BEHAVIOR Responses of Rhyzopertha dominica (Coleoptera: ) to Its Aggregation Pheromones as Influenced by Trap Design, Trap Height, and Habitat

1, 2 1 3 PETER A. EDDE, THOMAS W. PHILLIPS, AND MICHAEL D. TOEWS

Environ. Entomol. 34(6): 1549Ð1557 (2005) ABSTRACT Lindgren multiple funnel traps and Japanese traps captured more lesser grain borers, Rhyzopertha dominica F. (Coleoptera: Bostrichidae), than did Pherocon II sticky traps or bucket traps when all were baited with the same aggregation pheromones. Bucket traps captured six-fold fewer than Lindgren four-unit traps. Retentions of captured R. dominica were not signiÞcantly different in traps with soapy water or dry insecticide as killing agents for either trap design but were signiÞcantly higher than those retained in traps lacking a killing agent. Lindgren eight-unit funnel traps captured a similar number of R. dominica compared with the four-unit funnel traps. More beetles were captured near grain storage facilities than in forests or in open Þelds. Trap height (1, 2, or 4 m above the ground) had no detectable effect across habitat but signiÞcantly interacted with habitat. Traps placed 1 or 2 m high near grain elevators and in open Þelds captured similar numbers of beetles and yielded higher catches of R. dominica than traps placed 4 m high in these habitats. The reverse was true in forest habitats. Captured R. dominica were similarly female-biased in all trap designs, and the proportion of females to males did not differ among trap heights or habitats in which they were trapped.

KEY WORDS stored-product , lesser grain borer, pheromones

THE LESSER GRAIN BORER, Rhyzopertha dominica (F.) Current management strategies for R. dominica in- (Coleoptera: Bostrichidae), is a serious pest of stored volve the use of broad-spectrum insecticides, partic- grain worldwide. Life history and development of R. ularly organophosphates or pyrethroid grain pro- dominica were described in early works (e.g., Potter tectants such as malathion, pirimiphos-methyl, 1935, Crombie 1941). R. dominica is highly polypha- chlorpyrifos-methyl, and deltamethrin (Arthur 1996). gous and has been recorded feeding on diverse food Insecticides are effective in many cases, but insecti- crops such as legumes, tubers, bulbs, cereals, and pack- cide resistance is evident in many populations of R. aging material made from wood (Potter 1935). Sus- dominica, because of excessive use (Benhalima et al. ceptible cereal crops include , maize, rice, and 2004). Additionally, insecticides can be harmful to (Hagstrum et al. 1999). R. dominicaÐinfested nontarget species and may pollute the environment grain decreases in value as a function of live insects, (Lorini and Galley 1999). Thus, there is a need to damaged kernels, or insect fragments in milled develop sustainable and environmentally friendly pest products. Unlike most primary stored-grain pests, R. management tactics. dominica is not known to attack cereals in the Þeld, but Effective control of R. dominica, as well as other it is a strong ßier and has been found infesting grain, stored-grain pests, with minimal insecticide use re- in supposedly clean stores, within weeks or months quires an integrated management approach combin- after storage (Gates 1995). This rapid colonization ing sanitation, monitoring, and other preventive prac- behavior, strong ßight ability and broad polyphagy, tices, including use of pheromone-baited traps. coupled with the fact that R. dominica has been Pheromone traps can detect pests, monitor their dis- trapped in diverse environments, including wood- tributions in storage facilities, and possibly manipulate lands substantial distances from grain stores (Cogburn their populations (Phillips et al. 2000). The two male- 1988), led us to suspect movement of this pest be- produced aggregation pheromones of R. dominica tween potentially natural habitats and grain storage are (S)-(ϩ)-1-methylbutyl(E)-2-methyl-2-pentenoate facilities. (dominicalure-1 [DL-1]) and (S)-(ϩ)-1-methylbutyl (E)-2,4-dimethyl-2-pentenoate (dominicalure-2 [DL-2]) 1 Oklahoma State University, Department of Entomology and Plant (Williams et al. 1981). Both pheromones are equally Pathology, 127 Noble Research Center, Stillwater, OK 74078. attractive to both sexes of the beetle in the Þeld and 2 Corresponding author, e-mail: [email protected]. 3 USDAÐARS Grain Marketing and Production Research Center, in the laboratory. Sensitive and reliable pheromone- 1515 College Ave., Manhattan, KS 66502. baited traps are needed for R. dominica.

0046-225X/05/1549Ð1557$04.00/0 ᭧ 2005 Entomological Society of America 1550 ENVIRONMENTAL ENTOMOLOGY Vol. 34, no. 6

Table 1. Summary of study sites

Habitat Geographic location Site description Forest sites (I) 36Њ03Ј N; 097Њ10' W; 313 m a.s.l. 607 ha of naturally regenerating woodland dominated by Eastern Red Cedar Juniperus virginiana L. (Cupressaceae), Post oak Quercus stellata Wangenh (Fagaceae), and Slippery elm Ulmus rubra Muhl. (Ulmaceae) at Ϸ348 trees/ha (II) 36Њ07'N; 097Њ W; 300 m a.s.l 268 ha of naturally regenerating woodland dominated by Chinkapin Oak Quercus muhlenbergii Engelm (Fagaceae), Hackberry Celtis occidentalis L. (Ulmaceae), Post oak Quercus stellata Wangenh (Fagaceae), Loblolly pine Pinus taeda L. (Pinaceae) and Redbud Cercis Canadensis L. (Fabaceae) at Ϸ340 trees/ha Open Þeld (I) 36Њ07' N; 097Њ06' W; 274 m a.s.l. An open Þeld of Ϸ34 ha used annually for small-grain breeding, soil fertilizing, variety evaluations and forage research. Several ofÞce buildings and grain storage facilities are located on site. More than half of the area was used for variety evaluations of wheat and hay crops during the study period. The remaining area was left fallow. (II) 36Њ07' N; 097Њ07' W; 267.61 m a.s.l. 30 ha of open Þeld used annually for hay production. Field was fallow for the duration of the study. Grain elevator (I) 36Њ07' N; 097Њ08' W; 276 m a.s.l. A training and grain storage facility having 58 steel bins with combined capacity of 1,143 metric tons of grains but holding Ϸ327 metric tons of newly harvested hard red winter wheat Triticum aestivum L. during the experimental period. In an unrelated study, 12 ea 4.6-metric ton bins at the study site, holding wheat at full capacity, were each infested weekly from May through June 2002 with 500 each of unsexed adults of Rhyzopertha dominica (Coleoptera: Bostrichidae), Cryptolestes ferrugineus (Coleoptera: Laemophloeidae) and Tribolium castaneum (Coleoptera: Tenebrionidae). A mid- sized feed mill is located south and adjacent to the test site. (II) 36Њ09' N; 097Њ38' W; 1035 m a.s.l. A grain elevator with four commercial bins with combined capacity of 14,800 metric tons of grains but held Ϸ3,000 metric tons of newly harvested hard red winter wheat T. aestivum L. during the duration of the study

a.s.l., above sea level.

Factors that may affect efÞcacy of pheromone- tion, and a large plot size to accommodate the large baited traps include target species, trap design, trap number of traps in the experiment. Trap types eval- height, time of day, dosage of pheromone per trap, and uated in the Þrst study were the Lindgren four-unit habitat (Barak et al. 1991, De Groot and DeBarr 1998, multiple funnel traps (PheroTech, British Columbia, Boucher et al. 2001). Consideration of these factors Canada), Japanese beetle traps (Tre´ce´, Inc., Salinas, would enhance consistency and efÞciency of a trap- CA), Pherocon II traps (Tre´ce´, Inc.), and Unitrap or ping program for R. dominica. Our objectives, there- bucket traps (Agrisense/Biosys, Columbia, MD; Fig. fore, were to evaluate the effects of trap design, trap 1). Lindgren funnel traps consist of a series of verti- height, and habitat on capture of R. dominica in pher- cally connected black-plastic cone-shaped funnels omone-baited traps. terminating in a white-plastic collection jar at the bottom (Lindgren 1983). A pheromone lure was sus- pended by a wire at the midpoint inside the second Materials and Methods funnel from the bottom of each trap. Japanese beetle Pheromone Lures. Pheromone lures used in the traps have four-Þnned (omnidirectional) veins on the experiments were fabricated in our laboratory. Brießy, top of a tapered cone leading to a collection cup. number 11.5 sleeve stoppers (Fisher, Pittsburgh, PA; Japanese beetle traps used in this study were yellow. referred to here as rubber septa) were Þrst cleaned by A pheromone lure was afÞxed by a wire to the center soaking overnight in dichloromethane and allowed to of one Þn on each Japanese beetle trap. Pherocon II air dry under a fume hood for 24 h. Pheromones were traps were diamond-shaped white cardboard traps de- applied to the interior of a rubber septum by a 50% signed to capture ßying insects such as moths and hexane solution containing 5 mg each of DL-1 and beetles on an inner sticky surface. A pheromone lure DL-2. Treated rubber septa were afÞxed to traps as was simply placed on the center of the bottom sticky described below, and new lures were used for each surface of each pherocon II trap. Bucket traps consist trapping period in each experiment. of a funnel-shaped plastic receptacle with a lid and Experimental Series 1: Trap Comparisons. Four holder for attaching lures mounted over a bucket for trap designs for ßying insects were evaluated in 2000 retaining captured insects. A pheromone lure was sus- in an open Þeld near Stillwater, OK (Table 1; open pended within the funnel, attached at the lid, of each Þeld I). The study site was selected based on proximity bucket trap. Bucket traps used in the study had white- to our laboratory, sustained high R. dominica popula- colored receptacles and green lids. ModiÞcations to December 2005 EDDE ET AL.: TRAPPING OF R. dominica IN DIFFERENT HABITATS 1551

Experimental Series 2: Effect of Trap Height and Habitat on Capture of R. dominica. We tested the hypothesis that captures of R. dominica in pheromone- baited traps would vary because of differences in trap height and habitat, which may reßect optimal ßight height and habitat preference of R. dominica, by con- ducting an experiment from 7 July to 10 October 2002 at six locations (Table 1). Two of each of three dif- ferent habitats were used: a forest, an open agricul- tural Þeld, and an open Þeld adjacent to a grain storage facility, hereafter referred to as wooded, open Þeld, and grain elevator, respectively. Based on the results of Experimental Series 1, the Lindgren four-unit mul- tiple funnel traps were used in this experiment. Three traps baited with pheromone lures were deployed at each of the six locations, and each trap was assigned to one of three heights at each location. Trap heights were measured as the distance from the ground to the bottoms of the collection cups: 1, 2, or 4 m. The 1- and 2-m-high traps were hung from vertical polyvinyl chloride pipe stands equipped with a horizontal top arm. Those at 4 m were attached to ropes hung on the top vertical arms of 5-m vertical metal pipes inserted in the ground. Ropes were run through pulleys bolted to the arm of the pipe to facilitate trap servicing. The three traps at each location were spaced 15Ð20 m apart and arranged in an east-west orientation, perpendic- Fig. 1. Trap types. (A) Pherocon II trap, (B) Unitrap or ular to the prevailing southerly winds. Traps at the bucket trap, (C) Japanese beetle trap, and (D) Lindgren grain elevator sites were placed at least 6 m away from multiple funnel trap. grain bins. Soapy water was used in the collection cups to prevent captured insects from escaping. Trap po- traps included inserting a Þner screen in the collection sitions were rotated weekly at each location to min- cups of the Lindgren traps and replacing the collection imize positional effect on trap catch. The study was cup of the Japanese beetle traps with a 100-ml glass jar. organized as a two-factor experiment. The main fac- These modiÞcations were necessary because R. do- tors in the experiment were habitat types and trap minica could escape through the original equipment. height, each of which had three levels, and there were With the exception of Pherocon II traps that had a two replicates represented by the two habitat loca- sticky surface, captured insects were prevented from tions of each type. Trapping occurred at each location escaping by placing pieces of No-Pest Strip (United for a 1-wk period and was repeated for 15 wk, so that Industries, St. Louis, MO; active ingredient: dichlor- 30 replications were accumulated. vos) in trap receptacles. Traps were hung from vertical Experimental Series 3: Retention of Trapped In- polyvinyl chloride pipe stands we inserted into the sects. R. dominica captured in our trap comparison soil, Ϸ1.7 m above the ground, which placed traps study and in the habitat and trap height study were above the grass that generally grew to a height of Ϸ1 restrained with insecticide strips and soapy water, m. The Þrst study was conducted as a randomized respectively. However, it has been suggested that dif- complete block design from 30 May to 15 June 2000. ferent trap designs, supplied with different killing Eight experimental blocks of traps were deployed in agents, might be differentially effective in retaining the Þeld as separate trap lines in east-west orienta- captured insects (Morewood et al. 2002, De Groot and tions, perpendicular to the prevailing southerly winds, Nott 2003). We therefore studied retention of cap- in which each of the four trap types was represented tured R. dominica in pheromone-baited Lindgren once and randomly assigned a position in each block. four-unit funnel and Japanese beetle traps using dif- There were 15Ð20 m between traps in a block and at ferent killing agents. We chose the Lindgren four-unit least 50 m between blocks. funnel and Japanese beetle traps because results from A second experiment was conducted in the same Experimental Series 1 suggested that these two trap Þeld to compare the efÞcacy of Lindgren four-unit types were equally effective in capturing R. dominica, traps versus the longer eight-unit funnel traps of the and these captured more R. dominica than others. same basic design in capturing R. dominica. The ex- The killing agents tested were soapy water and periment was conducted from 16 August to 1 Septem- insecticide strips emitting dichlorvos. Approximately ber 2000 and was deployed as a completely random- 60 ml of soapy water (2% vol:vol of Palmolive washing ized design in which four traps of each of the two liquid soap; Colgate-Palmolive Company, New York, designs were randomly arranged in the Þeld with a NY) was added to the collection cups of both trap minimum of 80 m between traps. designs, hereafter referred to as wet traps. Dichlorvos 1552 ENVIRONMENTAL ENTOMOLOGY Vol. 34, no. 6 was gradually emitted froma2by3by0.8-cm blocks Results cut from a No-Pest strip placed in the collection cups Experimental Series 1: Trap Design. The number of of the tested trap designs, hereafter referred to as dry R. dominica captured differed signiÞcantly (F ϭ 32.6; traps. Control traps consisted of pheromone-baited df ϭ 3,28; P Ͻ 0.001) among trap types. Lindgren Lindgren funnel traps and Japanese beetle traps left four-unit funnel traps and Japanese beetle traps cap- blank, i.e., with neither soapy water nor dichlorvos. tured the most beetles (Fig. 2). Bucket traps captured Lindgren funnel traps used for wet trapping were further modiÞed by placing 100-ml plastic cups in the six-fold fewer beetles than Lindgren traps. Analyses of proportions of females only revealed no signiÞcant collection cups to prevent drainage of soapy water ϭ ϭ ϭ through the wire mesh in bottom of the cups. The differences among trap types (F 1.7; df 3,28; P 0.193); a similar Þnding was obtained when examining modiÞed Japanese beetle trap had no drainage holes; ϭ ϭ ϭ however, no rainfall occurred for the 6-d duration of proportions of males captured (F 1.8; df 3,28; P 0.17). Female to male ratios were signiÞcantly female the experiment (Stillwater weather data: http://www. ϭ ϭ Ͻ mesonet.org). biased (F 100.5; df 1,56; P 0.001). This ratio ranged from 0.67 Ϯ 0.08 to 0.78 Ϯ 0.04 but did not differ The bioassay was conducted from 6 to 12 September ϭ ϭ ϭ 2004 in forest habitats (Table 1). Based on the results among trap types (F 0.03; df 3,56; P 0.992). from Experimental Series 2, the forest habitat was Capture of R. dominica in Lindgren four-funnel and ϭ ϭ ϭ selected for the retention study. The traps, arranged in eight-funnel traps was similar (F 0.8; df 1,7; P Ϯ Ϯ east-west orientations, were hung from PVC pipes Ϸ2 0.391), with average captures of 55.8 8.4 and 59.6 m above ground and were spaced 15Ð20 m apart. The 4.8, respectively. experimental design was a completely randomized Experimental Series 2: Trap Height and Habitat. block in which each of the two forest sites represented There was a signiÞcant interaction between habitat ϭ ϭ Ͻ a block. Treatments were replicated three times in and trap height (F 5.7; df 4,225; P 0.001). Traps each block to yield six replications. placed 1 or 2 m high near grain elevators captured Sex Ratio. Sexes were determined from sampled more beetles than the other habitat and trap height insects in Experimental Series 1 and 2 to determine if combinations (Fig. 3). The next largest capture was in different trap designs, trap height, and habitats af- traps placed at 4 m high near grain elevators and in the fected sex ratio of captured R. dominica. Sex was de- forest and Þeld sites. Mean numbers of beetles cap- termined by squeezing the abdominal body region to tured were not signiÞcantly different between forest extrude their genitalia, which were viewed under a and open Þelds when traps were placed at 1 or 2 m dissecting microscope (Crombie 1941). Generally, high. Traps at 4 m high in open Þelds captured the 30% or more of the insects captured in each trap was fewest R. dominica (Fig. 3). There was no signiÞcant sexed, but this number differed based on insect con- main effect for the trap height factor (F ϭ 2.1; df ϭ dition. Insects in the Experimental Series 2 were sam- 2,225; P ϭ 0.120); but the main effect for habitat factor pled from all treatments at Þve different trap-check was signiÞcant (F ϭ 64.0; df ϭ 2,225; P Ͻ 0.001). Traps dates. Because the numbers of beetles sexed were not in areas adjacent to grain elevators signiÞcantly (F ϭ equal between treatments within experiments, data 27.5; df ϭ 2,135; P Ͻ 0.001; Fig. 4) captured more R. were standardized by converting the number of males dominica than those in forest or open Þelds. The open- and females sexed per treatment into proportion of the Þeld habitat yielded the fewest beetles, four-fold number of beetles sexed per treatment. Data on the fewer than near the grain elevator sites, and about proportions of males and female captured in Experi- one-half the number captured in the wooded sites mental Series 1 were analyzed as a two-factor exper- (Fig. 4). iment in which the main factors were trap design and Analysis of proportion of R. dominica sexes captured beetle sex. Similarly, data on the proportions of males using a three-factor analysis of variance (ANOVA) and female captured in Experimental Series 2 were test showed minimum interactions among habitat, analyzed as a three-factor experiment in which the height, and sex (F ϭ 2.3; df ϭ 4,90; P ϭ 0.065). Inter- main factors were habitat type, trap height, and beetle active effects between trap height and beetle sex (F ϭ sex. 1.3; df ϭ 2,90; P ϭ 0.267) and for habitat and trap height Data Analysis. Trap catch data were analyzed using combinations (F ϭ 0.01; df ϭ 4,90; P ϭ 0.999) were not SAS PROC MIXED (SAS Institute 2001). Blocks (trap signiÞcant. However, there were signiÞcant interac- design study) and locations and weeks (habitat, trap tions between habitat and sex (F ϭ 3.3; df ϭ 2,90; P Ͻ height study) were considered as random effects in 0.04). More females than males were captured within the respective mixed models and therefore included in each habitat, but the sex ratio was consistent within the RANDOM statement within the PROC MIXED habitats (Fig. 5). Trap height and habitat were not code. Before data analysis, count and percentage data signiÞcant in the model. Female to male ratios were were transformed using the log(x ϩ 1) and square- signiÞcantly female biased (F ϭ 202.2; df ϭ 1,90; P Ͻ root arcsine transformation methods (Zar 1999), re- 0.001), with a mean ratio of Ϸ3:2. spectively, to satisfy the assumptions of normality and Retention of captured R. dominica did not differ homogeneity of variance. Actual means and SEs are signiÞcantly between traps with soapy waters and presented in the text, tables, and Þgures. TukeyÕs Stu- those with dichlorvos for either trap design (Fig. 6), dentized range test was used to separate means but these were signiÞcantly higher than those retained (Tukey 1953). in control traps (F ϭ 7.2; df ϭ 5,25; P Ͻ 0.001). December 2005 EDDE ET AL.: TRAPPING OF R. dominica IN DIFFERENT HABITATS 1553

Fig. 2. Mean Ϯ SEM number of R. dominica captured per trap in different trap types baited with aggregation pheromones (DL-1 and DL-2) in Stillwater, OK, from 30 May 2000 to 15 June 2000. 4-funnel, Lindgren four-unit funnel traps; Jap. Beetle, Japanese beetle trap; Sticky, Pherocon II sticky trap; Bucket, Unitrap or bucket trap. N ϭ number of replications. Bars with the same letter are not signiÞcantly different (␣ ϭ 0.05).

Discussion possible that observed differences in captures of R. Trap design signiÞcantly affected outdoor trapping dominica were caused by differences in the size of trap of R. dominica. Lindgren four-unit funnel traps and openings. Trap openings refer to exposed portions of Japanese beetle traps were the most effective traps for the traps through which beetles gain unhindered ac- R. dominica. The bucket trap was least effective. It is cess into traps. This is Ϸ1,194.5, 810.5, 450.0, and 235.7

Fig. 3. Mean Ϯ SEM number of R. dominica captured per trap per week in wooded sites, outdoor near grain elevators, and open Þelds at different trap heights in Stillwater, OK, from 7 July 2002 to 10 October 2002. Grain, outdoor of grain storage facilities; Wooded, wooded habitat; Open, open Þeld. Bars with the same letter are not signiÞcantly different (␣ ϭ 0.05). 1554 ENVIRONMENTAL ENTOMOLOGY Vol. 34, no. 6

Fig. 4. Mean Ϯ SEM number of R. dominica captured per trap per week across locations in wooded sites, outdoor in grain elevators, and open Þeld in Stillwater, OK, from 7 July 2002 to 10 October 2002. N ϭ number of replications. Bars with the same letter are not signiÞcantly different (␣ ϭ 0.05). cm2 in Lindgren four-unit traps, Japanese beetle traps, and Japanese beetle traps, which may mimic the sil- Pherocon II traps, and bucket traps, respectively. houette of a vertical tree trunk, could provide visual Traps with larger openings, such as Lindgren multiple stimuli that work in concert with the chemical stimuli funnel traps and Japanese beetle traps, were likely easier for R. dominica to access and enter. Alterna- tively, the shapes of Lindgren multiple funnel traps

Fig. 6. Mean Ϯ SEM number of R. dominica captured in pheromone-baited Lindgren four-unit funnel and Japanese beetle traps with collection cups left blank, dry with insec- Fig. 5. Proportions Ϯ SEM of R. dominica sexes captured ticide, or partially Þlled with soapy water in Stillwater, OK, per trap in wooded sites, near grain elevators, and open Þelds from 6 September 2004 to 12 September 2004. Funnel, in Stillwater, OK, from 7 July 2002 to 10 October 2002. Grain, Lindgren four-unit funnel traps; Japanese, Japanese beetle outdoor near grain storage facilities; Wooded, wooded hab- trap; Wet, soapy water; Dry, insecticide; Con, blank collec- itat; Open, open Þeld. N ϭ number of replications. Bars with tion cups. N ϭ number of replications. Bars with the same the same letter are not signiÞcantly different (␣ ϭ 0.05). letter are not signiÞcantly different (␣ ϭ 0.05). December 2005 EDDE ET AL.: TRAPPING OF R. dominica IN DIFFERENT HABITATS 1555 to elicit beetle response. A tree-like shape might have mean numbers of beetles captured in the wooded provided R. dominica, which is from a family of wood habitats was higher than the numbers captured in boring beetles, with a cue for orientation that is lack- open Þelds, even though wooded habitats also lacked ing in the other trap designs. Other species of wood- a stored grain source of R. dominica. R. dominica boring beetles are known to respond to tree-like traps rarely, if ever, infests crops in the Þeld before harvest, (Borden et al. 1986, Flechtmann et al. 2000). based on thousands of samples collected at harvest in Although Lindgren eight-unit traps have twice as Oklahoma (P.A.E. and T.W.P., unpublished data). many identical openings as the four-unit funnel traps, This pattern contrasts to other stored grain pests such mean capture of R. dominica was not signiÞcantly as Sitophilus zeamais Motschulsky (Coleoptera: Cur- different between the two trap designs; this indicates culionidae) and Callosobruchus maculatus F. (Co- that no signiÞcant improvement in trapping efÞcacy leoptera: Bruchidae), which will commonly infest dry- may be achieved beyond the optimum trap openings ing crops in the Þeld before harvest (Rees 2004). required to maximize access by insects into the trap. However, R. dominica is known to disperse over dis- Our observations contrast with those reported for tances and has been observed attacking unprotected Dendroctonus ponderosae Hopkins (Coleoptera: Sco- grain in storage a few weeks after binning (Gates lytidae), in which trap capture of the beetles was 1995). The source of R. dominica attacking newly doubled by doubling the heights of conventional stored grain remains unknown. We cannot rule out the Lindgren multiple funnel traps (Borden et al. 1986). possibility that beetles trapped in the woods in this The difference in our observations and those of Bor- study originated from host material other than stored den et al. (1986) underscore the existence of species- grain. There are anecdotal reports of R. dominica tun- speciÞcity in attractiveness of silhouettes of different neling in various tree species in the wild (Potter 1935, lengths (i.e., some prefer longer traps, others prefer Linsley 1944, Mathew 1987). These beetles have also shorter traps). Lindgren eight-unit funnel traps are been reared successfully on several wild fruits and bulkier and more expensive than the four-unit traps; seeds in the laboratory (Wright et al. 1990), indicating therefore, it might be more economical to employ the that forest habitats may serve as a temporary niches or later or Japanese beetle traps for outdoor trapping of provide alternative food sources for R. dominica when R. dominica. preferred grains like wheat are not available. Results from Experimental Series 3 conÞrmed that Little is known about how R. dominica orients to R. dominica responded equally to Lindgren four-unit host material not accompanied by pheromones. Pre- funnel traps and Japanese beetle traps. Similarly, traps liminary outdoor trapping experiments using whole with dichlorvos pieces or traps with soapy water, for wheat and wheat extracts have failed to consistently both trap designs, were equally effective in retaining capture R. dominica, indicating that the pest may not captured R. dominica; about one-half as many insects respond to host plant volatiles from a distance (P.A.E., were retained in traps lacking a killing agent. This unpublished data). Fadamiro et al. (1998) obtained contrasts with the Þndings of Morewood et al. (2002) similar results with another bostrichid grain pest, Pro- and De Groot and Nott (2003) on some species of stephanus truncatus. We suspect, as has been proposed Cerambycidae and Buprestidae. These researchers for many Scolytidae (Borden 1982), that pioneer male observed that dry traps with or without insecticide R. dominica, dispersing from natal habitats, arrive by retained fewer insects than traps with soapy water. chance in grain warehouses where they feed and re- The larger size, longer legs, and greater agility of lease aggregation pheromones, to which conspeciÞc species of Cerambycidae and Buprestidae may have males and females are attracted. The requirement of enabled them to tolerate and escape from dry traps feeding before release of pheromones is well estab- with insecticide (De Groot and Nott 2003) where the lished for R. dominica (Mayhew and Phillips 1994, much smaller R. dominica could not. Traps with soapy Bashir et al. 2003) As newly arrived males begin to water may be advantageous over dry traps if sex de- feed, pheromones are produced, thus making grain termination of captured R. dominica using the squeez- storage facilities more attractive and easy for dispers- ing method is desired. We observed that beetles cap- ing adults to Þnd. It is likely that the higher numbers tured in traps with soapy water were softer and less of beetles captured near grain storages using synthetic likely to be damaged when squeezing the abdominal pheromones in this study might have resulted from body region to extrude beetle genitalia. recruitment of beetles already being attracted to these We found R. dominica populations were higher near sites by natural pheromone sources in the grain bins. grain elevators than in open Þelds or wooded habitats The phenomenon of increased attractiveness of in- (Fig. 4). Perhaps this ranking reßects the relative fested food sources is thought to occur in other stored ability of these habitats to sustain R. dominica popu- product insects and is well established in several spe- lations. Having wheat, a primary host plant, in the cies of bark beetles (Borden 1982, Likhayo and grain bins during the study may either have increased Hodges 2000). Alternatively, the higher number of attraction to those locations or served as a source of beetles observed near grain elevator sites, relative to beetles. However, absence of readily available food other habitats tested, might have resulted from emi- sources in the open Þeld habitats, and the relatively gration of endogenous beetle populations from within long distance of these traps from populations of R. the grain bins, i.e., products of earlier infestations and dominicaÐinfested grain, might be responsible for the their progeny. However, it is not known if R. dominica lower numbers of beetles in open Þelds. However, the would leave a source of “unlimited” food supply, as 1556 ENVIRONMENTAL ENTOMOLOGY Vol. 34, no. 6 represented by the grain bins in our study locations, to Acknowledgments respond to pheromone signals from our traps outside We thank the anonymous reviewers and the editor for the bins. Further studies are required to determine comments and contributions on this manuscript. We thank which, if any, of these explanations is adequate. P. Morton, J. Robertson, J. Metzger, and E. Bonjour for Captures of R. dominica in pheromone-baited traps technical support. W. Burkholder graciously supplied the R. were signiÞcantly affected by interactions between dominica pheromones. The USDA CSREES (Risk Avoidance habitat and trap height. The forest sites had a closed and Mitigation Program) under Agreement 2000-05385 canopy of trees at Ϸ6Ð10 m above the ground. Pher- funded this project. omone-baited traps placed near the vegetation cano- pies in our wooded sites captured more R. dominica than traps placed at lower heights. Information on References Cited optimal ßight height of dispersing R. dominica in dif- Arbogast, R. 1966. Migration of Agraulis vanillae in Florida. ferent habitats is limited. However, some insect spe- Fla. Entomol. 49: 141Ð145. cies are known to adopt a predetermined ßight height Arthur, F. 1996. Grain protectants: current status and pros- when dispersing and would avoid obstacles encoun- pects for the future. J. Stored Prod. Res. 32: 293Ð302. tered on their ßight path (Arbogast 1966). It is prob- Barak, A., W. E. Burkholder, and D. L. Faustini. 1991. Fac- tors affecting the designs of traps for stored-product in- able that the optimal ßight height of R. dominica re- sects. J. Kans. Entomol. Soc. 63: 466Ð485. sponding to pheromone-baited traps is Ͻ4 m above Bashir, T., R. J. Hodges, L. A. Birkinshaw, D. R. Hall, and D. I. the ground, however, on encountering obstacles such Farman. 2003. Phenotypic plasticity of Rhyzopertha do- as tree trunks in wooded habitats, they would attempt minica pheromone signaling: the effects of different hosts to ßy up and over these. This maneuvering may have and presence of conspeciÞc females on male produced brought the beetles into contact with pheromone aggregation pheromone. J. Chem. Ecol. 29: 945Ð959. plumes released from traps placed at higher heights (4 Benhalima, H, M. Q. Chaudhry, K. A. Mills, and N. R. Price. 2004. Phosphine resistance in stored-product insects col- m) and be captured in them. This hypothesis is sup- lected from various grain storage facilities in Morocco. J. ported by the Þnding that traps placed 1 or 2 m high Stored Prod. Res. 40: 241Ð249. near grain elevators and open Þelds performed simi- Borden, J. H. 1982. Aggregation pheromones, pp. 74Ð139. In larly and yielded higher trap catches of R. dominica J. B. Mitton and K. B. Sturgeon (eds.), Bark beetles in than traps placed 4 m high in these habitats. Unlike North American conifers. University of Texas Press, Aus- wooded habitats, trap perimeters in open and grain tin, TX. Borden, J. H., D.W.A. Hunt, D. R. Miller, and K. N. Slessor. elevators sites do not have objects that might pose 1986. Orientation in forest Coleoptera: an uncertain out- obstacles to approaching beetles. come of response by individual beetles to variable stimuli, Traps baited with aggregation pheromones of R. pp. 97Ð109. In T. L. Payne, M. C. Birch, and C.E.J. dominica captured signiÞcantly more females than Kennedy (eds.), Mechanism in insect olfaction. Claren- males, irrespective of the habitat or trap height treat- don Press, Oxford, UK. ments. This is in agreement with previous observations Boucher, T. J., R. A. Ashley, R. G. Adams, and T. F. Morris. on several other stored-product and wood-boring bee- 2001. Effect of trap position, habitat, and height on the capture of pepper maggot ßies (Diptera: Tephritidae). J. tles that use male-produced aggregation pheromones Econ. Entomol. 94: 455Ð461. (Plarre and Vanderwel 1999, Cronin et al. 2000, Phil- Cogburn, R. R. 1988. Detection, distribution and seasonal lips et al. 2000, De Groot and Nott 2001). One possible abundance of Sitotroga cerealella and Rhyzopertha do- explanation is that the primary function of male-pro- minica as indicated by pheromone-baited adhesive traps, duced pheromones in R. dominica is to attract females p. 451. Proceedings of the XVIII International Congress of as potential mates, but other males exploit the signal Entomology, 3Ð9 July 1988, University of British Colum- for locating assembled females and resources (Landolt bia, Vancouver, Canada. Crombie, A. C. 1941. On oviposition, olfactory conditioning 1997, Phillips 1997). In this sense, the aggregation and host selection in Rhyzopertha dominica Fab. (Insecta, pheromones of R. dominica may function more as sex Coleoptera). J. Exp. Biol. 18: 62Ð79. pheromones, as suggested for the stored product pest Cronin, J. T., J. L. Hayes, and P. Turchin. 2000. Evaluations Oryzaephilus surinamensis L. (Coleoptera: Silvanidae) of traps to monitor southern pine beetle aerial popula- (White and Chambers 1989), the bark beetle Den- tions and sex ratios. Agric. For. Entomol. 1: 69Ð76. droctonus terebrans Zimm. (Phillips et al. 1990), and De Groot, P., and G. L. DeBarr. 1998. Factors affecting other Scolytidae (Raffa et al. 1993). captures of white pine beetle, Conopthorus coniperda (Schwarz) (Col., Scolytidae) in pheromones traps. Trap design, trap height, and habitat are critical J. Appl. Entomol. 122: 281Ð286. factors that affect responses of R. dominica to phero- De Groot, P., and R. Nott. 2001. Evaluation of six trap de- mone-baited traps. Among the trap types tested, the signs to capture pine sawyer beetles (Coleoptera: Cer- Lindgren trap proved to be most effective in trapping ambycidae). Agric. For. Entomol. 3: 107Ð111. R. dominica. Optimum trap height for R. dominica De Groot, P., and R. Nott. 2003. Response of Monochamus varies with habitat. For example, traps should be (Col., Cerambycidae) and some Buprestidae to ßight intercept traps. J. Appl. Entomol. 127: 548Ð552. placed closer to the canopy vegetation in wooded Fadamiro, H. Y., I. Gudrups, and R. I. Hodges. 1998. Upwind habitats and from 1 to 2 m high in open habitats. These ßight of is mediated by aggrega- factors should be considered to optimize monitoring tion pheromone but not food volatiles. J. Stored Prod. Res. of R. dominica. 34: 151Ð158. December 2005 EDDE ET AL.: TRAPPING OF R. dominica IN DIFFERENT HABITATS 1557

Flechtmann, C.A.H., A.L.T. Ottah, and C. W. Berisford. Þcity of Dendroctonus terebrans (Coleoptera: Scolytidae) 2000. Comparison of four traps for Ambrosia beetles in to enantiomers of its sex pheromones. Ann. Entomol. Soc. Brazilian Eucalyptus stands. J. Econ. Entomol. 93: 1701Ð Am. 83: 251Ð257. 1707. Phillips, T. W., P. M. Cogan, and H. Y. Fadamiro. 2000. Gates, M. W. 1995. Population dynamics of lesser grain Pheromones, pp. 273Ð302. In Bh. Subramanyam and D. W. borer, rusty grain beetle, and Cephalonomia waterstoni in Hagstrum (eds.), Alternatives to pesticides in stored- commercial elevators. MS thesis, Oklahoma State Uni- product IPM. Kluwer Academic Publishers, Boston, MA. versity, Stillwater, OK. Plarre, R., and C. Vanderwel. 1999. Stored-product beetles, Hagstrum, D. W., C. Reed, and P. Kenkel. 1999. Manage- pp. 149Ð198. In J. Hardie and A. K. Minks (eds.), Pher- ment of stored wheat insect pests in the USA. Integr. Pest omones of non-lepidopteran insects associated with ag- Manag. Rev. 4: 127Ð142. ricultural plants. CABI Press, London, UK. Landolt, P. J. 1997. Sex attractant and aggregation phero- Potter, C. 1935. The biology and distribution of Rhyzopertha mones of male phytophagous insect. Am. Entomol. 43: dominica (Fab.). Trans. Roy. Entomol. Soc. 83: 449Ð482. 12Ð22. Raffa, K. F., T. W. Phillips and S. M. Salom. 1993. Mecha- Likhayo, P. W., and R. J. Hodges. 2000. Field monitoring nisms and strategies of host colonization by bark beetles, Sitophilus zeamais and Sitophilus oryzae (Coleoptera: pp. 103Ð128. In T. Schowalter and G. Filip (eds.), Inter- Curculionidae) using refuge and ßight traps baited with actions among bark beetles, pathogens and conifers in synthetic pheromone and cracked wheat. J. Stored Prod. North American forests. Academic, London, UK. Res. 36: 341Ð353. Lindgren, B. S. 1983. A multiple funnel trap for scolytid Rees, D. 2004. Insects of stored products. CSIRO Publish- beetles (Coleoptera). Can. Entomol. 115: 299Ð302. ing, Collingwood, Victoria, Australia. Linsley, E. G. 1944. Natural sources, habitat, and reservoir SAS Institute. 2001. SAS/STAT userÕs guide, version 8. SAS of insects associated with stored products. Hilgardia. 16: Institute, Cary, NC. 187Ð224. Tukey, J. W. 1953. The problem of multiple comparisons. Lorini, I., and D. J. Galley. 1999. Delthametrin resistance in Department of Statistics, Princeton University, Prince- Rhyzopertha dominica (F.) (Coleoptera: Bostrichidae), a ton, NJ. pest of stored grain in Brazil. J. Stored Prod. Res. 35: White, P. R., and J. Chambers. 1989. Saw-toothed grain bee- 37Ð45. tle Oryzaephilus surinamensis (L.) (Coleoptera: Silvani- Mathew, G. 1987. Insect borers of commercially important dae) antennal and behavioral responses to individual stored timber in state of Kerala, India. J. Stored Prod. Res. components and blends of aggregation pheromone. 23: 185Ð190. J. Chem. Ecol. 15: 1015Ð1031. Mayhew, T. J. and T. W. Phillips. 1994. Pheromone biology Williams, H. J., R. M. Silverstein, W. E. Burkholder, and A. of the lesser grain borer, Rhyzopertha dominica (Co- Khorramshahi. 1981. Dominicalure 1 and 2: compo- leoptera: Bostrichidae), pp. 541Ð544. In B. R. Champ, H. J. nents of aggregation pheromone from male lesser grain Banks, and E. Highley (eds.), Proceedings of the Sixth borer Rhyzopertha dominica (F.). J. Chem. Ecol. 7: 759Ð International Working Conference on Stored-Product 780. Protection. CABI Press, Canberra, Australia. Wright, V. F., E. E. Fleming, and D. Post. 1990. Survival of Morewood, W. D., K. E. Hein, K. J. Katinic, and J. H. Borden. Rhyzopertha dominica (Coleoptera, Bostrichidae) on 2002. An improved trap for wood-boring insects, with fruits and seeds collected from woodrat nests in Kansas. special reference to Monochamus scutellatus (Coleoptera: J. Kans. Entomol. Soc. 63: 344Ð347. Cerambycidae). Can. J. For. Res. 32: 519Ð525. Zar, J. H. 1999. Biostatistical analysis. Prentice-Hall, Engle- Phillips, T. W. 1997. Semiochemicals of stored product in- wood Cliffs, NJ. sects: current status and future perspective. J. Stored Prod. Res. 33: 17Ð30. Phillips, T. W., J. L. Nation, R. C. Wilkinson, J. L. Foltz, H. D. Received for publication 10 November 2004; accepted 14 Pierce, and A. C. Oehlschlager. 1990. Response speci- September 2005.