Estimation of Feeding Threshold for Homalodisca Vitripennis (Hemiptera: Cicadellidae) and Its Application to Prediction of Overwintering Mortality

PEST MANAGEMENT Estimation of Feeding Threshold for Homalodisca vitripennis (Hemiptera: Cicadellidae) and Its Application to Prediction of Overwintering Mortality

YOUNGSOO SON,1,2 RUSSELL L. GROVES,3,4 KENT M. DAANE,5 DAVID J. W. MORGAN,6 3 1,7 RODRIGO KRUGNER, AND MARSHALL W. JOHNSON

Environ. Entomol. 39(4): 1264Ð1275 (2010); DOI: 10.1603/EN09367 ABSTRACT The glassy-winged sharpshooter, Homalodisca vitripennis (Germar), vectors the bacterium Xylella fastidiosa that induces PierceÕs disease of grape. This study determined the effect of temperature on the feeding activity of H. vitripennis adults and the resulting production of excreta. The Logan type I model described a nonlinear pattern that showed excreta production increased up to an optimal temperature (33.1ЊC), followed by an abrupt decline near an estimated upper threshold (36.4ЊC). A temperature threshold for feeding, at or below which adults cease feeding, was estimated to be 10ЊC using a linear regression model based on the percentage of adults producing excreta over a range of constant temperatures. A simulated winter-temperature experiment using ßuctuating thermal cycles conÞrmed that a time period above the temperature threshold for feeding was a critical factor in determining adult survival. Using data from the simulated temperature study, a predictive model was constructed by quantifying the relationship between cumulative mortality and cooling degree-hours. In Þeld validation experiments, the model accurately predicted the temporal pattern of overwintering mortality of H. vitripennis adults held under winter temperatures simulating conditions in BakersÞeld and Riverside, California, in 2006Ð2007. Model prediction using winter temperature data from a Riverside weather station indicated that H. vitripennis adults would experience an average of 92% overwintering mortality before reproduction in the spring, but levels of mortality varied depending on winter temperatures. The potential for temperature-based indices to predict temporal and spatial dynamics of H. vitripennis overwintering is discussed.

KEY WORDS glassy-winged sharpshooter, feeding, temperature, overwintering, mortality

The glassy-winged sharpshooter, Homalodisca vitrip- located within infested areas (Blua et al. 1999, Sister- ennis (Germar), is an important vector of Xylella fas- son et al. 2008). Whereas both X. fastidiosa and native tidiosa Wells et al., a bacterial pathogen that causes vectors have been present in California for over 100 yr scorch-like diseases in agricultural and landscape (Freitag et al. 1952, Davis et al. 1980), the presence of plants (e.g., grape, almond, and oleander) (Hopkins H. vitripennis combined with the ubiquitous nature of and Purcell 2002, Redak et al. 2004). H. vitripennis was the pathogen (Shapland et al. 2006) posed a new and found in southern California in the mid-1980s (So- greater threat. Compared with leafhopper vectors na- rensen and Gill 1996), and later associated with an tive to California, H. vitripennis has a wide host range increase in PierceÕs disease incidence in vineyards (Hoddle et al. 2003) and long ßight period (Blua and Morgan 2003), which increases its dispersal capabili- 1 Department of Entomology, University of California, Riverside, ties. Additionally, some portion of the H. vitripennis CA 92521. population feeds on and transmits X. fastidiosa to the 2 Current address: California Department of Food and Agriculture, mature basal portion of the grape cane, which may 13720 Rockpile Road, Arvin, CA 93203. lead to higher X. fastidiosa overwintering survival be- 3 United States Department of Agriculture-Agricultural Research Service, San Joaquin Valley Agricultural Sciences Center, Parlier, CA cause bacterial infections cannot be eliminated by 93648. either winter curing or pruning (Almeida and Purcell 4 Current address: Department of Entomology, University of Wis- 2003, Park et al. 2006). consin, 1630 Linden Drive, Madison, WI 53706. In California, H. vitripennis populations are pres- 5 Department of Environmental Science, Policy, and Management, University of California, Berkeley, CA 94720. ently established in the southern coastal and interior 6 PierceÕs Disease Control Program, California Department of Food valleys, and the southern region of the San Joaquin and Agriculture, 4500 Glenwood Drive, Building E, Riverside, CA Valley (CDFA 2008). The northward range expansion 92501. of this pest is a concern because this poses a greater 7 Corresponding author: University of California Kearney Agricul- tural Center, 9240 South Riverbend Avenue, Parlier, CA 93648 risk for the northern California grape and almond (e-mail: [email protected]). industries. Previous studies reported that the native

0046-225X/10/1264Ð1275$04.00/0 ᭧ 2010 Entomological Society of America August 2010 SON ET AL.: FEEDING THRESHOLD OF Homalodisca vitripennis 1265 range of H. vitripennis includes the southeastern 3) validate the predictive model in a Þeld situation by United States and northeastern Mexico (Triapitsyn simultaneously monitoring temperature and overwin- and Phillips 2000). Based on predictions generated by tering survival. the CLIMEX program, Hoddle (2004) suggested that winter cold stress would prevent H. vitripennis from establishing permanent populations north of Califor- Materials and Methods nia. However, one drawback of using the CLIMEX model is that when there are no biological parameters Insects and Plants. Unless otherwise stated, labora- on the target species, the model parameters must be tory-reared H. vitripennis adults (Ϸ3 wk old) used in estimated from an inverse (or inferential) modeling this study were obtained from the California Depart- process by manually adjusting parameter values until ment of Food Agriculture Biocontrol Facility (Arvin, the simulated geographical distribution coincides CA). At the California Department of Food Agricul- with the observed distribution (Vera et al. 2002). Such was ture facility, the laboratory colony of H. vitripennis, the case with H. vitripennis and, therefore, the north- originally collected mainly at citrus orchards in Ven- ward expansion may be more or less limited by winter tura County, has been maintained to harvest H. vit- temperatures than predicted by current CLIMEX es- ripennis eggs to produce biological control agents (egg timates. Winter temperatures would also impact this parasitoids) against H. vitripennis. The adults were pestÕs population densities. Because H. vitripennis transferred and maintained in screened cages (Bio- adults overwinter without diapause (Turner and Pol- quip Products, Gardena, CA) provisioned with potted lard 1958), they are required to periodically feed on plants of cowpea (Vigna unguiculata [L.] Walp) and plants and their survival is likely affected by low tem- ÔFrost EurekaÕ lemon (Citrus limon [L.] Burm. f.) at an peratures, as seen in other insects (Bale 1991, Leather enclosed H. vitripennis laboratory at California State et al. 1993). Therefore, in addition to limiting H. vit- University, Fresno. Insecticide-free lemon plants were ripennisÕs geographic distribution, winter tempera- obtained from the United States Department of Ag- tures may reduce overwintering success in regions riculture-Agricultural Research Service San Joaquin where the pest is established and inßuence H. vitrip- Valley Agricultural Sciences Center (Parlier, CA), and ennis population densities the following year. individually transplanted into pots (3.78 liter), which In the laboratory, Son et al. (2009) showed that were then maintained in a greenhouse at the Univer- adult H. vitripennis survival was negatively inßuenced sity of California Kearney Agricultural Research and by prolonged exposure to temperatures Յ7.8ЊC, inde- Extension Center (Parlier, CA). Lemon trees were pendent of the availability of suitable host plants. They used as a feeding host because they were a common hypothesized that long-term exposure to low temper- year-round host for H. vitripennis in southern Califor- atures would impede H. vitripennis feeding and, as the nia (Perring et al. 2001) and did not show visible stress insect starves, it would suffer dehydration, depletion symptoms caused by low or high temperatures in a of energy reserves, and ultimately death. There are previous study (Son et al. 2009). also Þeld observations to support this hypothesis. In Determination of Feeding Thresholds: Constant southern California, adult H. vitripennis densities grad- Temperatures. Subsistence on nutrient-sparse (Ͻ5%) ually declined through the winter months (Blua et al. xylem requires H. vitripennis to ingest large volumes of 2001, Castle et al. 2005). Similarly, Pollard and Ka- xylem ßuid, and this behavior produces large volumes loostian (1961) observed that Homalodisca liturata of liquid excreta (Andersen et al. 1989), which has (Ball) adults remained sessile, without feeding, at tem- been used to measure feeding activity (Brodbeck et al. peratures Ͻ11.1ЊC. 1993). Similarly, we used excreta-collection sachets to Temperature is a major abiotic factor affecting var- estimate levels of H. vitripennis feeding activity. A ious aspects of an insectÕs biology (Gilbert and Raw- sachet (7.5 ϫ 6.5 cm) was made from ParaÞlm (VWR orth 1996), and its direct effects are often the key Lab, Batavia, IL) by the method of Pathak et al. (1982). determining factor for survival because insects are Adults were individually captured within a rearing poikilotherms with their metabolic rates strongly in- cage using a clear plastic vial (33 ml; Bioquip Products, ßuenced by ambient temperatures (Speight et al. Gardena, CA), and their gender was determined. 1999). However, we know of no published reports that Adult males and females were individually placed in describe the inßuences of temperature on H. vitrip- the sachet, which was then enclosed around the stem ennis feeding, other than our previous study (Son et al. of a potted lemon plant (1 m height) and sealed. Plants 2009). This information is needed to fully understand were then randomly allocated to constant tempera- this pestÕs feeding behavior in relation to pathogen ture treatments in environmental chambers set at 8.9, transmission, predict its seasonal population dynam- 13.3, 18.8, 21.7, 24.6, 31.1, 35.1, and 40.8 Ϯ 1ЊC, as ics, and delimitate its potential geographic distribu- monitored using a data logger (HOBO, Onset Com- tion. In this study, we investigated the impacts of puter, Bourne, MA), with a photoperiod of 10:14 temperature on adult H. vitripennis feeding and sur- (L:D) h. The lemon plants were watered daily until vival. The objectives were to: 1) determine the tem- the potting soil became fully saturated. After a 48-h perature threshold at or below which feeding does not feeding period, the excreta produced was measured occur; 2) develop a predictive model based on the (in mg) using an electronic balance (weight of sachet relationship between the low temperatures, feeding with excreta Ð initial weight of sachet). Adult survival activity, and the resulting H. vitripennis survival; and during the 48-h feeding period was also determined. 1266 ENVIRONMENTAL ENTOMOLOGY Vol. 39, no. 4

Statistical Analysis. Data presented in this work are Oakville (Napa County) in the North Coast region, mean values (ϮSEM). One insect per host plant was where H. vitripennis is likely to establish because of considered a replicate, and each treatment had 10Ð12 mild winters; and 3) Buntingville (Lassen County) in replicates per sex. Before statistical analysis, data on the Northeast Plateau region, where the pest is un- the weight of xylem excreta were transformed likely to establish as a result of harsh winter temper- ϩ (log10[x 1]) to normalize variance. Treatment ef- atures. The environmental chambers were set to sim- fects were determined using a mixed model analysis of ulate hourly cycles of a typical winter day (i.e., 1 variance (ANOVA) (SAS Institute 1995); treatment January 2006) for each location, excluding tempera- means were separated using the Student-Newman- tures at or below 0ЊC to ensure that we were measur- Keuls test. ␹2 analysis was used to determine the treat- ing the effects of low temperature inactivity instead of ment effect on the frequency of adults that produced freezing on the test insects. Hourly temperature data xylem excreta and survived the 48-h trial. Regression were obtained from the California Irrigation Manage- analysis was performed to estimate model parameters ment Information System website (http://www.cimis. for temperature effects on each response variable us- water.ca.gov). The geo-references (elevation, lati- ing the TableCurve 2D Curve Fitting Program (Jandel tude, and longitude) for the weather stations are ScientiÞc 1996). Riverside (311 m, 33Њ57Ј N, 117Њ20Ј W), Oakville (58 Temperature-Dependent Feeding Model. The rela- m, 38Њ26Ј N, 122Њ24Ј W), and Buntingville (1,221 m, tionship between temperature and the percentage of 40Њ17Ј N, 120Њ26Ј W). To record the actual treatment H. vitripennis adults that produced xylem excreta dur- temperatures experienced by the insects, a data logger ing the 48-h trial was described using a two-parameter (HOBO, Onset Computer, Bourne, MA) was installed Weibull distribution model (CockÞeld et al. 1994) as in each environmental chamber. follows: A cylindrical cage (15 cm diameter ϫ 75 cm height) made of clear plastic sheeting (PETG-Vivak) was ͑ ͒ ϭ Ϫ ϫ Ϫ͑ ␣͒␤ P T 100 100 exp[ T/ ] [1] placed over top of each of nine potted lemon plants where P(T) is the percentage of adults that produced used in the study. The cages had a screened top and excreta; ␣ and ␤ are scale and shape parameters, re- two ventilation holes (2.5 cm diameter) in the wall. spectively. By using the linear portion of the nonlinear Ten laboratory-reared young adults were transferred curve, the lower threshold temperature was estimated into each cylinder cage, and each cage was considered from a linear model (Campbell et al. 1974). a replicate. This experiment had three treatments The temperature-dependent trend for excreta pro- (three thermal proÞles) with three replicates per duction was quantiÞed as a function of temperature treatment. The number of surviving adults was using a combination of the nonlinear Logan type I checked every 2 d during 5 wk after exposure, and model (Logan et al. 1976) and the linear model. Before then every 7 d until all adults in all replications were regression, the amount of excreta produced hourly per dead. Treatment effects were determined using re- adult was calculated by dividing the excreta amount peated measures ANOVA (SAS Institute 1995). This per surviving adults by 48-h conÞnement time. The experiment was designed to determine the effect of Logan model was used to Þt the data over the full range ßuctuating and nonfreezing Þeld temperatures on of the test temperatures, and the mathematical ex- adult survival, and results were then used to develop pression was as follows: a predictive model (below), based on the relationship between duration of starvation below the feeding Y͑T͒ ϭ ␺͕exp͑␳T͒ Ϫ exp͓␳TL Ϫ ͑TL Ϫ T͒/⌬T͔͖ threshold and resulting survival. Cooling Degree-Hours Model. A quantitative [2] model was developed to describe the relationship be- where Y(T) is the amount of xylem excreta produced tween the period (hourly) of nonfeeding (i.e., star- at temperature T, ␺ is a measurable amount of the vation) below the estimated temperature threshold excreta production at an arbitrary base temperature, for feeding (10ЊC, see “Results”) and cumulative mor- ␳ can be interpreted as a composite Q10 value for tality of H. vitripennis adults. The mean values of enzyme-catalyzed biochemical reactions, TL is the le- mortality (%) obtained from the simulated tempera- thal maximum temperature, and ⌬T is the width of the ture experiment were plotted against the cumulative decline phase above the optimum temperature. cooling degree-hours (CDH), which accounts for Development of a Predictive Survival Model: Fluc- hours of starvation as a result of cold exposure below tuating Temperatures. To verify the validity of the the feeding threshold as a unit of time-temperature. feeding threshold estimated from constant tempera- The calculation method of CDH is as follows: tures, adults were held under ßuctuating temperature ϭ ⌺ regimes, selected to simulate realistic hourly cycles of CDHcumulative CDH [3] three California sites in winter. Under a photoperiod T Ϫ 10ЊC, If T Ͻ 10ЊC ϭ ͫ h h ͬ of 10:14 (L:D) h, temperature-controlled environ- CDH 0, If T Ͻ 10ЊC [4] mental chambers were programmed to simulate the h Њ hourly temperature cycles in three geographically dis- where Th is hourly temperature ( C) and CDHcumulative tant locations in California: 1) Riverside (Riverside is cumulative CDH. Using the TableCurve 2D Pro- County) in the Los Angeles Basin region where H. gram, a nonlinear regression model was developed to vitripennis populations are well established; 2) describe the relationship between CDH and percent August 2010 SON ET AL.: FEEDING THRESHOLD OF Homalodisca vitripennis 1267 mortality by Þtting the data to the Weilbull equation (equation 1 provided above). Before regression analysis, the mortality data were plotted against the cooling degree-hours for each treatment. Validation of the Predictive Model: Field Condi- tions. Field Monitoring of Overwintering Survival. To validate the CDH model, Þeld-cage experiments were conducted at two locations from late November 2006 to early March 2007 using H. vitripennis adults that were collected at citrus orchards at Fillmore in Ven- tura County, California, on 28 November. Upon collection, the adults were transferred to the California Department of Food Agriculture Arvin Facility and held in a cage containing potted cowpea plants at room temperature for1dtoeliminate individuals that Fig. 1. Survival (%) of H. vitripennis adults in 48-h feed- died because of stress during the Þeld collection. ing trial under constant temperatures using the ParaÞlm Twenty-Þve adults were collected, using an aspirator, sachet method. into each plastic vial (33 ml; Bioquip, Gardena, CA), which were then wrapped in paper towels and held in Prediction Versus Observation Values. For each a cooler (Ϸ15ЊC) during transportation to the study outer-tent cage, the temperature data were used to sites. As a result of regulatory protocols and growersÕ calculate the CDH to predict the mortality data be- concerns, these Þeld experiments were conducted in cause of differences in temperatures among the cages Ϯ BakersÞeld at the University of California Kern within the same site. Mean ( SEM) observed mor- County Extension OfÞce (elevation: 115 m, 35Њ20Ј 46Љ tality recorded from the three test cages inside each N and 118Њ57Ј 55Љ W) and Riverside at the University outer cage, and the predicted mortality data were of California Citrus Experiment Station (elevation: plotted against calendar date. Because temperature 300 m, 33Њ57Ј 59Љ N and 117Њ20Ј 46Љ W). H. vitripennis data from one of the data loggers at each site were unavailable because of technical problems, two sets of populations were naturally present at each site, but data per site were used to validate the predicted mor- winter temperatures differed between the sites. tality. Predicted and observed values on each obser- To meet areawide control concerns, insects were vation date were compared using Wilcoxon paired- double caged to prevent accidental escape. Test cages sample test after pooling four sets of data at the 5% were composed of cylindrical iron-wired frames (35 signiÞcance level. cm diameter ϫ 1.5 m height ϫ 4 cm mesh size) cov- Overwintering Mortality: Estimation Using Weather ered by a screen bag with a zipper to allow access. One Data. California Irrigation Management Information week before H. vitripennis were introduced, cages System weather station data were used to estimate the were attached to the top of individual pots (38 liter), overwintering survival of H. vitripennis in Riverside each planted with one grape (Vitis vinifera, ÔPinot (weather station at 310 m [elevation], 22Њ57Ј 54Љ N and NoirÕ) and one citrus (Citrus spp., ÔCamizoÕ) plant. 117Њ20Ј 08Љ W). The value of the lower temperature Three caged pots with host plants were placed inside threshold (10ЊC) was used to calculate the CDH using ϫ ϫ an outer screened tent (2.4 m 2.4 m 2.4 m), in hourly ambient air temperature during the H. vitrip- which a yellow sticky trap and a noncaged potted ennis overwintering period (i.e., November-Febru- citrus plant treated with imidacloprid were placed to ary) in each year, for nine winter seasons. The cumu- trap and kill any escaping insects. Groups of 50 adults lative CDH during this period was used as input data were released into each cage on 29 November 2006. A to estimate the overwintering mortality on corre- data logger (HOBO, Onset Computer, Bourne, MA) sponding calendar dates because reproduction of was placed inside a test cage in each outer-tent cage overwintered females begins in early March in Riv- to monitor hourly temperatures. Each site had a total erside (Hummel et al. 2006). of three outer tents and nine test cages. The number of surviving adults was monitored every 7 d until the end of February 2007. Insects were checked in the Results afternoon when warmer temperatures increased in- Constant Temperature Experiment. Adult Survival. sect activity and facilitated differentiating between Temperature signiÞcantly inßuenced adult survival live and dead insects. To aid visual inspection, white (␹2 ϭ 119.02, df ϭ 7, P Ͻ 0.0001), but insect sex was play sand was layered on the soil surface in the pots not a factor (␹2 ϭ 0.48, df ϭ 1, P Ͼ 0.05) (Fig. 1). and plant debris was removed during each observa- Survival of H. vitripennis adults was Ͼ80% at all tem- tion. Potted plants were irrigated routinely and did not peratures except 40.8ЊC, a temperature at which all of exhibit drought symptoms during the experimental the tested adults died in the 48-h trial. Forty of the 194 period. For data analysis, the mean numbers of sur- adults tested did not survive the trial. Among them, viving adults were transformed (log[x ϩ 1]) to nor- Þve adults drowned in their own excreta (1, 1, 1, and malize the variance, and treatment effects were com- 2 adults at 21.7, 24.6, 31.7, and 35.1ЊC, respectively) and pared using repeated measures ANOVA. these data were excluded from the analysis. 1268 ENVIRONMENTAL ENTOMOLOGY Vol. 39, no. 4

Table 1. Excreta production per Homalodisca vitripennis adult during 48-h feeding on ‘Eureka’ lemon plants at constant temperatures

Mean Ϯ SEM excreta (mg) Temp (ЊC) Na n Male N Female n Total 8.3 23 11 0.0 Ϯ 0.0c 10 0.0 Ϯ 0.0e 21 0.0 Ϯ 0.0e 13.3 22 10 0.0 Ϯ 0.0c 10 47.8 Ϯ 47.8de 20 23.9 Ϯ 23.9e 18.8 23 11 105.6 Ϯ 98.1c 12 340.7 Ϯ 323.3cd 23 228.2 Ϯ 173.1d 21.7 21 10 325.6 Ϯ 140.9b 10 712.2 Ϯ 310.0bc 20 518.9 Ϯ 171.6c 24.6 25 12 1,766.5 Ϯ 987.6a 12 2,302.0 Ϯ 1,027.6ab 24 2,034.2 Ϯ 699.2b 31.1 27 12 2,993.9 Ϯ 931.9a 12 6,932.0 Ϯ 2,384.3a 24 4,963.0 Ϯ 1,317.5a 35.1 25 10 4,156.6 Ϯ 1,286.4a 12 4,025.0 Ϯ 2,064.1ab 22 4,084.8 Ϯ 1,240.9ab 40.8 24 0 Ñb 0Ñ 0Ñ

Means followed by the same letter within each column are not signiÞcantly different (Student-Newman-Keuls test, P Ͻ 0.05). a Total number of tested adults per treatment. b Data unavailable due to 100% mortality.

Excreta Production. Among the temperature treat- ennis adults would produce excreta, respectively. A ments 13.3Ð35.1ЊC, the percentage of adults that pro- linear model was successfully Þtted to the data from duced excreta was signiÞcantly different (␹2 ϭ 131.58, the linear portion (8.3Ð24.6ЊC) of the nonlinear df ϭ 7, P Ͻ 0.0001), but did not vary between sex (␹2 ϭ Weibull curve (F ϭ 45.34; df ϭ 1, 4; P Ͻ 0.01) (Table 1.69, df ϭ 1, P Ͼ 0.05). At 24.6Ð35.1ЊC, all H. vitripennis 2) and provided an estimate (axis intercept) of a lower adults produced excreta. The percentage of adults feeding threshold temperature of 10.0ЊC. producing excreta declined as temperature decreased. Hourly excreta production, estimated by dividing At temperatures Յ13.3ЊC, only one of 41 adults tested total excreta production per adult by 48 h, was inßu- produced excreta. The excreta production per H. vit- enced by a temperature main effect (F ϭ 25.30; df ϭ ripennis adult surviving the 48-h trial was highly de- 6, 153; P Ͻ 0.001), but without a signiÞcant gender pendent upon temperature (F ϭ 234.9; df ϭ 6, 153; P Ͻ (F ϭ 0.15; df ϭ 6, 153; P Ͼ 0.05) or interaction (tem- 0.001), but there was no difference between sexes perature ϫ gender) effect (F ϭ 0.49; df ϭ 6, 153; P Ͼ (F ϭ 0.47; df ϭ 1, 153; P Ͼ 0.05) or interaction between 0.05). The effect of temperature on hourly excreta temperature and sex (F ϭ 0.50; df ϭ 6, 153; P Ͼ 0.05) production per adult was well described by the non- (Table 1). Therefore, data from males and females linear Logan model over the entire temperature range were pooled for regression analysis. No adults pro- (F ϭ 71.20; df ϭ 3, 6; P Ͻ 0.01) (Fig. 3; Table 3). This duced excreta at temperatures Յ8.3ЊC. The amount of nonlinear model estimated 33.1ЊC as the temperature excreta per H. vitripennis adult increased until it with the highest mean excreta production (117.8 peaked at 31.1ЊC. High variation in excreta production mg/h) and 36.4ЊC as the upper threshold temperature. among individuals was observed at 35.1ЊC, ranging Excreta production increased gradually up to 21.7ЊC from 7.2 to 25,241.7 mg. and then sharply increased to the optimal tempera- Temperature-Dependent Feeding Model. The ture, followed by an abrupt decline at temperatures Weibull model (equation 1) accurately described between the predicted optimal (33.1ЊC) and upper the nonlinear relationship between temperature and threshold (36.4ЊC) temperatures. A linear model gen- the percentage of H. vitripennis adults that produced erated from all data at temperatures Յ31.1ЊC esti- excreta over the temperature range from 8.3 to 35.1ЊC mated a lower threshold of 13.3ЊC with a signiÞcant (F ϭ 111.27; df ϭ 1, 6; P Ͻ 0.001) (Fig. 2; Table 2). At positive linear relationship with temperature (F ϭ 13.7, 15.9, 18.0, and 19.9ЊC, for instance, it was pre- 10.90; df ϭ 1, 5; P Ͻ 0.05) (Fig. 3; Table 3). dicted that 10, 25, 50, and 75% of surviving H. vitrip- Simulated Temperature Experiment. The mean (ϮSEM) temperatures for the 24-h proÞles were 12.0 Ϯ 0.3ЊC (range: 10.1Ð14.4ЊC), 10.5 Ϯ 0.4ЊC (range: 6.8Ð12.6ЊC), and 3.3 Ϯ 0.3ЊC (range: 1.3Ð7.9ЊC) for the temperature regimes simulating conditions in River- side, Oakville, and Buntingville, respectively (Fig. 4A). Xylem excreta production was observed under

Table 2. Parameter estimates of models to describe the relationship between temperature and the percentage of Homalodisca vitripennis adults that produced xylem excreta during 48-h trial

Model Parameter Estimate Ϯ SEM r2 Nonlinear Weibull model ␣ 18.9807 Ϯ 0.6441 0.957 (CockÞeld et al. 1994) ␤ 6.9183 Ϯ 2.5487 Linear modela a Ϫ63.7779 Ϯ 17.3926 0.938 Fig. 2. Temperature-dependent frequency (%) of H. vit- (Campbell et al. 1974) b 6.3986 Ϯ 0.9501 ripennis adults that produced the excreta for 48-h trial as Þtted to the nonlinear Weibull model. a Data range 8.3Ð24.6ЊC was chosen for the linear regression. August 2010 SON ET AL.: FEEDING THRESHOLD OF Homalodisca vitripennis 1269

model estimated that 10, 50, and 90% adult mortality would occur at 158.2, 683.2, and 1,736.5 CDH, respectively. Validation of the Predictive Model: Field Data. Re- peated measures ANOVA revealed a signiÞcant difference in the number of surviving adults between BakersÞeld and Riverside (F ϭ 7.04; df ϭ 1, 233; P ϭ 0.0173), time (F ϭ 144.58, df ϭ 12, P Ͻ 0.0001), and interaction of times and site (F ϭ 6.77, df ϭ 12, P Ͻ 0.0001). In BakersÞeld, 100% mortality was observed in early January 2007 (3 and 10 January for two different cages), whereas 99.3% mortality was observed in Riverside on 28 February 2007 (Fig. 6). There was Fig. 3. Temperature-dependent production of xylem ex- a difference in temperature condition between the creta (mg) per H. vitripennis adult on the lemon plant as two Þeld sites and also between the two cages at each Þtted to nonlinear Logan model (Logan et al. 1976). site. Accumulated CDH over the entire observation period (29 November 2006Ð28 February 2007) were the simulated winter temperatures for Riverside and 8,341.4 and 7,981.1 CDH at two cages in BakersÞeld, Oakville, but no excreta production was observed un- whereas they were 4,639.3 and 4,844.7 CDH at two der that of Buntingville, where temperature never cages in Riverside. The cooling degree-hours model surpassed 10ЊC. Under this regime, notably, 66.7 and provided a reliable prediction of the temporal pattern 100% of the insects were found inactive on the soil of overwintering mortality of H. vitripennis adults surface at 2- and 9-d exposure, respectively. In con- throughout the monitoring period at both sites, esti- trast, in the other simulated regimes at 9-d exposure, mating that time of 90% mortality would occur on 19 most of the adults (Ն80%) remained on the branches. December and 20 December for the two BakersÞeld Repeated measures ANOVA revealed that the num- cages and 30 December and 31 December for the two bers of surviving adults were inßuenced by tempera- Riverside cages (Fig. 6). Over the pooled data, there ture regime (F ϭ 29.2; df ϭ 2, 296; P Ͻ 0.001) and time was no signiÞcant difference between the observed (F ϭ 93.4, df ϭ 32, P Ͻ 0.001), with an interaction and predicted values of mortality (Wilcoxon paired- Ͼ between regime and time (F ϭ 15.3, df ϭ 64, P Ͻ 0.001) sample test, P 0.05), although some level of discrep- (Fig. 4B). The shape of the mortality curve displayed ancy was observed particularly in the early period of a nonlinear relationship against exposure time, which monitoring (Fig. 6). The deviation of predicted mor- was Þt to the Weibull equation (equation 1) for each tality (%) at the dates corresponding to observation Ͻ temperature regime to estimate the lethal time at the ranged from 0 to 28.5%, with 5% mean deviation Ϯ Ϯ 10, 50, and 90% mortality levels (Fig. 4B). It is note- (4.7 2.6%, mean SEM). worthy that 100% mortality occurred at 14 d when H. Validation of the Predictive Model Using Weather vitripennis adults were continuously exposed to Data. Accumulated CDH throughout winter also var- Buntingville thermal cycles (below 10ЊC), whereas it ied among years, which resulted in annual variation of took 147 and 161 d to reach 100% mortality under the the estimated overwintering mortality. The overwin- simulated temperatures of Riverside and Oakville, re- tering mortality in Riverside, estimated by the pre- spectively. The lethal times at 10, 50, and 90% mortality dictive model, showed a yearly variation (over nine under each simulated temperature were as follows: winter seasons, 1998Ð1999-2006Ð2007) that ranged 12.9, 56.6, and 145.1 d for that of Riverside; 9.3, 44.6, from 82.2 to 99.0% in the 1999Ð2000 and 2003Ð2004 and 145.1 d for that of Oakville; 2.2, 5.3, and 9.1 d for winters, respectively (Table 4). Mean percentage of that of Buntingville, respectively. overwintering mortality indicated that 92% of over- Cooling Degree-Hours Model. The Weibull model wintering H. vitripennis adults in Riverside would die quantiÞed the relationship between CDH and mean before reproduction in spring. In Riverside, Castle et mortality (F ϭ 990.8; df ϭ 1, 40; P Ͻ 0.001) and al. (2005) showed that the adult density in untreated Ϸ explained 96.2% variability of the data (Fig. 5). The lemon trees declined from 25 to 1 per sample unit (Ϸ96% mortality) in late October 2001 through late February 2002. Using temperature data from the Table 3. Parameter estimates of models to describe the rela- weather station, our model estimated 98% mortality tionship between temperature and hourly excreta production (mg) per Homalodisca vitripennis adults over the similar corresponding period (1 November 2001Ð28 February 2002). Model Parameter Estimate Ϯ SEM r2 Nonlinear model ␺ 0.082433 Ϯ 0.333639 0.987 Discussion (Logan et al. 1976) ␳ 0.25991 Ϯ 0.54435 Ϯ TL 36.3938 1.1532 Regression models in this work quantitatively de- ⌬T 2.83384 Ϯ 8.20987 scribed the temperature-dependent pattern of H. vit- Linear model a Ϫ57.0764 Ϯ 27.2256 0.731 (Davidson 1944)a b 4.2810 Ϯ 1.2973 ripennis feeding in terms of both xylem excreta and the percentage of adults producing excreta by applying a Data range (8.3Ð31.1ЊC) chosen for the linear regression. the models of Logan et al. (1976) and Weibull (Cock- 1270 ENVIRONMENTAL ENTOMOLOGY Vol. 39, no. 4

Fig. 4. Mortality under simulated winter temperatures at three locations in California: (A) actual hourly thermal proÞle to which the H. vitripennis adults were exposed and (B) percent mortality (mean Ϯ SEM) Þtted to Weibull equation for each treatment.

Þeld et al. 1994), respectively. We found that temper- (36.4ЊC) indicated that H. vitripennis feeding activity ature-dependent excreta production showed the typ- would be severely inhibited because of insect physi- ical nonlinear pattern commonly found in insect ological stress at temperatures equal or higher than the development, which is why the Logan model Þt the threshold. In fact, no adults tested in our study sur- pattern. The estimated upper threshold temperature vived the 48-h trial at 40.8ЊC. Similarly, we have ob-

Fig. 5. Cooling degree-hours model to describe the temperature-dependent percent mortality of H. vitripennis adults using the Weibull equation Parameter estimates (ϮSEM) to describe the relationship between cooling degree-hours and the percent mortality of H. vitripennis obtained from simulated temperature experiment were ␣ ϭ 908.323 Ϯ 31.152 and ␤ ϭ 1.286 Ϯ 0.075 (r2 ϭ 0.962). August 2010 SON ET AL.: FEEDING THRESHOLD OF Homalodisca vitripennis 1271

Fig. 6. Comparison of percent mortality of H. vitripennis adults predicted by the cooling degree-hours model with those adults observed in Þeld cages in BakersÞeld and Riverside. Four sets of data (two for each site) were used to validate the model. served that H. vitripennis adult longevity was nega- tions, as a result of thermoregulatory behavior. As a tively inßuenced by continuous exposure to high result of the relative large volume of xylem sap passing temperatures (Son et al. 2009). However, unlike the through the digestive system in a short period, these adults we conÞned in the sachet, adults in the Þeld insects also may be able to regulate body temperature might be capable of actively avoiding unfavorable according to feeding substrate versus ambient tem- microclimates upon exposure to the extreme condi- perature. Note that under Þeld conditions, plants are 1272 ENVIRONMENTAL ENTOMOLOGY Vol. 39, no. 4

Table 4. Estimated overwintering mortality of Homalodisca feeding or movement (Son et al. 2009). Under Þeld vitripennis in Riverside against CDH using weather station data situations, if temperature stays above the threshold long enough to trigger activity of this insect, plant Winter Mortality CDHa season (%) characteristics can be more inßuential on host selection behavior, particularly host acceptance and use (Bi 1998Ð1999 1,524.8 85.7 1999Ð2000 1,387.5 82.2 et al. 2005, Nadel et al. 2008, Krugner et al. 2009). 2000Ð2001 2,831.7 98.7 In addition to those plant variables reported in pre- 2001Ð2002 2,631.8 98.0 vious studies, this study clearly demonstrates that 2002Ð2003 1,477.5 84.6 adult feeding rate is highly dependent upon feeding 2003Ð2004 2,959.7 99.0 2004Ð2005 1,992.3 93.6 substrate and ambient temperatures. We showed that 2005Ð2006 1,662.3 88.7 the amount of xylem excreta per adult increased by 2006Ð2007 2,568.1 97.8 207-fold when temperature increased from 13.3 to Mean Ϯ SEM 2,115.0 Ϯ 210.9 92.0 Ϯ 2.3 31.1ЊC (Table 1). Results from our simulated temperature experiment indicated that prolonged exposure a Cooling degree-hours for each overwintering year, calculated using hourly ambient temperature data obtained from Riverside to moderately cold temperatures (below the thresh- weather station, were accumulated from Nov. through Feb. old, but above 0ЊC) increases mortality apparently because of starvation. For instance, 100% mortality occurred in 2 wk at the regime of 1.3Ð7.9ЊC versus 21 able to reduce leaf surface temperature by increasing wk at the regime of 10.1Ð14.4ЊC. Consistent with a transpiration (i.e., opening stomata). In our study, previous Þeld observation (Pollard and Kaloostian whole plants were kept in temperature cabinets, 1961), H. vitripennis adults maintained at thermal cy- which equalized food and ambient temperatures. Ad- cles continuously below the threshold dropped onto ditionally, duration of exposure to such extreme tem- the soil surface. The similarity in H. vitripennis survival peratures would be relatively short under ßuctuating recorded from the simulated temperature regimes for Þeld temperatures. Nonetheless, it is evident that H. Riverside and Oakville suggests that the insect may vitripennis feeding was signiÞcantly inßuenced by obtain adequate resources to survive over winter as temperature because all tested adults fed at Ն24.6ЊC. long as the temperature cycles remain above the lower In the linear regression analysis, the estimated lower threshold for a certain period of time even when threshold (13.3ЊC), based on the hourly excreta winter feeding is restricted to relatively short periods. amount, was higher than that of 10.0ЊC, which was Nonetheless, the longevity of the sharpshooters under based on the percentage of adults producing excreta. both winter-simulated conditions was much shorter, The value 10.0ЊC appeared to be more appropriate for compared with those (100% mortality at 36 wk) kept two reasons, as follows: 1) one adult was still capable under constant 23.4ЊC in the previous study (Son et al. of feeding at 13.3ЊC, and 2) the pattern of excreta 2009). production was poorly explained by the linear model By linking temperature cycles, feeding activity, and (r2 ϭ 0.731), which displayed a typical nonlinear survival rates, the model based on the lower threshold shape than that of the percentage of adults with ex- required for feeding explained 96.2% of the variance cretion (r2 ϭ 0.938). Thus, the lower threshold associated with H. vitripennis survivorship, thereby (10.0ЊC) estimated from the linear regression on the supporting our Þnding that starvation resulting from feeding frequency was more reasonable. insect inactivity during the cold period is a critical Little attention has been focused on how H. vitrip- factor affecting mortality. The period of inactivity ennis feeding is affected by abiotic environmental fac- caused by low temperatures as measured in terms of tors. In contrast, the impacts of host plant character- cooling degree-hours was a reliable predictor of over- istics (e.g., nutrient composition, plant species, and wintering mortality in H. vitripennis adults. Previous xylem tension) have been well studied (Andersen et studies indicated that H. vitripennis adults overwinter al. 1992; Brodbeck et al. 1993, 1995, 2004). Increased in some areas of California, and they ßew under warm xylem tension within a nonirrigated host plant nega- conditions during the winter, based on the number of tively inßuenced feeding (in terms of excreta produc- adults captured on yellow panel traps (Park et al. tion) of H. vitripennis adults (Andersen et al. 1992). 2006). As a winter-active species, H. vitripennis adults Feeding rates of the adults were correlated with the would be more directly inßuenced by increases in diurnal cycles in xylem ßuid chemistry, such as the winter temperatures than winter-inactive species (in ratio of nutrients (i.e., ratio of amides to total organic a diapausing status) because of enhanced feeding and compounds) (Andersen et al. 1992) or the concen- consequential reduced chances of mortality (Battisti tration of nutrients (i.e., total amino acid concentra- et al. 2005). For winter-active species, winter warming tion) (Brodbeck et al. 1993). The Þndings in this study appeared to be responsible for range expansion (to imply that effects of these plant variables on the in- higher latitudes and elevations), for example, as ob- sectÕs feeding could be more evident when the insects served in the lepidopterans Atalopedes campestris are able to feed under conditions where the ambient (Boisduval) (Crozier 2003) and Thaumetopoea pityo- temperature stays near optimal or at least above the campa (Denis and Schiffermu¨ ller) (Battisti et al. threshold. At low temperatures (Յ7.8ЊC) below the 2005), and coleopteran Dendroctonus frontalis Zim- threshold, independent of host plant availability, mermann (Williams and Liebhold 2002). Therefore, as the insects were physiologically inactive without any in these species, winter warming caused by climate August 2010 SON ET AL.: FEEDING THRESHOLD OF Homalodisca vitripennis 1273 change may allow H. vitripennis to colonize beyond its of the following early spring population can be esti- current distribution by increasing the overwintering mated (Bale 1991). Therefore, population dynamics success, which may further result in changes in the risk during late summer and early fall periods must be of PierceÕs disease epidemics and other diseases integrated to provide a complete picture on the sur- caused by X. fastidiosa. vival potential of H. vitripennis. In summary, our study The cooling degree-hours model was a reliable pre- clearly demonstrated that the feeding activity of H. dictor to describe the temporal pattern of overwin- vitripennis adults is strongly inßuenced by tempera- tering mortality of H. vitripennis under Þeld conditions ture. The predictive model described in this work has in BakersÞeld and Riverside. The model somewhat good potential to predict the temporal and spatial underestimated the mortality in the early weeks after pattern of overwintering success for H. vitripennis Þeld release, which seemingly resulted from insect adults based on current annual temperature proÞles. stress before Þeld release (i.e., collection, aspiration, and transportation). Although the Þeld-cage study conÞrmed our laboratory results, extrapolation of Acknowledgments Þndings from the Þeld-conÞnement experiment to an open-Þeld situation should be done with caution for We thank Martha Gerik (University of California, River- several reasons. First, the microclimate inside the Þeld side) for technical assistance and Raymond Yokomi (United cages appeared to be cooler than that in the open Þeld, States Department of Agriculture-Agricultural Research Ser- and this was the result of double screening of direct vice, San Joaquin Valley Agricultural Sciences Center, Par- lier, CA) for the citrus plants. This work was supported in part sunlight. Second, the adults conÞned within the cage by the California Department of Food and Agriculture had limited access to more favorable host plants. In a PierceÕs Disease/Glassy-Winged Sharpshooter Research Þeld situation, they may be capable of dispersing and Program and a United States Department of Agriculture- actively seeking better host plants (Blua and Morgan Agricultural Research Service SpeciÞc Cooperative Agree- 2003, Bi et al. 2005, Krugner et al. 2009). Third, removal ment provided to M.W.J. of plant litter on the soil surface in this study eliminated shelter that potentially might have mitigated direct chilling injury or cold shock during winter References Cited nights (Leather et al. 1993). In the Þeld, low vegetation cover and plant litter often exist on the ground Almeida, R.P.P., and A. H. Purcell. 2003. Transmission of where H. vitripennis adults drop during cold periods. Xylella fastidiosa to grapevines by Homalodisca coagulata (Hemiptera: Cicadellidae). J. Econ. Entomol. 96: 264Ð Temperature data acquired from a local weather 271. station were used in the model to estimate overwin- Andersen, P. C., B. V. Brodbeck, and R. F. Mizell III. 1989. tering mortality of H. vitripennis. Results estimated Metabolism of amino acids, organic acids and sugars ex- that an average of 8% of the H. vitripennis adults in tracted from the xylem ßuid of four host plants by adult Riverside would be able to survive and reproduce in Homalodisca coagulata. Entomol. Exp. Appl. 50: 149Ð159. spring, although annual variation of overwintering Andersen, P. C., B. V. Brodbeck, and R. F. Mizell III. 1992. mortality existed because of yearly differences in win- Feeding by the leafhopper, Homalodisca coagulata, in ter temperatures. These results indicate that the relation to xylem ßuid chemistry and tension. J. Insect model would also estimate winter mortality of H. vit- Physiol. 38: 611Ð622. Bale, J. S. 1991. Implications of cold hardiness for pest man- ripennis populations at multiple geographical loca- agement, pp. 461Ð498. In R. E. Lee and D. L. Delinger tions where winter climate condition varies, which (eds.), Insects at low temperatures. Chapman & Hall, would be useful to compare the overwintering success New York, NY. among locations. By using cooling-degree days based Battisti, A., M. Stastny, S. Netherer, C. Robinet, A. Schopf, A. on the feeding threshold of 10ЊC and temperature data Roques, and S. Larsson. 2005. Expansion of geographic from weather stations, Johnson et al. (2008) devel- range in the pine processionary moth caused by increased oped geographic information system mapping of winter temperatures. Ecol. Appl. 15: 2084Ð2096. postwinter mortality of H. vitripennis adults and Bi, J. L., S. J. Castle, F. J. Byrne, S. J. Tuan, and N. C. Toscano. showed that the mortality in most winters would range 2005. Inßuence of seasonal nitrogen nutrition ßuctua- tions in orange and lemon trees on population dynamics 80Ð90% in most agricultural areas of the Central Val- of the glassy-winged sharpshooter (Homalodisca coagu- ley. The prediction results estimated by using weath- lata). J. Chem. Ecol. 31: 2289Ð2308. er-station temperature data have yet to be validated, Blua, M. J., and D.J.W. Morgan. 2003. Dispersion of Homa- possibly because of the difference in temperature data lodisca coagulata (Hemiptera: Cicadellidae), a vector of between the weather station and the actual micro- Xylella fastidiosa, into vineyards in southern California. J. habitat where this species actively disperses over win- Econ. Entomol. 96: 1369Ð1374. ter (Blua and Morgan 2003, Park et al. 2006), although Blua, M. J., P. A. Phillips, and R. A. Redak. 1999. A new both temperatures tend to be closely correlated (Lam sharpshooter threatens both crops and ornamentals. and Pedigo 2000). Calif. Agric. 53: 22Ð25. Blua, M. J., R. A. Redak, D.J.W. Morgan, and H. S. Costa. The predictive model in this study allowed the es- 2001. Seasonal ßight activity of two Homalodisca species timation of the survival of overwintered adults by (Homoptera: Cicadellidae) that spread Xylella fastidiosa accumulating cooling degree-hours throughout win- in southern California. J. Econ. Entomol. 94: 1506Ð1510. ter. If one knows the density of H. vitripennis adults Brodbeck, B. V., R. F. Mizell, and P. C. Andersen. 1993. entering the overwintering period, the relative density Physiological and behavioral adaptations of three species 1274 ENVIRONMENTAL ENTOMOLOGY Vol. 39, no. 4

of leafhoppers in response to the dilute nutrient content Leather, S. R., K.F.A. Walters, and J. S. Bale. 1993. The of xylem ßuid. J. Insect Physiol. 39: 73Ð81. ecology of insect overwintering. Cambridge University Brodbeck, B. V., P. C. Andersen, and R. F. Mizell III. 1995. Press, Cambridge, United Kingdom. Differential utilization of nutrients during development Logan, J. A., D. J. Wollkind, S. C. Hoyt, and L. K. Tanigoshi. by the xylophagous leafhopper, Homalodisca coagulata. 1976. Analytic model for description of temperature-de- Entomol. Exp. Appl. 75: 279Ð289. pendent rate phenomena in arthropods. Environ. Ento- Brodbeck, B. V., P. C. Andersen, R. F. Mizell III, and S. Oden. mol. 5: 1133Ð1140. 2004. Comparative nutrition and developmental biology Johnson, M. W., K. Lynn-Patterson, S. Mark, and R. Groves. of xylem-feeding leafhoppers reared on four genotypes of 2008. Assessing the post-winter threat of glassy-winged Glycine max. Environ. Entomol. 33: 165Ð173. sharpshooter populations, pp. 22Ð27. In T. Esser, P. Blin- Campbell, A., B. D. Frazer, N. Gilbert, A. P. Gutierrez, and coe, D. West, S. Veling, R. Randhawa, and J. LeMaster M. Mackauer. 1974. Temperature requirements of some (eds.), Proceedings, PierceÕs Disease Research Sympo- aphids and their parasites. J. Appl. Ecol. 11: 431Ð438. sium, 15Ð17 December 2008, San Diego, CA. California Castle, S. J., F. J. Byrne, J. L. Bi, and N. C. Toscano. 2005. Department of Food and Agriculture, Copeland Printing, Spatial and temporal distribution of imidacloprid and Sacramento, CA. thiamethoxam in citrus and impact on Homalodisca co- Nadel, H., R. Seligmann, M. W. Johnson, J. R. Hagler, D. C. agulata populations. Pest Manag. Sci. 61: 75Ð84. Stenger, and R. L. Groves. 2008. Effects of citrus and [CDFA] California Department of Food and Agriculture. avocado irrigation and nitrogen-form soil amendment 2008. PierceÕs disease control program maps. (http:// on host selection by adult Homalodisca vitripennis www.cdfa.ca.gov/pdcp/map index.html). (Hemiptera: Cicadellidae). Environ. Entomol. 37: 787Ð Cockﬁeld, S. D., S. L. Butkewich, K. S. Samoil, and D. L. 795. Mahr. 1994. Forecasting ßight activity of Sparganothis Park, Y.-L., T. M. Perring, R. Yacoub, D. W. Bartels, and D. sulfureana (Lepidoptera: Tortricidae) in cranberries. J. Elms. 2006. Spatial and temporal dynamics of overwin- Econ. Entomol. 87: 193Ð196. tering Homalodisca coagulata (Hemiptera: Cicadellidae). Crozier, L. 2003. Winter warming facilitates range expan- J. Econ. Entomol. 99: 1936Ð1942. sion: cold tolerance of the butterßy Atalopedes campestris. Pathak, P. K., R. C. Saxena, and E. A. Heinrichs. 1982. Oecologia 135: 648Ð656. ParaÞlm sachet for measuring honeydew excretion by Davidson, J. 1944. On the relationship between tempera- Nilaparvata lugens (Homoptera: Delphacidae) on rice. J. ture and rate of development of insects at constant tem- Econ. Entomol. 75: 194Ð195. peratures. J. Anim. Ecol. 13: 26Ð28. Perring, T. M., C. A. Farrar, and M. J. Blua. 2001. Proximity Davis, M. J., S. V. Thomson, and A. H. Purcell. 1980. Etio- to citrus inßuences PierceÕs disease in Temecula Valley logical role of the xylem-limited bacterium causing vineyards. Calif. Agric. 55: 13Ð18. PierceÕs disease and almond leaf scorch. Phytopathology Pollard, H. N., and G. H. Kaloostian. 1961. Overwintering 70: 472Ð475. habits of Homalodisca coagulata, principal natural vector Freitag, J. H., N. W. Frazier, and R. A. Flock. 1952. Six new of phony peach disease virus. J. Econ. Entomol. 54: 810Ð leafhopper vectors of PierceÕs disease virus. Phytopathol- 811. ogy 42: 533Ð534. Redak, R. A., A. H. Purcell, J.R.S. Lopes, M. J. Blua, R. F. Gilbert, N., and D. A. Raworth. 1996. Insects and temperature: a general theory. Can. Entomol. 128: 1Ð13. Mizell III, and P. C. Andersen. 2004. The biology of Hoddle, M. S. 2004. The potential adventive geographic xylem ßuid-feeding insect vectors of Xylella fastidiosa range of glassy-winged sharpshooter, Homalodisca coagu- and their relation to disease epidemiology. Annu. Rev. lata and the grape pathogen Xylella fastidiosa: implica- Entomol. 49: 243Ð270. tions for California and other grape growing regions of SAS Institute. 1995. SAS userÕs guide: statistics. SAS Insti- the world. Crop. Prot. 23: 691Ð699. tute, Cary, NC. Hoddle, M. S., S. V. Triapitsyn, and D. J. Morgan. 2003. Shapland, E. B., K. M. Daane, G. Y. Yokota, C. Wistrom, J. H. Distribution and plant association records for Homalo- Connell, R. A. Duncan, and M. A. Viveros. 2006. Ground disca coagulata (Hemiptera: Cicadellidae) in Florida. Fla. vegetation survey for Xylella fastidiosa in California al- Entomol. 86: 89Ð91. mond orchards. Plant Dis. 90: 905Ð909. Hopkins, D. L., and A. H. Purcell. 2002. Xylella fastidiosa: Sisterson, M. S., R. Yacoub, G. Montez, E. E. Grafton-Card- cause of PierceÕs disease of grapevine and other emergent well, and R. L. Groves. 2008. Distribution and manage- diseases. Plant Dis. 86: 1056Ð1066. ment of citrus in California: implications for management Hummel, N. A., F. G. Zalom, N. C. Toscano, P. Burman, and of glassy-winged sharpshooter. J. Econ. Entomol. 101: C.Y.S. Peng. 2006. Seasonal patterns of female Homalo- 1041Ð1050. disca coagulata (Say) (Hemiptera: Cicadellidae) repro- Son, Y., R. L. Groves, K. M. Daane, D.J.W. Morgan, and M. W. ductive physiology in Riverside, California. Environ. En- Johnson. 2009. Inßuences of temperature on Homalo- tomol. 35: 901Ð906. disca vitripennis (Germar) (Hemiptera: Cicadellidae) Jandel Scientiﬁc. 1996. TableCurve 2D: automated curve survival under various feeding conditions. Environ. En- Þtting and equation discovery, version 4.0. Jandel Scien- tomol. 38: 1485Ð1495. tiÞc, San Rafael, CA. Sorensen, J. T., and R. J. Gill. 1996. A range extension of Krugner, R., R. L. Groves, M. W. Johnson, A. P. Flores, J. R. Homalodisca coagulata (Say) (Hemiptera: Clypeorrhyn- Hagler, and J. G. Morse. 2009. Seasonal population dy- cha: Cicadellidae) to southern California. Pan-Pac. En- namics of Homalodisca vitripennis (Hemiptera: Cicadel- tomol. 72: 160Ð161. lidae) in sweet orange trees maintained under continuous Speight, M. R., M. D. Hunter, and A. D. Watt. 1999. Ecology deÞcit irrigation. J. Econ. Entomol. 102: 960Ð973. of insects: concepts and applications. Blackwell, Oxford, Lam, W.K.F., and L. P. Pedigo. 2000. A predictive model for United Kingdom. the survival of overwintering bean leaf beetles (Co- Triapitsyn, S. V., and P. A. Phillips. 2000. First record of leoptera: Chrysomelidae). Environ. Entomol. 29: 800Ð Gonatocerus triguttatus (Hymenoptera: Mymaridae) 806. from eggs of Homalodisca coagulata (Homoptera: Ci- August 2010 SON ET AL.: FEEDING THRESHOLD OF Homalodisca vitripennis 1275

cadellidae) with notes on the distribution of the host. Fla. Tephritidae), with emphasis on Argentina and Australia. Entomol. 83: 200Ð203. Environ. Entomol. 31: 1009Ð1022. Turner, W. F., and H. N. Pollard. 1958. Life histories and Williams, D. W., and A. M. Liebhold. 2002. Climate change behavior of Þve insect vectors of phony peach disease. and the outbreak ranges of two North American bark U.S. Dep. Agric. Tech. Bull. 1188: 1Ð57. beetles. Agric. Forest Entomol. 4: 87Ð99. Vera, M. T., R. Rodriguez, D. F. Segura, J. L. Cladera, and R. W. Sutherst. 2002. Potential geographical distribution of the Mediterranean fruit ßy, Ceratitis capitata (Diptera: Received 15 December 2009; accepted 1 April 2010.