Hearne Scientific Software Pty Ltd.,Melbourne, Australia.

SYSTAT. 2000. SYSTAT. Version 10. SPSS Inc., Chicago, Ill.van Klinken, R.D., Fichera, G., and Cordo, H. 2003. Targeting bio-

logical control across diverse landscapes: the release, establish-ment and early success of two insects on mesquite (Prosopisspp.) insects in Australian rangelands. Biol. Control, 26(1): 8–20. doi:10.1016/S1049-9644(02)00107-X.

Venette, R.C., and Cohen, S.D. 2006. Potential climatic suitabilityfor establishment of Phytophthora ramorum within the contigu-ous United States. For. Ecol. Manage. 231(1–3): 18–26. doi:10.1016/j.foreco.2006.04.036.

Villedieu, Y., and van Frankenhuyzen, K. 2004. Epizootic occur-rence of Entomophaga maimaiga at the leading edge of an ex-panding population of the gypsy moth (Lepidoptera:Lymantriidae) in north-central Ontario. Can. Entomol. 136(6):875–878. doi:10.4039/N04-002.

Webb, R.E., White, G.B., Thorpe, K.W., and Talley, S.E. 1999.Quantitative analysis of a pathogen-induced premature collapseof a ‘‘leading edge’’ gypsy moth (Lepidoptera: Lymantriidae)population in Virginia. J. Entomol. Sci. 34: 84–100.

Webb, R.E., White, G.B., Podgwaite, J.D., D’Amico, V., Slavicek,J., Swearingen, J., Onken, B., and Thorpe, K.W. 2005. Compar-ison of aerially-applied Gypchek strains against gypsy moth (Le-pidoptera: Lymantriidae) in the presence of an Entomophagamaimaiga epizootic. J. Entomol. Sci. 40: 446–460.

Weseloh, R.M., and Andreadis, T.G. 1992. Mechanisms of trans-mission of the gypsy moth (Lepidoptera: Lymantriidae) fungus,Entomophaga maimaiga (Zygomycetes: Entomophthoraceae)and effects of site conditions on its prevalence. Environ. Ento-mol. 21: 901–906.

Weseloh, R.M., and Andreadis, T.G. 1997. Persistence of restingspores of Entomophaga maimaiga, a fungal pathogen of thegypsy moth, Lymantria dispar. J. Invertebr. Pathol. 69(2): 195–196. doi:10.1006/jipa.1996.4645. PMID:9056470.

Williamson, M. 1996. Biological invasions. Chapman and Hall,New York.

1970 Can. J. For. Res. Vol. 39, 2009

Published by NRC Research Press