A Thirteen-Year Survey of the Aphidophagous Insects of Alfalfa

Prairie Nat, 22(2):87-96. 1990. A Thirteen-year Survey of the Aphidophagous Insects of Alfalfa

N. C. ELLIOTTand R. W. KIECKHEFER USDA, ARS, NPA, Northern Grain Insects Research Laboratory, Rural Route #3, Brookings, SD 57006

ABSTRACT - Predatory insect populations were sampled by sweepnet from 1973 to 1985 in alfalfa fields in eastern South Dakota. Eight species of aphidophagous predators commonly occurred in alfalfa in the following order of decreasing relative abundance: Nabis americojerus Carayon, Hip- podomia convergens Guerin-Meneville, Chrysoper/ap/orabundo (Fitch), H. tredecimpunctata tibia/is (Say), H. parenthesis (Say), Co/eomegi//a macu/ata (DeGeer), Coccine//a transversoguttata Falder- mann, and Cyc/oneda munda (Say). Adults of most species were present throughout the growing season. The cumulative number of degree-days (DD) associated with the first occurrence of adults of most species in alfalfa was consistent from year to year. The DD associated with the first occurrence of immature N. americoferus and C. p/orabunda were consistent among years; however, the DD associated with the appearance of immature coccinellids were highly variable. Nab:s americoferus and C. p/orabunda developed through two generations in alfalfa, whereas coccinellids typically produced only a single generation and in some years did not reproduce in the crop. Abundances of each species varied over the 13 years; abundance of N. americoferus varied least (approximately twofold), while abundance of most species varied by one order of magnitude or more. The abundance of a particular species in a year was unrelated to its abundance in the previous year or to the abundance of other species in the same year. The composition of insect communities in crops often varies markedly in time, yet sometimes exhibits a degree of constancy when viewed at an appropriate temporal scale (Liss et al. 1986). Knowledge of levels of community change over time provides insight into the degree of predictability of community development. With respect to species assemblages of aphidophagous predatory insects in alfalfa (hereafter referred to as aphidophagous insect communities), knowledge of variation in their structure over time provides insight into the potential for biological control of aphids in alfalfa, although several factors other than natural enemy abundance are important in determining the effectiveness of biological control (Murdoch et al. 1985). Alfalfa is an important component of the agricultural production system of eastern South Dakota. In terms of total land area, alfalfa is the fourth most abundant crop grown in South Dakota, at just over two million hectares, or about 11% of the total acreage of the major crops (Jewell and Johnson 1986). Alfalfa grown as hay is typically in place for at least three years. In areas of intensive agriculture in eastern South Dakota, alfalfa is the only major crop that is relatively temporally stable, although hay harvest may precipitously disrupt the habitat and reduce abundance of some groups of arthropods (Graber and Sprague 1935, Howell and Pienkowski 1971).

Present address: USDA, ARS, SPA, Plant Science and Water Conservation Laboratory, P.O. Box 1029, Stillwater, OK 74076.

87 To our knowledge there is no published information on long-term trends in aphidophagous insect communities associated with alfalfa in eastern South Dakota. To fill this void we report on a study in which aphidophagous insect communities were sampled in fields of alfalfa at three geographically separated locations in eastern South Dakota for 13 consecutive years. We determined the species inhabiting the crop and levels of variability in their populations during the growing season and among years.

MATERIALS AND METHODS Relative estimates of aphidophagous predator populations were made in alfalfa fields throughout the growing seasons of 1973-1985. One field was sampled for predators during each year at each of three locations in eastern South Dakota: south of Brookings (Moody County), near Castiwood (Hamlin Coun- ty), and east of Clear Lake (Deuel County). Fields at the same sites (often the same fields) were sampled at these locations throughout the years of study. Estimates of predator abundance were taken by collecting six subsamples, each consisting of 50 pendular sweeps with a 38-cm diameter sweepnet (total of 300 sweeps). Each subsample was taken along a transect into the field from the field edge; the initial point along the field edge at which each subsample was begun was chosen randomly. Insects colected were chloroformed in the net, transferred to containers, and taken to the laboratory where they were counted and iden- tified to species. Sampling was done in the early afternoon on sunny days, after the ambient temperature had reached at least 15°C, wind velocity was under 24 kph, and foliage was dry. Each site was sampled at intervals ranging from one to two weeks. A median of twelve 300-sweep samples was taken from each field each year. Weather data were obtained from records of the NOAA recording station nearest each site. Distances of recording stations from sites ranged from about 6 to 14 km. Cumulative degree-days (DD) (Baskerville and Emin 1969) were calculated for each site each year beginning at 1 April using a lower developmental threshold of 8.9°C. A single lower developmental threshold was used for convenience, but 8.9°C is probably within 2.0°C of the true developmental thresholds of all species (Honek and Kocourek 1988) and provided more consistent patterns of seasonal abundance of species from year to year than calen- dar time. For each predator species, the average number of individuals per 50 sweeps (abundance) and relative abundance were calculated on daily, annual, and 13-year bases. Autocorrelation coefficients were calculated for each species using PROC ARIMA 2 (SAS Institute 1984) to determine if average annual abundance was related to abundance in previous years. Because only 13 years of data were available for each species, only a time lag of one year was considered. For each pair of species, Kendalls tau coefficients of association were calculated from average annual abundances over the 13 years using PROC CORR 2 (SAS

Mention of a commercial product does not constitute endorsement by the USDA.

88 Institute 1985) to determine whether populations of different species tended to fluctuate in unison among years.

RESULTS AND DISCUSSION Species Composition of Predator Communities Eight species of aphidophagous predators commonly occurred in alfalfa fields in eastern South Dakota. Aphidophagous predators other than these eight were occasionally observed but only in incidental numbers. The eight species were as follows: the common damsel bug, Nabis americoferus Carayon (Hemiptera: Nabidae); the common green lacewing, Chysoper/a p/orabunda3 (Fitch) (Neuroptera: Chrysopidae); the convergent lady beetle, Hippodamia convergens Guerin-Meneville; the thirteen spotted lady beetle, H. tredecimpunctata tibia/is (Say); H. parenthesis (Say); Coleomegilla mauIata (DeGeer); Coccine/la transversoguttata Faldermann; and Cyc/oneda munda (Say) (Coleoptera: Coccinellidae). All of the above mentioned species feed on aphids; Hippodamsa spp., Cycloneda munda, and Coccine/la transversoguttata are oligophagous and feed primarily on aphids, whereas N. americoferus and C. plorabunda feed on a broad range of prey (Hodek 1973, Neuenschwander et al. 1975). Coleomegi/la maculata feeds primarily on aphids and pollen (Hodek 1973). Sweepnet sampling is a problematic method of sampling insect populations because the efficiency with which insects are captured may be influenced by several biotic and abiotic variables. In alfalfa, both Fenton and Howell (1957) and Pruess et al. (1977) found that sweepnet sampling provided consistent estimates of abundances and relative abundances of predatory insects; their obser- vations alone do not insure that our sweepnet samples yielded abundance estimates with the above-mentioned properties, because some of the species we sampled differed from those encountered in their studies. We have found (Elliott, Kieckhefer, and Kauffman unpubl. data) that sweepnet samples from small grain fields yield consistent estimates of abundances of aphidophagous predators which are only slightly affected by abiotic and biotic factors. Results of the studies mentioned above lead us to suggest that our sweepnet samples from alfalfa may have yielded consistent estimates of predator abundances. However, we sampled each field at regular intervals throughout the growing season each year, so that even if abundance estimates based on individual 300-sweep samples lacked consistency due to variation in sampling efficiency, inconsistencies should have been mini- mized in our data by averaging across numerous samples each year. The cumulative percent of the total number of species collected in 300-sweep samples versus the cumulative number of subsamples collected on each sampling occasion (taken in the order in which they occurred in the field) was calculated

According to Oliver S. Flint (U.S. National Museum, Washington. DC) and Philip A. Adams (Califor- nia State University, Fullerton. CA), the common green lacewing of North America, formerly call. ed Chrysopa cornea Stephens, is currently designated Chtysoper/ap/orabunda (Fitch); C. plorabundo may ultimately prove to be a species complex. The name Chrysoperla cornea (Stephens) is applied to the common green lacewing of Europe.

89 NO I- z LU L) cc u-i

0 0 50 100 150 200 250 300 CUMULATIVE SWEEPS

Figure 1. Average cumulative proportion of total species contained in sweepnet samples from alfalfa fields versus the cumulative number of 50 sweep subsamples. for each 300-sweep sample. An average cumulative percent of species was then calculated based on all samples (Fig. 1). The curve increases only very slightly after 200 sweeps and standard errors are very small, indicating that in most in- stances all species that were collected from a field in a 300-sweep sample were already present in the sample by 200 sweeps. Thus, 300-sweep samples yielded

Table 1. Abundance (No./ 50 sweeps) and relative abundance of predators (com- bined life stages unless otherwise indicated) in sweepnet samples from alfalfa fields averaged over 13 years (1973-1985).

Relative Predator Abundance Abundance

Nabis americoferus 5.53 (3.39-7.59) 0.62 (0.40-0.82) Ch,ysoperla plorabunda (adults) 0.50 (0.10-1.55) 0.06 (0.01-0.12) C. plorabunda (immatures) 0.32 (0.05-0.62) 0.04 (0.01-0.09) Hippodamia convergens 1.36 (0.14-3.19) 0.14 (0.02-0.31) H. tredecimpuncta.ea 0.55 (0.08-1.13) 0.06 (0.01-0.13) H. parenthesis 0.48 (0.09-1.03) 0.06 (0.01-0.13) Coleomegiia maculata 0.19 (0.00-0.43) 0.02 (0.00-0.05) Cyc/oneda munda 0.01 (0.00-0.06) 0.001 (0.00-0.01) Coccinella rransversoguttota 0.04 (0.00-0.22) 0.003 (0.00-0.02)

Range of annual estimates in parentheses.

90 meaningful estimates of the species composition of aphidophagous insect communities in alfalfa fields. Nabis americoferus was the dominant predator in alfalfa in each of the 13 years of the study, accounting for a low of 40% of all predators in alfalfa in 1985 to a high of 82% in 1984 (Table 1). Averaged over the 13 years, N. americoferus accounted for 62% of all aphid predators in alfalfa. Nabis spp. are dominant predators in several crops at various geographic locations: alfalfa in New York (Wheeler 1977), soybeans in South Carolina (Shepard et al. 1974), winter wheat in Texas (Lopez and Teetes 1976); however, they are general feeders (Wheeler 1977, Guppy 1986) and do not demonstrate a consistent numerical response to variation in aphid density (Neuenschwander et al. 1975). Hippodamia convergens was the next most abundant predator, with relative abundance ranging from 2% to 31% and averaging 14%, followed in decreasing rank order of relative abundance by Chrysoperia plorabunda, H. tredecimpzinctata, H. parenthesis, Coccinelia transversoguttata, and Cyc/oneda munda.

Seasonal Abundance Patterns In many years, adults of most species were collected in alfalfa throughout the growing season (Fig. 2). However, Cycloneda munda and Coccine/la transver-

6.0 L_nN. americ- 2.0 7 H. tredeci oferus CL 4.0 1.5

LJ 2.0

LC) Lo 1:a

D ______:C I LLJ kHn.coT, UT.::.

CL 0.0 L 0 500 1000 1500 2000 0 500 1000 1500 2000 DEGREE—DAYS (base 8.9°C) Figure 2. Abundance of predators in sweepnet samples from alfalfa fields during the growing season averaged over 13 years. To facilitate pictorial representa- tion, data were grouped at 111 DD intervals (1-111 DD, 112-222 DD, etc.) and each sample in an interval was considered to have been taken at the mid- point of the interval. Mean values of samples from each interval were plotted. In all figures the solid line represents adults and the dashed line represents immatures.

91 soguttata occurred sporadically (their seasonal abundance patterns are not illustrated). Considering the frequency of sampling (1- to 2-week intervals), the DD associated with the first appearance of adults of most predator species in samples from alfalfa fields was relatively consistent from year to year (Table 2). However, the first appearance of Coleomegilla macu/ata and Cyc/onedz munda varied widely among years. The DD associated with the first occurrence of immature N. americoferus and Chyrsoperla plorabunda in samples was relatively consistent from year to year (Table 2), differing by approximately 350 DD from the earliest to the leatest appearance of immatures in the crop; a portion of this variation was undoubtedly related to the widely spaced sampling interval employed. First occurrence of immature coccinellids varied widely among years.

Table 2. The average number degree-days associated with the first occurrence of predators each in sweepnet samples, and number of years predators were recovered in sweepnet samples taken from alfalfa fields in eastern South Dakota over 13 years (1973-1985).

Number of years Degree-days of first occurrence recovered Predator Adult Immature Adult Immature

Nabis americoferus 152 (86-278) 392 (171-531) 13 13 Chrjsoper/a plorabunda 217 (107-367) 384 (246-619) 13 13 H:ppodamza convergens 186 (86-299) 592 (316-1116) 13 11 H. tredecimpunctata 154 (86-278) 605 (178-1525) 13 12 H. parenthesis 208 (107-424) 867 (206-1775) 13 9 Coleomegi/la macu/ala 338 (87-1290) 671 (343.1543) 12 5 Cyc/oneda munda 683 (292 . 1333) - 10 0 Coccine/la transversoguttata 298 (131-682) - 10 0

Range of annual estimates in parentheses.

In each year of the study, there were two distinct peaks in abundance of immature N. amerzcoferus at most sites, one at an average of 668 DD (SE = 21.1) typically early July) and another at 1221 DD (SE = 25.0) (typically mid- August). These peaks are somewhat obscured in Fig. 2 due to year-to-year variation in their time of occurrence. The presence of two peaks suggests that N. americoferus developed through two generations in alfalfa in eastern South Dakota. This species develops through two generations in alfalfa in eastern On- tario (Guppy 1986). Two distinct peaks in abundance of immature Chrysoperla plorabunda were also observed, the first at an average of 602 DD (SE = 23.8) and the second at 1199 DD (SE = 37.2), indicating that C. plorabunda probably also produced two generations in alfalfa. A single peak in abundance of immature coccinellids typically occurred. For all coccinellids the peak most fre- quently occurred at approximately 610 DD, but was quite variable, and in serveral years immature coccinellids were not recovered in samples from alfalfa (Table 2). Two widely spaced peaks in abundance of immatures of the three Hippodamia

92 species occurred in 1974, suggesting that these species produced two generations in alfalfa that year.

Annual Abundance Patterns Abundance of predators varied annually over the course of the study; abundance varied approximately 2- to 100-fold among years depending on species (Table 1). Abundance of N. americoferus varied least (in a proportional sense), from 3.39 to 7.59 individuals per 50 sweeps, while Co/eomegilla macu/ata, Cycloneda munda, and Coccine/la transversoguttata were virtually absent from alfalfa fields in one or more years. Chrysoper/ap/orabunda, H. convergens, H. tredecimpunctata, and H. parenthesis were consistently present in alfalfa, but abundance of these species varied by about one order of magnitude or more among years. Reproduction by N. amencoferus and Chrysoperla plorabunda occurred in each of the 13 years of the study (Table 2). Immatures of the three Hippodamia species were collected in samples in 9 to 12 years depending on species, indicating that reproduction by these species did not occur, or occurred at very low levels in alfalfa fields in some years. Immature Cyc/oneda munda and Coccinella transversoguttata were never obtained in samples. Although species abundances varied markedly from year to year and standard errors were relatively small (Fig. 3), there were no trends in annual abundance for any species. Sample autocorrelation coefficients ranged from -0. 13 to

10.0 75 N. americoferus H. parenes 11 20 K lv NMI 2orabund4 1.0 C. maculata Q Ui

LU 0.10 0 C. munda 1-n 4.0 Aver9e7\J cc 0 I I- 0.25 Cl 1.5, , H. tredecimpunctata 0.20 0.15 C. transverstiguttata LU 0.10 Ii 0.05 0 0.00 73 76 79 82 85 73 76 79 82 85 rd MIZ YEAR Figure 3. Annual abundance (± SE) of aphidophagous insects in alfalafa fields in each of 13 years.

93 0.32 and were nonsignificant (P>0.05) for all species. The absence of trends in abundance may have been caused by the influences of variable processes (such as weather) on survival, reproduction, and phenology of predators. Alternative- ly, the absence of trends may have resulted from unpredictable variation in patterns of prey abundance among habitats over time, causing fluctuations in predator populations in alfalfa as predators responded, by movement among habitats, to the variable prey populations (Honek 1982, Karieva 1986, Kring and Gilstrap 1986). There was little evidence to suggest that populations of different species fluctuated in unison from year to year. Kendalls tau coefficients of association ranged form -0.43 to 0.59 and averaged -0.01; only two out of a total of 28 coefficients (one coefficient for each pairwise combination of species) differed significantly from zero. With a significance level of 0.05 about 1.4 out of 28 significant coefficients would be expected on average simply by chance. Thus, populations of different species fluctuated nearly independently over time.

GENERAL DISCUSSION Aphidophagous predators inhabit a variety of plant communities, though many species exhibit distinct habitat preferences (Honek 1982). Hippodamia convergens, H. tedeczmpunctata, and Coleomegilla maculata, for example, are abundant in maize and small grains in eastern South Dakota and Minnesota (Ewert and Chiang 1966, Elliott and Kieckhefer 1990). Results of this study in- dicate that populations of these species also persist in alfalfa throughout the summer. Small grains and maize are the dominant crops grown in eastern South Dakota in terms of total acreage (Jewell and Johnson 1986). Spring and winter small grains typically mature and become uninhabitable by phytophagous insects and hence also by their predators in early July, before aphid and other prey populations typically develop in maize fields in the region. In addition to a variety of noncrop habitats, alfalfa provides an acceptable habitat for several predatory species when small grains and maize are uninhabitable. Year-to-year change in species populations, and consequently change in the structure of aphidophagous predator communities in alfalfa, is unpredictable. The unpredictable fluctuations of populations over time are probably the result of influences of variable processes, such as weather, on both predator and prey populations in alfalfa and in other habitats utilized by predators for feeding, reproduction, overwintering, and other activities. The lack of association among populations of different species suggests that different factors are involved in the population dynamics of different predator species or that factors differ in either their qualitative or quantitative effects on predator populations. The time of colonization of the crop each year and the time at which reproduction first occurred also fluctuated widely among years for several species. Thus, superim- posed on variation in community structure among years is temporal variability in community development within years. Nabis americoferus was the dominant predator in alfalfa fields and exhibited the most stable populations. This species is a general feeder, and the relative stability of its populations may reflect its ability to utilize a variety of prey species

94 depending on their availability (Neuenschwander et al. 1975). Even though N. americoferus feeds on a wide range of prey, it may be an important aphid predator in alfalfa because it is usually abundant in alfalfa fields early in the growing season when migrant aphids generally colonize the crop.

ACKNOWLEDGMENTS Several people assisted in collecting and processing samples; we thank each of them. We also thank Sandy Hubbard for processing and editing the manuscript, and Bob Burton, Frank Gilstrap, andJohn Obrycki for their helpful comments on the initial version of the manuscript.

LITERATURE CITED Baskerville, G. L., and P. Emin. 1969. Rapid estimation of heat accumulation from maximum and minimum temperatures. Ecology 50:514-522. Elliott, N. C., and R. W. Kieckhefer. 1990. Dynamics of aphidophagous coc- cinellid assemblages in small grain fields in eastern South Dakota. Eviron. Entomol. In press. Ewert, M. A., and H. C. Chiang. 1966. Dispersal of three species of coccinellids in cornfields. Can. Entomol. 98:999-1003. Fenton, F. A., and D. E. Howell. 1957. A comparison of five methods of sampling alfalfa fields for arthropod populations. Ann. Entomol. Soc. Am. 50:606-611. Graber, L. F., and V. G. Sprague. 1935. Cutting treatments of alfalfa in relation to infestations of leafhoppers. Ecology 16:48-59. Guppy, J. C. 1986. Bionomics of the damsel bug, Nabisamericoferus Carayon (Hemiptera: Nabidae), a predator of the alfalfa blotch leafminer (Diptera: Agromyzidae). Can. Entomol. 118:745-751. Hodek, I. 1973. Biology of Coccinellidae. Academia Press, Prague. Honek, A. 1982. Factors which determine the composition of field communities of adult aphidophagous Coccinellidae. Z. Angew. Entomol. 94:157-168. Honek, A., and F. Kocourek. 1988. Thermal requirements for development of aphidophagous coccinellidae (Coleoptera), Chrysopidae, Hemerobiidae (Neuroptera), and Syrphidae (Diptera), some general trends. Oecologia 76:45 5-460. Howell, J. D., and R. L. Pienkowski. 1971. Spider populations in alfalfa with notes on spider prey and effects of harvest. J . Econ. Entomol. 14:456-461. Jewell, L. D., and E. Johnson. 1986. Agricultural statistics. U.S. Government Printing Office, Washington, DC. Kareiva, P. 1986. Patchiness, dispersal, and species interactions: Consequences for communities of herbivorous insects. Pp. 192-206 in Community ecology (J. Diamond and T. J . Case, eds.). Harper and Row, New York. Kring, T. J . , and F. E. Gilstrap. 1986. Beneficial role of corn leaf aphid, Rhopalosip hum maidis (Fitch) (Homoptera: Aphididae), in maintaining Hip- podamia spp. (Coleoptera: Coccinellidae) in grain sorghum. Crop Protec- tion 5:125-128.

95 Liss, W. J., L. J. Gut, P. H. Westigard, and C. E. Warren. 1986. Perspectives on arthropod community structure, organization, and development in agricultural crops. Annu. Rev. Entornol. 31:455-478. Lopez, E. G., and G. L. Teetes. 1976. Selected predators of aphids in grain sorghum and their relation to cotton. J . Econ. Entomol. 69:198-204. Murdoch, W. W., J . Chesson, and P. L. Chesson. 1985. Biological control in theory and practice. Am. Nat. 125:344-366. Neuenschwander, P., K. S. Hagen, and R. F. Smith. 1975. Predation on aphids in California alfalfa fields. Hilgardia 43:53-78. Pruess, K. P., M. L. Saxena, and S. Koinzan. 1977. Quantitative estimation of alfalfa insect populations by removal sweeping. environ. Entomol. 6:705-708. SAS Institute. 1984. SAS/ETS users guide. SAS Institute, Cary, NC. SAS Institute. 1985. SAS users guide: Basics. SAS Institute, Cary, NC. Shepard, M., G. R. Garner, and G. S. Turpinseed. 1974. Seasonal abundance of predaceous arthropods in soybeans. environ. Entomol. 3:985-988. Wheeler, A. G., Jr. 1977. Studies on the arthropod fauna of alfalfa VII. Predaceous insects. Can. Entomol. 109:423-427. Received 5 May 1989. Accepted 15 February 1990.