Development of a Sampling Plan in Winter Wheat that Estimates Cereal Aphid Parasitism Levels and Predicts Population Suppression

Transcription

1 SAMPLING AND BIOSTATISTICS Development of a Sampling Plan in Winter Wheat that Estimates Cereal Aphid Parasitism Levels and Predicts Population Suppression KRISTOPHER L. GILES, DOUGLAS B. JONES, TOM A. ROYER, NORMAN C. ELLIOTT, 1 AND S. DEAN KINDLER 1 Department of Entomology and Plant Pathology, Oklahoma State University, Stillwater, OK 74078Ð3033 J. Econ. Entomol. 96(3): 975Ð982 (2003) ABSTRACT From 1998 to 2001, the relationship between the proportion of tillers with 0 mummiþed aphids (P tm ) and the proportion of cereal aphids parasitized (P p ) was estimated on 57 occasions in Þelds of hard red winter wheat located in central and western Oklahoma. Both original (57 Þelds) and validation data (34 Þelds; 2001Ð2002) revealed weak relationships between P tm and P p, however, when P tm 0.1, P p always exceeded the recommended parasitism natural enemy threshold of 0.2. Based on the relationship between P tm and P p, upper (P tm1 ) and lower (P tm0 ) decision threshold proportions were set at 0.1 and 0.02, respectively. We monitored cereal aphid populations in 16Ð25 winter wheat Þelds over time, and based on the upper and lower decision threshold proportions (P tm1 0.1, P tm0 0.02), predicted whether aphid intensities (# per tiller) would increase above or be maintained below selected economic thresholds (3, 9, and 15 aphids per tiller). Results of this validation study revealed that aphid intensity exceeded an economic threshold in only one Þeld when predicted to remain below P tm 0.1, but aphid intensity reached a maximum of only four aphids per tiller. The sampling plan developed during this study allowed us to quickly classify P tm, and independent of initial cereal aphid intensities, very accurately predict suppression of populations by parasitoids. Sequential sampling stop lines based on sequential probability ratio tests for classifying proportions were calculated for P tm1 0.1 and P tm A minimum of 26 tiller samples are required to classify P tm as above 0.1 or below Based on the results of this study, we believe that simultaneous use of aphid and parasitoid sampling plans will be efþcient and useful tools for consultants and producers in the southern plains and decrease the number of unnecessary insecticide applications. KEY WORDS cereal aphids, Lysiphlebus testaceipes, sequential sampling, natural enemy threshold GREENBUG (Schizaphis graminum Rondani) and birdcherry oat aphid (BCOA, Rhopalosiphum padi L.) are the most abundant and important insect pests of winter wheat in the southern plains. In nearly all of the 6Ð7 million acres of winter wheat grown in Oklahoma, variable levels of greenbug and BCOA cause damage from early fall until late spring before the crop matures. These aphids are often the most signiþcant factor limiting proþtable winter wheat production (Burton et al. 1985, Kieckhefer and Kantack 1988, Webster 1995, Riedell et al. 1999, Kindler et al. 2002). The hymenopteran parasitoid Lysiphlebus testaceipes (Cresson) is the most abundant and, arguably, the most important primary parasitoid of cereal aphids throughout the southern plains (Jackson et al. 1970, Gilstrap et al. 1984, Patrick and Boring 1990, Royer et al. 1998). Studies in Oklahoma have demonstrated that L. testaceipes attacks aphids throughout the winter wheat growing season (OctoberÐMay) and often effectively controls aphid populations (Jones 2001, 1 USDAÐARS-PSWCRL, Stillwater, OK N.C.E., unpublished data). Information from these studies clearly suggests that integrating natural enemy thresholds (estimates of parasitism and subsequent predictable mortality) into aphid management decisions will reduce the use of insecticides and increase proþt potential in wheat production. Natural enemy thresholds are based on predictable relationships between varying natural enemy/pest proportions (Nyrop and van der Werf 1994, Wilson 1994). The most clearly deþned natural enemy thresholds have been developed in several orchard systems for phytophagous mites and their predators and allow producers to evaluate the potential for pest populations to be biologically suppressed. These thresholds have evolved from sampling procedures requiring density ratio estimates from counts of phytophagous mites and their predators per leaf, to the development of efþcient binomial (presence-absence) classiþcation procedures for estimating numbers and subsequent predator/prey ratios (Tanigoshi et al. 1983, Wilson et al. 1984, Flaherty et al. 1992, Nyrop and van der Werf 1994, Flaherty and Wilson 1999) /03/0975Ð0982$04.00/ Entomological Society of America

2 976 JOURNAL OF ECONOMIC ENTOMOLOGY Vol. 96, no. 3 Patrick and Boring (1990) and Royer et al. (1998) suggest that cereal aphid populations in winter wheat Þelds will be quickly suppressed or prevented from increasing to economically damaging levels if the proportion of aphids parasitized (P p )byl. testaceipes reaches 0.2. Jones (2001) conþrmed the effectiveness of L. testaceipes in a Þeld cage study; when the P p was 0.05, aphids exceeded economic thresholds regardless of initial aphid intensity. Currently, accurate estimates of the proportion of aphids parasitized (P p )in winter wheat require outdated and time consuming sampling procedures (aphid and mummy counts) and subsequent rearing of collected specimens for 3Ð7 d (Royer et al. 1998; Giles et al. 2000a, 2000b; Royer et al. 2002a, 2000b). The most efþcient use of the recommended parasitism threshold (P p conservatively set at 0.20) would be to effectively incorporate it into newly developed binomial sequential sampling plans that classify aphid intensities from the proportion of tillers with 0 aphids (Giles et al. 2000a, 2000b; Royer et al. 2002a, 2000b). Simultaneous classiþcation of aphid and parasitism levels in Þelds requires accurate estimates of P p from measures of mummiþed aphids. In an investigation of Sitobion avenae F. and the parasitoid Aphidius ervi Haliday on spring wheat, Feng et al. (1993) observed a strong relationship (r ) between the proportion of mummiþed aphids and the proportion of tillers with 0 mummiþed aphids ( P tm ). This type of predictable relationship allows for an accurate estimation of the proportion of mummiþed aphids in wheat Þelds, but does not account for parasitized but nonmummiþed aphids, and therefore will often underestimate actual levels of parasitism (P p ). A more usable parasitism sampling scheme could be developed if a predictable relationship exists between P tm and P p. If this relationship is predictable, the counting and rearing processes for estimating parasitism (Patrick and Boring 1990, Royer et al. 1998) could then be eliminated allowing for simultaneous classiþcation of aphid intensity and parasitism. The overall goal of this research was to develop a sampling plan that predicts the potential for control of cereal aphids by parasitoids in winter wheat in the southern plains. The Þrst objective of this study was to examine and validate the relationship between the proportion of wheat tillers with 0 mummiþed aphids (P tm ) and the proportion of aphids parasitized (P p )in winter wheat Þelds. In particular, we were interested in identifying a threshold value for P tm that reliably predicted a P p value 0.2 (the recommended parasitism threshold for effective biological control). The second objective was to determine if a natural enemy threshold, estimated from a P tm threshold, could accurately predict suppression of aphid populations below economic thresholds. The third objective was to develop a binomial sequential sampling plan for classifying P tm (based on the P tm to P p relationship) as above or below a critical proportion that would accurately predict aphid population changes. Materials and Methods Sampling. From 1998 to 2001, the relationship between the proportion of cereal aphids parasitized (P p ) and the proportion of tillers with 0 mummiþed aphids (P tm ) was estimated on 57 occasions in Þelds of hard red winter wheat located in central and western Oklahoma. On each sampling date, between 0800 and 1700 hours, 100 tillers (stems) were individually collected in each Þeld. Tillers were collected by traveling a zig-zag U-shaped pattern through each Þeld and carefully clipping an individual tiller at ground level approximately every 10 m. Sampling occurred throughout the winter wheat growing season (OctoberÐMay); individual Þelds were sampled up to Þve times during a growing season, however, sampling times were separated by a minimum of 7 d when temperatures exceeded developmental requirements for greenbugs and L. testaceipes. All Þelds were 2ha in size, but cultivar, planting date, and crop management practices were not controlled. Wheat plant growth stage ranged from early shoot development to boot stage (Nelson et al. 1988). We counted the number of aphids (all species and stages) and mummiþed aphids on each tiller in each Þeld. Additionally, a minimum of 40 nonmummiþed aphids were randomly collected and placed on greenhouse grown wheat seedlings to monitor development of mummies and emergence of adult parasitoids. Adult parasitoids were identiþed to species after emergence. In each Þeld, P p was calculated by summing the total number of mummiþed aphids and the estimated number of parasitized nonmummiþed aphids (calculated from greenhouse rearing data), then dividing this value by the sum of mummiþed and nonmummiþed aphids. Independent data were obtained from 1998 to 2002 in central and western Oklahoma hard red winter wheat Þelds to: (1) validate the observed relationship between P tm to P p (34 Þelds), (2) determine whether a natural enemy threshold (P p 0.2), estimated from a reliable P tm threshold value (based on the P tm to P p relationship), accurately predicts suppression of aphid populations below economic thresholds, and (3) develop and evaluate a binomial sequential sampling plan for classifying P tm as above or below a critical value that would accurately predict whether aphid populations would be suppressed. Documenting the Relationship Between P tm and P p. For the original (57 Þelds) and validation (34 Þelds; 2001Ð2002) data sets, we documented the linear relationship between the proportion of tillers infested with 0 mummies (P tm ) and the proportion of aphids parasitized from each Þeld (P p ). Because of weak linear relationships between P tm and P p, data were plotted and visually inspected to document the lowest value for P tm that reliably predicted a P p value 0.2 (the recommended parasitism threshold). Predicting Aphid Suppression from P tm. During individual growing seasons from 1998 to 2002, winter wheat Þelds were monitored over time to evaluate the error rates for selected decision threshold proportions of P tm. Based on data documenting the relationship

3 June 2003 GILES ET AL.: SAMPLING FOR APHID PARASITOIDS IN WHEAT 977 between P tm to P p (previous section), the upper decision threshold proportion (P tm1 ) was set at 0.1 tillers with 0 mummiþed aphids. This upper P tm threshold predicts that P p will always exceed 0.20; the value selected is suggested as a parasitism threshold for aphids in winter wheat (Patrick and Boring 1990, Royer et al. 1998). Proportions above P tm1 predict that parasitism would be sufþcient to prevent aphid intensity from exceeding selected economic thresholds (ET; aphids per tiller). The lower decision threshold proportion (P tm0 ) was set at 0.02 tillers with a mummiþed aphid; a linear analysis of data points between P p 0.2 and P tm 0.1 revealed an intercept of Proportions below P tm0 predict that parasitism would be insufþcient to prevent aphid intensity from exceeding selected ETs. Proportions between P tm0 and P tm1 would fall within the nondecision (continue sampling) zone. Sampling procedures in each Þeld were identical to those previously described, however, 120 tillers were collected. After the Þrst sampling date, individual Þelds were sampled every 4Ð14 d and from 3 to 13 times until aphid intensities were permanently suppressed below three per tiller. The variability in number of times Þelds were sampled reßected an effort to initiate sampling during each month of winter wheat growth (October - May). Choosing variable sampling initiation dates allowed us to test whether the decision threshold proportions (P tm1 0.1, P tm0 0.02) for estimating the effects of parasitism were applicable throughout the winter wheat growing season in Oklahoma. This approach is important because calculated economic thresholds increase with increasing wheat growth (Giles et al. 2000a, 2000b; Elliott et al. 2002). During initial sampling dates, aphid populations were occasionally higher than the targeted aphid thresholds of 3, 9, and 15 per tiller (Giles et al. 2000a, 2000b). Therefore, cereal aphid populations in winter wheat were monitored over time in 16, 22, and 25 Þelds for ETs of 3, 9, and 15 aphids per tiller, respectively. Individual Þelds were sampled and P tm was recorded. For each Þeld, the upper (P tm1 0.1) and lower (P tm0 0.02) decision threshold proportions were used as decision points for predicting whether aphid populations would exceed ETs of 3, 9, and 15 aphids per tiller. Giles et al. (2000 a, 2000b) developed binomial sequential sampling plans for this range of aphid intensities in Oklahoma winter wheat. The accuracy of decisions for predicting potential aphid abundance from estimates of P tm (relative to selected ETs) was determined by recording the number of decision errors. Decision errors occurred when (1) aphid intensity remained below the ET when predicted to exceed, or (2) when aphid intensity exceeded the ET when predicted to remain below. Additionally, we were interested in comparing the error rates associated with using this newly developed upper threshold of P tm 0.1 to error rates associated with using P p 0.2 as the upper decision threshold. When P tm 0.1, corresponding P p values are often well in excess of 0.2. Therefore, we predicted that error rates associated with using an upper threshold of P tm 0.1 would be lower than those associated with using P p 0.2 as an upper decision threshold. For this comparison, published guidelines (Patrick and Boring 1990, Royer et al. 1998, Jones 2001) for upper (P p 0.20) and lower (P p 0.05) decision threshold proportions were used as decision points for predicting whether aphid populations would exceed or remain below ETs. Development of a Binomial Sequential Sampling Plan using P tm. Using the binomial distribution formulas for classiþcation of a proportion based on WaldÕs (1947) sequential probability ratio test, we calculated sampling plan parameters (slopes and intercepts) for predicting the effects of parasitoids on cereal aphid populations in winter wheat from estimates of P tm (Jones 1994, Young and Young 1998). The upper (P tm1 ) and lower (P tm0 ) decision threshold proportions for P tm were 0.1 and 0.02, respectively. Type 1( ) and Type 2 ( ) error rates were set at 0.1. Additionally, we computed the operating characteristic curve, corresponding average sample number (ASN) requirements, and minimum sample size (Jones 1994, Young and Young 1998). The operating characteristic and ASN curves consider only variability resulting from the binomial probability distribution for speciþed values of and. Statistical Analyses. Linear regression models were Þt using the PROC REG procedure of SAS (SAS Institute 1996) at a signiþcance level of P Before analyses, P tm and P p values were transformed [arcsine (square root of value)] to normalize data (Steel and Torrie 1980). Results and Discussion Parasitoid Species Recovered. Similar to previous surveys of cereal aphid parasitoids in the southern plains (Gilstrap et al. 1984), L. testaceipes was the most commonly (95%) recovered primary parasitoid from central and western Oklahoma wheat Þelds, followed by Diaeretiella rapae M c Intosh (5%). Hyperparasites (species complex) were recovered during late spring months as wheat was maturing, but accounted for only 12% of all specimens recovered. Documenting the Relationship Between P tm and P p. SigniÞcant but weak linear relationships were detected between P tm and P p for both original (r ; df 1, 55; P 0.001) and validation data (r ; df 1, 32; P 0.001). For both original and validation data, there were several Þeld sites with very low P tm values but very high corresponding P p values (Fig. 1). Obviously, for these Þelds, a high percentage of the collected nonmummiþed aphids were actually parasitized (based on greenhouse rearing data). Locations with low P tm but high P p values were not commonly observed during any speciþc period of the winter wheat growing season; multiple Þelds with these data points were observed during fall, winter, and spring months. Fields with low P tm but high P p values could occur based on our timing of sampling or a number of interesting interactions we have observed between cereal aphids and L. testaceipes (D.B.J. and

4 978 JOURNAL OF ECONOMIC ENTOMOLOGY Vol. 96, no. 3 Fig. 1. Proportion of tillers infested with 0 mummiþed aphids (P tm ) versus proportion of aphids parasitized (P p )in winter wheat. N.C.E., unpublished data). Some Þelds were sampled more than once and occasionally as populations of parasitoids were increasing. Emerging parasitoids were heavily attacking nonparasitized aphids and we occasionally observed dramatic increases in P p from one date to the next. Potentially, although uncon- Þrmed, parasitoid populations may also rapidly colonize Þelds from adjacent areas and attack nearly all of the aphids present. Alternatively, in wheat Þelds that were sampled during cold weather (DecemberÐ January), low P tm but high P p values could result from heavily parasitized cereal aphid populations that were not developing or continued foraging and attacks by parasitoids as observed by Jones (2001). Our data set with winter wheat Þelds having low P tm but high P p values would prevent precise predictions of actual levels of parasitism (P p ) values from binomial counts of mummiþed aphids per tiller (P tm ). This observation is problematic for Þtting a linear model, but does not affect the usefulness of the Þndings for developing a sampling approach. The distribution of data points (Fig. 1) reveal that a minimum P p value of 0.20 (natural enemy threshold recommended by Patrick and Boring 1990 and Royer et al. 1998) is certain to occur if P tm in the Þeld exceeds Clearly, the frequency of tillers with a mummiþed aphid (P tm ) allows for prediction of a minimum level of actual parasitism (P p ) that is useful as a natural enemy threshold. Predicting Aphid Suppression from P tm. Because any future use of a parasitism sampling plan based on P tm requires validation on predicted aphid suppression, we evaluated cereal aphid populations over time in winter wheat. Using upper and lower threshold proportions set at P tm1 0.1 and P tm0 0.02, respectively, aphid intensity exceeded the economic threshold when predicted to remain below in only one Þeld (ET 3, Table 1). In this Þeld, however, aphid intensity reached a maximum of four per tiller before decreasing to nearly undetectable levels within 7 d. In

5 June 2003 GILES ET AL.: SAMPLING FOR APHID PARASITOIDS IN WHEAT 979 Table 1. Validation of decision threshold methods for predicting effects of parasitoids over a range of economic thresholds (ET) for cereal aphids in Oklahoma winter wheat fields, Method Decision thresholds a ET b n c Predictions Lower Upper Errors Correct d 2 ET 1 ET 1 ET 2 ET No decisions e Proportion tillers with a mummy Proportion parasitized Proportion parasitism (P p ) calculated from Þeld-collected and greenhouse reared aphids. a Values below lower decision threshold proportion (P tm ) predict that parasitism would be insufþcient to prevent aphid intensity from exceeding the ET. Values above the upper decision threshold proportion predict that parasitism would be sufþcient to prevent aphid intensity from exceeding the ET. b Economic thresholds expressed as number of aphids per tiller (intensity). c Number of winter wheat Þelds evaluated over time. d Aphid intensity exceeded or remained below ET as predicted. e Proportions from Þeld samples were between lower and upper decision thresholds. three Þelds, at each targeted ET, we observed that aphid intensities remained below ETs when predicted to exceed (Table 1). Based on the number of Þelds with low P tm but high P p values from original and validation data (Fig. 1), this relatively high error rate would be expected. Indeed, in two of these Þelds, P p was well above 0.20 despite corresponding P tm values For most of the Þelds evaluated, aphid intensity was correctly predicted to increase above or be maintained below respective ETs. In 1Ð5 Þelds, P tm was between 0.1 and 0.02, therefore no prediction was made regarding aphid suppression. In these no-decision Þelds, the ET was exceeded Þve of 11 times (Table 1). This observed 50:50 ratio of exceeding or remaining below the ET justiþes the deþned no-decision zone between 0.1 and In evaluating the precision of parasitism thresholds based on P tm, we evaluated recommended upper (0.20) and lower (0.05) decision threshold proportions for P p (Patrick and Boring 1990, Royer et al. 1998, Jones 2001) as decision points for predicting whether aphid populations would exceed ETs of 3, 9, and 15 aphids per tiller. Because P tm 0.1 corresponds to P p values that are often well in excess of 0.2, we would predict that error rates associated with using an upper threshold of P tm 0.1 would be lower than those associated with using P p 0.2 as an upper decision threshold. Our predictions appear to be correct; when using P p 0.2 as an upper decision threshold, aphid intensities exceeded ETs when predicted to remain below in 8Ð19% of the Þelds (Table 1). This Þnding suggests that when ET 3 aphids per tiller, the upper decision threshold of P p 0.2 may be too low. This apparent problem was avoided by using P tm1 0.1 as an upper decision threshold (Table 1); in Þelds where P tm 0.1, nearly all associated values for P p were above 0.3. Development of a Binomial Sequential Sampling Plan Using P tm. The binomial sequential sampling plan parameters and estimated minimum sample size for predicting suppression of cereal aphids by the proportion of wheat tillers with a mummiþed aphid (P tm ) were developed with type 1 and 2 error rates set at 0.1 (Table 2). The calculated minimum tiller sample size of 26 is far below the 40 recommended for aphid binomial sequential sampling plans in winter wheat (Giles et al. 2000a, 2000b). The operating characteristics and ASN requirements are presented in Fig. 2. The rapidly decreasing slope of the operating characteristic curve indicates that variations in P tm above 0.1 or below 0.02 would be quickly classiþed. The ASN required to classify P tm as above 0.1 or below 0.02 peaked at 37 (P tm 0.036). Binomial distribution formulas for classiþcation of a proportion based on the sequential probability ratio test would be expected to yield maximum ASNs between P tm1 and P tm0 (Wald 1947, Jones 1994, Young and Young 1998). Interestingly, the peak ASN required to classify P tm is similar to the minimum recommended for classifying cereal aphids in winter wheat (Giles et al. 2000a, 2000b). Clearly, future integration of aphid and parasitoid sampling schemes is possible. The sampling plan developed during this study allows for rapid classiþcation of P tm, and independent of initial population intensities, accurately predicts suppression of cereal aphid populations by parasitoids. Suppression of cereal aphid populations by parasitoids was relatively common in Þelds with P tm values below Table 2. Binomial sequential sampling plan parameters for predicting effects of aphid parasitoids in winter wheat from the proportion of wheat tillers infested with >0 mummified aphids (P tm ) Decision and thresholds a Sampling stop lines Minimum sample size b P tm0 P tm1 Intercepts Slope and are type 1 and type 2 error rates, respectively. a Proportion of tillers infested with 0 mummiþed aphids. Values below lower proportion predict that parasitism would be insufþcient to prevent aphid intensity from exceeding the ET. Values above upper proportion predict that parasitism would be sufþcient to prevent aphid intensity from exceeding the ET. b MSS intercept /slope.

6 980 JOURNAL OF ECONOMIC ENTOMOLOGY Vol. 96, no. 3 Fig. 2. Binomial sequential sampling plan operating characteristics and average sample number functions for proportion of tillers infested with 0 mummiþed aphids (P tm ). Upper and lower threshold proportions were set at P tm1 0.1 and P tm0 0.02, respectively. Type 1 ( ) and Type 2 ( ) error rates were set at 0.1. the lower decision threshold proportion of 0.02, however, in these Þelds, corresponding P p values ranged from 0.09 to These decision errors would not affect aphid management programs or sampling plans. In each of these Þelds, binomial sequential sampling plans for aphids in wheat would have required continued sampling and eventually would have concluded that insecticide control would not have been necessary (Royer et al. 2002a, 2002b). Sampling Plan Recommendations. We recommend the use of the binomial sequential sampling plan developed during this study for classifying the proportion of tillers with a mummiþed aphid (P tm ) that accurately estimates a minimum proportion of cereal aphids parasitized (0.2) in winter wheat Þelds in the southern plains. There is a clear need to formally integrate this sampling plan for estimating the effects of parasitoids with binomial sequential sampling plans already developed for cereal aphids in winter wheat (Elliott et al. 1994; Giles et al. 2000a, 2000b; Royer et al. 2002a, 2002b). The integrated sampling approach will require the use of sequential sampling plan parameters for estimating P tm, and parameters for classifying aphid intensities. A minimum and maximum of 40 and 130 tillers, respectively, should be evaluated throughout a Þeld for the presence of mummiþed aphids and nonmummiþed aphids in winter wheat Þelds 40 ha. The initial management decision should be based on classiþcation of P tm.ifp tm is classiþed above 0.1, aphid populations are expected to remain below or would be quickly suppressed below the ET. Under this scenario, according to the Þndings of our study, classiþcation of aphid intensity would not be necessary. If P tm cannot be classiþed (between 0.1 and 0.02) or is classiþed below 0.02, the sequential sampling plan for aphids should be used for management

7 June 2003 GILES ET AL.: SAMPLING FOR APHID PARASITOIDS IN WHEAT 981 decisions. If aphids are classiþed as above a deþned threshold, treatment is recommended. If aphids cannot be classiþed (no-decision zone) or are classiþed as below a threshold, Þelds should be resampled according to developed protocols (Royer et al. 2002a, 2002b). It is important to note that population suppression by parasitoids may occur when P tm is classiþed below 0.1. When Þelds need to be resampled, prior information on P tm may be useful for predicting the effects of parasitoids. We believe that this sampling plan which allows for simultaneous evaluation of aphid intensity and parasitism will be efþcient and useful for consultants and producers in the southern plains and decrease the number of unnecessary insecticide applications. Acknowledgments We thank A. Mackay, R. Fuentes, F. Tao, T. Johnson, W. French, J. Spurlin, J. Shelburne, E. Ismail, and D. Kastl for technical help. We also thank R. Berberet and T. Phillips for their critical review of this manuscript. Mention of trade names or commercial products in this article is solely for the purpose of providing speciþc information and does not imply recommendation or endorsement by the U.S. Department of Agriculture. This work was approved for publication by the Director of the Oklahoma Agricultural Experiment Station, and supported in part under project OKL02334 and OCAST grant AR References Cited Burton, R. L., D. D. Simon, K. J. Starks, and R. D. Morrison Seasonal damage by greenbugs (Homoptera: Aphididae) to a resistant and a susceptible variety of wheat. J. Econ. Entomol. 78: 395Ð401. Elliott, N. C., G. L. Hein, and B. M. Shepard Sampling arthropod pests of wheat and rice, pp. 627Ð666. In L. Pedigo and G. D. Buntin [eds.], Handbook of sampling methods for arthropods in agriculture. CRC, Boca Raton, FL. Elliott, N. C., K. L. Constein, T. R. Royer, K. L. Giles, S. D. Kindler, and D. A. Waits Greenbug Pest Management Decision Support System. Pest Management Software, USDAÐARS. okstate. edu/ greenbug/index. htm. Feng, M. G., R. M. Nowierski, and Z. Zeng Populations of Sitobion avenae and Aphidius ervi on spring wheat in the northwestern United States. Entomol. Exp. Appl. 67: 109Ð117. Flaherty, D. L., L. T. Wilson, S. C. Welter, C. D. Lynn, and R. Hanna Spider mites. In D. L. Flaherty, L. P. Christensen, W. T. Lanini, J. J. Marois, P. A. Phillips, and L. T. Wilson [eds.], Grape pest management. Univ. of California, Berkeley, Agric. Sci. Publ Flaherty, D. L., and L. T. Wilson Biological control of insects and mites on grapes, pp. 853Ð869. In T. S. Bellows, and T. W. Fisher [eds.], Handbook of biological control. Academic, San Diego. Giles, K. L., T. A. Royer, N. C. Elliott, and S. D. Kindler. 2000a. Development and validation of a binomial sequential sampling plan for the greenbug (Homoptera: Aphididae) infesting winter wheat in the southern plains. 93: 1522Ð1530. Giles, K. L., T. A. Royer, N. C. Elliott, and S. D. Kindler. 2000b. Binomial sequential sampling plan of Rhopalosiphum Padi in winter wheat in the southern plains. Southwestern Entomol. 25: 191Ð199. Gilstrap, F. E., T. J. Kring, and G. W. Brooks Parasitism of aphids (Homoptera: Aphididae) associated with Texas Sorghum. J. Econ. Entomol. 13: 1613Ð1617. Kieckhefer, R. W., and B. H. Kantack Yield losses in winter grains caused by cereal aphids (Homoptera: Aphididae) in South Dakota. J. Econ. Entomol. 81: 317Ð 321. Kindler, S. D., N. C. Elliott, K. L. Giles, T. A. Royer, R. Fuentes-Grandaos, and F. Tao Effect of greenbug (Homoptera: Aphididae) on yield loss of winter wheat. J. Econ. Entomol. 95: 89Ð95. Jackson, H. B., L. W. Coles, E. A. Wood, and R. D. Eikenbary Parasites reared from the greenbug and corn leaf aphid in Oklahoma in 1968 and J. Econ. Entomol. 63: 733Ð736. Jones, V. P Sequential estimation and classiþcation procedures for binomial counts, pp. 175Ð205. In L. Pedigo and G. D. Buntin [eds.], Handbook of sampling methods for arthropods in agriculture. CRC, Boca Raton, FL. Jones, D. J Natural enemy thresholds for greenbug, Schizaphis graminum (Rondani), on winter wheat. M. S. thesis, Oklahoma State Univ., Stillwater, OK. Nelson, J. E., K. D. Kephart, A. Bauer, and J. F. Conner Growth staging of wheat, barley and wild oat: a strategic step to timing of Þeld operations. NDSU Extension Service. PE North Dakota State University, Fargo, ND. Nyrop, J. P., and W. van der Werf Sampling to predict and monitor biological control. In L. Pedigo and D. G. Buntin [eds.], Handbook of sampling methods for arthropod pests in agriculture. CRC, Boca Raton, FL. Patrick, C. D., and E. P. Boring III Managing insect and mite pests of Texas small grains. Texas Agricultural Extension Service Publication B-1251, College Station, TX. Riedell, W. E., R. W. Kieckhefer, S. D. Haley, M. C. Langham, and P. D. Evenson Winter wheat responses to bird cherry-oat aphids and barley yellow dwarf virus infection. Crop Sci. 30: 158Ð163. Royer, T. A., K. L. Giles, and N. C. Elliott Small grain aphids in Oklahoma. Oklahoma Cooperative Extension Service, Oklahoma State University Extension Facts, F-7183, Stillwater. OK. Royer, T. A., K. L. Giles, and N. C. Elliott. 2002a. Glance Õn go sampling for greenbugs in winter wheat: spring edition. Oklahoma Cooperative Extension Service, Oklahoma State University Extension Facts, L-306, Stillwater, OK. Royer, T. A., K. L. Giles, and N. C. Elliott. 2002b. Glance Õn go sampling for greenbugs in winter wheat: fall edition. Oklahoma Cooperative Extension Service, Oklahoma State University Extension Facts, L-307, Stillwater, OK. SAS Institute Version SAS Institute, Cary, NC. Steel, R.G.D., and J. H. Torrie Principles and procedures of statistics: a biometrical approach. McGraw-Hill, New York. Tanigoshi, L. K., S. C. Hoyt, and B. A. Croft Basic biological and management components for mite pests and their natural enemies, pp. 153Ð202. In B. A. Croft, and S. A. Hoyt [eds.], Integrated management of insect pests of pome and stone fruits. Wiley-Interscience, New York. Wald, A Sequential analysis. Wiley, New York. Webster, J. A Economic impact of the greenbug in the western United States: 1992Ð3 Volume Publication No Great Plains Agricultural Council, Stillwater, OK. Wilson, L. T., M. A. Hoy, F. G. Zalom, and J. M. Smilanick Sampling mites in almonds. I. Within-tree distri-

8 982 JOURNAL OF ECONOMIC ENTOMOLOGY Vol. 96, no. 3 bution and clumping pattern of mites with comments on predator-prey interaction, Hilgardia. 52: 1. Wilson, L. T Estimating abundance, impact, and interactions among arthropods in cotton agroecosystems. In L. Pedigo and D. G. Buntin (eds.), Handbook of Sampling Methods for Arthropod Pests in Agriculture. CRC, Boca Raton, FL. Young, L. J., and J. H. Young Statistical ecology. Kluwer Academic Publishers, Boston. Received for publication 18 July 2002; accepted 16 December 2002.