Comparative dispersal and larvicidal activity of exotic and Azorean isolates of entomopathogenic nematodes against Popilliajaponica ( Coleoptera : Scarabaeidae )

The Japanese beetle, Popillia japonica Newman, is an introduced pest on Terceira, one of nine islands in the Azorean Archipelago. Research conducted on Terceira indicates that entomopathogenic nematodes in the families Steinemematidae and Het­ erorhabditidae provide good to excellent control of Japanese beetle larvae, but the species that have been evaluated are not native to the Azores. An efficacious species that is native to the archipelago might provide increased capabilities for persisting and recycling in Azorean soil and weather conditions. Surveys on the islands of Terceira and Santa Maria resulted in the isolation of two Hetero­ rhabditis strains (Sao Mateus and Praia Formosa) with good larvicidal activity for P. japonica. Initial bioassays conducted with Steinernema glaseri (Steiner) originally from North Carolina against P. japonica third instar larvae and pupae produced LC50 values of 3.2 x 105 infectivejuveniles (IJs)/m2 and 0.9 x 105 IJs/m2, respectively. Comparative bioassays of the native isolates and S. glaseri against P. japonica revealed similar larvicidal activity. The LC5os of the Sao Mateus and Praia Formosa isolates against third instar larvae were 3.64 x 105 and 4.44 x 105 IJs/m2, respectively. The LC50 of S. glaseri ranged from 3.2 to 5.5 x 105 IJs/m2 . The higher larvicidal activity of the Azorean Heterorhabditis isolates for P. japonica indicates that native nematodes are as effective as S. gla­ seri. Heterorhabditid species also have demonstrated ability for persistence and apparent recycling under conditions where sustain­ able control of this introduced pest is needed. Studies comparing the dispersal behavior of the Heterorhabditis bacteriophora Poinar Sao Mateus isolate with that of S. glaseri and native and exotic strains of Steinernema carpocapsae (Weiser) revealed that the H. bacteriophora isolate demonstrated a greater propensity to disperse than other strains in the presence or absence of P. japonica lar­ vae. In the presence of a host, a greater proportion of H. bacteriophora and S. glaseri dispersed than either of the two S. carpocapsae strains.


INTRODUCTION
The Japanese beetle, Popillia japonica Newman, is an introduced pest on Terceira Island (Azores, Portugal).Adults are present from June until October and are most abundant in July and August (Simoes & Martins, 1985).The larvae and pupae inhabit the soil of pastures and lawns for the remainder of the year.Since its accidental introduction in the early 1970s into Terceira, it has spread to virtually every part of the island (Martins et al., 1988).Attempts to control the beetle with chemical insecticides have provided only temporary local suppression.Other sustainable methods for the long-term control of the beetle are needed.Biological control through the use of parasites and pathogens of P. japonica offers significant potential for sustainable control on an island habitat (Martins & Simoes, 1988;Lacey et al., 1994).Entomo pathogenic nematodes in the families Steinemematidae and Heterorhabditidae offer excellent potential for control of insects in soil habitats.(Klein et al., 2000).Research conducted on Terceira (Simoes et al., 1993;Lacey et al., 1994) and in the USA (Klein & Georgis, 1992) indicates that entomopathogenic nematodes provide effective con trol of Japanese beetle larvae and, under certain condi tions, persist and recycle in the host population.The nematode species that have been previously evaluated on Terceira were isolated from areas outside of the Azores.The most promising of these, Steinernema glaseri (Stei ner), was originally isolated from Japanese beetle larvae in the United States (Glaser & Fox, 1930).
Surveys conducted around the world, including other island habitats, reveal a diverse fauna of entomopatho genic nematode species.Rather than introducing a non native entomopathogenic nematode for Japanese beetle larval control, an efficacious species that is native to the Azorean Archipelago might provide increased capabilities for persisting and recycling under soil and climatic condi tions in Terceira.
A major determinant for selection of candidates for insect control is the ability of the nematode to find and penetrate its host as quickly as possible.The searching and dispersal behavior of entomopathogenic nematodes has been investigated by numerous researchers for several nematode species.The propensity for nematodes to disperse is strongly correlated with the strategy for host finding (Kaya & Gaugler, 1993).The "ambushers" tend to wait for passing mobile host insects, while "cruisers" actively seek out hosts.A variety of environmental fac tors including the effect of temperature, soil type, depth and moisture, presence and type of vegetation, and prox imity of host insects affect the distance and rate of ver tical and horizontal movement of infective juveniles (Georgis & Poinar, 1983a, 1983b;Molyneux & Bedding, 1984;Choo & Kaya, 1991).
The objectives of our study were to determine the effectiveness of S. glaseri against third instars and pupae of P. japonica, to compare the efficacy of two native heterorhabditid isolates with S. glaseri against third instar P. japonica, and to study the comparative dispersal of endemic and exotic nematodes in the presence and absence of an insect host.

MATERIALS AND METHODS
Source of nematodes.Native Heterorhabditis sp.(Praia For mosa, Az 24 and Sao Mateus, Az 162) and Steinernema carpo capsae (Weiser) (Agualva, Az 20) were isolated from the islands of Santa Maria and Terceira using the Galleria mellonella (L.) bait method (Bedding & Akhurst, 1975;Kaya & Stock, 1997) during 1990 as part of a general survey of entomo pathogenic nematodes in the Azorean Archipelago (Rosa et al., 2000).Koch's postulates showed that the IJs obtained using the baiting technique were infectious to G. mellonella larvae.A sample of adult nematodes from the Sao Mateus AZ162 isolate was identified by Dr. George Poinar, Oregon State University, Corvallis, Oregon as Heterorhabditis bacteriophora Poinar.The Praia Formosa Az 24 isolate was not identified to species.The NC1 isolate of S. glaseri originally isolated in North Carolina, USA and the All strain of S. carpocapsae originally isolated in Georgia, USA were obtained from cultures that had been main tained at the University of the Azores, Ponta Delgada.
Nematode culture.The methods of production of S. glaseri used in bioassays were published previously (Lacey et al., 1993).IJs were produced on a pork kidney medium as described by Bedding (1981).They were harvested 13-15 days after nematode inoculation and stored in distilled water at 8-10°C.IJs of the Azorean isolates of Heterorhabditis species were pro duced in G. mellonella larvae according to Dutky et al. (1964) and stored in distilled water at 8-10°C until used in bioassays.All IJs were used within 2 weeks after harvest.For dispersal studies, all IJs were produced in G. mellonella larvae as described above.
Bioassays.Comparative bioassays of S. glaseri NC1 isolate were conducted against third instars and pupae of the Japanese beetle.Larvae and pupae of P. japonica were collected in the field and held overnight in soil to eliminate those that were damaged during collection.Larvae and pupae were then set up in soil in plastic containers.Bioassays were conducted in pas ture soil that had been heat treated at 65°C for 24 h and stored at room temperature for at least 5 days before use.The soil was mixed with distilled water to bring the moisture content to 15%.Six hundred grams of soil were added to each 0.95 liter plastic container.The surface area of the soil was 122.7 cm2 and soil depth was 5 cm.Ten P. japonica larvae were added to each con tainer into separate depressions in the soil (2 cm deep, 1 cm diam.) and covered with soil.Ten P. japonica pupae per con tainer were placed in other containers into separate depressions in the soil (1.5-2 cm deep) and covered with soil.Groups of three containers were treated with one of five concentrations of nematodes ranging from 105 to 106 IJs/m2 (1,230-12,300 IJs per container).The IJs were applied in suspensions in 10 ml of water over the surface of the soil in each container.Ten ml of water was also applied to each of three control containers.The containers were covered with a perforated plastic lid and held at 24°C for 7 days, at which time larval and pupal mortality was assessed.The bioassays were repeated 4 times over a 5-week period.
Similar studies as described above were conducted with fieldcollected third instars of the Japanese beetle using S. glaseri and two Azorean isolates of the Heterorhabditis species.Groups of three containers were treated with one of five concentrations of nematodes ranging from 105 to 106 IJs/m2 (1,230-12,300 IJs per container), and mortality was assessed 7 days later.Bioassays were repeated on six separate dates over an 18-week period.The relative infectivity ratios of S. glaseri and the Heterorhabditis species were calculated using LC50 values.
Dispersal assay.Comparative dispersal activity was assessed for native isolates of S. carpocapsae (Az 20) and H. bacteriophora (Sao Mateus, Az 162) and the exotic isolates of S. glaseri (NC1) and S. carpocapsae (All).Nematode dispersal was assessed in plastic tubes, 8 cm in length, each consisting of 4 sections, 2 cm length and 1 cm inner diameter, joined together with adhesive tape.Four such 8 cm tubes were connected to a central chamber (4 cm3) (Fig. 1).The tubes were each filled with 10.5 g of washed and autoclaved sand that had been mois tened with distilled water (10% w/w).The central chamber was filled with 5 g of sand.The distal ends of 4 different tubes were each inoculated with 1000 IJs of one of the 4 isolates of nema todes in 150 pl water.Assays were conducted with and without a third instar of P. japonica enclosed in the central chamber.
After 24 h at 23°C, the plastic tubes were carefully separated and the number of nematodes in different sections was deter mined by washing the sand and counting all live nematodes using a dissecting microscope.The experiment was replicated 12 times with each of the 4 isolates, with and without P. japon ica.
Statistical analysis.Mortality data from all assays were ana lyzed using probit analysis (LeOra Software, 1987).Students t-test and linear regression analyses were performed using SAS software (Ver. 6.12, 1996).Nematode migration data were sub jected to analysis of variance (ANOVA) and Tukey's studentized range test at the P = 0.05 level (SAS, 1996) ), was 11.5 ± 2.0% (larvae); 14.2 ± 2.1% (pupae).The relative infectivity values were calculated from the ratios of the LC50s with the LC50 for larvae producing the baseline value of 1.0.The LC50 value for larvae divided by that for pupae produced the relative infectivity value for pupae.
values were normalized using arcsine transformation before analysis.

RESULTS
Bioassays.Pupal P. japonica had a considerably lower LC50 than larvae after exposure to S. glaseri, although there was slight overlap in the 95% C.I. (Table 1).Com parison of the mortality response of pupae and larvae to the discriminating concentration of 5 x 105 IJs/m2 using Students t-test indicated a highly significant difference in mortality (85.9 ± 5.2% versus 55.9 ± 5.7%, respectively, (t = 3.92, df = 6, P = 0.008).There was a strong concentration-mortality response for both larvae and pupae of P. japonica exposed to increasing concentrations of S. glaseri IJs (Fig. 2).The larvicidal activity of the two Azorean isolates of Heterorhabditis was comparable to or slightly better than that observed for S. glaseri (Table 2).Although the Sao Mateus Az 162 isolate of H. bacteriophora appears to be more virulent than S. glaseri based on the LC50 ratios, there was overlap in the 95% C.I. indicating the differ ence was not significant.Also the larval mortality data for 1 2.5   1 were close to that observed for the Heterorhabditis isolates (Table 2).There was a strong concentration-mortality response of P. japonica larvae exposed to increasing concentrations of the two Hetero rhabditis isolates (Fig. 3).In the bioassays of S. glaseri that were conducted concomitantly with the Heterorhab ditis isolates, the concentration mortality response of P. japonica larvae was less pronounced than that observed for the Heterorhabditis isolates (Fig. 3) and the S. glaseri bioassays with larvae and pupae (Fig. 2).Dispersal studies.The distance covered by the four strains was significantly different among nematode spe cies (F = 169.58;df = 44; P = 0.0001).The percentage of the two S. carpocapsae strains dispersing beyond 2 cm within 24 h was minimal and not significantly different from one another (Fig. 4) in the presence or absence of a host insect.Less than 1% of IJs of the two S. carpocapsae strains dispersed 8 cm in the presence or absence of a host (Table 3).A significantly higher proportion of the native H. bacteriophora and NC1 S. glaseri dispersed farther than 2 cm in the presence of a host insect compared to the two S. carpocapsae strains (F = 211.70;d f = 44; P = 0.0001).The percentage of IJs dispersing more than 2 cm was significantly higher in H. bacteriophora than in S. glaseri in the presence and absence of a host insect (Fig. 4).In the absence of a host, the percentage of response in S. glaseri was not significantly different from that of the two S. carpocapsae strains.The percentage of IJs dis persing 8 cm was significantly higher for both H. bacteriophora and S. glaseri with a host insect present than in the absence of a host (Table 3).The mean (± s.e.m.) overall recovery rate of NC1, All, Az 20, and Az 162 IJs was 95.7 ± 1.0%, 94.6 ± 1.7%, 94.9 ± 1.2%, and 84.7 ± I. 5%, respectively.Mortality of the P. japonica larvae used in the dispersal studies was not monitored due to the brevity of the experiments.

DISCUSSION
The larvicidal activity of the Azorean Heterorhabditis isolates against P. japonica demonstrates that suitable biological control agents are available within the archipel ago.The heterorhabditids are as effective as S. glaseri; moreover, they have the ability to persist and recycle as indicated by their natural occurrence in the Azores (Rosa et al., 2000).The two isolates used in our study were col lected within 200 m of the sea.Sao Mateus is a port vil lage and the Praia Formosa site is a sandy beach with clumps of grasses.However, most pastures that are posi tive for P. japónica larvae occur above 150 m elevation.Rosa et al. (2000) found that the majority of sites (70%) positive for Heterorhabditis species were below 150 m elevation in soil ranging from pH of 5.6 to 6.3 and more often in association with native vegetation than intro duced vegetation.In contrast, the Steinernema species were more prevalent above 300 m and not found in asso ciation with native vegetation.Thus, further field trials with promising native isolates of Heterorhabditis and Steinernema species should be conducted where P. japónica larvae are causing serious problems to determine environmental conditions that might influence the effi cacy and persistence of these native species.
The susceptibility of insect pupae to entomopathogenic nematodes is variable ranging from highly susceptible to resistant to infection (Kaya & Hara, 1980;Henneberry et al., 1995).The high susceptibility of pupal P. japónica might be misleading because in nature, prior to pupation, mature third instar larvae make an earthen cell in which to pupate (Vittum et al., 1999).Although S. glaseri IJs readily infected the pupae, their entrance into the pupal cells under natural conditions could be restricted by the barrier of the earthen cell surrounding the pupa.
The dispersal behavior of the two S. carpocapsae strains, H. bacteriophora and S. glaseri was consistent with the host seeking behavior previously observed for each of these species.S. carpocapsae did not disperse very well in sand in the presence or absence of a host and confirmed the earlier dispersal studies conducted with this species (Kaya, 1990).In contrast, S. glaseri and H. bacte riophora demonstrated a high degree of dispersal behav ior.Our results with S. glaseri were similar to observa tions made by Lewis et al. (1992) who documented that this species responds strongly to host chemo-attractants, but in the absence of a host insect, its dispersal response was minimal.On the other hand, over 25% of the Azorean H. bacteriophora IJs dispersed beyond 2 cm within a 24 h period in the absence of host cues.With a host present, a significantly greater proportion of IJs (45%) dispersed beyond 2 cm.Overall, these data show that the heterorhabditids are more active than S. glaseri.
The behavior of the nematodes is critical to effective control of Japanese beetle larvae.Georgis & Gaugler (1991) evaluated 380 field treatments of entomopatho genic nematodes against this insect and concluded that S. carpocapsae was ill-adapted to infect Japanese beetle larvae under a wide range of conditions.One of the main factors is that S. carpocapsae is an ambusher strategist and is ineffective against an insect pest that occurs deep in the soil profile (Kaya et al., 1993).H. bacteriophora was effective against Japanese beetle larvae, but only when the hatch depth and soil moisture, temperature and type were favorable (Georgis & Gaugler, 1991).Another critical factor is that H. bacteriophora is a "cruiser" strategist and is found throughout the soil profile (Camp bell et al., 1996;Gaugler et al., 1997) and is well adapted to find, infect and kill Japanese beetle larvae (Wang et al., 1995) .Our studies indicate that the two Azorean heterorhabditid isolates are excellent biological control agent candidates for the Japanese beetle.
Native nematodes offer alternatives to the introduction of exotic species such as S. glaseri in island habitats where increasing public concern and/or legislation may restrict or prohibit such introductions.The documented and potential negative impact of introduced biological control agents into island habitats has been reviewed by Howarth (1991), Follett & Duan (1999) and others.In addition to untoward permanent effects of introduced spe cies on nontarget insects, there is also the potential for antagonistic effects on endemic species of entomopathogenic nematodes (Barbercheck & Millar, 1999).Accord ingly, wherever feasible, the use of native entomopathogenic nematodes against soil insect pests should be explored before introducing exotic ones.

Fig. 2 .
Fig.2.The mortality response of larvae and pupae of Popillia japonica exposed to several concentrations of infective juveniles of Steinernema glaseri.Control mortalities in larvae and pupae were 11.5 ± 2.0% and 14.2 ± 2.1%, respectively.

Fig. 4 .
Fig. 4. Influence of insect presence on the dispersal of infec tive juveniles of four nematode strains 2 cm or farther in hori zontal sand columns, 24 hours after adding 1000 IJ to the distal end of each column.Bars with the same letter are not signifi cantly different from one another at the P = 0.05 level (Tukey's multiple range test).(F = 194.8;df = 88; P < 0.01).Az20= Azorean Steinernema carpocapsae; All=exotic S. carpocapsae; Az162=Heterorhabditis bacteriophora; NC1=S.glaseri.

Table 3 .
The percentage of infectivejuveniles dispersing in horizontal sand columns in 24 h period in the presence and absence of an insect host.Means in the same row followed by the same lowercase letter are not significantly different from one another at the 0.05 level (Tukey's multiple range test).Means in the column in the same treatment group (presence or absence of P. japónica) followed by the same uppercase letter are not significantly different from one another at the 0.05 level (Tukey's multiple range test).