OUP user menu

Prairie Grasses as Hosts of the Western Corn Rootworm (Coleoptera: Chrysomelidae)

Isaac O. Oyediran, Bruce E. Hibbard, Thomas L. Clark
DOI: http://dx.doi.org/10.1603/0046-225X-33.3.740 740-747 First published online: 1 June 2004


We evaluated 21 prairie grass species thought to be among those dominant 200 yr ago in the western Great Plains as larval hosts of the western corn rootworm, Diabrotica virgifera virgifera LeConte. Maize, Zea mays L., and sorghum, Sorghum bicolor L., were included as positive and negative controls. Twenty pots of each test species were planted, and each pot was infested 5 wk later with 20 neonate western corn rootworm larvae. Four pots within each of four replications were randomly assigned a sample date for larval extraction. The remaining pot from each replication was used to monitor adult emergence. At 5, 10, 15, and 20 d after infestation, pot contents from assigned pots were placed in Tullgren funnels equipped with 60 W-lights for extraction of larvae. The percentage of larvae recovered, larval head capsule width, and adult emergence varied significantly among the grass species. The percentage of larvae recovered from western wheatgrass, Pascopyrum smithii (Rydb.); pubescent wheatgrass, Elytrigia intermedia (Host); and side-oats grama, Bouteloua curtipendula Michx., was not significantly different than that from maize when sample dates were combined. The number of adults produced from pubescent wheatgrass was not significantly different than the number produced from maize. The average dry weight and head capsule width of adults produced from grass species were not significantly different than the head capsule widths and dry weights of those adults from maize. Overall, adults were produced from 14 of the 23 species evaluated. The results from this study are discussed in relation to the potential ancestral hosts of western corn rootworm larvae and in relation to resistance management of transgenic maize.

  • prairie grasses
  • alternate hosts
  • resistance management
  • Diabrotica virgifera virgifera


The western corn rootworm, Diabrotica virgifera virgifera LeConte, is one of the most severe pests of maize, Zea mays L., in the primary maize-growing regions of the United States. This beetle belongs to a diverse genus, divided into three distinct species groups (virgifera, fucata, and signifera) based on morphology (Wilcox 1972, Krysan and Smith 1987) and mitochondrial genes (Clark et al. 2001). Damage is caused by larvae feeding on the roots of the maize plants; with annual losses in terms of control costs and yield reduction estimated at $1 billion annually (Metcalf 1986).

Control tactics, such as crop rotation, application of soil and aerial insecticides, and use of resistant maize cultivars, have been used over the years. However, some of these control tactics are no longer effective, at least in certain regions. In parts of the eastern Corn Belt, the western corn rootworm adapted to crop rotation by laying eggs in fields adjacent to maize (Levine et al. 2002). In areas where continuous maize is grown, insecticide is the most common management tactic (Mayo 1986). The western corn rootworm has developed resistance to organochlorine, organophosphate, and carbamate insecticides in parts of the United States (Ball and Weekman 1962, Meinke et al. 1998). Because of the problems of development of resistance and the evolution of a new strain, additional control tactics, such as the use of transgenic maize that expresses specific endotoxin(s) from the bacterium Bacillus thuringiensis Berliner (Bt) have been developed by several seed companies (Moellenbeck et al. 2001, Ellis et al. 2002) to control damage from larvae of the western and northern corn rootworm, Diabrotica barberi Smith & Lawrence. Because of the behavioral and genetic plasticity of western corn rootworm, adaptation to transgenic control tactics is a concern. The U.S. Environmental Protection Agency has mandated that all registrants for Bt crops have a resistance management plan in place. Efforts to model adaptation of corn rootworms to transgenic maize with resistance to larval feeding (Onstad et al. 2001, Storer 2003) have renewed interest in the basic biology of these insects. Studies on nontarget effects (Lundgren and Wiedenmann 2002, Duan et al. 2002, Al-Deeb and Wilde 2003), adult movement and sampling (Nowatzki et al. 2002, Whitworth et al. 2002), larval movement (Hibbard et al. 2003, 2004), and alternate hosts (Clark and Hibbard 2004) have all been stimulated to some extent to meet the demand for additional information on the basic biology of these important pests. Interestingly, though, some of the most basic information of all still eludes Diabrotica workers, such as the original host(s) of the western corn rootworm.

Given the distribution of maize in the 1860s (Weatherwax 1954), it is not unreasonable to assume that D. v. virgifera survived on hosts other than maize in the region of western Kansas where LeConte first collected it (LeConte 1868). According to Goodman (1987), “At the time of European colonization of the New World, maize was being grown from southern Canada to central Chile, although little was grown in the grassy plains or savannas of the central United States.” In fact, the distribution map of Weatherwax (1954), shows that maize was not grown in western Kansas at that time. However, Branson and Krysan (1981) suggested that “D. virgifera became a specialist on corn in the tropics or subtropics and ʻfollowedʼ the diffusion of corn into the temperate United States.” Krysan and Smith (1987) stated that “it is reasonable to conclude that the presence of D. v. virgifera in the United States does not predate the presence of corn.” Despite the fact that neither Branson and Krysan (1981) nor Krysan and Smith (1987) provided any evidence that maize was grown in western Kansas in the 1860s, their view is the generally accepted dogma today. The primary justification that Branson and Krysan (1981) and Krysan and Smith (1987) used in suggesting maize as an original host for the western corn rootworm was the fact that maize is by far the best host found to date (Branson and Ortman 1967a,b, 1970, Clark and Hibbard 2004). However, their reasoning is partially flawed because “best host” status does not necessarily correlate to original host. In fact, another virgifera group species, Diabrotica longicornis (Say), does not feed on maize when feral, but actually develop as larvae at a rate equal or even faster on maize than larvae of its sibling species, which specialize on maize, the northern corn rootworm (Golden and Meinke 1991). In another example, Horton et al. (1988) found that even though local populations of the Colorado potato beetle, Leptinotarsa decemlineata (Say), had been adapted to specialize on a local host and were not agricultural pests, these same local populations did very well when forced onto cultivated potato, Solanum tuberosum L. Although we cannot know for certain the ancestral host of the western corn rootworm, the goal of the current experiment was to evaluate potential hosts that were likely present in western Kansas in the 1860s for their ability to support western corn rootworm development.

Materials and Methods

Plants, Insects, and Experimental Design

The experiment was conducted in a greenhouse in 2002. Prairie grasses of 21 species, along with maize and sorghum as checks, were evaluated for their suitability as larval hosts of the western corn rootworm. The seed source for each species was documented (Table 1). Seeds were planted in pots containing 2:1 mixture of autoclaved soil/peat-based growing medium (Promix, Premier Horticulture LTEЀ, Quebéc, Canada). Drainage openings in the pots were fitted with a fine (114 μ m per opening) stainless steel mesh (TWP, Inc., Berkley, CA) to prevent larval escape (Clark and Hibbard 2004). The experimental design was a split-plot arrangement of treatments, and evaluated survival and growth parameters among and within species over time. Five pots of each plant species were randomly placed together on each of four separate tables (five pots per species per table). Each table was regarded as a replication. The five pots of each species within each replication were randomly assigned a sample date for larval extraction or adult emergence. Five weeks after planting, each pot was infested with 20 neonate western corn rootworm larvae by gently transferring larvae with a moistened camelʼs-hair brush into a hole (1 by 3 cm in depth) that was dug in each pot. After infestation, each hole was gently covered with excess soil. A photoperiod of 14:10 (L:D) h was maintained with 1000-W sodium bulb (GE Lighting, Cleveland, OH). Temperature was maintained at ≈25°C A temperature recorder was available during a portion of the experiment and temperatures remained at a daily average of 25 ±2°C

View this table:
Table 1.

Larval Growth and Development

At 5, 10, 15, and 20 d after infestation, the contents (soil mixture, roots, and larvae) of the randomly assigned pots were individually placed in Tullgren funnels equipped with a 60 W light bulb (Great value, Soft white, Wal-Mart Company) for the extraction of larvae. Collection jars containing water were placed under the funnels and checked daily for larvae for 4 d. Larvae recovered were counted and transferred to individual-labeled scintillation vials containing 95% ethanol. The head capsule width of each larva was measured using an ocular micrometer (10x/21, Wild Co., Heerbrugg, Switzerland) mounted on a microscope (M3Z, Wild Co., Heerbrugg, Switzerland). The average head capsule width of the population of neonate larvae used to infest pots was 0.22 mm. The change in larval head capsule width was as the measured head capsule width of the recovered larvae in question minus 0.22 mm. When no larvae were recovered from a particular replication, the change in head capsule width and the average weight for the replicate were considered missing values. Dry weights of the larvae were determined using an analytical scale (ER-182A, A & D Co., Tokyo, Japan) after placing the larvae in a desiccating oven (Thelco model 16, GCA/Precision Scientific Co., Chicago, IL) at 80°C for 48 h.

Beetle Emergence

On day 28 after infestation, the vegetation from the fifth pot was cut close to the soil level and the pots were covered with a mesh to prevent beetle escape. Each pot was checked a minimum of twice per week for adult emergence. All the adults collected were stored in 95% ethanol. The adults were identified to sex and the head capsule width and dry weight were measured as described earlier. When no adult was collected from a particular replicate, head capsule width and average weight were considered missing values.

Statistical Analysis

The larval data were analyzed as arandomized complete block split-plot design by using the PROC GLM procedure of the statistical package SAS (SAS Institute 1990). The model contained the main plot of species, the subplot of sampling time, and their interaction. Least significant differences (LSDs, α =0.05) were calculated for intraspecific growth and recovery through time and interspecific growth and recovery at a given sample time as described by Steel and Torrie (1980) by using mean squares and degrees of freedom parameters from the PROC GLM procedure. A separate analysis was done for percentage of larval recovery, average dry weight, and head capsule width. Adult data were a randomized complete block design. A separate analysis was for adult number, head capsule width, average dry weight, and percentage of females. Although untransformed data are shown in the tables, to meet the assumptions of the analysis, all data were transformed by square root (√x +0.5) except for percentage data, which were transformed with arcsine (√x) (Snedecor and Cochran 1989). The data were transformed because the residuals of the untransformed data were distributed in a Poisson and not in a normal distribution fashion. The (√x) transformation was chosen because it is the most effective if values are small, and it also stabilizes the variance more effectively (Snedecor and Cochran 1989).


Larval Recovery

The percentage of western corn rootworm larvae recovered from root masses placed in Tullgren funnels was significantly different among the species evaluated (F =8.88; df =22, 207; P < 0.0001) as well as between sample days (F =6.95; df =3, 207; P =0.0002). The interaction between species evaluated and sample date was not significant (F =1.04, df =66, 207; P =0.4007). The highest percentage of larvae recovered was from maize, averaging 24.6 ±4.7%. The highest percentage of larvae recovered was from maize, averaging 24.6 ±4.7% overall. When sample dates were combined, there was no significant difference between the percentage of larvae recovered from maize and the percentage recovered from pubescent wheatgrass, side-oats grama, slender wheatgrass, or western wheatgrass (Table 2). No larvae were recovered from sorghum on any sample date. Trends within sample dates were generally similar to the combined analysis, except for side-oats grama on the last sample date that had significantly fewer larvae

View this table:
Table 2.

Head Capsule Width of Larvae

Average head capsule width (HCW) of larvae recovered from the prairie grasses varied significantly among species (F =6.08; df =20, 46; P < 0.0001) and sampling dates (F =120.11; df =3, 46; P < 0.0001). The interaction between grass species and sample date (F =3.21; df =44, 46; P < 0.0001) also affected HCW. The HCW of field-collected neonate western corn rootworms ranges from 0.2 to 0.26 mm, whereas second and third instars ranged from 0.3 to 0.4 mm and 0.44–0.56 mm, respectively (Hammack et al. 2003). Centers of fitted normal curves describing frequency distributions of HCW for first, second, and third instars were 0.216, 0.329, and 0.499 mm (Hammack et al. 2003). On sample day 10, most larvae recovered from maize, Canada wild rye, needle-and-thread grass, slender wheatgrass, witch grass, lndian grass, and side-oats grama, had reached the second instar, although the HCW of larvae recovered from maize on this date was significantly larger than that of larvae from any other species except needle-and-thread grass (Table 3). Most of the larvae recovered from maize, pubescent wheatgrass, Canada wild rye, porcupine grass, green needlegrass, slender wheatgrass, squirreltail grass, witch grass, side-oats grama, and prairie cordgrass were third instars on the third sample date. The change in head capsule width of larvae recovered from pubescent wheatgrass and slender wheatgrass was not significantly different from that of larvae recovered from maize on this date (Table 3). By the fourth sample date, most larvae recovered from several additional species, including western wheatgrass, needle-and-thread grass, buffalo grass, purple threeawn, and tall dropseed, also reached the third instar.

View this table:
Table 3.

Average Dry Weight of Larvae

Average dry weight varied significantly among sample dates (F =15.58; df =3, 45; P< 0.0001), but not species (F= 0.90; df =20,45; P =0.5863) (Table 4). The interaction between treatments and sampling dates was not significant. The average dry weight generally increased over time, but was highest for maize, pubescent wheatgrass, porcupine grass, needle-and-thread grass, and witch grass (Table 4). Witch grass, however produced the heaviest larvae 15 d rather than 20 d after infestation.

View this table:
Table 4.

Adult Recovery

The total number of adults collected varied significantly among the prairie grasses (F =1.78; df =22, 66; P =0.0300). Adult production was not significantly different between maize and pubescent wheatgrass and was also comparatively high in Canada wild rye, western wheatgrass, slender wheatgrass, prairie cordgrass, and squirreltail grass (Table 5). Only nine of the 23 species evaluated did not produce an adult when artificially infested under our greenhouse conditions (Table 5). The adults produced from maize were 58.3% female. All other species produced a numerically lower percentages of females (<25%), but the differences among species were not statistically significant (F =1.59; df =22, 66; P =0.0755).

View this table:
Table 5.

Adult Head Capsule Width and Dry Weight

Plant species did not significantly impact adult dry weight (F =2.31; df =13, 10; P =0.0947), nor head capsule width (F =2.42, df =13,10; P =0.0843 (Table 5). The fewer degrees of freedom associated with the analysis of average dry weight and HCW is because of the large number of missing values where no adults were collected. When adults were collected, average head capsule width ranged between 1.05 ±0.01 mm for witchgrass and 1.23 ±0.02 mm for maize and dry weight ranged from 0.43 mg for side-oats grama to 1.81 ±0.30 mg for maize.


Despite some anecdotal evidence, the widespread distribution of maize production in the United States, makes it difficult to assess whether larvae of the western corn rootworm currently use hosts other than maize. One of us (T.L.C.) has collected western corn rootworm beetles on Cucurbita foetidissima Humboldt, Bonpland & Kunth in areas of New Mexico, Kansas, and Colorado, which were >60 km from any maize production. As noted, LeConte (1868)first collected the western corn rootworm in western Kansas on “wild gourd,” probably also C. foetidissima, when that whole region of the western Great Plains was devoid of maize production (Weatherwax 1954). Despite these indirect observations, production of western corn rootworm beetles from areas beyond maize fields has not been documented other than when artificial infestations of alternate hosts have been made. The larval hosts of beetles that LeConte (1868) collected are unknowable. Results from the current study suggest that prairie grasses other than maize are capable of supporting western corn rootworm populations. Indicators of developmental vigor, such as head capsule width and average dry weight for adults produced from maize, western wheatgrass, pubescent wheatgrass, slender wheatgrass, and galleta, were not significantly different, further suggesting that western corn rootworm populations are capable of inhabiting regions devoid of maize.

Although natural western corn rootworm populations using hosts other than maize are not documented, its subspecies, the Mexican corn rootworm, Diabrotica virgifera. zeae Krysan & Smith, can develop on other hosts. Branson et al. (1982) collected Mexican corn rootworm beetles in emergence traps placed over a mixture of four grass species: Brachiaria plan-taginea (Link), Eleusine indica (L.), Eragrostis indica (Hornem.), and Digitaria ciliaris (Retz.). They also collected larvae and pupae from the grasses B. plan-taginea and Panicum hallii Vasey and the sedge Cype-rus macrocephalus Liebm. in Mexico. At least this subspecies of D. virgifera can be maintained on species beyond maize. Other examples exist suggesting that this was not an isolated incident. Krysan and Smith (1987)cite an example in Sutton County, TX, where the Mexican corn rootworm was found at a density of two beetles per plant in a maize field in its third year of production (the second generation of potential beetle production from maize) despite >250 km of isolation from any other maize production area. The quick establishment of an economically damaging population (Witkowski et al. 1986), despite isolation strongly, suggested that alternate larval hosts may have maintained a Mexican corn rootworm population in that area as well.

Results from the current study document that adult production was not significantly different between maize and pubescent wheatgrass. Head capsule width and average dry weight of adults produced from maize were not significantly different than any of the other grass species for which adults were produced (Table 5). Canada wild rye, western wheatgrass, slender wheatgrass, squirreltail grass, and prairie cordgrass also produced a comparatively high number of adults.

Only nine of the 23 species evaluated did not produce an adult when artificially infested under the greenhouse conditions of this experiment (Table 5). Clark and Hibbard (2004) evaluated larval survivorship and growth parameters on the roots of 29 plant species comprised of maize, maize-field weeds, forage grasses, and small grain crops. Larvae survived at least 6 d after infestation on 27 species and 24 d on 23 plant species. Growth and development of larvae on plant species was significantly slower on most species other than maize; however, 18 of the species had larvae develop to the second instar, whereas larvae on 14 species developed to the third instar. Adults were recovered from five plant species in addition to maize. Although a smaller percentage of species produced adults in the study of Clark and Hibbard (2004)than in the current study, it is possible that some of this difference was due to procedural differences. Regardless, it is clear that western corn rootworm larvae are capable of using most grass species as hosts for at least some larval development and the roots of many grass species are capable of producing western corn rootworm adults. These results contrast with that of Branson and Ortman (1967b, 1970) who initially evaluated naked grass roots in petri dishes without soil, which may account for the low percentage of grasses which were considered hosts in their studies.

Most of the better hosts from the current study are perennial sod-producing species and remain among the more dominant grasses throughout western Kansas (Weaver and Albertson 1956) where western corn rootworm was first found. It is possible that older perennial plants with larger root systems may be even better hosts than the seedlings evaluated in the current experiment. Our data, combined with distribution of maize in 1868, imply that the adults collected by LeConte in 1868likely developed as larvae on native grasses, such as pubescent wheatgrass.

The implications of our data for resistance management of rootworm-resistant transgenic maize remain unclear. Nick Storer (Dow Agrosciences, Indianapolis, IN) has modeled adaptation of corn rootworms to rootworm-resistant Bt maize (Storer 2003). Unpublished output of the model predicts that if as few as 0.5% of the adults come from spatially well-distributed nonmaize hosts, the onset of resistance would be significantly delayed in a system with a poorly distributed 5% fixed location refuge, although this delay is not significant under more conservative refuge deployment scenarios, such as the 20% refuge being required for the product which is currently registered (N. Storer, personal communication). We have documented significant adult production with alternate hosts, but not what percentage, if any, of the adults commonly found in maize agroecosystems developed as larvae on hosts other than maize.


We thank Matt Higdon (Department of Entomology, University of Missouri) and Arnulfo Antonio (USDA–ARS, Plant Genetics Unit, Columbia, MO), for technical assistance in this research. Mark Ellersieck (University of Missouri, Agricultural Experiment Station), assisted in the analysis. We thank Ted Wilson, Matt Higdon, Yvonne Schweikert, (University of Missouri) and Larry Darrah (USDA–ARS, Plant Genetics Unit, Columbia, MO) for suggestions on an earlier version of this manuscript. Funding, in part, was provided by USDA–CSREES–NRI–CGP Project Award Nos. 2001-35316-10000 and 2002-35316-12282.


  • This article reports the results of research only. Mention of a proprietary product does not constitute an endorsement or recommendation for its use by the USDA or the University of Missouri.

This is an Open Access article distributed under the terms of the Creative Commons Attribution Non-Commercial License (http://creativecommons.org/licenses/by-nc/3.0/), which permits non-commercial reuse, distribution, and reproduction in any medium, provided the original work is properly cited. For commercial re-use, please contact journals.permissions{at}oup.com

References Cited

View Abstract