- Email: [email protected]
Persistence of the fungal pathogen Entomophaga maimaiga and its impact on native Lymantriidae
Biological Control 30 (2004) 466–473 www.elsevier.com/locate/ybcon
Persistence of the fungal pathogen Entomophaga maimaiga and its impact on native Lymantriidae Ann E. Hajek,a,* John S. Strazanac,b Micheal M. Wheeler,a Francoise M. Vermeylen,c and Linda Butlerb b
a Department of Entomology, Cornell University, Ithaca, NY 14853-0901, USA Division of Plant and Soil Sciences, West Virginia University, Morgantown, WV 26506-6108, USA c Division of Nutritional Sciences, Cornell University, Ithaca, NY 14853, USA
Received 18 September 2003; accepted 6 February 2004
Abstract The entomopathogenic fungus Entomophaga maimaiga has been reported infecting gypsy moth, Lymantria dispar, larvae in the United States since 1989. Laboratory bioassays demonstrated that larvae of numerous species of native tussock moths (Lymantriidae) are susceptible to E. maimaiga. From 1997 to 2001, larvae of gypsy moth and native lymantriids were collected in long-term plots in national forests in Virginia and West Virginia to evaluate infection levels. Throughout this time, gypsy moth populations were low although increasing the last 2 years, whereas native lymantriids remained uncommon throughout. Among gypsy moth larvae, infection by E. maimaiga first occurred during 2000 and increased in both areas during 2001 as gypsy moth populations increased. Seven species of native lymantriids were collected but only three were infected: Dasychira obliquata, Dasychira vagans, and Orgyia leucostigma. Infection of these native lymantriid species did not occur in 1997–1999 and was always <50% per species during 2000 and 2001. Ecological studies were conducted to evaluate when and where native lymantriids could be exposed to E. maimaiga inoculum. Field bioassays demonstrated that E. maimaiga resting spores were active in soil at bases of trees beginning in April, especially when soil was moist, but few gypsy moth larvae were infected by airborne conidia from resting spores early in the season. In 2000 and 2001, cadavers that produced conidia were collected throughout the season and conidia were the predominant spore form produced for at least the first month of the field season. Resting spore densities in soil were high during 1997 but titers declined through the study, with the exception of an increase following E. maimaiga infections in increasing gypsy moth populations in West Virginia during 2000. Ó 2004 Elsevier Inc. All rights reserved. Keywords: Gypsy moth; Entomophaga maimaiga; Lymantriidae; Epizootiology; Nontarget effects; Host specificity; Conservation; Risk assessment
1. Introduction The fungal pathogen Entomophaga maimaiga Humber, Shimazu et Soper was initially described infecting larvae of the gypsy moth, Lymantria dispar (L.), in Japan and was first found in introduced gypsy moth populations in North America in 1989 (Hajek, 1999). Epizootics caused by this pathogen in North America have subsequently occurred in populations of gypsy moth ranging from low to outbreak densities. Land
* Corresponding author. Fax: 1-607-255-0939. E-mail address: [email protected] (A.E. Hajek).
1049-9644/$ - see front matter Ó 2004 Elsevier Inc. All rights reserved. doi:10.1016/j.biocontrol.2004.02.005
managers have been interested in using this pathogen for biological control and therefore need to understand the extent to which E. maimaiga might impact nontarget invertebrate fauna in forests. Under ideal conditions during laboratory bioassays, E. maimaiga caused low levels of infection in larvae from a diversity of lepidopteran families (Hajek et al., 1995). The only species of Lepidoptera consistently highly susceptible to E. maimaiga belonged to the tussock moth family, Lymantriidae (Hajek et al., 1995; Soper et al., 1988), which also includes the gypsy moth. In the northeastern US, the tussock moth family includes 15 native species in 3 genera (Ferguson, 1978) so this is a relatively small family. While some of the
A.E. Hajek et al. / Biological Control 30 (2004) 466–473
species of lymantriids in North America are outbreak species, most years these species are generally uncommon (e.g., Butler, 1992; Butler and Kondo, 1991), as are the majority of native lepidopteran species. During a field study of nontarget effects in Virginia, although many gypsy moth larvae collected from foliage were infected, only one larva from each of two abundant species (one lasiocampid and one noctuid) from among 1790 lepidopteran larvae collected and reared became infected (Hajek et al., 1996a). All seven individuals of three species of native lymantriids that were collected and reared were not infected. We hypothesized that the peculiar behavior of later instar gypsy moth larvae, resting in the leaf litter during the day, particularly exposes them to E. maimaiga inoculum because the long-lived resting spores ( ¼ azygospores) of E. maimaiga persist at high densities in the soil (Hajek, 2001). The depauperate lepidopteran fauna in the leaf litter was collected and reared to see if species occupying this potentially risky habitat became infected. Although 37% of the gypsy moth larvae that were collected in the litter were infected, only one gelechiid larva (out of 84 collected) and one larva of a relatively abundant noctuid species were infected (Hajek et al., 2000). No lymantriids aside from gypsy moth were found in the leaf litter during this study. Although these studies suggested that E. maimaiga was having extremely little impact on nontargets, there was the nagging concern that few native lymantriids had been collected during quantitative field studies. Over a 7-year period (1989–1995) when multitudes of E. maimaiga-killed gypsy moth larvae from the field were diagnosed, the only E. maimaiga-infected nontargets received were 5 field-collected cadavers of lymantriids, belonging to three species in the genus Dasychira (Hajek et al., 1996a). In addition, among 17 lymantriid larvae (total of 5 native species) randomly collected from central New York and central Massachusetts during the gypsy moth larval field seasons between 1998 and 2000, only the three individuals of Dasychira basiflava (Packard) were infected by E. maimaiga (G. Boettner, pers. commun.; A.E.H., unpubl. data). We report results from a systematic sampling program developed to evaluate the effect of E. maimaiga on native lymantriid species over 5 years, with accompanying ecological studies to investigate when and where native lymantriids would be exposed to E. maimaiga inoculum. Because larvae of several of the native lymantriids are present in the field earlier than gypsy moth larvae, we evaluated the potential for infection early in the season. We also evaluated production of airborne conidia through the season in association with seasonality of native lymantriids because these larvae are not known to rest in the leaf litter. The densities of E. maimaiga resting spores in the soil of sample sites were quantified to evaluate persistence of this pathogen in relation to both nontarget and target impact.
467
2. Materials and methods 2.1. Study sites Each year from 1997 to 2001, nine plots were sampled in xeric mixed oak and pine forests in the George Washington National Forest (GWNF), Augusta, Virginia, and nine plots were sampled in more mesic forests in the Monongahela National Forest (MNF), Pocahontas, West Virginia (Butler and Strazanac, 2000b). Gypsy moth populations were just moving into that part of the MNF in 1996 and caused defoliation in 1995 and 1996 in GWNF. The increasing gypsy moth populations crashed in 1996 with reports that infection by E. maimaiga had been abundant (L.B., unpubl. data). Study plots of 200 ha were all oak-dominated. Each year the density of gypsy moth populations was quantified by counting all egg masses in 28 0.01 ha subplots per plot (Liebhold et al., 1994). In each plot, 30 cm wide green canvas bands were stapled 1.5 m above the ground around 12 dominant or co-dominant trees, with 6 bands near a ridgetop and 6 in a watershed. In each plot, among the 12 trees banded, 10 were oaks (Quercus spp.), one was a maple (Acer spp.), and one was a hickory (Carya spp.). Sampling began in early May during early leaf expansion and continued each week through August (14– 15 sample periods). Plots were sampled by collecting any lymantriid larvae beneath bands (Butler and Strazanac, 2000a), and larvae were also collected from foliage samples clipped by pole pruners from the mid- and lower canopies of trees (Butler and Strazanac, 2000b). For canopy sampling, pole pruners were equipped with metal rings to which plastic bags were attached to catch samples. Each week, five foliage samples were taken from each plot with 3 samples from oak, 1 from hickory, and 1 from maple. For oak and maple, a sample consisted of 21 branch tips and for hickory, a sample was 15 branch tips. After collection, foliage was examined for macrolepidopteran larvae. Larvae were placed individually in 96.1 ml plastic cups (7.5 cm diam. top 5.2 cm diam base 3.5 cm h) and were provided with the species of foliage on which they had been collected. Lymantriid larvae have distinctive morphologies (Ferguson, 1978) and were readily identified to species. All native lymantriid larvae and the majority of gypsy moth larvae collected (82%) were reared to detect entomophthoralean infections. E. maimaiga is the only entomophthoralean known to infect gypsy moth. However, at the time of this study, we could not confirm that infections of native lymantriids were caused by E. maimaiga because we had not yet developed the necessary molecular methods and bioassays using gypsy moth were not possible. However, infection of both native lymantriids and gypsy moth collected at the same times and locations strongly suggests that E. maimaiga
468
A.E. Hajek et al. / Biological Control 30 (2004) 466–473
was the entomophthoralean species infecting native lymantriid larvae during this study (see Section 4). For 14 days after collection, larvae were provided with fresh foliage as necessary at 20–25 °C and 14:10 (L:D) and were monitored daily for death. Cadavers were placed on 1.5% water agar at 20 °C and for 3 days were observed daily for fungal outgrowth. Cadavers remained at 20 °C for 7 more days to allow maturation of resting spores. Cadavers were then stored at 4 °C and subsequently dissected and examined using a compound microscope to detect E. maimaiga resting spores. Daily temperature and rainfall from nine sites surrounding the GWNF plots and six sites surrounding the MNF plots were averaged (US Environmental Data Service, 1997–2001a, 1997–2001b). For each year, we included weather data from the 7 days before larval sampling began, to include conditions when the first infections would have occurred, through to the end of larval sampling. 2.2. Quantifying E. maimaiga resting spores in soil samples Soil samples were collected in both December 1996 and April 1997 to determine whether resting spore titers changed over winter. For 1998–2001, each year soil samples were collected only from March through May. At three areas within each plot, five co-dominant oaks were chosen. The organic layer of soil was collected up to 3 cm in depth at 60°, 180°, and 300°, from within 10 cm of the trunks of the five trees. Approximately 20 g of soil were sampled per tree for a total of 300 g/plot. We attempted not to sample from the bases of the same trees each year. Quantifying resting spores in soil is more accurate when higher titers of resting spores are present (Hajek and Wheeler, 1994). Therefore, for seasonal comparisons in 1996–1997, we only quantified resting spores in the three plots in GWNF and the three plots in MNF with the highest gypsy moth egg mass densities hatching in 1995. After collection, soil samples were maintained at 4 °C until quantification. Resting spores in samples were quantified following soil extraction by wet sieving and use of Percoll discontinuous density gradients, as described in Hajek and Wheeler (1994). Resting spore densities during spring, when resting spores germinate, were quantified yearly for the same six plots used for seasonal winter/spring comparisons. Based on increased densities of gypsy moth in the MNF, counts from three more MNF plots were subsequently added using spring soil samples for all years.
antriid larvae in early spring. Field bioassays were conducted at the most northern plot in GWNF, Elliot Springs Run (Plot 1), during early spring 1997. Gypsy moth larvae were obtained as neonates from USDA, APHIS, Otis Methods Development Center, Otis ANGB, Massachusetts and were reared following standard procedures (Bernon, 1995). Early fourth instars were placed in cages made with 20 20 mesh aluminum window screening (23 31 cm) and containing a cube of artificial diet. Twenty early fourth instar larvae were placed in each cage and one cage was placed at the base of each of three co-dominant oaks. The soil beneath the cages and around the bases of each tree was watered once per week with 3.8 liters to promote resting spore activity in case it did not rain. After the first two exposure periods, cages were placed at the bases of three more trees beneath which the soil was not watered. Resting spores actively eject infective germ conidia so we evaluated whether larvae above the ground became infected by germ conidia during early spring when conidia from cadavers would not yet be present. The same cage design was used to contain gypsy moth larvae on hardware cloth platforms within 10 cm of the trunks of the three watered oaks at 2, 5, 10, and 50 cm above the ground as well as at 2 m height, hanging from a nearby branch. For each exposure period, cages remained in the field for 48 h, after which larvae were removed from cages and provided with artificial diet in groups of 10 in 236.6 ml plastic cups with paper lids at room temperature. Larvae were monitored daily for 10 days to detect mortality and dead larvae were treated as described above to detect E. maimaiga infections. Exposures began on 4 April and ended on 8 May, for a total of 15 exposure periods at watered trees and 13 exposures at nonwatered trees. While larvae were caged in the field, electronic weather recording equipment (Omnidata, Logan, UT) quantified leaf wetness at 1 m in the understory and soil temperature and moisture (centibars) at 2 cm depth every 30 min. We evaluated the potential for production of airborne conidia throughout the season by quantifying production of conidia versus resting spores from cadavers of gypsy moth larvae dying from E. maimaiga infections. This was only possible in 2000 and 2001 because no infections were found from 1997 to 1999. Sampling began 7–9 May and continued weekly while larvae were present in the field. All cadavers were treated as described above and instars were recorded as well as whether conidia only, resting spores only or both spore types were produced in or on cadavers. 2.4. Data analysis
2.3. Conidial production by E. maimaiga Field bioassays were conducted to evaluate the extent to which E. maimaiga could be infecting native lym-
Statistical analysis could not be used to evaluate levels of infection among native lymantriid larvae because densities were too low. Season-long infection rates
A.E. Hajek et al. / Biological Control 30 (2004) 466–473
for gypsy moth larvae were calculated as in Hajek et al. (1990). Proportions of survival for each week were multiplied by each other to estimate the proportion of larvae surviving to pupation that year, after which proportion E. maimaiga infection was estimated as 1 ) (proportion survival). To compare E. maimaiga resting spore density in winter versus spring, soil samples and changes in resting spore density across years, Poisson regression models adjusted for overdispersion of data were used (SAS Institute, 1999). Post hoc comparisons among years within states used least square means with Bonferroni corrections.
3. Results 3.1. Entomophaga maimaiga infections among lymantriids Throughout this study, gypsy moth populations remained extremely low considering that this is a species that can characteristically increase to >1000 egg masses/ ha during outbreaks. In 1995 and 1996, egg mass densities in GWNF and some northern MNF plots averaged from 267 to 697/ha but populations crashed during the 1996 field season. From 1997 to 2001, maximum densities of gypsy moth in GWNF averaged 3 3 egg masses/ha (mean SE; eggs hatching in 1999) and in MNF averaged 25 23 egg masses/ha (eggs hatching in 2001). At such low densities, larval densities rather than egg mass densities are more indicative of changes in population level. Counts of gypsy moth larvae from canvas bands and pruning demonstrated a trend of increase during 2000 and 2001, especially in MNF (Fig. 1).
Fig. 1. Average number of gypsy moth larvae (SE) collected per plot using standardized collection on foliage and under canvas bands in George Washington National Forest, Virginia and Monongahela National Forest, West Virginia, 1997–2001.
469
Native lymantriids found in plots were also not abundant, with only 16 and 21 total larvae per year collected during 1997 and 1998, respectively, increasing to 49 larvae collected during 1999. Although during 2000 and 2001 gypsy moth populations steadily increased, native lymantriids did not increase significantly, with 15 and 50 total larvae collected, respectively (Table 1). Throughout the 5 years, gypsy moth larvae were predominantly first instars with a few second instars when sampling began 5–12 May, while during this same sampling week, some of the native lymantriids [D. basiflava, Dasychira meridionalis (Barnes & McDunnough), D. obliquata] were predominantly instars 4–6. None of the gypsy moth larvae collected in 1997– 1999 were infected by E. maimaiga (Fig. 2). During these years, rainfall during the period of larval sampling ranged from a low of 23.1 cm total in 1999 to 44.5 (1997) and 54.0 cm (1998). Total rainfall throughout the larval collection periods in 2000–2001 was more similar to 1998 with 53.1 (2000) and 56.8 cm (2001) (US Environmental Data Service, 1997–2001a, 1997–2001b). In 2000, the first infected gypsy moth larvae were collected during the early fifth instar (4–5 June, fifth week of sampling), and infections were thereafter found weekly. During 2000, all gypsy moth infections were found in MNF, where gypsy moth populations were more abundant. A total of 3 larvae were infected out of 15 native lymantriids collected in 2000 (Table 1). Two of these infected larvae were collected in two plots in MNF (19–20 June; 85.2% infection among gypsy moth larvae collected that week in MNF) and one was collected from GWNF (June 13) where no infected gypsy moth larvae were collected from 1997 to 2000. During 2001, gypsy moth populations were more abundant (Fig. 1) and E. maimaiga infections among gypsy moth larvae increased and became abundant in both national forests (Fig. 2). Only 4 of the 50 native lymantriids collected became infected. In 2001, infected native lymantriids were all collected as later instars on the same sample date (June 4) at three different plots in GWNF (32.0% infection among gypsy moth larvae collected that week in GWNF). Among the 7 species of native lymantriids collected during this study, infections only occurred in three species (Table 1). The species with most infection was D. obliquata although we always found <50% infection among larvae collected during any 1 year and for 3 of the 5 years, none of the D. obliquata larvae collected were infected. Surprisingly, a species previously found infected with E. maimaiga, D. basiflava, was never found infected throughout this 5-year study although it was the most abundant of the native lymantriids. The 7 individuals of native lymantriids that were infected all were 4–6 instar. Six of the 7 were collected under canvas bands, so they had wandered from the foliage. This
470
A.E. Hajek et al. / Biological Control 30 (2004) 466–473
Table 1 Number of gypsy and tussock moth (Lymantriidae) larvae collected in George Washington National Forest, Virginia and Monongahela National Forest, West Virginia from 1997 to 2001 and number infected by the fungal entomopathogen E. maimaiga Species
1997
1998
1999
2000
2001
n
No. inf.
n
No. inf.
n
No. inf.
n
No. inf.
n
No. inf.
Lymantria dispar (L.) Dasychira basiflava (Packard) Dasychira dorsipennata (Barnes & McDunnough) Dasychira meridionalis (Barnes & McDunnough) Dasychira obliquata (Grote & Robinson) Dasychira vagans (Barnes & McDunnough) Orgyia definita Packard Orgyia leucostigma (J.E. Smith)
19 6
0 0
34 9
0 0
174 40
0 0
599 3
50 0
1077 19
223 0
5
0
4
0
2
0
0
—
0
—
1
0
4
0
4
0
3
0
15
0
3
0
3
0
3
0
7
2
7
3
0
—
0
—
0
—
0
—
3
1
1 0
0
0 1
—
0 0
—
0 2
—
0
1
1 5
0 0
Total native lymantriids
16
0
21
0
49
15
3
50
4
—
—
0
3.2. Resting spore persistence
Fig. 2. Resting spores/g dry soil (SE) during early spring associated with total yearly percent infection among gypsy moth larvae collected throughout the field season from 1997 to 2001 in nine plots in (A) George Washington National Forest, Virginia and (B) Monongahela National Forest, West Virginia.
behavior was not unusual because for the total 151 native lymantriids collected, 86.0% were found under canvas bands rather than on foliage.
A significant difference was found between resting spore densities in December 1996 and April 1997 (t ¼ 2:58; p < 0:0157) but, contrary to our hypotheses, the densities of resting spores across plots in spring (6359.1 1788.2) were greater than the densities in winter (4154.9 1067.8). These data included counts from MNF plot 15 that were >8 times higher in spring than fall. Spring counts from plot 15 were also much higher than counts from other plots where winter and spring counts were more similar. We repeated this analysis considering plot 15 as an outlier and statistical results changed. Without plot 15, resting spore densities did not differ between winter and spring (t ¼ 0:61; p > 0:5464). GWNF counts were greater than MNF counts (t ¼ 3:48; p < 0:0064), as we hypothesized because gypsy moth populations had been present longer in GWNF and had been higher before crashing (so more resting spores would potentially have been produced). Comparing spring resting spore densities across all years, overall results were similar with or without the high counts from plot 15 in April 1997. Patterns of densities differed for sites in GWNF versus MNF (Fig. 2). In GWNF, resting spore densities declined after 1998 while, in MNF, resting spore titers showed a trend of decrease from 1997 through 2000 but then increased again in 2001. 3.3. Conidial production by E. maimaiga Bioassays conducted in 1997 confirmed that the resting spores at bases of trees in Plot 1, GWNF were E. maimaiga. The very first infections seen were few, only occurring among larvae caged on the ground under watered trees from 4 to 6 April. Gypsy moth eggs in this
A.E. Hajek et al. / Biological Control 30 (2004) 466–473
area began hatching slightly later, ca. 14 April (R. Webb, pers. commun.). During the exposure period from 11 to 13 April, due to a rain event leaf wetness was >50 for 21 h (conditions necessary for conidial production and infection by airborne conidia) but no infections occurred, with only a few infections among larvae caged directly on soil during the following exposure period (13–15 April). For cages on watered soil, infections started in earnest beginning 25 April (Fig. 3). Soil moisture remained high throughout the study under watered trees (never >1.8 centibars; average ¼ 0.7 centibars) and therefore would not have limited resting spore germination early in the season. Under trees where the soil was not watered, infections were only seen during four exposure periods, between 25 April and 4 May, and averaged only 6.2 2.2%. During this period, infections were only >3% when soil moisture was <1.6 centibars and, during intervals when infections did not occur after 25 April, soil moisture always averaged >2 centibars. High variability in infection was seen among trees, with no infections at all occurring among larvae caged under one unwatered tree. Larvae caged above the soil that could only be infected by airborne conidia were very seldom infected. One larva out of 60 was infected at 2 cm height during three intervals (4–6 April, 25–27 April, and 6–8 May), one larva was infected at 10 cm height (2–4 May) and none of the larvae caged at 5 or 50 cm and 2 m height were infected throughout the study. Collections of gypsy moth larvae during the field season demonstrated that early in the season conidia were exclusively produced and resting spore production only increased late in the season (Fig. 4). Fourth instars were predominantly collected in weeks 4–5 in both 2000 and 2001 and fourth instar cadavers never contained resting spores. Resting spore production only began
471
Fig. 4. Percent production of conidia versus resting spores by E. maimaiga in cadavers of gypsy moth larvae collected throughout the field season in George Washington National Forest, Virginia and Monongahela National Forest, West Virginia. (A) 2000 and (B) 2001.
when fifth and later instars were present although cadavers that produced conidia as well as resting spores were found until the end of the season. To connect this with presence of lymantriid larvae, E. maimaiga-infected native lymantriids were only collected weeks 6 and 7 in 2000 when gypsy moths were P 5th instar and week 5 in 2001 when 4–5 instar gypsy moth larvae predominated.
4. Discussion
Fig. 3. Percent infection (mean SE) among L. dispar larvae caged for 2 days on top of soil around bases of either watered or nonwatered oaks at Elliot Springs Run, George Washington National Forest, Virginia, 4 April–8 May, 1997.
Native lymantriids remained at low densities throughout this study and only became infected by E. maimaiga during years when gypsy moth populations were more abundant and gypsy moth larvae were infected (2000 and 2001). The highest percent infection among native lymantriids was found among D. obliquata (5 of 23 larvae) although this was not the most abundant native lymantriid species collected. One cadaver of an E. maimaiga-killed D. obliquata larva had been collected in both 1989 and 1995 but D. obliquata larvae collected during epizootics in gypsy moth in
472
A.E. Hajek et al. / Biological Control 30 (2004) 466–473
Virginia in 1994 were not infected (Hajek et al., 1996a). Among all native lymantriid species collected during this study, infection was never >50% for any 1 year and, in fact, only approached this level for one species out of seven, during 1 year out of 5. During this study, the presence and activities of different types of E. maimaiga inoculum were studied to evaluate potential impact of this pathogen on native lymantriids. E. maimaiga resting spores were abundant in soil, especially during the early years of this study. Unless resting spores are added, resting spore titers generally decrease in density to some extent each year because some resting spores germinate whether gypsy moth larvae are present or not although many persist as a soil reservoir (Hajek and Eastburn, 2001). In agreement, between 1997 and 1999, resting spore titers declined although gypsy moth larvae were scarce. Field bioassays conducted during 1997 demonstrated the importance of soil moisture for resting spores to germinate. During 1997 and 1998, rain fell on >50% of the days when larvae were present in the field although rainfall amount and frequency decreased in 1999. Rainfall in 1998, when no infections were seen, was similar to 2000– 2001, when infections occurred. Therefore, we hypothesize that the reasons we did not find infected gypsy moth or native lymantriid larvae from 1997 to 1999 was not due to moisture limiting resting spore germination, especially in 1998. The present study documented for the first time a rapid increase in resting spore titer after E. maimaiga infections were abundant during one season (2000). This increase was seen in West Virginia plots in 2001 (Fig. 2), and occurred although gypsy moth density in the plots was very low (9 4 egg masses/ha in 2000). For E. maimaiga to begin infection cycles each season, resting spores in the soil germinate to actively eject infective germ conidia that can cause infections ( ¼ primary infection). Although germ conidia are actively ejected, no one has studied to what extent they become airborne. Results from our 1997 early season studies provide little indication that germ conidia infect many larvae above the ground level. After gypsy moth larvae become infected and die, then infective conidia actively ejected from larval cadavers cause Ôsecondary infectionsÕ and these conidia definitely become airborne (Hajek et al., 1999). Infections among larvae while they are in the tree and shrub canopy are most likely due to secondary infections. Models have shown that Ôsecondary infectionsÕ are responsible for the exponential increases in infection characteristic of epizootics (e.g., Hajek et al., 1993). Lymantriid larvae other than gypsy moth are not known to commonly rest in the leaf litter as is characteristic of later instar gypsy moth larvae. Therefore, native lymantriids would have much less risk of infection from the bank of resting spores in the soil compared with gypsy moth larvae. Native lymantriid larvae would
principally be exposed to infection initiated by E. maimaiga resting spores only if they traveled on the forest floor when the soil was moist. We hypothesize that it is more likely that native lymantriids become infected from conidia produced from cadavers while they are in the understory or tree canopy. During this study, cadavers of gypsy moth larvae predominantly produced conidia through much of the season, until late instars were present, at which time many late instar cadavers produced resting spores. We hypothesize that E. maimaiga infections were not found from 1997 to 1999 among native lymantriids because gypsy moth densities were too low to produce abundant conidia for secondary infections. In addition, because high levels of E. maimaiga infection in gypsy moth populations often occur late in the season (Hajek, 1999) and many of the native lymantriids are earlier than gypsy moth and would have pupated by the time gypsy moth are late instars, we hypothesize that relative seasonality of these species would result in native lymantriids largely escaping periods when airborne conidia of E. maimaiga might be abundant. Unfortunately, during this study we could not confirm that each of the infections in native lymantriids was caused by E. maimaiga. E. maimaiga belongs to the same species complex as the morphologically identical Entomophaga aulicae (Reichardt in Bail) Humber, which is native to the US, and both species infect lepidopteran larvae. Although E. aulicae does not infect gypsy moth, it has been reported infecting the lymantriids Calliteara pudibunda (L.) ( ¼ Dasychira pudipunda) (Lakon, 1919), Euproctis chrysorrhoea (L.), and Orgyia antiqua nova Fitch (reported as O. nova) (Speare and Colley, 1912), Orgyia vetusta Boisduval (Hajek et al., 1996b), and O. leucostigma (van Frankenhuyzen et al., 2002). In the field E. maimaiga has been confirmed as infecting the North American lymantriids D. basiflava, Dasychira leucophaea (J.E. Smith), D. obliquata (Hajek et al., 1996a; A.E.H., unpubl. data) and O. leucostigma (Hajek et al., 2000). We assume that infections of native lymantriids during this study were caused by E. maimaiga because this fungus has previously been shown to infect native lymantriids in the field, the 1996 gypsy moth populations in Virginia and West Virginia crashed and resting spores in soil were subsequently abundant and high percentages of gypsy moth larvae caged over soil became infected by E. maimaiga in 1997, and abundant infections were found among gypsy moth larvae collected at the same times and locations as native lymantriids in 2000 and 2001.
Acknowledgments We thank B. Knoblauch, C. Eastburn, D. Blue, M. Bertoia, N. Bertoia, J. McNeil, A. Savage, T. Kun-Lin
A.E. Hajek et al. / Biological Control 30 (2004) 466–473
and K. Zuniga for assistance with rearing and monitoring larvae, D. Curry for counting gypsy moth egg masses and T. Carrington, M. Whitman, R. Braud, K. Grubb, and J. Parrish for collecting larvae and soil samples in Virginia and West Virginia during the field season. L. Liebherr helped with data summarization, A. Burke assisted with graphics and C. Nielsen assisted with comments on the manuscript. We thank R. Reardon for his continued interest in this subject. Studies were funded through USDA, Forest Service Cooperative Agreements to L.B. and A.E.H.
References Bernon, G.L., 1995. Status of mass-reared gypsy moths: protocol. USDA, For. Serv. Northeast. For. Exp. Stn. Gen. Tech. Rpt. NE213, p. 54. Butler, L., 1992. The community of Macrolepidopterous larvae at CooperÕs Rock State Forest, West Virginia: a baseline study. Can. Entomol. 124, 1149–1156. Butler, L., Kondo, V., 1991. Macrolepidopterous moths collected by blacklight trap at CooperÕs Rock State Forest, West Virginia: a baseline study. WV Agric. For. Exp. Stn. Bull. 705, 25 pp. Butler, L., Strazanac, J., 2000a. Macrolepidopteran larvae sampled by tree bands in temperate mesic and xeric forests in eastern United States. Proc. Entomol. Soc. Wash. 102, 188–197. Butler, L., Strazanac, J., 2000b. Occurrence of Lepidoptera on selected host trees in two central Appalachian national forests. Ann. Entomol. Soc. Am. 93, 500–511. Ferguson, D.C., 1978. Noctuoidea (in part), Lymantriidae. In: Dominick, R.B., et al. (Eds.), The moths of North America. Fasc. 22.2. E. B. Classey and The Wedge Entomological Research Foundation, London, pp. 1–110. Hajek, A.E., 1999. Pathology and epizootiology of the Entomophaga maimaiga infections in the forest Lepidoptera. Microbiol. Mol. Biol. Rev. 63, 814–835. Hajek, A.E., 2001. Larval behavior in Lymantria dispar increases risk of fungal infection. Oecologia 126, 285–291. Hajek, A.E., Eastburn, C.C., 2001. Effect of host insects on activation of Entomophaga maimaiga resting spores. J. Invertebr. Pathol. 77, 290–291. Hajek, A.E., Wheeler, M.M., 1994. Application of techniques for quantification of soil-borne entomophthoralean resting spores. J. Invertebr. Pathol. 64, 71–73. Hajek, A.E., Butler, L., Wheeler, M.M., 1995. Laboratory bioassays testing the host range of the gypsy moth fungal pathogen Entomophaga maimaiga. Biol. Control 5, 530–544.
473
Hajek, A.E., Olsen, C.H., Elkinton, J.S., 1999. Dynamics of airborne conidia of the gypsy moth (Lepidoptera: Lymantriidae) fungal pathogen Entomophaga maimaiga (Zygomycetes: Entomophthorales). Biol. Control 16, 111–117. Hajek, A.E., Larkin, T.S., Carruthers, R.I., Soper, R.S., 1993. Modeling the dynamics of Entomophaga maimaiga (Zygomycetes: Entomophthorales) epizootics in gypsy moth (Lepidoptera: Lymantriidae) populations. Environ. Entomol. 22, 1172–1187. Hajek, A.E., Walsh, S.R.A., Strong, D.R., Silver, J.C., 1996b. A disjunct Californian strain of Entomophaga aulicae infecting Orgyia vetusta. J. Invertebr. Pathol. 68, 260–268. Hajek, A.E., Butler, L., Liebherr, J.K., Wheeler, M.M., 2000. Risk of infection by the fungal pathogen Entomophaga maimaiga among Lepidoptera on the forest floor. Environ. Entomol. 29, 645–650. Hajek, A.E., Humber, R.A., Elkinton, J.S., May, B., Walsh, S.R.A., Silver, J.C., 1990. Allozyme and restriction fragment length polymorphism analyses confirm Entomophaga maimaiga responsible for 1989 epizootics in North American gypsy moth populations. Proc. Natl. Acad. Sci. USA 87, 6979–6982. Hajek, A.E., Butler, L., Walsh, S.R.A., Silver, J.C., Hain, F.P., Hastings, F.L., ODell, T.M., Smitley, D.R., 1996a. Host range of the gypsy moth (Lepidoptera: Lymantriidae) pathogen Entomophaga maimaiga (Zygomycetes: Entomophthorales) in the field versus laboratory. Environ. Entomol. 25, 709–721. Lakon, G., 1919. Die Insektenfeinde aus der Familie der Entomophthoreen. Z. Angew. Entomol. 5, 161–216. Liebhold, A., Thorpe, K., Ghent, J., Lyons, D.B., 1994. Gypsy moth egg mass sampling for decision-making: a userÕs guide. USDA, Forest Service, Forest Health Protection, NA-TP-04-94, 12 pp. SAS Institute, 1999. SAS/STAT UserÕs Guide, Version 8. SAS Institute, Cary, NC, USA. Soper, R.S., Shimazu, M., Humber, R.A., Ramos, M.E., Hajek, A.E., 1988. Isolation and characterization of Entomophaga maimaiga sp. nov., a fungal pathogen of gypsy moth, Lymantria dispar, from Japan. J. Invertebr. Pathol. 51, 229–241. Speare, A.T., Colley, R.H., 1912. The Artificial Use of the Brown-Tail Fungus in Massachusetts with Practical Suggestions for Private Experiments, and a Brief Note on a Fungus Disease of the Gypsy Caterpillar. Wright & Potter, Boston, MA. US Environmental Data Service, 1997–2001a. Climatological Data: Virginia. Vols. 107–111. US Environmental Data Service, 1997–2001b. Climatological Data: West Virginia. Vols. 105–109. van Frankenhuyzen, K.S., Ebling, P., Thurston, G., Lucarotti, C., Royama, T., Guscott, R., Georgeson, E., Silver, J., 2002. Incidence and impact of Entomophaga aulicae (Zygomycetes: Entomophthorales) and a nucleopolyhedrovirus in an outbreak of the whitemarked tussock moth (Lepidoptera: Lymantriidae). Can. Entomol. 134, 825–846.
Persistence of the fungal pathogen Entomophaga maimaiga and its impact on native Lymantriidae Ann E. Hajek,a,* John S. Strazanac,b Micheal M. Wheeler,a Francoise M. Vermeylen,c and Linda Butlerb b
a Department of Entomology, Cornell University, Ithaca, NY 14853-0901, USA Division of Plant and Soil Sciences, West Virginia University, Morgantown, WV 26506-6108, USA c Division of Nutritional Sciences, Cornell University, Ithaca, NY 14853, USA
Received 18 September 2003; accepted 6 February 2004
Abstract The entomopathogenic fungus Entomophaga maimaiga has been reported infecting gypsy moth, Lymantria dispar, larvae in the United States since 1989. Laboratory bioassays demonstrated that larvae of numerous species of native tussock moths (Lymantriidae) are susceptible to E. maimaiga. From 1997 to 2001, larvae of gypsy moth and native lymantriids were collected in long-term plots in national forests in Virginia and West Virginia to evaluate infection levels. Throughout this time, gypsy moth populations were low although increasing the last 2 years, whereas native lymantriids remained uncommon throughout. Among gypsy moth larvae, infection by E. maimaiga first occurred during 2000 and increased in both areas during 2001 as gypsy moth populations increased. Seven species of native lymantriids were collected but only three were infected: Dasychira obliquata, Dasychira vagans, and Orgyia leucostigma. Infection of these native lymantriid species did not occur in 1997–1999 and was always <50% per species during 2000 and 2001. Ecological studies were conducted to evaluate when and where native lymantriids could be exposed to E. maimaiga inoculum. Field bioassays demonstrated that E. maimaiga resting spores were active in soil at bases of trees beginning in April, especially when soil was moist, but few gypsy moth larvae were infected by airborne conidia from resting spores early in the season. In 2000 and 2001, cadavers that produced conidia were collected throughout the season and conidia were the predominant spore form produced for at least the first month of the field season. Resting spore densities in soil were high during 1997 but titers declined through the study, with the exception of an increase following E. maimaiga infections in increasing gypsy moth populations in West Virginia during 2000. Ó 2004 Elsevier Inc. All rights reserved. Keywords: Gypsy moth; Entomophaga maimaiga; Lymantriidae; Epizootiology; Nontarget effects; Host specificity; Conservation; Risk assessment
1. Introduction The fungal pathogen Entomophaga maimaiga Humber, Shimazu et Soper was initially described infecting larvae of the gypsy moth, Lymantria dispar (L.), in Japan and was first found in introduced gypsy moth populations in North America in 1989 (Hajek, 1999). Epizootics caused by this pathogen in North America have subsequently occurred in populations of gypsy moth ranging from low to outbreak densities. Land
* Corresponding author. Fax: 1-607-255-0939. E-mail address: [email protected] (A.E. Hajek).
1049-9644/$ - see front matter Ó 2004 Elsevier Inc. All rights reserved. doi:10.1016/j.biocontrol.2004.02.005
managers have been interested in using this pathogen for biological control and therefore need to understand the extent to which E. maimaiga might impact nontarget invertebrate fauna in forests. Under ideal conditions during laboratory bioassays, E. maimaiga caused low levels of infection in larvae from a diversity of lepidopteran families (Hajek et al., 1995). The only species of Lepidoptera consistently highly susceptible to E. maimaiga belonged to the tussock moth family, Lymantriidae (Hajek et al., 1995; Soper et al., 1988), which also includes the gypsy moth. In the northeastern US, the tussock moth family includes 15 native species in 3 genera (Ferguson, 1978) so this is a relatively small family. While some of the
A.E. Hajek et al. / Biological Control 30 (2004) 466–473
species of lymantriids in North America are outbreak species, most years these species are generally uncommon (e.g., Butler, 1992; Butler and Kondo, 1991), as are the majority of native lepidopteran species. During a field study of nontarget effects in Virginia, although many gypsy moth larvae collected from foliage were infected, only one larva from each of two abundant species (one lasiocampid and one noctuid) from among 1790 lepidopteran larvae collected and reared became infected (Hajek et al., 1996a). All seven individuals of three species of native lymantriids that were collected and reared were not infected. We hypothesized that the peculiar behavior of later instar gypsy moth larvae, resting in the leaf litter during the day, particularly exposes them to E. maimaiga inoculum because the long-lived resting spores ( ¼ azygospores) of E. maimaiga persist at high densities in the soil (Hajek, 2001). The depauperate lepidopteran fauna in the leaf litter was collected and reared to see if species occupying this potentially risky habitat became infected. Although 37% of the gypsy moth larvae that were collected in the litter were infected, only one gelechiid larva (out of 84 collected) and one larva of a relatively abundant noctuid species were infected (Hajek et al., 2000). No lymantriids aside from gypsy moth were found in the leaf litter during this study. Although these studies suggested that E. maimaiga was having extremely little impact on nontargets, there was the nagging concern that few native lymantriids had been collected during quantitative field studies. Over a 7-year period (1989–1995) when multitudes of E. maimaiga-killed gypsy moth larvae from the field were diagnosed, the only E. maimaiga-infected nontargets received were 5 field-collected cadavers of lymantriids, belonging to three species in the genus Dasychira (Hajek et al., 1996a). In addition, among 17 lymantriid larvae (total of 5 native species) randomly collected from central New York and central Massachusetts during the gypsy moth larval field seasons between 1998 and 2000, only the three individuals of Dasychira basiflava (Packard) were infected by E. maimaiga (G. Boettner, pers. commun.; A.E.H., unpubl. data). We report results from a systematic sampling program developed to evaluate the effect of E. maimaiga on native lymantriid species over 5 years, with accompanying ecological studies to investigate when and where native lymantriids would be exposed to E. maimaiga inoculum. Because larvae of several of the native lymantriids are present in the field earlier than gypsy moth larvae, we evaluated the potential for infection early in the season. We also evaluated production of airborne conidia through the season in association with seasonality of native lymantriids because these larvae are not known to rest in the leaf litter. The densities of E. maimaiga resting spores in the soil of sample sites were quantified to evaluate persistence of this pathogen in relation to both nontarget and target impact.
467
2. Materials and methods 2.1. Study sites Each year from 1997 to 2001, nine plots were sampled in xeric mixed oak and pine forests in the George Washington National Forest (GWNF), Augusta, Virginia, and nine plots were sampled in more mesic forests in the Monongahela National Forest (MNF), Pocahontas, West Virginia (Butler and Strazanac, 2000b). Gypsy moth populations were just moving into that part of the MNF in 1996 and caused defoliation in 1995 and 1996 in GWNF. The increasing gypsy moth populations crashed in 1996 with reports that infection by E. maimaiga had been abundant (L.B., unpubl. data). Study plots of 200 ha were all oak-dominated. Each year the density of gypsy moth populations was quantified by counting all egg masses in 28 0.01 ha subplots per plot (Liebhold et al., 1994). In each plot, 30 cm wide green canvas bands were stapled 1.5 m above the ground around 12 dominant or co-dominant trees, with 6 bands near a ridgetop and 6 in a watershed. In each plot, among the 12 trees banded, 10 were oaks (Quercus spp.), one was a maple (Acer spp.), and one was a hickory (Carya spp.). Sampling began in early May during early leaf expansion and continued each week through August (14– 15 sample periods). Plots were sampled by collecting any lymantriid larvae beneath bands (Butler and Strazanac, 2000a), and larvae were also collected from foliage samples clipped by pole pruners from the mid- and lower canopies of trees (Butler and Strazanac, 2000b). For canopy sampling, pole pruners were equipped with metal rings to which plastic bags were attached to catch samples. Each week, five foliage samples were taken from each plot with 3 samples from oak, 1 from hickory, and 1 from maple. For oak and maple, a sample consisted of 21 branch tips and for hickory, a sample was 15 branch tips. After collection, foliage was examined for macrolepidopteran larvae. Larvae were placed individually in 96.1 ml plastic cups (7.5 cm diam. top 5.2 cm diam base 3.5 cm h) and were provided with the species of foliage on which they had been collected. Lymantriid larvae have distinctive morphologies (Ferguson, 1978) and were readily identified to species. All native lymantriid larvae and the majority of gypsy moth larvae collected (82%) were reared to detect entomophthoralean infections. E. maimaiga is the only entomophthoralean known to infect gypsy moth. However, at the time of this study, we could not confirm that infections of native lymantriids were caused by E. maimaiga because we had not yet developed the necessary molecular methods and bioassays using gypsy moth were not possible. However, infection of both native lymantriids and gypsy moth collected at the same times and locations strongly suggests that E. maimaiga
468
A.E. Hajek et al. / Biological Control 30 (2004) 466–473
was the entomophthoralean species infecting native lymantriid larvae during this study (see Section 4). For 14 days after collection, larvae were provided with fresh foliage as necessary at 20–25 °C and 14:10 (L:D) and were monitored daily for death. Cadavers were placed on 1.5% water agar at 20 °C and for 3 days were observed daily for fungal outgrowth. Cadavers remained at 20 °C for 7 more days to allow maturation of resting spores. Cadavers were then stored at 4 °C and subsequently dissected and examined using a compound microscope to detect E. maimaiga resting spores. Daily temperature and rainfall from nine sites surrounding the GWNF plots and six sites surrounding the MNF plots were averaged (US Environmental Data Service, 1997–2001a, 1997–2001b). For each year, we included weather data from the 7 days before larval sampling began, to include conditions when the first infections would have occurred, through to the end of larval sampling. 2.2. Quantifying E. maimaiga resting spores in soil samples Soil samples were collected in both December 1996 and April 1997 to determine whether resting spore titers changed over winter. For 1998–2001, each year soil samples were collected only from March through May. At three areas within each plot, five co-dominant oaks were chosen. The organic layer of soil was collected up to 3 cm in depth at 60°, 180°, and 300°, from within 10 cm of the trunks of the five trees. Approximately 20 g of soil were sampled per tree for a total of 300 g/plot. We attempted not to sample from the bases of the same trees each year. Quantifying resting spores in soil is more accurate when higher titers of resting spores are present (Hajek and Wheeler, 1994). Therefore, for seasonal comparisons in 1996–1997, we only quantified resting spores in the three plots in GWNF and the three plots in MNF with the highest gypsy moth egg mass densities hatching in 1995. After collection, soil samples were maintained at 4 °C until quantification. Resting spores in samples were quantified following soil extraction by wet sieving and use of Percoll discontinuous density gradients, as described in Hajek and Wheeler (1994). Resting spore densities during spring, when resting spores germinate, were quantified yearly for the same six plots used for seasonal winter/spring comparisons. Based on increased densities of gypsy moth in the MNF, counts from three more MNF plots were subsequently added using spring soil samples for all years.
antriid larvae in early spring. Field bioassays were conducted at the most northern plot in GWNF, Elliot Springs Run (Plot 1), during early spring 1997. Gypsy moth larvae were obtained as neonates from USDA, APHIS, Otis Methods Development Center, Otis ANGB, Massachusetts and were reared following standard procedures (Bernon, 1995). Early fourth instars were placed in cages made with 20 20 mesh aluminum window screening (23 31 cm) and containing a cube of artificial diet. Twenty early fourth instar larvae were placed in each cage and one cage was placed at the base of each of three co-dominant oaks. The soil beneath the cages and around the bases of each tree was watered once per week with 3.8 liters to promote resting spore activity in case it did not rain. After the first two exposure periods, cages were placed at the bases of three more trees beneath which the soil was not watered. Resting spores actively eject infective germ conidia so we evaluated whether larvae above the ground became infected by germ conidia during early spring when conidia from cadavers would not yet be present. The same cage design was used to contain gypsy moth larvae on hardware cloth platforms within 10 cm of the trunks of the three watered oaks at 2, 5, 10, and 50 cm above the ground as well as at 2 m height, hanging from a nearby branch. For each exposure period, cages remained in the field for 48 h, after which larvae were removed from cages and provided with artificial diet in groups of 10 in 236.6 ml plastic cups with paper lids at room temperature. Larvae were monitored daily for 10 days to detect mortality and dead larvae were treated as described above to detect E. maimaiga infections. Exposures began on 4 April and ended on 8 May, for a total of 15 exposure periods at watered trees and 13 exposures at nonwatered trees. While larvae were caged in the field, electronic weather recording equipment (Omnidata, Logan, UT) quantified leaf wetness at 1 m in the understory and soil temperature and moisture (centibars) at 2 cm depth every 30 min. We evaluated the potential for production of airborne conidia throughout the season by quantifying production of conidia versus resting spores from cadavers of gypsy moth larvae dying from E. maimaiga infections. This was only possible in 2000 and 2001 because no infections were found from 1997 to 1999. Sampling began 7–9 May and continued weekly while larvae were present in the field. All cadavers were treated as described above and instars were recorded as well as whether conidia only, resting spores only or both spore types were produced in or on cadavers. 2.4. Data analysis
2.3. Conidial production by E. maimaiga Field bioassays were conducted to evaluate the extent to which E. maimaiga could be infecting native lym-
Statistical analysis could not be used to evaluate levels of infection among native lymantriid larvae because densities were too low. Season-long infection rates
A.E. Hajek et al. / Biological Control 30 (2004) 466–473
for gypsy moth larvae were calculated as in Hajek et al. (1990). Proportions of survival for each week were multiplied by each other to estimate the proportion of larvae surviving to pupation that year, after which proportion E. maimaiga infection was estimated as 1 ) (proportion survival). To compare E. maimaiga resting spore density in winter versus spring, soil samples and changes in resting spore density across years, Poisson regression models adjusted for overdispersion of data were used (SAS Institute, 1999). Post hoc comparisons among years within states used least square means with Bonferroni corrections.
3. Results 3.1. Entomophaga maimaiga infections among lymantriids Throughout this study, gypsy moth populations remained extremely low considering that this is a species that can characteristically increase to >1000 egg masses/ ha during outbreaks. In 1995 and 1996, egg mass densities in GWNF and some northern MNF plots averaged from 267 to 697/ha but populations crashed during the 1996 field season. From 1997 to 2001, maximum densities of gypsy moth in GWNF averaged 3 3 egg masses/ha (mean SE; eggs hatching in 1999) and in MNF averaged 25 23 egg masses/ha (eggs hatching in 2001). At such low densities, larval densities rather than egg mass densities are more indicative of changes in population level. Counts of gypsy moth larvae from canvas bands and pruning demonstrated a trend of increase during 2000 and 2001, especially in MNF (Fig. 1).
Fig. 1. Average number of gypsy moth larvae (SE) collected per plot using standardized collection on foliage and under canvas bands in George Washington National Forest, Virginia and Monongahela National Forest, West Virginia, 1997–2001.
469
Native lymantriids found in plots were also not abundant, with only 16 and 21 total larvae per year collected during 1997 and 1998, respectively, increasing to 49 larvae collected during 1999. Although during 2000 and 2001 gypsy moth populations steadily increased, native lymantriids did not increase significantly, with 15 and 50 total larvae collected, respectively (Table 1). Throughout the 5 years, gypsy moth larvae were predominantly first instars with a few second instars when sampling began 5–12 May, while during this same sampling week, some of the native lymantriids [D. basiflava, Dasychira meridionalis (Barnes & McDunnough), D. obliquata] were predominantly instars 4–6. None of the gypsy moth larvae collected in 1997– 1999 were infected by E. maimaiga (Fig. 2). During these years, rainfall during the period of larval sampling ranged from a low of 23.1 cm total in 1999 to 44.5 (1997) and 54.0 cm (1998). Total rainfall throughout the larval collection periods in 2000–2001 was more similar to 1998 with 53.1 (2000) and 56.8 cm (2001) (US Environmental Data Service, 1997–2001a, 1997–2001b). In 2000, the first infected gypsy moth larvae were collected during the early fifth instar (4–5 June, fifth week of sampling), and infections were thereafter found weekly. During 2000, all gypsy moth infections were found in MNF, where gypsy moth populations were more abundant. A total of 3 larvae were infected out of 15 native lymantriids collected in 2000 (Table 1). Two of these infected larvae were collected in two plots in MNF (19–20 June; 85.2% infection among gypsy moth larvae collected that week in MNF) and one was collected from GWNF (June 13) where no infected gypsy moth larvae were collected from 1997 to 2000. During 2001, gypsy moth populations were more abundant (Fig. 1) and E. maimaiga infections among gypsy moth larvae increased and became abundant in both national forests (Fig. 2). Only 4 of the 50 native lymantriids collected became infected. In 2001, infected native lymantriids were all collected as later instars on the same sample date (June 4) at three different plots in GWNF (32.0% infection among gypsy moth larvae collected that week in GWNF). Among the 7 species of native lymantriids collected during this study, infections only occurred in three species (Table 1). The species with most infection was D. obliquata although we always found <50% infection among larvae collected during any 1 year and for 3 of the 5 years, none of the D. obliquata larvae collected were infected. Surprisingly, a species previously found infected with E. maimaiga, D. basiflava, was never found infected throughout this 5-year study although it was the most abundant of the native lymantriids. The 7 individuals of native lymantriids that were infected all were 4–6 instar. Six of the 7 were collected under canvas bands, so they had wandered from the foliage. This
470
A.E. Hajek et al. / Biological Control 30 (2004) 466–473
Table 1 Number of gypsy and tussock moth (Lymantriidae) larvae collected in George Washington National Forest, Virginia and Monongahela National Forest, West Virginia from 1997 to 2001 and number infected by the fungal entomopathogen E. maimaiga Species
1997
1998
1999
2000
2001
n
No. inf.
n
No. inf.
n
No. inf.
n
No. inf.
n
No. inf.
Lymantria dispar (L.) Dasychira basiflava (Packard) Dasychira dorsipennata (Barnes & McDunnough) Dasychira meridionalis (Barnes & McDunnough) Dasychira obliquata (Grote & Robinson) Dasychira vagans (Barnes & McDunnough) Orgyia definita Packard Orgyia leucostigma (J.E. Smith)
19 6
0 0
34 9
0 0
174 40
0 0
599 3
50 0
1077 19
223 0
5
0
4
0
2
0
0
—
0
—
1
0
4
0
4
0
3
0
15
0
3
0
3
0
3
0
7
2
7
3
0
—
0
—
0
—
0
—
3
1
1 0
0
0 1
—
0 0
—
0 2
—
0
1
1 5
0 0
Total native lymantriids
16
0
21
0
49
15
3
50
4
—
—
0
3.2. Resting spore persistence
Fig. 2. Resting spores/g dry soil (SE) during early spring associated with total yearly percent infection among gypsy moth larvae collected throughout the field season from 1997 to 2001 in nine plots in (A) George Washington National Forest, Virginia and (B) Monongahela National Forest, West Virginia.
behavior was not unusual because for the total 151 native lymantriids collected, 86.0% were found under canvas bands rather than on foliage.
A significant difference was found between resting spore densities in December 1996 and April 1997 (t ¼ 2:58; p < 0:0157) but, contrary to our hypotheses, the densities of resting spores across plots in spring (6359.1 1788.2) were greater than the densities in winter (4154.9 1067.8). These data included counts from MNF plot 15 that were >8 times higher in spring than fall. Spring counts from plot 15 were also much higher than counts from other plots where winter and spring counts were more similar. We repeated this analysis considering plot 15 as an outlier and statistical results changed. Without plot 15, resting spore densities did not differ between winter and spring (t ¼ 0:61; p > 0:5464). GWNF counts were greater than MNF counts (t ¼ 3:48; p < 0:0064), as we hypothesized because gypsy moth populations had been present longer in GWNF and had been higher before crashing (so more resting spores would potentially have been produced). Comparing spring resting spore densities across all years, overall results were similar with or without the high counts from plot 15 in April 1997. Patterns of densities differed for sites in GWNF versus MNF (Fig. 2). In GWNF, resting spore densities declined after 1998 while, in MNF, resting spore titers showed a trend of decrease from 1997 through 2000 but then increased again in 2001. 3.3. Conidial production by E. maimaiga Bioassays conducted in 1997 confirmed that the resting spores at bases of trees in Plot 1, GWNF were E. maimaiga. The very first infections seen were few, only occurring among larvae caged on the ground under watered trees from 4 to 6 April. Gypsy moth eggs in this
A.E. Hajek et al. / Biological Control 30 (2004) 466–473
area began hatching slightly later, ca. 14 April (R. Webb, pers. commun.). During the exposure period from 11 to 13 April, due to a rain event leaf wetness was >50 for 21 h (conditions necessary for conidial production and infection by airborne conidia) but no infections occurred, with only a few infections among larvae caged directly on soil during the following exposure period (13–15 April). For cages on watered soil, infections started in earnest beginning 25 April (Fig. 3). Soil moisture remained high throughout the study under watered trees (never >1.8 centibars; average ¼ 0.7 centibars) and therefore would not have limited resting spore germination early in the season. Under trees where the soil was not watered, infections were only seen during four exposure periods, between 25 April and 4 May, and averaged only 6.2 2.2%. During this period, infections were only >3% when soil moisture was <1.6 centibars and, during intervals when infections did not occur after 25 April, soil moisture always averaged >2 centibars. High variability in infection was seen among trees, with no infections at all occurring among larvae caged under one unwatered tree. Larvae caged above the soil that could only be infected by airborne conidia were very seldom infected. One larva out of 60 was infected at 2 cm height during three intervals (4–6 April, 25–27 April, and 6–8 May), one larva was infected at 10 cm height (2–4 May) and none of the larvae caged at 5 or 50 cm and 2 m height were infected throughout the study. Collections of gypsy moth larvae during the field season demonstrated that early in the season conidia were exclusively produced and resting spore production only increased late in the season (Fig. 4). Fourth instars were predominantly collected in weeks 4–5 in both 2000 and 2001 and fourth instar cadavers never contained resting spores. Resting spore production only began
471
Fig. 4. Percent production of conidia versus resting spores by E. maimaiga in cadavers of gypsy moth larvae collected throughout the field season in George Washington National Forest, Virginia and Monongahela National Forest, West Virginia. (A) 2000 and (B) 2001.
when fifth and later instars were present although cadavers that produced conidia as well as resting spores were found until the end of the season. To connect this with presence of lymantriid larvae, E. maimaiga-infected native lymantriids were only collected weeks 6 and 7 in 2000 when gypsy moths were P 5th instar and week 5 in 2001 when 4–5 instar gypsy moth larvae predominated.
4. Discussion
Fig. 3. Percent infection (mean SE) among L. dispar larvae caged for 2 days on top of soil around bases of either watered or nonwatered oaks at Elliot Springs Run, George Washington National Forest, Virginia, 4 April–8 May, 1997.
Native lymantriids remained at low densities throughout this study and only became infected by E. maimaiga during years when gypsy moth populations were more abundant and gypsy moth larvae were infected (2000 and 2001). The highest percent infection among native lymantriids was found among D. obliquata (5 of 23 larvae) although this was not the most abundant native lymantriid species collected. One cadaver of an E. maimaiga-killed D. obliquata larva had been collected in both 1989 and 1995 but D. obliquata larvae collected during epizootics in gypsy moth in
472
A.E. Hajek et al. / Biological Control 30 (2004) 466–473
Virginia in 1994 were not infected (Hajek et al., 1996a). Among all native lymantriid species collected during this study, infection was never >50% for any 1 year and, in fact, only approached this level for one species out of seven, during 1 year out of 5. During this study, the presence and activities of different types of E. maimaiga inoculum were studied to evaluate potential impact of this pathogen on native lymantriids. E. maimaiga resting spores were abundant in soil, especially during the early years of this study. Unless resting spores are added, resting spore titers generally decrease in density to some extent each year because some resting spores germinate whether gypsy moth larvae are present or not although many persist as a soil reservoir (Hajek and Eastburn, 2001). In agreement, between 1997 and 1999, resting spore titers declined although gypsy moth larvae were scarce. Field bioassays conducted during 1997 demonstrated the importance of soil moisture for resting spores to germinate. During 1997 and 1998, rain fell on >50% of the days when larvae were present in the field although rainfall amount and frequency decreased in 1999. Rainfall in 1998, when no infections were seen, was similar to 2000– 2001, when infections occurred. Therefore, we hypothesize that the reasons we did not find infected gypsy moth or native lymantriid larvae from 1997 to 1999 was not due to moisture limiting resting spore germination, especially in 1998. The present study documented for the first time a rapid increase in resting spore titer after E. maimaiga infections were abundant during one season (2000). This increase was seen in West Virginia plots in 2001 (Fig. 2), and occurred although gypsy moth density in the plots was very low (9 4 egg masses/ha in 2000). For E. maimaiga to begin infection cycles each season, resting spores in the soil germinate to actively eject infective germ conidia that can cause infections ( ¼ primary infection). Although germ conidia are actively ejected, no one has studied to what extent they become airborne. Results from our 1997 early season studies provide little indication that germ conidia infect many larvae above the ground level. After gypsy moth larvae become infected and die, then infective conidia actively ejected from larval cadavers cause Ôsecondary infectionsÕ and these conidia definitely become airborne (Hajek et al., 1999). Infections among larvae while they are in the tree and shrub canopy are most likely due to secondary infections. Models have shown that Ôsecondary infectionsÕ are responsible for the exponential increases in infection characteristic of epizootics (e.g., Hajek et al., 1993). Lymantriid larvae other than gypsy moth are not known to commonly rest in the leaf litter as is characteristic of later instar gypsy moth larvae. Therefore, native lymantriids would have much less risk of infection from the bank of resting spores in the soil compared with gypsy moth larvae. Native lymantriid larvae would
principally be exposed to infection initiated by E. maimaiga resting spores only if they traveled on the forest floor when the soil was moist. We hypothesize that it is more likely that native lymantriids become infected from conidia produced from cadavers while they are in the understory or tree canopy. During this study, cadavers of gypsy moth larvae predominantly produced conidia through much of the season, until late instars were present, at which time many late instar cadavers produced resting spores. We hypothesize that E. maimaiga infections were not found from 1997 to 1999 among native lymantriids because gypsy moth densities were too low to produce abundant conidia for secondary infections. In addition, because high levels of E. maimaiga infection in gypsy moth populations often occur late in the season (Hajek, 1999) and many of the native lymantriids are earlier than gypsy moth and would have pupated by the time gypsy moth are late instars, we hypothesize that relative seasonality of these species would result in native lymantriids largely escaping periods when airborne conidia of E. maimaiga might be abundant. Unfortunately, during this study we could not confirm that each of the infections in native lymantriids was caused by E. maimaiga. E. maimaiga belongs to the same species complex as the morphologically identical Entomophaga aulicae (Reichardt in Bail) Humber, which is native to the US, and both species infect lepidopteran larvae. Although E. aulicae does not infect gypsy moth, it has been reported infecting the lymantriids Calliteara pudibunda (L.) ( ¼ Dasychira pudipunda) (Lakon, 1919), Euproctis chrysorrhoea (L.), and Orgyia antiqua nova Fitch (reported as O. nova) (Speare and Colley, 1912), Orgyia vetusta Boisduval (Hajek et al., 1996b), and O. leucostigma (van Frankenhuyzen et al., 2002). In the field E. maimaiga has been confirmed as infecting the North American lymantriids D. basiflava, Dasychira leucophaea (J.E. Smith), D. obliquata (Hajek et al., 1996a; A.E.H., unpubl. data) and O. leucostigma (Hajek et al., 2000). We assume that infections of native lymantriids during this study were caused by E. maimaiga because this fungus has previously been shown to infect native lymantriids in the field, the 1996 gypsy moth populations in Virginia and West Virginia crashed and resting spores in soil were subsequently abundant and high percentages of gypsy moth larvae caged over soil became infected by E. maimaiga in 1997, and abundant infections were found among gypsy moth larvae collected at the same times and locations as native lymantriids in 2000 and 2001.
Acknowledgments We thank B. Knoblauch, C. Eastburn, D. Blue, M. Bertoia, N. Bertoia, J. McNeil, A. Savage, T. Kun-Lin
A.E. Hajek et al. / Biological Control 30 (2004) 466–473
and K. Zuniga for assistance with rearing and monitoring larvae, D. Curry for counting gypsy moth egg masses and T. Carrington, M. Whitman, R. Braud, K. Grubb, and J. Parrish for collecting larvae and soil samples in Virginia and West Virginia during the field season. L. Liebherr helped with data summarization, A. Burke assisted with graphics and C. Nielsen assisted with comments on the manuscript. We thank R. Reardon for his continued interest in this subject. Studies were funded through USDA, Forest Service Cooperative Agreements to L.B. and A.E.H.
References Bernon, G.L., 1995. Status of mass-reared gypsy moths: protocol. USDA, For. Serv. Northeast. For. Exp. Stn. Gen. Tech. Rpt. NE213, p. 54. Butler, L., 1992. The community of Macrolepidopterous larvae at CooperÕs Rock State Forest, West Virginia: a baseline study. Can. Entomol. 124, 1149–1156. Butler, L., Kondo, V., 1991. Macrolepidopterous moths collected by blacklight trap at CooperÕs Rock State Forest, West Virginia: a baseline study. WV Agric. For. Exp. Stn. Bull. 705, 25 pp. Butler, L., Strazanac, J., 2000a. Macrolepidopteran larvae sampled by tree bands in temperate mesic and xeric forests in eastern United States. Proc. Entomol. Soc. Wash. 102, 188–197. Butler, L., Strazanac, J., 2000b. Occurrence of Lepidoptera on selected host trees in two central Appalachian national forests. Ann. Entomol. Soc. Am. 93, 500–511. Ferguson, D.C., 1978. Noctuoidea (in part), Lymantriidae. In: Dominick, R.B., et al. (Eds.), The moths of North America. Fasc. 22.2. E. B. Classey and The Wedge Entomological Research Foundation, London, pp. 1–110. Hajek, A.E., 1999. Pathology and epizootiology of the Entomophaga maimaiga infections in the forest Lepidoptera. Microbiol. Mol. Biol. Rev. 63, 814–835. Hajek, A.E., 2001. Larval behavior in Lymantria dispar increases risk of fungal infection. Oecologia 126, 285–291. Hajek, A.E., Eastburn, C.C., 2001. Effect of host insects on activation of Entomophaga maimaiga resting spores. J. Invertebr. Pathol. 77, 290–291. Hajek, A.E., Wheeler, M.M., 1994. Application of techniques for quantification of soil-borne entomophthoralean resting spores. J. Invertebr. Pathol. 64, 71–73. Hajek, A.E., Butler, L., Wheeler, M.M., 1995. Laboratory bioassays testing the host range of the gypsy moth fungal pathogen Entomophaga maimaiga. Biol. Control 5, 530–544.
473
Hajek, A.E., Olsen, C.H., Elkinton, J.S., 1999. Dynamics of airborne conidia of the gypsy moth (Lepidoptera: Lymantriidae) fungal pathogen Entomophaga maimaiga (Zygomycetes: Entomophthorales). Biol. Control 16, 111–117. Hajek, A.E., Larkin, T.S., Carruthers, R.I., Soper, R.S., 1993. Modeling the dynamics of Entomophaga maimaiga (Zygomycetes: Entomophthorales) epizootics in gypsy moth (Lepidoptera: Lymantriidae) populations. Environ. Entomol. 22, 1172–1187. Hajek, A.E., Walsh, S.R.A., Strong, D.R., Silver, J.C., 1996b. A disjunct Californian strain of Entomophaga aulicae infecting Orgyia vetusta. J. Invertebr. Pathol. 68, 260–268. Hajek, A.E., Butler, L., Liebherr, J.K., Wheeler, M.M., 2000. Risk of infection by the fungal pathogen Entomophaga maimaiga among Lepidoptera on the forest floor. Environ. Entomol. 29, 645–650. Hajek, A.E., Humber, R.A., Elkinton, J.S., May, B., Walsh, S.R.A., Silver, J.C., 1990. Allozyme and restriction fragment length polymorphism analyses confirm Entomophaga maimaiga responsible for 1989 epizootics in North American gypsy moth populations. Proc. Natl. Acad. Sci. USA 87, 6979–6982. Hajek, A.E., Butler, L., Walsh, S.R.A., Silver, J.C., Hain, F.P., Hastings, F.L., ODell, T.M., Smitley, D.R., 1996a. Host range of the gypsy moth (Lepidoptera: Lymantriidae) pathogen Entomophaga maimaiga (Zygomycetes: Entomophthorales) in the field versus laboratory. Environ. Entomol. 25, 709–721. Lakon, G., 1919. Die Insektenfeinde aus der Familie der Entomophthoreen. Z. Angew. Entomol. 5, 161–216. Liebhold, A., Thorpe, K., Ghent, J., Lyons, D.B., 1994. Gypsy moth egg mass sampling for decision-making: a userÕs guide. USDA, Forest Service, Forest Health Protection, NA-TP-04-94, 12 pp. SAS Institute, 1999. SAS/STAT UserÕs Guide, Version 8. SAS Institute, Cary, NC, USA. Soper, R.S., Shimazu, M., Humber, R.A., Ramos, M.E., Hajek, A.E., 1988. Isolation and characterization of Entomophaga maimaiga sp. nov., a fungal pathogen of gypsy moth, Lymantria dispar, from Japan. J. Invertebr. Pathol. 51, 229–241. Speare, A.T., Colley, R.H., 1912. The Artificial Use of the Brown-Tail Fungus in Massachusetts with Practical Suggestions for Private Experiments, and a Brief Note on a Fungus Disease of the Gypsy Caterpillar. Wright & Potter, Boston, MA. US Environmental Data Service, 1997–2001a. Climatological Data: Virginia. Vols. 107–111. US Environmental Data Service, 1997–2001b. Climatological Data: West Virginia. Vols. 105–109. van Frankenhuyzen, K.S., Ebling, P., Thurston, G., Lucarotti, C., Royama, T., Guscott, R., Georgeson, E., Silver, J., 2002. Incidence and impact of Entomophaga aulicae (Zygomycetes: Entomophthorales) and a nucleopolyhedrovirus in an outbreak of the whitemarked tussock moth (Lepidoptera: Lymantriidae). Can. Entomol. 134, 825–846.