Diuraphis Noxia Reproduction and Development with a Comparison of Intrinsic Rates of Increase to Other Important Small Grain Aphids: a Meta-Analysis

QUANTITATIVE ECOLOGY Diuraphis noxia Reproduction and Development With a Comparison of Intrinsic Rates of Increase to Other Important Small Grain Aphids: A Meta-Analysis

1 SCOTT C. MERRILL, THOMAS. O. HOLTZER, AND FRANK B. PEAIRS

Department of Bioagricultural Sciences and Pest Management, Colorado State University, Campus Delivery 1177, Fort Collins, CO 80523

Environ. Entomol. 38(4): 1061Ð1068 (2009) ABSTRACT The Russian wheat aphid, Diuraphis noxia (Kurdjumov), is a signiÞcant pest of small grains in the United States and worldwide. There is an increasing need for quality population dynamic

models to aid in development of integrated pest management strategies. Unfortunately, there exists Downloaded from high variability in published data regarding basic life history traits that frequently direct model parameterization. Metadata were analyzed to develop relationships between temperature and reproductive and developmental traits of D. noxia. SpeciÞcally, functions were developed between temperature and the following traits: lifespan, fecundity, fecundity rate, pre-nymphipositional period, reproductive period, and intrinsic rate of increase. Lower and upper temperature reproductive Њ

thresholds were calculated as 0.6 and 36.9 C, respectively. The lower temperature developmental http://ee.oxfordjournals.org/ threshold was calculated as Ϫ0.69ЊC. Modeled longevity reached its maximum at Ϸ80 d. Meta-analysis indicates maximum fecundity at Ϸ18.5ЊC, with a maximum fecundity rate of Ϸ2.1 nymphs per day over the nymphipositional period. The calculated maximum total fecundity was Ϸ55 nymphs per female. The maximum reproductive period was calculated to be 29.9 d. Compared with other aphid species, as temperature increased, the intrinsic rate of increase of D. noxia increased more slowly relative to Schizaphis graminum (Rondani) and Rhopalosiphum padi L., but at a similar rate to Sitobian avenae (F.).

KEY WORDS Russian wheat aphid, reproduction, development, intrinsic rate of increase, meta- by guest on March 16, 2016 analysis

Aphid phenology and population growth rates are aphid, Diuraphis noxia (Kurdjumov) (no male D. largely dictated by climate and weather variables. For noxia have been found to date in North America). example, temperature has been correlated with spring Since its introduction into the United States in the migration ßight date in cereal aphids (Turl 1980, mid-1980s, this aphid has caused losses of more than Walters and Dewar 1986). The severity of winter one billion dollars to the small grain industry, prompt- weather has been directly related to aphid overwin- ing efforts to develop an effective integrated pest tering success (Dewar and Carter 1984, Messina 1993, management (IPM) strategy (Webster et al. 1994, Armstrong and Peairs 1996). Many species have mech- Morrison and Peairs 1998). Localized overwintering anisms that mitigate population mortality during pe- success has been linked to increased crop losses. That riods of harsh climatic conditions. For example, the is, D. noxia populations that survive the overwinter egg-laying stage of holocyclic aphid species is more period in a given Þeld typically cause signiÞcantly resistant to stressful conditions and thus enables a more damage than populations that build up after relatively robust strategy for survival during harsh migration from outside sources (Peairs et al. 2006). overwinter or oversummer periods. However, some Population mortality has been found to be correlated species rely largely on an anholocyclic strategy, which with snow, precipitation, and temperature variables can give them the advantage of early and rapid pop- (Armstrong and Peairs 1996), with temperature vari- ulation growth during mild conditions but does not ables having the largest impact (Merrill and Holtzer protect them from increased mortality resulting from unpublished work). Variables that drive population exposure to extreme conditions by life stages that are dynamics (e.g., temperature) must be robustly quan- ill-equipped to deal with them. tiÞed if we are to develop accurate population dy- An important pest species with an anholocyclic life- namic prediction models (e.g., outbreak prediction cycle within North America is the Russian wheat models and risk assessment models for use in IPM). Prediction models frequently rely on accurate bound- 1 Corresponding author, e-mail: [email protected]. ary conditions as well as rate and stage functions quan-

0046-225X/09/1061Ð1068$04.00/0 ᭧ 2009 Entomological Society of America 1062 ENVIRONMENTAL ENTOMOLOGY Vol. 38, no. 4 tifying aphid development and reproduction. How- Table 1. List of metadata sources ever, results from studies examining the inßuence of temperature on development and reproduction of D. Citation F RR TDT L NP PNP rm noxia are highly variable. For example, Nowierski et al. Aalbersberg et al. (1987) x x x x x x (1995) found a reproductive period of Ϸ41 d for D. Behle and Michels (1990) x x x x x x x Њ Girma et al. (1990) x x x x x x x noxia at 19.5 C, whereas Girma et al. (1990) found a Hawley et al. (2003) x x x reproductive period of9dat19.7ЊC. Kieckhefer and Elliott (1989) x x x x Because of high variability observed between stud- Merrill et al. (2008) x x x x x x x ies (stemming from differences in sampling tech- Michels and Behle (1988) x x x x x x Michels and Behle (1989) x x x x niques to sampling conditions), individual experi- Miller et al. (2003) x x x x x ments may have limited value if they cannot be Nowierski et al. (1995) x x x x generalized to address the overall questions of interest Randolph et al. (2007) x x x x x (Gurevitch and Hedges 1993). Combining results Webster and Starks (1987) x x x x from multiple experiments typically allows for greater Webster et al. (1993) x x generalization than the use of a single experiment. Of The x indicates that metadata were obtained from the source for the available methods for combining results from in- each category. Metadata either were extracted directly from text or dependent experiments, meta-analyses are encour- calculated from data provide in the source. aged because they provide statistically defendable F, fecundity; RR, reproductive rate; TDT, temperature developmental threshold; L, longevity; NP, nymphipositional period; PNP, quantiÞcation of the analysis results, in contrast to a pre-nymphipositional period; r , intrinsic rate of increase. Downloaded from number of other methods currently used (Gurevitch m et al. 1992, Gurevitch and Hedges 1993). For example, one potentially misleading method for synthesizing this meta-analysis can be used to improve D. noxia results from independent studies is called vote count- “ model parameterization and in the development of ing. Vote counting is frequently used to determine ” IPM strategies.

whether an effect is signiÞcant by Þrst categorizing http://ee.oxfordjournals.org/ multiple studies where the effect has been studied by whether the effect was found to be signiÞcant or not Materials and Methods signiÞcant. Vote counting results (signiÞcant or not) are determined simply by declaring the category that We used the following references as D. noxia meta- received the most votes the winner. Unfortunately, data sources: Aalbersberg et al. (1987), Webster and vote counting and other similar methods can be mis- Starks (1987), Michels and Behle (1988, 1989), Kiec- leading because signiÞcance varies based on both ef- khefer and Elliott (1989), Behle and Michels (1990), fect and sample sizes (Gurevitch et al. 1992). The Girma et al. (1990), Webster et al. (1993), Nowierski

meta-analysis we conducted enables a useful synthesis et al. (1995), Hawley et al. (2003), Miller et al. (2003), by guest on March 16, 2016 of research from numerous studies, resulting in a more Randolph et al. (2008), and Merrill et al. (2008). Table robust understanding of the relationships between 1 indicates which metadata (e.g., fecundity or longev- temperature and life history traits of D. noxia com- ity metadata) were provided by each reference. pared with the use of results from a single data set. Parameters for development and reproduction vari- The inßuence of weather and climate (particularly ables were extracted from each of the aforementioned temperature) on D. noxia biology has been extensively research papers. Parameterized variables of interest studied (Webster and Starks 1987). Temperature were longevity, total fecundity, daily fecundity rate, seems to be a dominant factor inßuencing reproduc- reproductive period (or nymphipositional period), tive and development rates (Aalbersberg et al. 1987). pre-nymphipositional period (i.e., the period from However, many other variables play strong roles. Ex- birth to the production of the aphidÕs Þrst offspring), amples of factors inßuencing reproductive and devel- and intrinsic rate of increase. Ideally, measures of opmental rates include host plant growth stage (Girma variance would be reported across studies, which et al. 1990), host quality (Merrill et al. 2008), aphid would allow for temperature data to be weighted biotype (Randolph et al. 2008), population density based on a measure of their variance (e.g., the SE). (Michaud et al. 2006), and feeding site on the plant Unfortunately, central tendency measurements were (Stary 1999). Much more of the variation associated not uniformly reported. However, sample number was with reproduction and development could be ex- uniformly reported and can be related to the SE, plained if all these covariates were included in a meta- which is equal to the SD of the sample population analysis. However, our analysis was restricted because divided by the square root of the number of samples. of inconsistencies among studies (e.g., different pro- Therefore, because an increased sample size reduces tocols), inadequately reported data, or data were con- the SE, metadata were weighted by the square root of sidered too variable to be included (e.g., unexplained the sample size. SpeciÞcally, the higher the number of variation resulting from cultivar effects). Rather, we samples used to obtain a data point, the more weight concentrated on temperature effects and ignored the was conferred on that data point. If multiple temper- variability associated with other covariates. atures were studied within a single research paper, Our goal was to provide robust quantiÞcation of the each temperature was considered a data point and was relationships between temperature and life history weighted by the sample size used to develop the tem- traits of D. noxia. QuantiÞed relationships provided by perature-speciÞc data. August 2009 MERRILL ET AL.: META-ANALYSIS OF D. noxia REPRODUCTION AND DEVELOPMENT 1063

We modeled the relationships between tempera- slightly different values, this component of variation ture, reproduction, and development using simple was not included in our analysis. mathematical relationships. Functions for describing arthropod temperature-dependent rates may Þt a non- Results symmetric curve around an optimal value (Logan et al. 1976). For example, high temperatures may create a Fecundity and Temperature. Temperature has a rapidly declining curve, possibly attributed to enzyme strong effect on lifetime fecundity, deÞned as the total denaturization or desiccation, in many temperature- number of nymphs produced over the lifetime of an related functions (Logan et al. 1976). However, until aphid. Exceptionally high and low temperatures result aphid data collection is more uniform across studies, in no reproduction. After a preanalysis, maximum fe- a parsimonious approach may be warranted: that is, cundity was estimated to occur at Ϸ20ЊC with upper assuming linearity or symmetry until variability in data and lower thresholds occurring at Ϸ0 and Ϸ35ЊC. collection across studies is reduced. Therefore, we Given these assumptions, we assumed that the fecun- modeled relationships between temperature, repro- dityÐtemperature relationship could be Þt by a Gauss- duction, and development using linear, Gaussian, ian or Weibull function. The Gaussian function had a Weibull, and power (as suggested by exponentially lower AIC value than the Weibull function (Gaussian ϭ ϭ increasing population growth rates) functions based function AIC 358.95, Weibull function AIC on reasonable biological expectations (e.g., fecundity 360.16). These AIC values are very close, making rate was not expected to linearly increase with tem- model selection complex. That is, the Gaussian func- Downloaded from perature, and therefore linear functions were not Þt to tion had a likelihood of 0.65 and the Weibull function the fecundity data). We used an information theoretic had a likelihood of 0.35, indicating strong likelihoods approach for model selection and, where necessary, for each model being the best approximating model model averaging (Burnham and Anderson 2002, 2004). (Burnham and Anderson 2004). Figure 1 depicts mod- Comparison of Intrinsic Rates of Increase Between eled lifetime fecundity using both the Guassian and Important Small Grain Aphids and D. noxia. Metadata Weibull functions versus temperature metadata. A http://ee.oxfordjournals.org/ were assembled for three other important aphid pests Gaussian function was Þt to the data using a weighted of small grains on the Great Plains: Rhopalosiphum least mean squares error approach and is calculated as: padi L. (Dean 1974, Michels and Behle 1989, Selhorst Gaussian function: nymphs/female ϭ 55.3 1994, Sengonca et al. 1994, Asin and Pons 2001, Nyaanga et al. 2005); Sitobion avenae (F.) (Dean 1974, ϫ eϪ(T Ϫ 18.4)2/92.5ˆ [1] Acreman and Dixon 1989, Sengonca et al. 1994, Asin where T is the temperature. The Gaussian function and Pons 2001); and Schizaphis graminum (Rondani) calculates a maximum fecundity of Ϸ55.3 nymphs (at (Wadley 1936, Webster et al. 1987, Walgenbach et al. Њ

18.4 C), well under the maximum reported average of by guest on March 16, 2016 1988, Michels and Behle 1989, Nyaanga et al. 2005). 81.5 nymphs (at 17.25ЊC) (Aalbersberg et al. 1987). Reliable intrinsic rate of increase metadata could not Ͼ Њ The Weibull function was Þt to the data using a be found for high temperatures (e.g., 30 C) for all weighted least squares approach and is calculated as: aphid species. Therefore, the temperatureÐintrinsic rate of increase relationship could not be compara- Weibull function: nymphs/female ϭ 156.5 tively modeled when an increase in temperature cor- 3.0 ͓͑ϪT/18.5͒3.0ˆ ͔ related with a decrease in the aphidÕs intrinsic rates of ϫ ͑T/18.5͒ ϫ e [2] increase. However, reliable data were found for rela- where T is the temperature. The Weibull function tionships between temperature and intrinsic rate of calculates a maximum fecundity of Ϸ57.6 nymphs (at increase for temperatures in a range where the intrin- 18.5ЊC). sic rate of increase was increasing with temperature The temperature reproduction thresholds are de- from the lower temperature development threshold Þned as the upper and lower temperature limits of toward a hypothetical maximum intrinsic rate of in- reproduction (i.e., temperatures at which reproduc- crease. To limit the meta-analysis to metadata within tion per female is less than one nymph per female). these constraints, data were discarded if the intrinsic Because both the Gaussian and Weibull functions had rate of increase was less than half of the maximum support in the data (i.e., strong likelihoods for being reported value at temperatures higher than the tem- the best approximating model), a model-averaged ap- perature at the maximum reported intrinsic rate of proach was used to calculate the reproductive thresh- increase. As an example, Asin and Pons (2001) re- olds (Burnham and Anderson 2002, 2004). That is, ported an intrinsic rate of 0.15 at 27.5ЊC for S. avenae. upper and lower reproduction threshold values were This data point is likely on the decreasing portion of calculated for both the Gaussian and Weibull func- the intrinsic rate of increase curve and was therefore tions, and the AIC weight (or in this case the likeli- discarded. Across all studies, each of the temperature hood) was used to average the values. The Gaussian by intrinsic rate of increase values was treated as a lower and upper reproduction threshold values were datum in the meta-analysis and weighted by the Ϫ0.9 and 37.7ЊC, whereas the Weibull lower and upper square root of the sample size. Although different threshold values were 3.4 and 35.4ЊC, respectively. methods of calculation of the intrinsic rate of increase The model-averaged lower and upper thresholds were (Birch 1948, Wyatt and White 1977) may result in 0.6 and 36.9ЊC, respectively. 1064 ENVIRONMENTAL ENTOMOLOGY Vol. 38, no. 4 Downloaded from Fig. 1. Metadata and Þtted Weibull and Gaussian functions for the average number of nymphs produced per female at temperatures from 5 to 30ЊC. Black triangles indicate the upper and lower reproductive threshold values.

Fecundity rate (Fig. 2), deÞned as the average num- nymphs per day at 20ЊC, which is similar to the value ber of nymphs per reproductive day, has a very similar obtained by Behle and Michels (1990). relationship with temperature as average lifetime fe- Temperature Development Threshold. The tem- http://ee.oxfordjournals.org/ cundity. However, unlike lifetime fecundity, the perature development threshold, deÞned as the low- Gaussian function (AIC ϭϪ130.6, likelihood Ͼ 0.99) est temperature where development occurs, has been had a much lower AIC value than the Weibull function calculated for D. noxia in numerous studies: Aalbers- ϭϪ Ͻ (AIC 120.3, likelihood 0.01). Therefore, only berg et al. (1987) reported 0.54ЊC, Girma et al. (1990) the Gaussian function was reported. Fecundity rate, reported a value of Ϫ1.57ЊC, and Kieckhefer and El- modeled using a Gaussian function, was Þt with a liott (1989) reported a lower developmental threshold weighted least squares approach: limit of 4.1ЊC for development of immature D. noxia. Nowierski et al. (1995) reported 5.2ЊC on excised Gaussian function: fecundity rate by guest on March 16, 2016 barley leaves. The average lower developmental Ϫ Ϫ ˆ ϭ 2.1 e (T 20.6)2/131.0 [3] threshold for these four studies was 2.06ЊC. where T is the temperature. This function calculates Using weighted metadata, the temperature devel- Ϫ Њ a maximum of 2.1 nymphs per day (at 20.6ЊC) during opment threshold (Fig. 3) was calculated as 0.69 C the reproductive period. The maximum reported av- using a linear regression of the inverse of develop- erage nymphs per day was 6.0 at 26.5ЊC (Haile et al. mental time from birth to adult against temperature 1999). However, when we considered this data point for all metadata (Aalbersberg et al. 1987, Kieckhefer as an outlier, because it was approximately twice as and Elliott 1989, Girma et al. 1990). large as the next highest value, and excluded it from Longevity and Temperature. Although many stud- the analysis, we obtained a maximum value of 3.1 ies have indicated reduced longevity at subzero tem-

Fig. 2. Metadata and Þtted Weibull and Gaussian functions for the average number of nymphs produced per female per day during the reproductive period at temperatures from 5 to 30ЊC. August 2009 MERRILL ET AL.: META-ANALYSIS OF D. noxia REPRODUCTION AND DEVELOPMENT 1065

Fig. 3. Linear regression of the inverse of the pre-nymphipositional time versus temperature metadata providing the temperature development threshold (Ϫ0.69ЊC). Downloaded from peratures (Knight et al. 1986, Butts 1992), metadata for increases as temperatures decrease toward the repro- aphids born and held at temperatures under their duction threshold, these data can be modeled using developmental threshold do not (and arguable can- either a linearly decreasing function or a power func- not) exist. We modeled longevity as linearly increas- tion. Alternatively Gaussian or Weibull function could ing as temperatures approached the lower tempera- be Þt to the data if we assume that the aphidÕs repro- http://ee.oxfordjournals.org/ ture development threshold. Longevity was modeled ductive period drops to zero when temperatures are as the following linear function of temperature using under the reproductive threshold. a weighted least squares approach (Fig. 4): The Gaussian function generated the best Þt with ϭ ϭ Ϫ ϫ ϩ the lowest AIC and highest likelihood (AIC 221.4, Linear function: longevity 2.0 T 79.6 likelihood ϭ 0.66) compared with the Weibull func- [4] tion (AIC ϭ 223.2, likelihood ϭ 0.27), linear function (AIC ϭ 225.9, likelihood ϭ 0.07), and the power func- where T is the temperature. Modeled longevity tion (AIC ϭ 237.1, likelihood ϭ 0.002). The Gaussian

Ϸ by guest on March 16, 2016 reached its maximum at 80 d, with longevity increas- function (Fig. 5) for the reproductive period versus ing as temperature decreased toward the develop- temperature relationship is given by the following mental threshold. equation: Reproductive Period and Temperature. Reproduc- tive period (nymphipositional period) is deÞned as Gaussian function: reproductive period the number of days from the Þrst to last birth per ϭ Ϫ(T Ϫ 12.0)2/297.1ˆ mother aphid. These data (Fig. 5) are quite variable 29.9 e [5] among studies and a pattern is not easily discerned. where T is the temperature. Results indicate a max- However, if we assume that the reproductive period imum reproductive period of 29.9 d at Ϸ12.0ЊC. The

Fig. 4. Linear regression of the inverse of longevity versus temperature metadata providing the maximum longevity (Ϸ80 d). 1066 ENVIRONMENTAL ENTOMOLOGY Vol. 38, no. 4

Interestingly, if one models pre-nymphipositional period as a power function (Fig. 6, Formula 6), a second estimate of the reproduction threshold can be obtained. The second estimate is based on the assump- tion that when the pre-nymphipositional period is equal to the maximum lifespan, no reproduction can occur. Substituting maximum lifespan (79.6 d as calculated using Formula 4) for the pre-nymphipositional period in Formula 6, the lower temperature reproduction threshold was calculated to be 3.7ЊC: Pre-nymphipositional period ϭ max lifespan ϭ 353.6 ϫ T(Ϫ1.14) [7] Fig. 5. Metadata and Þtted Weibull, Gaussian, and linear Comparison of Intrinsic Rates of Increase Between functions for the length of the reproductive period in days Important Small Grain Aphids and D. noxia. Results versus temperature. indicate that the slope of the relationship between temperature and intrinsic rate of D. noxia was rela- ϭ minimum reported average reproductive period re- tively low compared with R. padi (P 0.0244, Downloaded from ported in the original publications was6dbyMich- t-value ϭ 2.29) and S. graminum (P ϭ 0.0854, t-value ϭ els and Behle (1988) at their highest temperature 1.74) but likely not different than S. avenae (P ϭ regimen (30ЊC). The maximum reproductive period 0.4176, t-value ϭϪ0.81). The slope of the relationship reported was 56.44 d at 13ЊC (Aalbersberg et al. between temperature and intrinsic rate of R. padi was 1987). higher than S. avenae (P ϭ 0.0016, t-value ϭϪ3.25) but ϭ ϭ Pre-nymphipositional Period and Temperature. not different than S. graminum (P 0.8942, t-value http://ee.oxfordjournals.org/ Pre-nymphipositional period is deÞned as the number Ϫ0.13). The slope of the relationship between tem- of days from birth of an aphid to the birth of its Þrst perature and the intrinsic rate of increase values was offspring. Pre-nymphipositional period provides a higher in S. graminum than S. avenae (P ϭ 0.0148, necessary variable for calculating the intrinsic rate of t-value ϭϪ2.48). In general, the aphid species D. increase. SpeciÞcally, the delay between birth and the noxia and S. avenae were similar to each other regard- potential for the newly born aphid to give birth is ing changes to their intrinsic rates of increase in re- essential for calculating generation time. Data indicate sponse to changes in temperature (Fig. 7). R. padi and that the number of days until Þrst birth increases as S. graminum showed similar responses in their intrin-

temperature decreases toward the lower developmen- sic rates of increase to changes in temperature. by guest on March 16, 2016 tal threshold (Fig. 6). Pre-nymphipositional period was modeled using linear and power functions. The Discussion power function had a much better Þt (AIC ϭ 132.3, likelihood Ͼ 0.99) than the linear function (AIC ϭ This meta-analysis was designed to build a more 233.2, likelihood Ͻ 0.01): comprehensive, more robust understanding of the in- teraction between life history traits of D. noxia and Power function: pre-nymphipositional period temperature. The relationships derived from this re- ϭ 353.6 ϫ T(Ϫ1.14) [6] search have greater utility as guidelines for model parameterization and for informing IPM strategy than where T is the temperature. do individual data sets used alone or parameterization derived from simple averaging procedures. Our meta- analysis showed that the maximum lifespan of D. noxia is Ͻ3 mo, which indicates that aphid populations from the fall will not survive through to the spring unless reproduction occurs. Reproduction is unlikely below Ϸ1ЊC and above Ϸ37ЊC. D. noxia seems to have its highest fecundity occurring around 18.5ЊC. Longevity is increased as temperatures decrease toward the developmental threshold of approximately Ϫ1ЊC, under which damage may occur to the aphid (e.g., chill coma; Knight et al. 1986). As longevity increases, a reproductive penalty is incurred. That is, fecundity decreases as longevity increases. Although not quantiÞed in this study, D. noxia reproductive and development rates seem to be opti- mimal at lower temperatures compared with other Fig. 6. Metadata and the Þtted power function for pre- important small grain aphid pests on the Great Plains. nymphipositional period versus temperature. Therefore, typical conditions in the early spring August 2009 MERRILL ET AL.: META-ANALYSIS OF D. noxia REPRODUCTION AND DEVELOPMENT 1067

Fig. 7. Metadata and linear functions for the daily intrinsic rates of increase versus temperature for many important small grain aphids. Downloaded from

(cooler temperatures) should favor D. noxia popula- Armstrong, J. S., and F. B. Peairs. 1996. Environmental pa- tion development over many other aphid pests. For rameters related to winter mortality of the Russian wheat example, Harvey and Martin (1988) studying aphids in aphid (Homoptera: Aphididae): basis for predicting mor- Kansas found that D. noxia was more cold tolerant and tality. J. Econ. Entomol. 89: 1281Ð1287. better adapted to overwinter conditions than S. grami- Asin, L., and X. Pons. 2001. Effect of high temperature on http://ee.oxfordjournals.org/ num. the growth and reproduction of corn aphids (Homoptera: Aphididae) and implications for their population dynam- Additional meta-analysis of important aphid pests of ics on the northeastern Iberian peninsula. Environ. En- small grains on the Great Plains would be valuable for tomol. 30: 1127Ð1134. increasing our understanding of aphid population re- Behle, R. W., and G. J. Michels. 1990. Russian wheat aphid sponses to different temperature regimens. Such (Homoptera, Aphididae) development, reproduction, meta-analysis would also have implications for pest and survival on wheat and rye grown in 4 host-plant forecasting and for predicting aphid response to cli- media. Southwest. Entomol. 15: 109Ð121. mate change. Meta-analyses have a distinct advantage Birch, L. C. 1948. The intrinsic rate of natural increase of an over decisions made using a selection of a single “best” insect population. J. Anim. Ecol. 17: 15Ð26. by guest on March 16, 2016 study or a simple average of multiple studies in that Burnham, K. P., and D. R. Anderson. 2002. Model selection they allow for mathematical and statistical formulation and multimodel inference: the practical information the- incorporating data from across a wider scope of con- oretic approach. Springer, New York. Burnham, K. P., and D. R. Anderson. 2004. Multimodel in- ditions, which produces more robust results. ference: understanding AIC and BIC in model selection. Our meta-analysis provides improved parameter- Sociol. Methods Res. 33: 261Ð304. ization for D. noxia population dynamic models. For Butts, R. A. 1992. Cold hardiness and its relationship to example, degree-day models would be enhanced by overwintering of the Russian wheat aphid (Homoptera, inclusion of derived temperature dependencies, and Aphididae) in southern Alberta. J. Econ. Entomol. 85: outbreak prediction models would be improved by 1140Ð1145. more robust parameterization of the effects of tem- Dean, G. J. 1974. Effect of temperature on cereal aphids perature on aphids. Metopolophium dirhodum (Wlk), Rhopalosiphum padi (L), and Macrosiphum avenae (F) (Hem Aphididae). Bull. Entomol. Res. 63: 401Ð409. Acknowledgments Dewar, A. M., and N. Carter. 1984. Decision trees to assess the risk of cereal aphid (Hemiptera, Aphididae) out- This study was funded by USDAÐNRI Grant 2000-02992 breaks in summer in England. Bull. Entomol. Res. 74: “Effective Russian wheat aphid management by modeling 387Ð398. climate and natural enemies.” Girma, M., G. Wilde, and J. C. Reese. 1990. Inßuence of temperature and plant-growth stage on development, re- References Cited production, life-span, and intrinsic rate of increase of the Russian wheat aphid (Homoptera, Aphididae). Environ. Aalbersberg, Y. K., F. Dutoit, M. C. Vanderwesthuizen, and Entomol. 19: 1438Ð1442. P. H. Hewitt. 1987. Development rate, fecundity and Gurevitch, J., and L. V. Hedges. 1993. Meta-analysis. Com- lifespan of apterae of the Russian wheat aphid, Diuraphis bining the results of independent experiments, pp. 378Ð noxia (Mordvilko) (Hemiptera, Aphididae), under con- 398. In J. Gurevitch and S. M. Scheiner (eds.), Design and trolled conditions. Bull. Entomol. Res. 77: 629Ð635. analysis of ecological experiments. Chapman & Hall, New Acreman, S. J., and A.F.G. Dixon. 1989. The effects of tem- York. perature and host quality on the rate of increase of the Gurevitch, J., L. L. Morrow, A. Wallace, and J. S. Walsh. grain aphid (Sitobion-Avenae) on wheat. Ann. Appl. Biol. 1992. A meta-analysis of competition in Þeld experi- 115: 3Ð9. ments. Am. Nat. 140: 539Ð572. 1068 ENVIRONMENTAL ENTOMOLOGY Vol. 38, no. 4

Haile, F. J., L. G. Higley, X. Z. Ni, and S. S. Quisenberry. alarysym of three aphid species. Discov. Innov. 17: 117Ð 1999. Physiological and growth tolerance in wheat to 121. Russian wheat aphid (Homoptera: Aphididae) injury. Peairs, F. B., G. L. Hein, and M. J. Brewer. 2006. High plains Environ. Entomol. 28: 787Ð794. integrated pest management. Colorado State University, Harvey, T. L., and T. J. Martin. 1988. Relative cold tolerance Fort Collins, CO. of Russian wheat aphid and Biotype-E Greenbug (Ho- Randolph, T. L., S. C. Merrill, and F. B. Peairs. 2008. Re- moptera, Aphididae). J. Kans. Entomol. Soc. 61: 137Ð140. productive rates of Russian wheat aphid (Hemiptera: Hawley, C. J., F. B. Peairs, and T. L. Randolph. 2003. Cat- Aphididae) biotypes 1 and 2 on a susceptible and a re- egories of resistance at different growth stages in halt, a sistant wheat at three temperature regimes. J. Econ. En- winter wheat resistant to the Russian wheat aphid (Ho- tomol. 101: 955Ð958. moptera: Aphididae). J. Econ. Entomol. 96: 214Ð219. Selhorst, T. 1994. Inßuence of temperature on the fecundity Kieckhefer, R. W., and N. C. Elliott. 1989. Effect of ßuctu- of Rhopalosiphum padi. J. Plant Dis. Protect. 101: 654Ð661. ating temperatures on development of immature Russian Sengonca, C., A. Hoffmann, and B. Kleinhenz. 1994. Inves- wheat aphid (Homoptera, Aphididae) and demographic- tigations on development, survival and fertility of the statistics. J. Econ. Entomol. 82: 119Ð122. cereal aphids Sitobion avenae (F) and Rhopalosiphum Knight, J. D., J. S. Bale, F. Franks, S. F. Mathias, and J. G. padi (L) (Hom, Aphididae) at different low-tempera- Baust. 1986. Insect cold hardiness: supercooling points tures. J. Appl. Entomol. 117: 224Ð233. and prefreeze mortality. Cryo-Lett. 7: 194Ð203. Stary, P. 1999. Distribution and ecology of the Russian Logan, J. A., D. J. Wollkind, S. C. Hoyt, and L. K. Tanigoshi. wheat aphid, Diuraphis noxia (Kurdj.), expanded to cen- 1976. Analytic model for description of temperature-de- tral Europe (Hom.: Aphididae). J. Pest Sci. 72: 25Ð30.

pendent rate phenomena in arthropods. Environ. Ento- Turl, L.A.D. 1980. An approach to forecasting the incidence Downloaded from mol. 5: 1133Ð1140. of potato and cereal aphids in Scotland. Bull. OEPP Eur. Merrill, S. C., F. B. Peairs, H. R. Miller, T. L. Randolph, J. B. Mediterr. Plant Protect. Organ. 10: 135Ð141. Rudolph, and E. E. Talmich. 2008. Reproduction and Wadley, F. M. 1936. Development-temperature correlation development of Russian wheat aphid Biotype 2 on crested in the green bug, Toxoptera graminum. J. Agric. Res. 53: wheatgrass, intermediate wheatgrass, and susceptible and 0259Ð0266. resistant wheat. J. Econ. Entomol. 101: 541Ð545. Walgenbach, D. D., N. C. Elliott, and R. W. Kieckhefer. Messina, F. J. 1993. Winter mortality of the Russian wheat 1988. Constant and ßuctuating temperature effects on http://ee.oxfordjournals.org/ aphid (Homoptera, Aphididae) on range grasses in north- developmental rates and life table statistics of the Green- ern Utah. Great. Basin. Nat. 53: 310Ð313. bug (Homoptera, Aphididae). J. Econ. Entomol. 81: 501Ð Michaud, J. P., J. L. Jyoti, and J. A. Qureshi. 2006. Positive 507. correlation of Þtness with group size in two biotypes of Walters, K.F.A., and A. M. Dewar. 1986. Overwintering Russian wheat aphid (Homoptera: Aphididae). J. Econ. strategy and the timing of the spring migration of the Entomol. 99: 1214Ð1224. cereal aphids Sitobion avenae and Sitobion fragariae. Michels, G. J., and R. W. Behle. 1988. Reproduction and J. Appl. Ecol. 23: 905Ð915. development of Diuraphis-noxia (Homoptera, Aphidi- Webster, J. A., and K. J. Starks. 1987. Fecundity of Schizaphis dae) at constant temperatures. J. Econ. Entomol. 81: graminum and Diuraphis noxia (Homoptera, Aphididae) 1097Ð1101. at 3 temperature regimes. J. Kans. Entomol. Soc. 60: 580Ð by guest on March 16, 2016 Michels, G. J., and R. W. Behle. 1989. Inßuence of temper- 582. ature on reproduction, development, and intrinsic rate of Webster, J. A., R. L. Burton, and K. J. Starks. 1987. Diuraphis increase of Russian wheat aphid, Greenbug, and Bird noxia, a new United-States aphid pest of small grains. J. cherry-oat aphid (Homoptera, Aphididae). J. Econ. En- Kans. Entomol. Soc. 60: 483Ð483. tomol. 82: 439Ð444. Webster, J. A., F. Dutoit, and T. W. Popham. 1993. Fecun- Miller, H. R., T. L. Randolph, and F. B. Peairs. 2003. Cat- dity comparisons of the Russian wheat aphid (Ho- egories of resistance at four growth stages in three wheats moptera, Aphididae) in Bethlehem, South Africa, and in resistant to the Russian wheat aphid (Homoptera: Aphi- Stillwater, Oklahoma. J. Econ. Entomol. 86: 544Ð548. didae). J. Econ. Entomol. 96: 673Ð679. Webster, J. A., S. Amosson, L. Brooks, G. L. Hein, G. D. Morrison, W. P., and F. B. Peairs. 1998. Response model Johnson, D. E. Legg, W. Massey, P. Morrison, F. B. Peairs, concept and economic impact, pp. 1Ð11. In S. S. Quisen- and M. Weiss. 1994. Economic impact of the Russian berry and F. B. Peairs (eds.), Response model for an wheat aphid in the western United States: 1992Ð1993. introduced pest: the Russian wheat aphid. Thomas Say Russian Wheat Aphid Task Force to the Great Plains Publications in Entomology, Lanham, MD. Agricultural Council, Stillwater, OK. Nowierski, R. M., Z. Zeng, and A. L. Scharen. 1995. Age- Wyatt, I. J., and P. F. White. 1977. Simple estimation of speciÞc life table modeling of the Russian wheat aphid intrinsic increase rates for aphids and tetranychid mites. (Homoptera, Aphididae) on barley grown in benzimid- J. Appl. Ecol. 14: 757Ð766. azole agar. Environ. Entomol. 24: 1284Ð1290. Nyaanga, J. G., A. W. Kamau, and J. K. Wanjama. 2005. The inßuence of temperature on population increase and Received 3 April 2008; accepted 1 May 2009.