The Effect of Temperature and Host Plant Resistance on Population Growth of the Soybean Aphid Biotype 1 (Hemiptera: Aphididae)

Environmental Entomology Advance Access published December 26, 2016 Plant Insect Interactions Environmental Entomology, 2016, 1 10 doi: 10.1093/ee/nvw160 Research Article The Effect of Temperature and Host Plant Resistance on Population Growth of the Soybean Aphid Biotype 1 (Hemiptera: Aphididae) Ashley R. Hough, 1 James R. Nechols, 1,2 Brian P. McCornack, 1 David C. Margolies, 1 Brett K. Sandercock, 3 Donglin Yan, 4 and Leigh Murray 4 1 Department of Entomology, Kansas State University, 1603 Old Claflin Place, Manhattan, KS 66506 (ahough31@ksu.edu; jnechols@ksu.edu; mccornac@ksu.edu; dmargoli@ksu.edu), 2 Corresponding author, e-mail: jnechols@ksu.edu, 3 Division of Biology, Kansas State University, 116 Ackert Hall, 1717 Claflin Rd., Manhattan, KS 66506 (bsanderc@ksu.edu), and 4 Department of Statistics, Kansas State University, 101 Dickens Hall, 1116 Mid-Campus Dr. North, Manhattan, KS 66506 (donglinyan2@gmail.com; lmurray@ksu.edu) Subject Editor: Yasmin Cardoza Received 30 August 2016; Editorial decision 7 November 2016 Abstract A laboratory experiment was conducted to evaluate direct and indirect effects of temperature on demographic traits and population growth of biotype 1 of the soybean aphid, Aphis glycines Matsumura. Our objectives were to better understand how temperature influences the expression of host plant resistance, quantify the individual and interactive effects of plant resistance and temperature on soybean aphid population growth, and generate thermal constants for predicting temperature-dependent development on both susceptible and resistant soybeans. To assess indirect (plant-mediated) effects, soybean aphids were reared under a range of temperatures (15 30 C) on soybean seedlings from a line expressing a Rag1 gene for resistance, and life history traits were quantified and compared to those obtained for soybean aphids on a susceptible soybean line. Direct effects of temperature were obtained by comparing relative differences in the magnitude of life-history traits among temperatures on susceptible soybeans. We predicted that temperature and host plant resistance would have a combined, but asymmetrical, effect on soybean aphid fitness and population growth. Results showed that temperature and plant resistance influenced preimaginal development and survival, progeny produced, and adult longevity. There also appeared to be a complex interaction between temperature and plant resistance for survival and developmental rate. Evidence suggested that the level of plant resistance increased at higher, but not lower, temperature. Soybean aphids required about the same number of degree-days to develop on resistant and susceptible plants. Our results will be useful for making predictions of soybean aphid population growth on resistant plants under different seasonal temperatures. Key words: Aphis glycines, plant-mediated temperature effect, thermal stress Temperature influences population growth of insects in several ways, including determining rates of growth, development, and reproduction (Laudien 1973, Logan et al. 1976, Sharpe and Demichele 1977, Bauerfeind and Fischer 2013); regulating the timing and duration of dormancy in diapausing and nondiapausing species (Tauber et al. 1986); and by impacting survival (Precht 1973, Turnock et al. 1983, Chown and Nicolson 2004). In addition to direct effects, temperature may influence insect fitness indirectly by altering food abundance (Denno and McClure 1983) or food quality (Schalk et al. 1969, Tang et al. 1999). For insect herbivores, temperature effects on host plant quality are well documented and may result from changes in the production of primary products (e.g., sugars and amino acids) or secondary defense compounds (Went 1953, Maxwell and Jennings 1980). With respect to the use of host plant resistance, previous studies have shown that the expression of plant resistance is influenced by temperature, especially at lower or upper extremes, which may trigger genetic or physiological changes related to plant stress (Maxwell and Jennings 1980). In some cases, increases or decreases in temperature elevate the level of resistance (Thindwa and Teetes 1994), while in others resistance can be reduced or lost (Wood and Starks 1972, Salim and Saxena 1991, Harvey et al. 1994, Chen et al. 2014). Unfortunately, for most pest species, whether or how temperature affects host plant resistance is either unknown or information is incomplete. The relationship between temperature and developmental rate in insects is well known, and thermal constants have been derived for a large number of species including several plant pests (Logan et al. VC The Authors 2016. Published by Oxford University Press on behalf of Entomological Society of America. All rights reserved. For Permissions, please email: journals.permissions@oup.com 1

2 Environmental Entomology, 2016, Vol. 0, No. 0 1976). Having this information has enabled pest managers to develop accurate predictive models of pest occurrence and population growth based on degree-day accumulations (Pruess 1983). However, degree-day models generally are not available for predicting insect development on resistant plants or for comparing differences in temperature-dependent responses between resistant and susceptible plants. The soybean aphid, Aphis glycines Matsumura, is an invasive pest from Asia that has increased its distribution and pest status in the United States since it was first observed in 2000 (Hill et al. 2001, Tilmon et al. 2013). Host plant resistance is a management option for soybean aphid that continues to evolve as new aphid biotypes emerge (LaBarge 2011, Varenhorst et al. 2015). Most breeding programs for resistance to the soybean aphid employ genes in the rag family. For biotype 1, which is one of the most widely distributed soybean aphid biotypes, the Rag1 gene has been shown to confer resistance to the soybean aphid based on a combination of antibiosis and antixenosis (Diaz-Montano et al. 2006, 2007; Hill et al. 2012). Limited work has examined the effect of temperature on Rag1 resistance to soybean aphid biotype 1, but data from two studies suggest that resistance may weaken at lower temperature (Richardson 2011, Chirumamilla et al. 2014). Direct effects of temperature on soybean aphid biology have also been studied. Both Hirano et al. (1996) and McCornack et al. (2004) evaluated biotype 1 on a susceptible soybean cultivar, comparing development, survival, reproduction, and the intrinsic rate of population increase for different but overlapping ranges of temperature. They also estimated the lower thermal threshold for development, but did not generate a thermal (degreeday) constant. Neither study included a resistant soybean variety as an experimental treatment. To clarify the suggested relationship between temperature and soybean resistance involving the Rag1 gene, we conducted an experiment with biotype 1 soybean aphids to address four objectives: 1) evaluate the direct effect of temperature on soybean aphid fitness using a different susceptible variety than in previous studies; 2) investigate possible changes in the level of Rag1 resistance relative to susceptible plants at both lower and higher temperatures; 3) generate lower thermal threshold and degree-day constants for soybean aphids on resistant and susceptible soybeans; and 4) test the prediction that rates of temperature-dependent aphid development not only will differ between plant types, but will vary in a nonuniform manner over the range of temperatures tested. Our overall goal was to generate data that could be used to develop a more accurate and comprehensive predictive model of soybean aphid population growth and pest risk under the thermally dynamic conditions experienced by this pest in the Midwest. To address our goal and objectives, we measured four demographic responses development, survival, fecundity, and adult longevity and calculated thermal constants and life table statistics. Materials and Methods Plant and Insect Cultures Plants and insects used in experiments were maintained inside an environmental growth chamber at Kansas State University under 25.6 6 0.1 C, 75.6 6 1.6% relative humidity, a photoperiod of 16:8 (L:D) h, and a light intensity of 273.1 6 33.0 W/m 2. Seedlings of two soybean lines were used as host plants: one line is known to be susceptible to soybean aphid (SD01-76R) and one line contains the Rag1 gene for soybean aphid resistance (LD(05)-16060). Rag1 acts as a single dominant gene. The Rag1-resistant soybean line we used was developed by others by crossing Loda with Dowling (carrying the Rag1 allele) and then making three crosses to the susceptible line we tested (SD01-76R). The resistant line was a BC 2 F 2 -derived line that carries the Rag1 allele (Hesler et al. 2013). Therefore, the resistant and susceptible soybean lines used in our experiments had a very similar genetic background except for the Rag1 gene. Plants were grown from seed in pots (7.6 by 10.2 cm [H by D]) containing Metro-Mix 360 (Sun Gro Horticulture, Bellevue, WA). Four seeds were sown per pot and after germination we thinned each pot, so it contained only one healthy seedling. Plant propagation was done for 7 to 9 d, a period that was sufficient for seedlings to reach the V-0 stage. This stage is reached when plants are 5 cm tall and have cotyledons that are cupped around the unifoliate leaves (McCornack et al. 2004). During propagation and throughout the experiment, both susceptible and resistant seedlings were checked daily; individual plants were watered when the soil surface began to dry, but when the underlying soil layer was still slightly moist to the touch at a depth of about 1.5 cm. Biotype 1 soybean aphids were obtained from samples collected from soybean fields in Nebraska in July 2008, from which a colony was established on a susceptible soybean cultivar (KS4202), in environmental growth chambers ranging from 20 to 30 C. During the experiment (March through November, 2014), the soybean aphid colony was maintained in a sealed cylinder cage on the susceptible soybean, SD01-76R, in the same growth chamber and environmental conditions described above for the plant colonies. Specimens used in this research are deposited as voucher number 236 in the KSU Museum of Entomological and Prairie Arthropod Research. Experimental Procedures The experiment was repeated four times between 14 March 2014 and 6 November 2014. In each replicate of the experiment, potted seedlings from the resistant and susceptible (control) soybean lines were inoculated by transferring two adult soybean aphids from the colony to each seedling and leaving them for 24 h after which the adults and all but one nymph were removed. Aphids were transferred using a fine camel hair paintbrush (#000). Equal numbers of each type of seedling (resistant and susceptible) were then placed in four environmental growth chambers (Percival Scientific, Inc., Model I36VLC8, Perry, IA, or Conviron, Model A1000, Winnipeg, MB) assigned to either 15, 20, 25, or 30 C. The actual temperatures measured in the growth chambers were 15.3 6 0.02, 20.7 6 0.05, 25.3 6 0.02, and 30.0 C 6 0.03 and average light intensity was 83.3 6 9.9 W/m 2 for the Percival chambers and 341.0 6 17.1 W/m 2 for the Conviron chambers. Thereafter seedlings were inspected daily to determine if nymphs were alive and, if so, at which life stage. It was difficult to distinguish among individual soybean aphid instars, so three age classes were used: 1st 2nd, 3rd 4th, and adult. To maintain consistently high plant quality, every three days the soybean seedlings in each temperature treatment were replaced with new V-0 stage seedlings by transferring individual soybean aphids from the old to new seedlings. When soybean aphids began to produce progeny, the number of offspring per reproductive was recorded and removed from the seedling daily. Monitoring was continued until the adult soybean aphids died. Thus, we collected data on four demographic parameters: preimaginal survival, development time of soybean aphids that survived from 1st instar nymph to adult emergence, total number of progeny per adult, and adult longevity of individual soybean aphids.

Environmental Entomology, 2016, Vol. 0, No. 0 3 Experimental Design and Statistical Analyses The experiment was arranged in a Latin Square Design (LSD) with a split-plot and replication on the split-plot. Temperature was the whole-plot treatment factor (four levels: 15, 20, 25, and 30 C) and the LSD row and column factors were growth chamber and time repeat, each with four levels. Plant type (two levels: resistant and susceptible) was the split-plot treatment factor with 20 pots per plant type, thus giving a total of 40 pots per growth chamber for a total of 160 pots. In trial 1, contamination by lacewing larvae eliminated soybean aphids from several susceptible plants early in the experiment. Missing samples were replaced with an excess of three plants to ensure that the minimum number was maintained, which resulted in 23 (vs. 20) replications for a total of 163 pots for the entire experiment. The slight imbalance did not affect the statistical analysis. Temperature and soybean plant type were considered fixed effects, whereas chamber and replicate were considered as random effects. Based on results of a previous study, location within growth chamber (shelf) was ignored in the analysis. Numerical responses (rate and duration of soybean aphid development, preimaginal development time from 1st instar nymph to adult emergence, total number of progeny per adult, and adult longevity) were analyzed using the MIXED procedure in SAS, V9.4 (PROC MIXED, SAS Institute 2013) with the Restricted Maximum Likelihood (REML) method of model fitting. Residuals were evaluated and found to be somewhat nonnormal but symmetric. Therefore, given the large sample size, our results based on the normal distribution are considered valid. The proportions of aphids surviving to adult emergence were analyzed using the GLIMMIX procedure (PROC GLIMMIX V9.4, SAS Institute 2013) with the default pseudo-likelihood method, using the binomial distribution and the logit function. The GLIMMIX s default pseudo-likelihood method of model fitting gives the same results as the MIXED REML method when the normal distribution is being used. Results from MIXED and GLIMMIX analyses included REML estimates and approximate Wald test statistics for the random effects variance components, Type III F-tests to test fixed effects of the temperature and conditioning main effects and their interaction, lsmeans and standard errors for the fixed effects, and pairwise comparisons for a significant temperature effect using unadjusted (LSD) P-values at P 0.05 level. In addition, because of the interest in knowing whether differences in soybean aphid responses between the resistant and susceptible plant types were influenced by temperature, simple effects (using the slice option) were used to test the difference between the two plant types at each temperature. The variance components for both growth chambers and time repeat were never significant, indicating that variation in aphid responses did not change due to differences in growth chambers or the experiment being run four times over the course of several months. Developmental times of soybean aphid for each plant type, and at each of the four temperatures, were converted to rates (1/d). Developmental rates (four per temperature corresponding to the four replicates) were then regressed against their respective temperature using a least squares linear regression procedure (http://www. alcula.com/calculators/statistics/linear-regression/, accessed 18 November 2016). For each plant type, a linear curve was produced (r ¼ 0.90 and 0.91 for resistant and susceptible plants, respectively) and the lower thermal threshold temperature (t) was estimated at the point where the line crossed the X-axis. The degree-day constant (K) for aphids on each plant type were computed using the standard formula from Logan et al. (1976): K ¼ dðt tþ where K is the degree-day constant, d is the number of days for preimaginal development (averaged over the four replicates), T is the growth chamber temperature, and t is the lower thermal threshold temperature. Means and standard errors were generated using the K values computed at each of the four temperatures. To determine whether the degree-day constants for soybean aphids were significantly different when they were reared on resistant versus susceptible plants, the individual K values obtained for each temperature were used as a source of variation and a two-tailed t-test was run on the means and variances using the GraphPad Software and the QuickCalcs automatic t-test (http://www.graphpad.com/quickcalcs/ ttest1.cfm). Demographic response data were used to generate three life table statistics using the popbio package of Program R (R i386 3.1.2): finite rate of population change (k), net reproductive rate (R o ), and mean generation time (T). The input data consisted of the age, p x (age-specific survival rate from age x to age x þ 1), m x (number of female offspring per female of age x), and the standard error for both values. The output consisted of the mean, standard deviation, and confidence intervals for each set of data to be tested. Program R was also used to calculate P-values to separate treatment using bootstrap values for each value. Bootstrap estimate were based on a minimum of 100 iterations. This particular method was chosen over analysis of variance because it is based on resampling the data and no asymptotic assumptions were required. Standard errors for the transition rates (p) were then calculated based on the sample sizes of soybean aphids (n) that survived to adulthood where SE(p) ¼ p(1 p)/n. The precision of the estimates was maximized by pooling trials for the largest possible sample size. The variance terms of the demographic parameters were then generated by resampling the original data via bootstrapping. Mean and SEM values for all life table statistics were plotted to compare treatment differences using SigmaPlot (Systat Software, Inc., San Jose, CA). In addition, we plotted the mean schedules of survivorship and reproduction for aphids exposed to different temperatures and plant types. Results Development Temperature (F ¼ 24.22; df ¼ 3, 6; P < 0.0009) and plant type (F ¼ 8.67; df ¼ 1, 12; P < 0.0005) both had a highly significant effect on the rate of soybean aphid development, but the interaction was not significant (F ¼ 2.11; df ¼ 3, 12; P < 0.15), suggesting that the relative difference in developmental rate across temperatures was the same for both plant types. However, the simple effect tests of differences between plant types at each temperature showed significant differences in development rate between susceptible and resistant soybeans at 25 and 30 C(P < 0.0003 and P < 0.03, respectively), but not at the two lower temperatures (Fig. 1). For both susceptible and resistant soybeans, a direct relationship was present between temperature and the rate of soybean aphid development. Temperature and plant type also significantly affected the number of days required for soybean aphids to develop from the first nymphal stage to adult (temperature: F ¼ 105.2; df ¼ 3, 6; P < 0.0001; plant type: F ¼ 8.67; df ¼ 1, 12; P < 0.012). However, no significant temperature by plant type interaction was observed (F ¼ 0.41; df ¼ 3, 12; P < 0.75), indicating that the relative effect of temperature on development was the same for both plant types. Development time became progressively shorter as temperature increased, but differences were only significant between 15 and

4 Environmental Entomology, 2016, Vol. 0, No. 0 Fig. 1. Mean 6 SE rate of development for biotype 1 soybean aphid nymphs to reach adulthood on susceptible or Rag1-resistant soybean seedlings under 15, 20, 25, or 30 C. Range of observations (low high) for the four trials for each treatment combination: 15 C: susceptible: 7 13, resistant: 1 6, 20 C: susceptible: 14 15, resistant: 6 9, 25 C: susceptible: 15 19, resistant: 8 10, and 30 C: susceptible: 16 19, resistant: 2 4. Temperatures with different capital letters between plant types indicate a significant difference in development rate (Simple effect tests, P < 0.05). Fig. 2. Mean 6 SE number of days required for biotype 1 soybean aphid nymphs to reach adulthood on susceptible or Rag1-resistant soybean seedlings under 15, 20, 25, or 30 C. Range of observations (low high) for the four trials for each treatment combination: 15 C: susceptible: 7 13, resistant: 1 6, 20 C: susceptible: 14 15, resistant: 6 9, 25 C: susceptible: 15 19, resistant: 8 10, and 30 C: susceptible: 16 19, resistant: 2 4. Capital letters denote statistical difference among temperatures with data pooled for the two plant types (Temperature main effect, P < 0.05). Lower case letters denote statistical difference between plant types within each temperature (Simple effect tests, P < 0.05). 25 C (Fig. 2). Soybean aphid development was longer by 1 d among temperatures on resistant versus susceptible soybeans, but the difference was only significant at 25 C(F ¼ 6.75; df ¼ 1, 12; P < 0.02). The mean number of days to develop from first-instar nymph to adult ranged from 5 to 12 d on susceptible and resistant plants. Survival Temperature (F ¼ 12.51; df ¼ 3, 6; P < 0.0054) and plant type (F ¼ 117.91; df ¼ 1, 12; P < 0.0001) had a significant effect on soybean aphid survival. A significant temperature by plant type interaction was also observed (F ¼ 5.26; df ¼ 3, 12; P < 0.0151). At all temperatures, percentage survival was significantly higher on susceptible plants than on resistant plants (Fig. 3). On average, soybean aphid survival was 44% higher on susceptible plants than on resistant plants. On both plant types, soybean aphid survival increased between 15 and 25 C(Fig. 3). However, whereas survival on resistant soybeans decreased significantly between 25 and 30 C, on susceptible soybeans survival was uniformly high at 25 and 30 C and the difference between the two temperatures was not significant (Fig. 3). Lifetime Progeny Production Plant type (F ¼ 140.21; df ¼ 1, 12; P < 0.0001) and temperature (F ¼ 6.08; df ¼ 3, 6; P < 0.0299) both had a significant effect on the number of soybean aphid progeny produced, but the plant type by temperature interaction was not significant (F ¼ 2.66; df ¼ 3, 12; P < 0.0957). The number of progeny produced on resistant soybean plants was consistently and significantly much lower than on susceptible plants. On both plant types, there was a trend for the fewest progeny at 30 C although the difference was not statistically significant from 15 C(Fig. 4). The pattern of progeny production among temperatures also appeared to differ depending on plant type. On resistant soybeans, the mean number of progeny produced was similar (17 18 nymphs) between 15 and 25 C but about 8-fold lower (2) at 30 C, whereas on

Environmental Entomology, 2016, Vol. 0, No. 0 5 Fig. 3. Mean 6 SE proportion of biotype 1 soybean aphid nymphs surviving to adulthood on susceptible or Rag1-resistant soybean seedlings under 15, 20, 25, or 30 C. Range of observations (low high) for the four trials for each treatment combination: 15 C: susceptible: 7 13, resistant: 1 6, 20 C: susceptible: 14 15, resistant: 6 9, 25 C: susceptible: 15 19, resistant: 8 10, and 30 C: susceptible: 16 19, resistant: 2 4. Lower case letters denote statistical difference among temperatures within plant types (Simple effect tests, P 0.002). Capital letters denote statistical difference between plant types within each temperature (Simple effect tests, P < 0.05). Fig. 4. Mean 6 SE number of progeny produced in a lifetime by biotype 1 soybean aphid adults on susceptible or Rag1-resistant soybean seedlings under 15, 20, 25, or 30 C. Range of observations (low high) for the four trials for each treatment combination: 15 C: susceptible: 7 13, resistant: 1 6, 20 C: susceptible: 14 15, resistant: 6 9, 25 C: susceptible: 15 19, resistant: 8 10, and 30 C: susceptible: 16 19, resistant: 2 4. Capital letters denote statistical difference among temperatures with data pooled for the two plant types (Temperature main effect, P < 0.05). Lower case letters denote statistical difference between plant types within each temperature (Simple effect tests, P < 0.05). susceptible plants, the trend in number of progeny was highest at the two middle temperatures (20 and 25 C; Fig. 4). Adult Longevity Temperature had a significant effect on the length of adult life of soybean aphids (F ¼ 8.84; df ¼ 3, 6; P < 0.0127), as did plant type (F ¼ 45.39; df ¼ 1, 12; P < 0.0001). However, the interaction between temperature and plant type was not significant (F ¼ 1.03; df ¼ 3, 12; P < 0.4155). A trend for adult longevity to be reduced as temperature increased was apparent. However, the only significant difference was between 15 and 30 C(Fig. 5). Adult life spans were significantly shorter at all temperatures on resistant soybean seedlings, with mean longevity ranging from 3 to 14 d on resistant plants and from 8 to 21 d on susceptible plants. The differences in soybean aphid longevity between susceptible and resistant plants, averaged over all temperatures, was 8 d. Thermal Constants The number of degree-days (DD) for soybean aphids to complete development on resistant (61.6 6 14.6) and susceptible (62.8 6 10.0) soybean seedlings was not significantly different (t ¼ 0.07; df ¼ 6; P < 0.95). The lower thermal threshold for development was slightly higher on resistant plants (13 C) than on susceptible plants (11.75 C). Life Table Statistics Finite rates of population change (k) were positive on both resistant and susceptible soybeans at all four temperatures (Fig. 6). However, on resistant soybeans, population growth was consistently lower (significantly so at the three highest temperatures), and there were smaller differences among temperatures, than on susceptible plants. Differences were highly significant at all temperatures (P < 0.001), except for 15 C where no significant difference was observed between plant types (Fig. 6). On both plant types there were slight

6 Environmental Entomology, 2016, Vol. 0, No. 0 Fig. 5. Mean 6 SE number of days biotype 1 soybean aphids lived after reaching adulthood on susceptible or Rag1-resistant soybean seedlings under 15, 20, 25, or 30 C. Range of observations (low high) for the four trials for each treatment combination: 15 C: susceptible: 7 13, resistant: 1 6, 20 C: susceptible: 14 15, resistant: 6 9, 25 C: susceptible: 15 19, resistant: 8 10, and 30 C: susceptible: 16 19, resistant: 2 4. Capital letters denote statistical difference among temperatures with data pooled for the two plant types (Temperature main effect, P < 0.05). Lower case letters denote statistical difference between plant types within each temperature (Simple effect tests, P < 0.05). increases in k values between 15 and 25 C, but then a decrease at 30 C(Fig. 6). The optimal temperature for soybean aphid population growth on both plant types was 25 C. The net reproductive rates (R o ) of soybean aphid on susceptible plants were all positive, ranging from 20 to 60 females per female per generation; they were highest at 20 and 25 C(Fig. 6). In contrast, on resistant plants mean R o values were consistently low (< 10 females per female), and at 30 C the net reproductive rate was essentially zero. R o values were significantly higher on susceptible soybeans compared to resistant soybeans at all temperatures (P < 0.001; Fig. 6). The generation time (T) of soybean aphid declined with increasing temperature on both susceptible and resistant plants. Mean generation times ranged from 7 to 14 d. As temperature increased, the decline in mean generation time occurred at a lower rate on susceptible plants than on resistant plants. Differences in T between resistant and susceptible soybeans were significant, but only at 20 and 25 C(Fig. 6). Survivorship and Reproductive Schedules A comparison of lifetime survivorship curves indicated that soybean aphids died sooner at all temperatures on resistant soybeans than on susceptible soybeans (Fig. 7A). On resistant plants, soybean aphid survivorship decreased sharply within the first week at all temperatures, reaching 50% survivorship between 2 and 7 d, whereas on susceptible plants, the decrease in survivorship was more gradual, with 50% survivorship occurring between 10 and 25 d, depending on temperature. A comparison of survivorship among temperatures revealed two patterns. First, survivorship was lowest on both plant types at 30 C(Fig. 8A and B). Second, beginning at about 21 d there was a wider separation in soybean aphid survivorship among temperatures on the susceptible line compared to the resistant line. Specifically, on susceptible plants survivorship was inversely related to temperature (Fig. 8A). In contrast, survivorship curves on resistant plants were similar and overlapping between 15 and 25 C and markedly lower only at 30 C(Fig. 8B). A comparison of reproductive schedules revealed apparent differences between resistant and susceptible soybeans (Fig. 7B) and among temperatures (Fig. 8C and D). At 15 C, both the magnitude and schedule of reproduction were similar, whereas at higher temperatures mean daily progeny produced was generally higher on susceptible soybeans than on resistant soybeans, and the differences appeared to increase with increasing temperature (Fig. 7B). In addition, progeny production lasted longer on susceptible plants compared to resistant plants. On both plant types, the reproductive period was longest at the lower two temperatures and became shorter as temperature increased (Figs. 7B and 8C and D). Discussion Soybean aphids responded to resistant soybean plants with the Rag1 gene as expected, with lower rates of population growth observed at all temperatures compared to susceptible soybeans. Our study provides the first experimental evidence for an interaction between host plant resistance and temperature on demographic performance of soybean aphid. Reduced population growth on resistant plants appeared to be most strongly linked with lower preadult survival of soybean aphid but also was likely influenced by fewer progeny associated with shorter adult life spans, and slower developmental times. Differences in development time between resistant and susceptible plants were not statistically significant at all temperatures, but probably contributed to differences in rates of population change. Overall, our findings indicate that host plant resistance had a depressive effect on multiple aspects of soybean aphid life history. Responses to temperature differed between plant types and among life history traits. With the exception of progeny production, which peaked in the middle range of temperatures (20 and 25 C), the other responses showed a linear relationship with temperature, but only between 15 and 25 C. Within this range, as temperature increased development was faster and survival higher, but adult longevity was shorter. The life table parameters derived from the four demographic responses also varied with temperature. Finite rates of increase and net reproductive rates increased with a peak at 25 C, then decreased at 30 C; mean generation times progressively decreased with increasing temperature. However, for all parameters larger differences were observed among temperatures on susceptible soybeans than on resistant soybeans. We attribute the lower mean values and reduced variation among temperatures to effects of plant resistance. Specifically, suppressive effects of Rag1 resistance on soybean aphid likely obscured differences in life table parameters resulting from different temperatures.

Environmental Entomology, 2016, Vol. 0, No. 0 7 Fig. 6. Demographic statistics for biotype 1 soybean aphids on susceptible or Rag1-resistant soybean seedlings under 15, 20, 25, or 30 C. For finite rates of change and net reproductive rates, means above dashed line indicate an increase; below the line represents a decrease. Mean generation times (days). Number of observations: finite rates of change for both plant types and all temperatures, 80 83; net reproductive rates: susceptible, 15 C, 38; 20 C, 58; 25 C, 69; 30 C, 68; resistant, 15 C, 14; 20 C, 28; 25 C, 36; 30 C, 12; mean generation times: susceptible, 15 C, 38; 20 C, 58; 25 C, 69; 30 C, 68; resistant, 15 C, 14; 20 C, 28; 25 C, 36; 30 C, 12. Within temperatures, conditioning times separated by *, **, or *** are significant at the P < 0.05, 0.01, and 0.001 level, respectively. Population growth of soybean aphid was greatest on both plant types at 25 C and lowest at 30 C, findings that are consistent with McCornack et al. (2004) and Hirano et al. (1996) who studied effects of temperature on biotype 1 soybean aphid that were reared on susceptible soybean lines. The combined experimental data support evidence from field observations that soybean aphids are notwell-adaptedtohightemperature.however,whenweexamined the response to temperature for individual traits, high temperature had either a positive or negative effect depending on the trait. For example, adult longevity and progeny production were lower on resistant and susceptible plants at 30 Cthanat25 C, suggesting that soybean aphid adults were affected directly and negatively by increased temperature. In contrast, survival to adult was strongly affected by the type of plant on which soybean aphids were reared. Specifically, whereas survival was uniformly high at both 25 and 30 C on susceptible soybeans, on resistant plants a large drop in survival occurred between 25 and 30 C. Our findings suggest that reduced survival of soybean aphids at 30 C was not a result of direct high temperature stress on the aphids; but, rather, an indirect effect of factors that enhanced host plant resistance. We also conclude that soybean aphid population growth is determined by a complex interaction involving direct and indirect effects of temperature as well as the type of host plant on which this pest occurs.

8 Environmental Entomology, 2016, Vol. 0, No. 0 Fig.7. Soybean aphid survivorship (A) and mean number of progeny produced (B) on Rag1-resistant and susceptible soybeans at 15, 20, 25, and 30 C.

Environmental Entomology, 2016, Vol. 0, No. 0 9 Fig. 8. Comparison of soybean aphid survivorship (A and B) and mean number of progeny produced (C and D) among temperatures (15, 20, 25, 30 C) on susceptible and Rag1-resistant plants, respectively. The enhanced level of host plant resistance to the soybeans aphid that we observed at higher temperature has been documented for other insects. For example, in the greenbug, Schizaphis graminum, Thindwa and Teetes (1994) found a decrease in population growth at 26 and 30 C compared to 21 C associated with slower preimaginal development times, lower fecundity, and reduced adult longevity. In contrast, resistance in wheat to the Hessian fly, Mayetiola destructor, broke down under higher temperatures (Sosa and Foster 1976, Tyler and Hatchett 1983, Chen et al. 2014). In separate studies with biotype 1 soybean aphids, Richardson (2011) and Chirumamilla et al. (2014) found no difference in population growth on Rag1-resistant and susceptible soybeans at moderate (21 C) versus relatively high (28 C) temperatures. However, our experiment was conducted at a higher temperature (30 C) than previous studies. Therefore, it is possible that some or all Rag1-resistant soybean lines respond only when temperatures have reached a critical upper threshold. Low temperatures can also affect the expression of plant resistance, in some cases strengthening it (Chen et al. 2014), in others causing resistance to break down (Wood and Starks 1972, Schweissing and Wilde 1979, Harvey et al. 1994). In soybean aphid, Richardson (2011) and Chirumamilla et al. (2014) observed a smaller difference in population numbers between resistant and susceptible soybeans at 14 C compared to 21 C and concluded that Rag1-resistance was reduced or lost at lower temperature. When we compared individual demographic responses of soybean aphids between resistant and susceptible plants, no consistent pattern existed to suggest that resistance either increased or decreased at our lowest temperature (15 C). However, the difference in the finite rate of population change between resistant and susceptible soybeans was much smaller at 15 C(Dk¼ 0.09) than at the higher temperatures (range: 0.42 0.62), suggesting that the cumulative effect of low temperature on Rag1-resistant soybeans may have been to reduce the overall expression of resistance to soybean aphid. Soybean aphid development on resistant soybeans was slightly longer than on susceptible plants. However, differences in development on the two plant types were not uniform across temperatures and were significant only at the two highest temperatures. These results suggest that relative differences in population growth rates on resistant and susceptible plants could vary over the range of ecological temperatures that soybean aphids experience in the field. On the other hand, when development was expressed as degree-days, there was no statistical difference between resistant and susceptible plants. Moreover, the estimated lower thermal threshold temperatures were similar for the two plant types, although about a degree lower on susceptible plants. A comparison of thermal constants for soybean aphids on the susceptible soybean line we tested with those derived from data published by McCornack et al. (2004) for a different susceptible soybean line revealed some differences and some similarities. Using their own data, and combining it with data published by Hirano et al. (1996), McCornack et al. did a linear regression of development rate against temperature and generated an estimate for the lower thermal threshold, but not a degree-day constant. We computed a degree-day constant from their combined data, which resulted in a mean (6 SE) of 68.4 6 4.0 DD, which is close to ours (62.8 6 10.0) and is not statistically significant (two-tailed t-test: P ¼ 0.59, t ¼ 0.56, df ¼ 7). This finding is not unexpected since, in at least some insects, thermal constants tend to be conserved geographically (Tauber et al. 1987). On the other hand, there was an apparent difference in lower thermal

10 Environmental Entomology, 2016, Vol. 0, No. 0 thresholds (8.6 C for the McCornack-Hirano data vs. 11.75 Cinour study). This difference was greater than the difference in threshold temperatures we found between susceptible and resistant soybeans. Based on our current data, soybean aphids could begin developing sooner in other locations compared to Kansas. However, without a statistical comparison, any apparent differences in threshold temperatures geographically are inconclusive. In terms of pest management applications, our study expands existing knowledge about how temperature affects host plant resistance for the soybean aphid. Specifically, we were able to show that resistance increases under higher temperature, and that reduced survival associated with plant resistance may interact with direct negative effects of high temperature on other soybean aphid traits to produce a larger combined suppressive impact on population growth than either factor by itself. A better understanding of the interactive effects of temperature and plant resistance is of practical value for soybean producers and pest managers who are interested in predicting crop risk under different thermal and host plant conditions. In addition, the thermal constants now available for both resistant and susceptible soybeans will allow producers to better predict soybean aphid seasonal occurrence and population growth regardless of whether the growers are planting a resistant or susceptible soybean variety. However, we recognize that thermal responses may differ on new soybean lines, especially cultivars with resistant properties. Acknowledgments We thank Xiaoli Wu for technical assistance, Ming Chen and William Schapaugh for the use of their environmental growth chambers, and John Ruberson for reviewing an earlier draft of this manuscript. This is contribution number 16-379-J from the Kansas Agricultural Experiment Station. References Cited Bauerfeind, S. S., and K. Fischer. 2013. Testing the plant stress hypothesis: stressed plants offer better food to an insect herbivore. Entomol. Exp. Appl. 149: 148 158. Chen, M. S., S. Wheeler, W. Haiyan, and J. R. Whitworth. 2014. Impact of temperatures on Hessian fly (Diptera: Cecidomyiidae) resistance in selected wheat cultivars (Poales: Poaceae) in the Great Plains region. J. Econ. Entomol. 107: 1266 1273. Chirumamilla, A., C. B. Hill, and G. L. Hartman. 2014. Stability of resistance in soybean across different temperatures. Crop Sci. 54: 2557 2563. Chown S.L., and S. W. Nicolson. 2004. Insect physiological ecology: mechanisms and patterns. Oxford University Press, Oxford, United Kingdom. Denno, J., and M. S. McClure. 1983. Variable plants and herbivores in natural and managed systems. Academic Press, New York, NY. Diaz-Montano, J., J. C. Reese, W. T. Schapaugh, and L. R. Campbell. 2006. Characterization of antibiosis and antixenosis to the (Hemiptera : Aphididae) in several soybean genotypes. J. Econ. Entomol. 99: 1884 1889. Diaz-Montano, J., J. C. Reese, J. Louis, L. R. Campbell, and W. T. Schapaugh. 2007. Feeding behavior by the (Hemiptera: Aphididae) on resistant and susceptible soybean genotypes. Entomol. Soc. Am. 100: 984 989. Harvey, T. L., G. E. Wilde, K. D. Kofoid, and P. J. Bramel-Cox. 1994. Temperature effects on resistance to greenbug (Homoptera: Aphididae) biotype 1 in sorghum. J. Econ. Entomol. 87: 500 503. Hesler,L.S.,M.V.Chiozza,M.E.O Neal,G.C.MacIntosh,K.J.Tilmon, D. I. Chandrasena, N. A. Tinsley, S. R. Cianzio, A. C. Costamagna, E. M. Cullen, et al. 2013. Performance and prospects of Rag genes for management of soybean aphid. Entomol. Exp. Appl. 147: 201 216. Hill, J. H., R. Alleman, D. B. Hogg, and C. R. Grau. 2001. First transmission of Soybean mosaic virus by Aphis glycines in the New World. Plant Dis. 85: 561. Hill, C. B., A. Chirumamilla, and G. L. Hartman. 2012. Resistance and virulence in the soybean-aphis glycines interaction. Euphytica 186: 635 646. Hirano, K., K. Honda, and S. Miyai. 1996. Effects of temperature on development, longevity, and reproduction of the, Aphis glycines (Homoptera; Aphididae). Appl. Entomol. Zool. 31: 178 180. LaBarge, G. 2011. Be cautious with resistance genes in Ohio. Ohio State University Extension. (http://maumeeag.osu.edu/2011/03/, accessed 18 November 2016). Laudien, H.1973. Changing reaction systems, pp. 355 372. In H. Precht, J. Christophersen, H. Hensel and W. Larcher (eds.), Temperature and life. Springer-Verlag. Berlin, Germany. Logan, J. A., D. J. Wollkind, S. C. Hoyt, and S. K. Tanigoshi. 1976. An analytical model for description of temperature dependent rate phenomena in arthropods. Environ. Entomol. 5: 1133 1140. Maxwell, F. G., and P. R. Jennings. 1980. Breeding plants resistant to insects. Wiley-Interscience. New York, NY. McCornack, B. P., D. W. Ragsdale, and R. C. Venette. 2004. Demography of (Homoptera : Aphididae) at summer temperatures. J. Econ. Entomol. 97: 854 861. Precht, H. 1973. Limiting temperatures and life functions, pp. 400 440. In H., Precht, J. Christophersen, H. Hensel, and W. Larcher (eds.), Temperature and life. Springer-Verlag, Berlin, Germany. Pruess, K. P. 1983. Day-degree methods for pest management. Environ. Entomol. 12: 613 619. Richardson, M. L. 2011. Temperature influences the expression of resistance of soybean (Glycine max) to the (Aphis glycines). J. Appl. Entomol. 136: 641 645. Salim, M., and R. C. Saxena. 1991. Temperature stress and varietal resistance in rice effects on whitebacked planthopper. Crop Sci. 31: 1620 1625. SAS Institute. 2013. SAS Institute Inc. Version 9.4. SAS Institute, Cary, NC. Schalk, J. M., S. D. Kindler, and G. R. Manglitz. 1969. Temperature and preference of the spotted alfalfa aphid for resistant and susceptible alfalfa plants. J. Econ. Entomol. 62: 1000 1003. Schweissing, F. C., and G. Wilde. 1979. Temperature and plant nutrient effects on resistance of seedling sorghum to the greenbug. J. Econ. Entomol. 72: 20 23. Sharpe, P. J. H., and D. W. Demichele. 1977. Reaction kinetics of poikilotherm development. J. Theor. Dev. 64: 649 670. Sosa, O., and J. E. Jr Foster. 1976. Temperature and expression of resistance in wheat to the Hessian fly. Environ. Entomol. 5: 333 336. Tang, Y. Q., S. I. LaPointe, L. G. Brown, and W. B. Hunter. 1999. Effects of host plant and temperature on the biology of Toxoptera citricida (Homoptera: Aphididae). Environ. Entomol. 28: 895 900. Tauber, C. A., M. J. Tauber, and J. R. Nechols. 1987. Thermal requirements for development in Chrysopa oculata: A geographically stable trait. Ecology 68: 1479 1487. Tauber, M. J., C. A. Tauber, and S. Masaki. 1986. Seasonal adaptations of insects. Oxford University Press, New York, NY. Thindwa, H. P., and G. L. Teetes. 1994. Effect of temperature and photoperiod on sorghum resistance to biotype C and E greenbug (Homoptera: Aphididae). J. Econ. Entomol. 87: 1366 1372. Tilmon, K., E. Hodgson, M. O Neal, and D. Ragsdale. 2013. NPIPM: Aphis glycines (soybean). Bugwood Wiki. (http://wiki.bugwood.org/ NPIPM:Soybean_aphid #Host_Plant_Resistance, accessed 18 November 2016). Turnock, W. J., R. J. Lamb, and R. P. Bodnaryk. 1983. Effects of cold stress during pupal diapause on the survival and development of Mamestra configurata (Lepidoptera: Noctuidae). Oecologia 56: 185 192. Tyler, J. M., and J. H. Hatchett. 1983. Temperature influence on expression of resistance to Hessian fly (Diptera: Cecidomyiidae) in wheat derived from Triticum tauschii. J. Econ. Entomol. 76: 323 326. Varenhorst, A. J., M. T. McCarville, and M. E. O Neal. 2015. Reduced fitness of virulent Aphis glycines (Hemiptera: Aphididae) biotypes may influence the longevity of resistance genes in soybean. PLoS ONE. 10: 1 14, http://dx. doi.org/10.1371/journal.pone.0138252. Went, F. W. 1953. The effect of temperature on plant growth. Annu. Rev. Plant Physiol. 4: 347 362. Wood, E., and K. Starks. 1972. Effect of temperature and host plant interaction on the biology of three biotypes of the greenbug. Environ. Entomol. 1: 230 234.