PEST MANAGEMENT AND SAMPLING Methamidophos Application Effects on elongatus (Coleoptera: Carabidae): An Update

NANCY E. MCINTYRE

Department of Biology and Graduate Degree Program in Ecology, Colorado State University, Fort Collins, CO 80523

Environ. Entomol. 24(3): 559-563 (1995) ABSTRACT Population densities of a predatory ground , Pasimaclws elongatlJ8 LeConte (Coleoptera: Carabidae), were examined at the Central Plains Experimental Range, Downloaded from https://academic.oup.com/ee/article/24/3/559/2394861 by guest on 01 October 2021 Colorado. Using distance sampling theory, I analyzed density estimates dating from before an application of the insecticide methamidophos (1988) and three beetle generations afterward (1994) to address the following questions. Have population densities of P. elongatl18 recovered to prespray levels? Are cun-ent population densities of P. elongatl18 in areas exposed to meth- amidophos in 1988 equal to those not exposed in 1988? How do population densities of P. elongatlls vary with habitat type in methamidophos exposed and nonexposed areas? This work updates earlier findings that P. elongatus population densities decreased after methamidophos applications. Lingering differences in P. elongatus densities were found among areas differing in methamidophos exposure; these differences were also related to habitat type. Higher pop- ulation densities were found in methamidophos exposed than in control areas; higher densities were also found in shrub dominated than in grass dominated sites, regardless of past meth- amidophos exposure. Because potential prey species were found to be equally abundant in methamidophos exposed ~md nonexposed areas despite the greater abundance of the carniv- orous Pasimaclllts in treatment areas, the P. elongatus density patterns seen are discllssed in relation to possible long-term dismption of the prey community by methamidophos. Other explanations for the observed patterns (including competitive release and physiological resistance) are also discussed.

KEY WORDS Colorado, methamidophos, Pasimachus elongatl1s

Pasillwclws elongatlls LECONTE is a predatory ecosystem, Ebert and Kondratieff (1992) found (Carabidae) that ranges from the that densities of P. elongatlls were negatively cor- eastern United States to Montana ~md from south- related with methamidophos levels. P. elongatlls ern Canada to New Mexico (Noonan 1992), inhab- numbers declined by =90% after a methamido- iting dry woodlands and prairies (Arnett 1968). P. ph os application in 1988 and had returned to only elongatlls is an abundant and conspicuous member =40% of their original levels 1 yr later. of the fauna of the shortgrass steppe of the western Long-term effects of insecticides such as meth- United States and feeds on other adults and amidophos are of increasing ecological concern. larvae, including several economically destnlCtive Since Ebert and Kondratieff's 1988-1989 study, crop pests, thereby acting in an agriculturally ben- three generations of P. elongatus (a biennial spe- eficial capacity (Blatchley 1910, Cress and Lawson cies) have passed. My objective was to address the 1971). Because P. elongatlls is an economically im- following questions as an update of their work. portant insect species and an effective predator, Have population densities of P. elongatus recov- knowledge of its biology is an important compo- ered to prespray levels? Are current population Iwnt in understanding the grassland fauna as a densities of P. elongatus in areas exposed to meth- whole. amidophos in 1988 currently equal to those not Methamidophos (0, S-dimethyl phosphorami- exposed in 1988? How do present population den- dothioate) is a preharvest insecticide used to con- sities of P. elongatus vary with habitat type in meth- trol agricultural pests (WHO 1993). Methamido- amidophos exposed and nonexposed areas? These phos has been shown by Ebert and Kondratieff questions address the potential long-term effects (1992) to reduce population densities of P. elon- of a pesticide application event on local insect pop- ulations, and their answers will be compared with gatus. Because P. elongatlls feeds on other , tile null hypothesis of no lingering differences with it can be affected by insecticides both directly treatment or habitat. (through direct mortality) and indirectly (through mortality of prey organisms). Following a series of methamidophos applications in a shortgrass steppe

0046-225X/95/0559-0563$02.00/O © 1995 Entomolo!l;ical Society of America 560 ENVIRONMENTAL ENTOMOLOGY Vol. 24, no. 3

Materials and Methods _ = trap Study Site. This study was conducted at the 80 traps total 6,280-ha Central Plains Experimental Range N (CPER). The CPER is a National Science Foun- dation Long-Term Ecological Research area coor- NW -_ NE dinated by the U.S. Department of Agriculture's - Agricultural Research Service. Approximately 1,650 m in elevation and averaging =300 mm in . - - annual precipitation, it is located in the shortgrass -- - -- steppe of north-central Colorado 61 km NE of ------_ - _ - center circle Fort Collins. The site is divided into fenced pas- \.1,m _ ..:..v diameter = 0.5 m tures grazed by cattle since 1939 (Van Cleve and Downloaded from https://academic.oup.com/ee/article/24/3/559/2394861 by guest on 01 October 2021 Martin 1991). Pastures 13SES, 15SWE, 22CN, W_------__ ~ E 22CS 24NE and 30NE were used in the current - stud/ (pastt:res 22CN and 22CS have recently -. . been merged to create one large pasture, 22C. . -- Similarly, 13SES and 15SWE are now parts of larg------. -.- er pastures [13SE and 15SW, respectively]. How- ever, these former sections are still recognizable ------SW -. SE and will be considered as discrete regions in this article, and the former pasture nomenclature will therefore be retained.) - S In 1988, methamidophos was aelially applied - once in pastures 22CN and 24NE (at 1.15 kglha) Fig. 1. Trapping web. Not drawn to scale. and twice in a 1-mo period in pastures 13SES and 15SWE (at 0.29 kg [AI]/ha per application to give there would have a capture probability of a total exposure of 0.58 kg(AI)/ha, a total dosage 1, a key assumption in the use of trapping webs that was half that of the other methamidophos ex- (Anderson et al. 1983). Trapping effort is concen- posed pastures). The insecticide was in aqueous trated at the web center and dissipates toward the solution; 1.53 liters (AI)/ha of this solution was ap- outer edge. plied to each treatment pasttlre. Pashlres 22CS Each pitfall trap consisted of a SOO-mlplastic and 30NE were not exposed to the methamido- cup set into the ground with the lip of the cup phos solution and served as controls (Ebert and flush with the ground surface. A smaller, remova- Kondratieff 1992). P elongatus population densi- ble plastic cup was placed inside each larger cup ties were estimated in each of these pastures 1 wk along with a plastic funnel to serve as an escape before and following the first methamidophos ap- guard and to provide shade. Drainage holes were plication, after the second application, and a year drilled into the bottom of the cups. Traps were later (Ebert 1990). closed with wooden lids when not in use. Captured Because it rapidly degrades to nontoxic prod- were marked with a small dot of enamel ucts, methamidophos is not persistent in the en- nail polish on the elytra. Sampling was conducted vironment and does not accumulate, even after re- three times in June 1994, once in mid-July, and peated applications. It has a half-life of 15.9 d in twice in early August, following the procedures of water and 7.5 d in silt (WHO 1993). Because it has Ebert and Kondratieff (1992) with one minor mod- such a short life-span and because no methami- ification-traps were checked every third day in- dophos or other chemicals have been applied in stead of every day during each sampling interval. these areas since 1988, no remaining methamido- This does not alter the population estimates ob- phos was expected to be actively present in the tained because recaptured animals were not count- system. ed more than once (Anderson et al. 1983) and the Sampling Regime. I captured P elongatus bee- same number of trap-nights was present as in the tles in each of the six pastures by pitfall trapping previous study. Data from 17,280 trap-nights were with a 0.0638-ha trapping web (Anderson et al. collected. 1983, Parmenter et al. 1989) from 20 May to 12 Habitat Analyses. The CPER is dominated by August 1994. Web dimensions followed the design perennial warm-season grasses, especially blue gra- used by Ebert and Kondratieff (1992) with minor ma, Bouteloua gracilis (von Humbolt, Bonpland & modifications to meet the statistical assumptions Kunth) Lagasca, and buffalograss, Buchloe dactlj- for analysis of trapping web data. Each web con- loides (Nuttall) Engelmann. However, shmbs, pri- sisted of eight arms separated by 45° radiating marily saltbush, Atriplex canescens (Pursh) Nutt, from a central point with 10 traps per arm (Fig. and pasture sagebrush, Artemisia frigida von 1). Traps were separated by 1.5 m except at the Willdenow, attain high densities in certain pastures center, where they were placed adjacent to one (Whicker and Tracy 1987, Crist et al. 1992). To another in a O.S-m-diameter circle to ensure that assess the habitat type of each trapping area, the June 1995 McINTYRE: METHAMIDOPHOS ApPLICATION EFFECTS ON Pasil1wchus 561

Tnblt, 1. Avt'rnjl;"I••·r •.•.ntn~•. of trnl'. will. .I.rub cov- Ebert and Kondratieff (1992) (Laake et a!. 1994). ,'r wilbin 9 ('11\ p"r lruPI,in~ wd, (pu.lure) DISTANCE is based on distance sampling thcory, whereby the spacing between captured organisms Avg % with Pastllrt' shmb cover is used to determine their density (Wilson and An- derson 1985). It should be noted that it is this 22CS 2.50 22C:\T 10.00 spacing between areas of high numbers of captures 15S\\'1'; 3.75 and not the numbers themselves that exerts the 2·jNE 46.25 greater influence on the resulting density estimate 13SES 65.00 (BuckhU1det al. 1993). Density estimates are ob- 30NE 42.50 tained from significant fitted models of detection probabilities (see Buckland et al. 1993, Laake et al. 1994). DISTANCE is a robust technique of cal- percentage of traps possessing shrub cover in a 9- culating population density, as verined by several Downloaded from https://academic.oup.com/ee/article/24/3/559/2394861 by guest on 01 October 2021 cnl-dianlC'ter circle around each pitfall trap was field studies (e.g., Anderson et a!. 1983, Parmenter calculated and the average number of traps pos- et aI. 1989). However, DISTANCE does not per- sessing nearby shrub cover per trapping web then form well when dealing with low capture rates determined. This technique provided a measure of (fewer than ""15 animals), giving biased results. habitat t)1Jeand indicated (confirmed by ocular in- Therefore, densities in webs where fewer than 15 spection) that three of the six trapping webs were animals were captured were estimated as number located in pastllfes with relatively high densities of of animals caught per trapping web area and have shrubs (pastllfes 13SES, 24NE, 30NE), and the no confidence intervals associated with them; this other three were in locations with few shnlbs procedure may give inflated estimates and is one (15SWE, 22CN, 22CS) (Table 1). Within each of the reasons for the development of DISTANCE main habitat type (grass and shmb), one pasture (Anderson et a!. 1983). Biases from low sample st'lvt'd as a control, one received a low dosage of sizes as well as from greater trap concentration in nwthamidophos (0.58 kg [AIl/ha), and one re- the web center are remedied during model fitting ceived a high dosage (1.15 kg [AIl/ha) in 1988. P by way of truncation (in the case of center over- clollgatrls is present in both habitat types. How- representation) (see Buckland et a!. 1993) or sim- l'ver, these two habitats were not examined sepa- ply estimating density as number of animals per ratt'ly by Ebert and Kondratieff (1992) and there- trapping web area (in the case of low sample size). fore they could not assess any differences in Estimated densities of P elongatus were com- population'density in methamidophos exposed and pared among areas differing in methamidophos ex- nonexposed grass and shmb habitats. To examine posure by comparing overlap in 95% CI and tested how this habitat variation affected the beetles, I against the null hypothesis of no difference with separated pastures according to their habitat type methamidophos exposure. These densities wcre as well as their level of 1l1ethamidophosexposure. also compared with those dating from before the Associated Prey .Analyses. P. elongatus is a 1988 methamidophos application and tested gelH'ralist predator (Cress and Lawson 1971) against the hypothesis of no lingering difference whost' distribution is associated with the distribu- with methamidophos treatment. Finally, current tion of its prey (Calkins and Kirk 1974). Because densities from different habitats (grass and shmbs) f(>wcomprehensive studies of this species exist, were compared and tested against the hypothesis howevpr, its diet is not well understood, particu- of no difference with habitat type. larly on the Colorado shortgrass steppe. To exam- ine the possibility that any differences present in Results P dOllgatlls numbers could be related to differ- t'nces in the abundance of their prey in metham- Current Versus Premethamidophos Popula- idophos l'xposed and nonexposed areas, overall ar- tion Densities and Present Differences in Meth- thropod abundances were compared by way of amidophos Exposed and Nonexposed Areas. analysis of variance (ANOVA) among areas differ- Current population densities of P elongatus are ing in 1l1('thamidophosexposure and in habitat (So- equivalent to prespray (1988) levels, as noted in kal and Rohlf 1981). were captured the overlap of the 95% CI (Table 2). However, with the same trapping web pitfall trap design used while no differences were found between treat- to trap P elollgatlls. Ants (Hymcnoptera: Formic- ment and control areas in 1988, current population idae) were not included in analysis because of dif- densities of P elongatus were higher in areas ex- ficulty of identification and lack of evidence that posed to methamidophos compared \vith unex- they are a food item for P elongatl1s. posed areas. No significant differences were found Data Analysis. The computer package DIS- in densities between areas exposed to heavy dos- TANCE was used to obtain P elongatlls density ages of insecticide and areas exposed to light dos- estimates (l'J\1)ressedas number of individuals per ages. hectare) and 95% CI (see Buckland et al. 1993, Differences with Habitat Type. Differences in pp. 88-89, for formula of confidence intervals); this current and prespray population densities with package replaced program TRANSECT used by methamidophos exposure were confounded with 562 ENVIRONMENTAL ENTOMOLOGY Vol. 24, no. 3

Table 2. Estimates of population densities of P. elon- Table 3. Estimates of population densities of P. elon- galus in areas exposed to varying amowlts of methami- . galus in areas differing in habitat type and i•• level of dophos, pooled across replicate sites and sampling dates methamilIol.hos exposure, pooled across replicate sites and sampling dates No. Level of individuals/ha, 95% Cl No. exposure 110. captures Habitat type individualslha, 95% CI no. eaptllres Pre-Exposure (1988) 1.15 kg/Ila (high) 979.41, 26 522.26-1,836.70 Pre-Exposure (1988) 0.58 kg/ha (low) 1,810.70, 16 944.09-3,473.00 Grass 212.21", 15 o (control) 4,191.10, 51 1,169.70-15,017.00 Shrub 7,179.80, 1I5 4,847.60-10,634.00 Grass/High 70.74",5 Current (1994) Grass/Low 84.88",6 1.15 kglha (high) 1,344.60, 69 721.41-2,506.00 Grass/Control 56.59",4

0.58 kglha (lnw) 12,734.00, 94 484.26-334,830.00 Shrub/High 943.74, 21 475.84-1,871.80 Downloaded from https://academic.oup.com/ee/article/24/3/559/2394861 by guest on 01 October 2021 o (control) 141.08",9 Shrub/Low 4,707.90, 49 257.34-86,128.00 Shmb/Control 7,995.50, 47 382.94-166,940.00 (, Densities ill webs where <15 animals were captured were es- timated as number of animals caught per trapping web area and Current (1994) have no confidence intervals associated with them. Grass 125.4oa, 8 5hmb l,538.40, 40 881.87-2,683.50 Grass/High 62.7oa,4 Grass/Low 47.00",3 differences in habitat types. Although current P Grass/Control 31.30",2 elongatus densities were approximately equivalent 5hmb/High 1,208.30, 65 623.46-2,341..60 to prespray levels, areas dominated by shrubs had Shrub/Low 13,248.00, 91 510.60-343,750.00 Shrub/Control 125.40",8 higher estimated densities than did grassy sites (Table 3). Shrub-dominated sites displayed higher Exposure levels correspond to values in Table 1. Number of variation in capture numbers than did grassy sites captures in habitats broken down by methamidophos exposure before and 6 yr after the methamidophos appli- level do not nec"essarily sum to overall habitat captures because cation. The highest prespray P elongatus density of data tmncation in model fitting. " Densities in webs where <15 animals were captured were es- occurred in shrubby habitat not exposed to meth- timated as number of animals caught per trapping web area and amidophos whereas the highest 1994 density oc- have no confidence intervals associated with them. curred in shrub dominated areas exposed to low insecticide doses. Associated Prey. ANOVA revealed that arthro- bers in methamidophos exposed versus nonexpos- pods were equally abundant in methamidophos ed areas. As such vegetative data do not exist for treatment and control areas; shrub dominated hab- my study area, however, this possibility remains to itats likewise supported as many arthropod species be proven. and individuals as grass dominated areas (a = 0.05; Although it is unknown if predators like P elon- Table 4). Approximately 75% of the arthropod spe- gatus regulate prey numbers in this system, they cies caught were captured in both habitats (i.e., probably exert some influence. The equal abun- there was little uniqueness in the fauna of a par- dance of prey in areas with high and low predator ticular habitat type). densities (Table 4) suggests that prey numbers would be even higher in methamidophos exposed areas if there were an absence of predators like P Discussion elongatus, which concurs with the situation de- Surprisingly, more P elongatus individuals were scribed above. found in areas that had been exposed to insecticide Differences in population numbers with habitat (Table 2). Several explanations for this pattern are type were also important. The data in Table 3 sug- possible, including physiological adaptation to gest that areas dominated by shmbs provide some methamidophos and release from competitive attractant (e.g., favorable microclimate) to P elon- pressure by elimination of other members of the predatory insect guild. Because methamidophos is Table 4. Results of ANOVAtest for differences in ar- a broad-spectrum insecticide, tllis pattern could thropod abwldunees (number of species and individuals) also be caused by a lingering disruption caused by in areas differing in habitat type and in level of meth- methamidophos-related fatalities to other mem- arnidophos exposure, pooled across replicate sites and bers of the arthropod community. Exposure to sampling dates methamidophos may have reduced the numbers of Source of variation df F P herbivorous insects present, which in tum may have reduced P elongatus numbers. A flourishing No. arthropod species of vegetation may have resulted from this reduc- Methamidophos treatment 2 0.68 0.5609 Habitat tion of herbivory. Abundant vegetation may have 1 3.53 0.4419 then acted as an attractant to herbivores and No. arthropod individuals thence to predators once insect population levels Methamidophos treatment 2 3.38 0.3171 Habitat 1 4.54 0.3611 recovered, resulting in higher P elongatus num- June 1995 McINTIRE: METHAMIDOPHOS ApPLICATION EFFECTS ON Pasimachus 563 gatlls that influences the effects of former meth- Cress, D. C. and F. A. Lawson. 1971. Life history of amidophos exposure. Interaction between insecti- . J. Kans. Entomol. Soc. 44: cide effectiveness and habitat heterogeneity merits 304-313. further study. Crist, T. 0., D. S. Guertin, J. A. Wiens, and B. T. Milne. 1992. movement in heterogeneous Because there are differences in current P elon- landscapes: an experiment with Eleodes beetles in gatlts population densities between exposed and shortgrass prairie. Funct. Ecol. 6: 536-544. nonexposed areas, this suggests that methamido- Ehert, T. A. 1990. Interactions between pesticide ap- l?hos use evokes long-term changes to ecosystem plications and food resources in a PasillUlchus elon- fauna. Direct changes in one faunal component gatus population. M.S. thesis, Colorado State Univer- may in turn result in important indirect community sity, Fort Collins. and ecosystem effects (Menge et al. 1986), possibly Ehert, T. A. and B. C. Kondratieff. 1992. Effects of in a cascading manner from one trophic level to a methamidophos application on PasirrUlchuselonga- another (Fairweather 1990), as may be the case tus LeConte (Coleoptera: Carabidae). J. Kans. Ento- Downloaded from https://academic.oup.com/ee/article/24/3/559/2394861 by guest on 01 October 2021 with this system. My results indicated that more mol. Soc. 65: 151-156. studies on long-term insecticide effects (including Fairweather, P. G. 1990. Is predation capable of in- indirect effects and effects mediated by habitat teracting with other community processes on rocky reefs? Aust. J. Ecol. 15: 453-464. variability) on arthropod populations and commu- Laake, J. L., S. T. Buckland, D. R. Anderson, and K. nities are warranted. P. Burnham. 1994. DISTANCE user's guide V2.1. Colorado Cooperative Fish and Wilcllife Research Acknowledgments Unit, Colorado State University, Fort Collins. Menge, B. A., J. Luhchenco, L. R. Ashkenas, and F. I thank Scott Maltzahn (Depmtment of Biology, Col- Ramsey. 1986. Experimental separation of effects orado StatE'University) for field assistance. I am grateful of consumers on sessile prey in the low zone of a to Dave Anderson (Department of Fishery and Wildlife rocky shore in the Bay of Panama: direct and indirect Biology, CSU), Ken Burnham (Depmtment of Fishery consequences of food web complexity. J. Exp. Mar. and Wildlife Biology,CSU), and Paul Stapp (Department BioI. Ecol. 100: 225-269. of Biology, CSU) for advice on using DISTANCE. Boris Noonan, G. R. 1992. Biogeographic patterns of the Kondratieff (Depmtment of Entomology, CSU), Jim Mil- montane Carabidae of North America north of Mex- lE'r(DE'partment of Biology, CSU), John Wiens (Depart- ico (Coleoptera: Carabidae), pp. 1-41. In G. R. Noon- ment of Biology, CSU), and two anonymous reviewers an, G. E. Ball, and N. E. Stork [eds.], The biogeog- provided useful critiques of the manuscript. This research raphy of ground beetles of mountains and islands. was supported by the National Science Foundation Intercept, Andover, UK. Shortgrass Steppe Long-Term Ecological Research Pro- Parmenter, R. R., J. A. MacMahon, and D. R. An- ject (BSR-9011659). The CPER is administered by the derson. 1989. Animal density estimation using a Great Plains Systems Research Unit of the United States trapping web design: field validation experiments. Department of Agriculture's Agricultural Research Ser- Ecology 70: 169-179. vice. Sokol, R. R. and F. J. Rohlf. 1981. Biometry, 2nd ed. Freeman, New York. Van Cleve, K. and S. Martin, [cds.] 1991. Long-term References Cited ecological research in the United States: a network of Anderson, D. R., K. P. Burnham, G. C. White, and research sites, 6th ed. Long-Term Ecological Re- D. L. Otis. 1983. Density estimation of smail-mam- search Network Office, University of Washington, Se- ma! populations using a trapping web and distance attle. sampling methods. Ecology 64: 647-680. Whicker, A. D. and C. R. Tracy. 1987. Tenebrionid Arnett, R. H. 1968. The beetles of the United States. beetles in the shortgrass prairie: daily and seasonal Amelican Entomological Institute, Ann Arbor, MI. patterns of activity and temperature. Ecol. Entomol. Blatchley, W. S. 1910. An illustrated descriptive cat- 12: 97-108. alogue of the Coleoptera \mown to occur in Indiana. Wilson, K. R. and D. R. Anderson. 1985. Evaluation Nature, Indianapolis, IN. of a density estimator based on a trapping web and Buckland, S. T., D. R. Anderson, K. P. Burnham, amI distance sampling theory. Ecology 66: 1185-1194. J. L. Laake. 1993. Distance sampling: estimating WHO. 1993, Methamidophos: health and safety guide. abundance of biological populations. Chapman & World Health Organization health and safety guide Hall, London. no. 79. World Health Organization, Geneva. Calkins, C. O. and V. M. Kirk. 1974. Temporal and spatial distlibution of Pasimachus elongatus (Coleop- tera: Carabidae), a predator of false wireworms. Ann. Received for publication 14 October 1994; accepted 22 Entomol. Soc. Am. 67: 913-914. Febmanj 1995.