See discussions, stats, and author profiles for this publication at: https://www.researchgate.net/publication/232666776

Cost-Benefit Analysis of Sorghum Midge, Stenodiplosis sorghicola 1 (Coquillett)-Resistant Sorghum Hybrid Research and Development i....

Article in Southwestern Entomologist · December 2009 DOI: 10.3958/059.034.0404

CITATIONS READS 11 130

3 authors:

Tebkew Damte Bonnie B Pendleton Ethiopian Institute of Agricultural Research West Texas A&M University

13 PUBLICATIONS 117 CITATIONS 72 PUBLICATIONS 185 CITATIONS

SEE PROFILE SEE PROFILE

Lal K. Almas West Texas A&M University

94 PUBLICATIONS 397 CITATIONS

SEE PROFILE

Some of the authors of this publication are also working on these related projects:

Greenbug View project

Ethiopian Institute of Agricultural research View project

All content following this page was uploaded by Tebkew Damte on 02 April 2019.

The user has requested enhancement of the downloaded file. VOL. 34, NO. 4 SOUTHWESTERN ENTOMOLOGIST DEC. 2009

CONTENTS

SEASONALITY AND MOVEMENT OF ADVENTIVE POPULATIONS OF THE ARUNDO WASP (HYMENOPTERA: EURYTOMIDAE), A BIOLOGICAL CONTROL AGENT OF GIANT REED IN THE LOWER RIO GRANDE BASIN IN SOUTH TEXAS Alexis E. Racelis, John A. Goolsby, and Patrick Moran…………………………….347

PRE-RELEASE ASSESSMENT OF IMPACT ON ARUNDO DONAX BY THE CANDIDATE BIOLOGICAL CONTROL AGENTS TETRAMESA ROMANA (HYMENOPTERA: EURYTOMIDAE) AND RHIZASPIDIOTUS DONACIS (HEMIPTERA: DIASPIDIDAE) UNDER QUARANTINE CONDITIONS John A. Goolsby, David Spencer, and Linda Whitehand……………………………359

ECONOMIC IMPLICATIONS FOR THE BIOLOGICAL CONTROL OF ARUNDO DONAX: RIO GRANDE BASIN Emily K. Seawright, M. Edward Rister, Ronald D. Lacewell, Dean A. McCorkle, Allen W. Sturdivant, Chenghai Yang, and John A. Goolsby…………...372

COST-BENEFIT ANALYSIS OF SORGHUM MIDGE, STENODIPLOSIS SORGHICOLA (COQUILLETT)-RESISTANT SORGHUM HYBRID RESEARCH AND DEVELOPMENT IN TEXAS Tebkew Damte, Bonnie B. Pendleton, and Lal K. Almas……………………………390

EFFECTIVENESS OF SPRING BURNING AS A PHYSICAL MANAGEMENT TACTIC FOR THRIPS IN PHLEUM PRATENSE L. (POALES: POACEAE) Dominic D. Reisig, Larry D. Godfrey, and Daniel D. Marcum………………………402

THRIPS (THYSANOPTERA: THRIPIDAE) ON COTTON IN THE LOWER RIO GRANDE VALLEY OF TEXAS: SPECIES COMPOSITION, SEASONAL ABUNDANCE, DAMAGE, AND CONTROL S. M. Greenberg, Tong-Xian Liu, and J. J. Adamczyk………………………………412

TRITROPHIC INTERACTIONS AMONG HOST PLANTS, WHITEFLIES, AND PARASITOIDS Shoil M. Greenberg, Walker A. Jones, and Tong-Xian Liu………………………….426

OVIPOSITION CHARACTERISTICS OF PECAN WEEVIL Michael W. Smith and Phillip G. Mulder………………………………………………442

PHYLOGENETIC ANALYSIS OF HEAT SHOCK PROTEINS IN GLASSY-WINGED SHARPSHOOTER, HOMALODISCA VITRIPENNIS Henry Schreiber IV, Daymon Hail, Wayne Hunter, and Blake Bextine……………452

i

A NEW METHOD FOR COLLECTING CLEAN STABLE FLY (DIPTERA: MUSCIDAE) PUPAE OF KNOWN AGE Dennis R. Berkebile, Anthony P. Weinhold, and David B. Taylor………………….464

ESTERASES IN AEDES ALBOPICTUS (SKUSE) FROM NORTHEASTERN MEXICO Gustavo Ponce-Garcia, Mohammad Badii, Mercado Roberto, and Adriana E. Flores………………………………………………………………………..472

SCIENTIFIC NOTE:

A VISUAL GUIDE FOR IDENTIFICATION OF EUSCHISTUS SPP. (HEMIPTERA: PENTATOMIDAE) IN CENTRAL TEXAS Jesus F. Esquivel, Roger M. Anderson, and Robert E. Droleskey………………….480

FORTUITOUS ESTABLISHMENT OF RHYZOBIUS LOPHANTHAE (COLEOPTERA: COCCINELLIDAE) AND APHYTIS LINGNANESIS (HYMENOPTERA: ENCYRTIDAE) IN SOUTH TEXAS ON THE CYCAD AULACASPIS SCALE, AULACASPIS YASUMATSUI (HEMIPTERA: DIASPIDIDAE) Daniel Flores and Jason Carlson………………………………………………...... 484

SUBJECT INDEX TO VOLUME 34…………………………………….……..………488

AUTHOR INDEX TO VOLUME 34…………………………………………………….496

APPLICATION FOR MEMBERSHIP………………………………………………….498

STATEMENT OF OWNERSHIP, MANAGEMENT, AND CIRCULATION………...500

ii VOL. 34, NO. 4 SOUTHWESTERN ENTOMOLOGIST DEC. 2009

Seasonality and Movement of Adventive Populations of the Arundo Wasp (Hymenoptera: Eurytomidae), a Biological Control Agent of Giant Reed in the Lower Rio Grande Basin in South Texas

Alexis E. Racelis, John A. Goolsby, and Patrick Moran

USDA-ARS, Beneficial Insects Research Unit, 2413 E. Highway 83. Weslaco, TX 78596

Abstract. The arundo wasp, Tetramesa romana Walker, has been permitted as a biological control agent for the invasive perennial grass, the giant reed, Arundo donax L. Evidence of adventive populations of the arundo wasp in the Lower Rio Grande Basin was confirmed with a spatio-temporal survey spanning more than 350 river miles. A total of 2,414 adult females of T. romana was collected during a 14- month period of study in 2008-2009. This study documents the initial locations and regional expansion of two adventive populations of T. romana, centered around the cities of Eagle Pass and Laredo, TX. Peaks in T. romana abundance in August 2008 and June 2009 indicate a region-wide positive association between abundance of T. romana and warm summer temperatures. Correlations between site-specific abundance data and weather suggest the presence of population- specific associations with both temperature and rainfall.

Introduction

Giant reed, Arundo donax L. (Poaceae), is a perennial grass introduced to the Americas by Spanish settlers from Mediterranean Europe, possibly in the 1600s (Dunmire 2004). It thrives in warm temperate and subtropical climates and has increased prolifically throughout the southwestern U.S. and in Mexico, especially along canals, riparian areas, roadsides, and other waterways. Giant reed is particularly invasive in the Lower Rio Grande Basin along the Mexico-U.S. border where it infests more than 30,000 ha of riparian habitat (Yang et al. 2009). Growing in dense stands along the Rio Grande, and capable of rapid stand expansion (Thornby et al. 2007, Spencer et al. 2008), giant reed poses an economic threat to water supplies (Seawright et al. 2009), riparian biodiversity, waterway access, and border security (reviewed in Goolsby and Moran 2009). Research to develop a biological control program for giant reed in North America has been underway since 2001. Several potential biological control agents have been identified and tested, including the arundo wasp, Tetramesa romana Walker (Hymenoptera: Eurytomidae) (Goolsby and Moran 2009), which was permitted for release in 2009. Tetramesa romana is a stem-galling eurytomid wasp native to Europe and North Africa that reproduces primarily via parthenogenesis (Askew 1984, Moran and Goolsby 2009). Male wasps are rare. Females deposit eggs into the stem of giant reed, and larval development induces gall formation, significantly impacting plant growth and development as documented in quarantine and small-scale field studies (Goolsby et

347 al. 2009b, J. Goolsby and A. Racelis, unpublished data). Although large inundative releases are planned, no significant populations of T. romana have been released to date. In December 2007, galls and exit holes indicative of larval damage and adult emergence by T. romana were found on patches of giant reed along Shoal Creek in central Austin, TX (Goolsby et al. 2009a). Exit holes and galls were also detected on giant reed along the Rio Grande River in Laredo, TX, in February 2008. Stems from these populations were collected and adult arundo wasps emerged, confirming the presence of T. romana at each location. These adventive populations provided a singular opportunity to monitor and study the movement and seasonality of the arundo wasp and how this biological control agent may be affected by key abiotic factors such as temperature or rainfall. Understanding the relationship between weather and biological control organisms can be helpful in the success of biological control programs (Goeden and Andres 1999, Goolsby et al. 2005). Many studies have corroborated that climate and weather can influence both movement and dispersal of insects and are important factors that affect biological processes and distribution of insects (Williams and Liebhold 2002, Peacock et al. 2006, Sims-Chilton et al. 2009). A substantial portion of these studies focuses on the effects of weather on population dynamics of invasive species (Christian and Keith 2008, Dale et al. 2009, Sutherst and Bourne 2009). Recently, particular attention has been placed on understanding possible climatic effects on invasive species and their biological control agents (Herrera et al. 2005, Corn et al. 2009), in light of anthropogenically-induced climate change in some cases (Beaumont et al. 2009). Spencer et al. (2008) found that regional differences in rainfall and temperature influence the phenology and production of giant reed. Because T. romana prefers phenologically labile shoot tips (Moran and Goolsby 2009), a similar trend in populations of T. romana could be expected because of both variation in availability of shoot tips for oviposition and the possible effects of temperature and moisture on development of gall tissues. Increased availability of moisture may lead to more robust development of apical and lateral shoots (Quinn and Holt 2007). For example, new shoots initiated in July 2008 in Laredo and Mission, TX, showed an increase in formation of side shoots in the following late winter/early spring (P. Moran, unpublished data). Additionally, Moran and Goolsby (2009) suggested that seasonal temperature variation can affect development of the final larval instar of T. romana. We therefore hypothesized that adventive populations of T. romana in the Lower Rio Grande Basin will have some relationship to fluctuations in temperature and rainfall. This paper reports data on abundance of adult wasps across a temporal and spatial gradient along the Rio Grande in southern Texas. We then used these correlations of wasp abundance with rainfall and temperature to provide an indication of when releases of biological control agents might be most effective.

Materials and Methods

Site Selection. Populations of T. romana were studied across seven counties that span the Lower Rio Grande Basin in South Texas (Table 1, Fig. 1). Sites were established within 50 m of the edge of the Rio Grande River, where giant reed was abundant and where access was granted by private and governmental landowners. At the beginning of the survey, we established 21 field sites from Del Rio to Brownsville, TX, but after one year of sampling we limited our survey to two locations within 80 km to locations (within ~50 miles) where T. romana populations

348 Table 1. Geographical Location of Field Sites Established in April 2008 in the Lower Rio Grande Basin Site number Site name County Location 1 Sycamore Creek Val Verde N 29°14.633 W 100°47.574 2 Las Moras Maverick N 29°01.369' W 100°39.329' 3 Eagle Pass Boat Ramp Maverick N 28°42.525' W 100°30.613' 4 Bill George Ranch Maverick N 28°35.168' W 100°23.930' 5 El Cinco Ranch Maverick N 28°30.287' W 100°21.003' 6 Comanche Ranch Webb N 28°14.184' W 100°13.574' 7 Farco Mines Webb N 27°50.055' W 99°52.511' 8 Colombia Bridge Webb N 27°42.007' W 99°44.662' 9 La Bota Ranch Webb N 27°36.201' W 99°34.875' 10 McNaboe Park Webb N 27°35.261' W 99°31.908' 11 Herbst River Vega Zapata N 27°13.052' W 99°26.348' 12 San Ygnacio Zapata N 27°05.574' W 99°25.671' 13 Roma Bridge Starr N 26°24.256' W 99°01.117' 14 Rio Grande Bridge Starr N 26°20.290' W 98°47.265'

had been detected. Estimates of density and dispersal for the 14 remaining sites are reported in this paper. Additional sites to the west (Comstock, TX; Amistad Dam and Del Rio Bridge, Del Rio, TX (Val Verde County)) and east (North American Butterfly Association Park, Mission, TX; Hidalgo County Irrigation District #9 Reservoir, Mercedes, TX; Los Indios, TX; and Brownsville, TX) of the adventive wasp area of distribution were also surveyed. No T. romana were detected at these sites for more than a year, and hence data are not reported from these sites. Data Collection. Five sticky traps (yellow, 23 by 14 cm, unbaited Pherocon® AM, Trécé, Inc., Adair, OK) were placed at each monitored site. Where possible, traps were placed along a trail perpendicular to the river (5-50 m from the edge of the water) to account for variations in microclimate and environment. A twist tie was used to suspend each trap from a giant reed plant approximately 1 to 1.5 m above the ground. Traps were collected and replaced monthly from April 2008 until June 2009, except for April 2009. Traps were taken to a laboratory where they were examined at 10X magnification to count all adult females of T. romana. Only a few male wasps were captured in traps and are not reported. Total numbers of adult female wasps per trap, and means across five traps were calculated for each site for each collection period. At some sites, traps were lost because of misplacement, fire, or inadvertent mowing of the stand. In these cases, the monthly data point was discarded unless at least three traps remained. To explore the relationship between temperature and T. romana across all 14 sites, we used weather data from Laredo, TX, as a proxy for the average climatic data for the region, because it correlates well with weather data from other areas in the study and because reliable weather data were not available at sites where the presence of T. romana was recorded. For sites in the vicinity of Eagle Pass and Laredo, TX, we calculated monthly average daily air temperatures and total monthly rainfall using data from nearby airport weather stations: Piedras Negras Airport (MMPG), Piedras Negras, Coahuila, Mexico, and Laredo International Airport (KRLD), Laredo, TX (N28°37.6, W100°32.1, N27°32.6, W99°27.7, respectively).

349

Fig. 1. Map of study sites in the Lower Rio Grande Basin. Dots refer to sites in the original survey, where T. romana populations went undetected for more than one year. Numbers correspond to sites described in Table 1.

Rainfall data for Eagle Pass was provided by R. Cooley Maverick Irrigation District, Eagle Pass, TX. Air temperature and rainfall were correlated with locations (SYSTAT 2002). Because Eagle Pass and Laredo were the original areas where adventive populations were first recorded and had steady populations throughout the study period, we used data from these weather stations and the two sites in each city to determine correlations between rainfall or temperature and local populations of T. romana. Both weather data and monthly observation data did not meet normality requirements, and thus we used Spearman rank order correlations (SYSTAT 2002) to explore the relationship between weather and fluctuations in abundance of T. romana. We also correlated population data with weather data both one month and two months prior.

350 Results and Discussion

Our study confirms the presence of established, adventive T. romana populations in the Lower Rio Grande Basin. During the 14-month survey, we captured and recorded a total of 2,414 T. romana females at 10 of the 14 sites (Table 2). Populations of T. romana were most readily detected around the city of Eagle Pass (Sites 3-5) in Maverick County, and near the city of Laredo (Sites 8-10) in Webb County, TX. Wasps were scarce or undetectable at two sites (Sites 6 and 7) between these two cities, verifying that Eagle Pass and Laredo are likely where adventive populations may have first established and from where populations have since radiated. No insects were detected at Site 6 (Comanche Ranch) until the final month of our survey. Additionally, no T. romana were collected for the first 10 months of observation in San Ygnacio (Site 13), but were detected in our traps at the end of our study. The results for Sites 6 and 13 suggest that T. romana dispersed about 25 river-km in 12-14 months. This detected expansion of T. romana may indicate that populations are slowly establishing southward from the original two adventive populations. We found that total T. romana populations across the entire study area peaked in August 2008 and began to peak again in June 2009, when we captured a total of 688 and 582 females, respectively. These collections correspond with summer months when the average daily air temperature was consistently warmer than 30°C (Fig. 2). Populations of T. romana were largely undetectable during the winter (January and February) at all sites (Table 2). This may reflect cool temperature- or short daylength-triggered diapause in T. romana, similar to that experienced in other biological control insects (Bean et al. 2007). However, Moran and Goolsby (2009) suggested that mildly cool temperatures could decelerate the development of late larval stages of non-diapausing T. romana, which may explain the fewer adults collected in the winter. Populations of T. romana were correlated with general temperature patterns across the entire study area in the Lower Rio Grande Basin (rs = 0.686, p < 0.01), but were not correlated to rainfall in Laredo (Table 3). Temperature at the four sites around Laredo and Eagle Pass was very correlated (rs = 0.975, p < 0.000001), similar to the region as a whole where peaks in temperatures tended to occur during the late spring and summer months. Rainfall was marginally correlated with average air temperature at Eagle Pass (rs = 0.532, p < 0.05), but not at Laredo (rs = 0.289, p = 0.308). We documented a 1-2 month lag between the June 2008 temperature peak and peak wasp abundance per trap (Fig. 3). Monthly collections at the Eagle Pass Boat Ramp and McNaboe Park correlated significantly with average air temperature (rs = 0.646, p < 0.05 and rs = 0.564, p < 0.05, respectively) (Table 3). Abundance at these four sites peaked in August 2008, followed by a drop to near- or completely-undetectable levels in the cooler winter months. A new peak in wasp abundance began in May-June 2009, corresponding to a large (± 10°C) increase in average daily temperature during this period. Although we would have expected a similar increase in 2009, abundance decreased in July, likely linked to historic, extreme temperatures in the Lower Rio Grande Basin where average air temperature for July was 34°C and average high daily temperatures exceeded 41°C. This may suggest that T. romana is not only sensitive to normal fluctuations in temperature, but also may have an upper-limit temperature threshold. Although warmer temperatures correlate with more T. romana, atypical extreme temperatures may negatively affect extant populations.

351

Table 2. Total Number of T. romana Collected in Sticky Traps at Each Field Site from April 2008 to June 2009 (See Table 1 for Location Detail) Apr May Jun Jul Aug Sep Oct Nov Dec Jan Feb Mar May Jun Total Site 08 08 08 08 08 08 08 08 08 09 09 09 09 09 by site 1 Sycamore Creek 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 2 Las Moras 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 3 Eagle Pass Boat Ramp 30 20 5 1 15 0 5a 0 0 0 1 0 7 46 130 4 Bill George Ranch 21 5 0 0 97 1 5 0 0 0 -- 0 0 0 129 5 El Cinco Ranch 0 1 0 3a 27 2 2 0 0 0 0 1 0 22 58

352 6 Comanche Ranch 0 0 0 0 0 0 0 0 0 0 0 0 0 3 3 7 Farco Mines 0 0 0 0 1 0 1 0 2 0 0 0 0 2 6 8 Columbia Bridge 24 65 18 4 67 0 33 14 22 0 0 18 27 74 366 9 La Bota Ranch 76 54 41 3 123 0 8a 18 16 6 0 31 21 99 496 10 McNaboe Park 126 44 25 136 341 1 9 26 13 1 0 40 55a 58 875 11 Herbst Vega -- 0 3 2 9 1 0 0a 0 0 0 26 32 259 332 12 San Ygnacio -- 0 0 0 0 0 0 0 0 0 0 2 3 14 19 13 Roma 0 0 0 0 0 0a 0 0 0a 0 0 0 0 0 0 14 Rio Grande City -- 0 0 0 0 0a 0 0 0 0 0 0 0 0 0 Total by Month 277 189 92 149 680 5 63 58 53 7 1 118 145 577 2,414 aAt least 3 of 5 traps were collected.

800 34 Total T. romana collected (all sites) Average Air Temp - KLRD 32

30 600 28

collected 26

400 24

T. romana 22 Temperature (C) 20 200 Total no. 18

16

0 14

8 8 8 8 8 8 8 8 8 9 9 9 9 9 0 0 0 0 0 0 0 0 0 0 0 0 0 0 r y e y g p t v c n b r y e p l c a a n u u e o e a e a n A O M u J A S N D J F M M u J J

Fig. 2. Total monthly collections of T. romana across all sites in the Lower Rio Grande Basin. Monthly average air temperatures from Laredo International Airport (KRLD).

Table 3. Spearman Rank Order Correlations Between Monthly Collections of T. romana Populations and Monthly Weather Data at Eagle Pass (Sites 3 and 4) and Laredo (Sites 9 and 10), and Across All Sites (1-14), Including 1 Month and 2-Month Time Lagged Correlations with Rainfall. Reported Are Values for Correlation Coefficient (rs) and Significance (p value) Average air Rainfall Site temperature Actual 1-month lag 2-month lag (3) Eagle Pass Boat Ramp 0.647 0.646 0.397 0.121 P (0.012)a (0.012)a (0.173) (0.699) (4) Bill George Ranch 0.177 0.501 0.269 0.433 P (0.532) (0.064) (0.362) (0.904) (9) La Bota Ranch 0.471 -0.037 0.065 -0.510 P (0.084) (0.892) (0.821) (0.084) (10) McNaboe Park 0.564 0.488 0.000 -0.294 P (0.034)* (0.072) (0.993) (0.340) (1-14) All sites 0.686 0.372 0.226 -0.301 P (0.0050)* (0.184) (0.447) (0.329) aSignificant at P = 0.05 by Spearman rank order correlation.

353 T. romana and wateravailabilityaffect phenology andgrowthofgiantreed(QuinnHolt development of Eagle Passrevealedthatrain mightalsobealocallyimportantfactorgoverning pertraptorainfallin Laredo and specific correlationsofmonthlywaspcaptures respectively. Airport (KRLD)andPiedrasNegras(MMPG)for Laredo andEaglePass, and EaglePass(bottom).Weatherdatawerecompiledfrom LaredoInternational Fig. 3.Fluctuationsinadventive Mean no. T. romana per trap +/- SD 100 120 140 160 20 40 60 80 10 20 30 40 50 0 0 Although a region-wide correlation to monthly rainfall wasnotapparent, Although aregion-widecorrelationto

probablywouldinvolveatime-lag ofonetotwomonths,becauserainfall A pr 08 Ma y T. romana 08 Ju ne 08 Ju ly 08 Au . Adirectcorrelationbetween rainfallandabundanceof g 08 S ep 08 Oct T. romana 08 N Total Rainfall-KRLD Rainfall-KRLD Total Average Air Temp-KLRD Park McNaboe (10) (9) La Bota Ranch ov Rainfall Total MMPG - Temp Air Average Ranch BillGeorge (4) BoatRamp EaglePass (3) 08

354 De c 08 J populationsaroundLaredo,TX(top) an 09 Fe b 09 M ar 09 M ay 09 Ju ne 09 14 16 18 20 22 24 26 28 30 32 34 12 14 16 18 20 22 24 26 28 30 32 34

Average Air Temperature (monthly, C) 0 2 4 6 8 10 12 14 16 18 20 0 2 4 6 8 10

Total Monthy Rainfall (cm) 2007), and increased development of apical and lateral shoots would lead to more abundant adult wasps from galls already initiated at the time of the rain event about one month after the rainfall peak (Moran and Goolsby 2009). However, neither one- month nor two-month lagged-time correlations revealed a significant relationship (Table 3). Only actual month correlations revealed a marginally significant relationship with monthly rainfall (rs = 0.646, p < 0.05 at Eagle Pass Boat Ramp). Site- or city-specific positive associations between rainfall and abundance of T. romana were likely obscured by the region-wide effects of temperature in our study. In addition, too much rainfall could have been directly detrimental to T. romana. Torrential rains in August 2008 preceded a precipitous drop in T. romana across all sites, although average daily temperatures in September and October 2008 remained within a few degrees of the mildly warm level (27°C) used in quarantine studies of T. romana (Moran and Goolsby 2009). Adult mortality via injury or drowning may have thus occurred during extreme rainfall events. In conclusion, this study documented the apparent initial establishment locations and regional expansion of two adventive populations of T. romana during a 14-month period of study in 2008-2009. The results indicated a region-wide positive association between abundance of T. romana and warm summer temperatures, and suggest the presence of population-specific associations with rainfall. The results show that T. romana is tolerant of and responds positively to the hot summer conditions of the Lower Rio Grande Basin, although it is still unclear how these wasps would respond to extended dry periods of extreme heat that often occur in this region. With the weather regime we observed, spring and early summer seem to be the best times for release of T. romana to maximize the likelihood of establishment and impact on exotic and invasive giant reed.

Acknowledgment

The authors acknowledge Matthew Rector, Reyes Garcia, Juan Ramos, and Bill Warfield for conducting the monthly surveys. The U.S. Border Patrol kindly provided monthly ground and boat access to several remote locations along the Rio Grande. La Bota Ranch, Laredo, TX, allowed frequent access to the river for the studies. Helpful comments to the manuscript were given by Scott Armstrong and Don Thomas, USDA-ARS, Weslaco, TX, and Tom Vaughan, Texas A&M International University, Laredo, TX.

References Cited

Askew, R. R. 1984. The biology of gall wasps, pp. 223-271. In T. N. Ananthakrishnan [ed.], The Biology of Gall Insects. Oxford and IBH Press, New Delhi, India. Bean, D. W., T. L. Dudley, and J. C. Keller. 2007. Seasonal timing of diapause induction limits the effective range of Diorhabda elongata deserticola (Coleoptera: Chrysomelidae) as a biological control agent for tamarisk (Tamarix spp.). Environ. Entomol. 36: 15-25. Beaumont, L. J., R. V. Gallagher, W. Thuiller, P. O. Downey, M. R. Leishman, and L. Hughes. 2009. Different climatic envelopes among invasive populations may lead to underestimations of current and future biological invasions. Diversity and Distributions 15: 409-420.

355

Corn, J. G., J. M. Story, and L. J. White. 2009. Comparison of larval development and overwintering stages of the spotted knapweed biological control agents Agapeta zoegana (Lepidoptera: Tortricidae) and Cyphocleonus achates (Coleoptera: Curculionidae) in Montana versus Eastern Europe. Environ. Entomol. 38: 971-976. Dale, V. H., K. O. Lannom, M. L. Tharp, D. G. Hodges, and J. Fogel. 2009. Effects of climate change, land-use change, and invasive species on the ecology of the Cumberland forests. Can. J. Forest Res. 39: 467-480. Dunmire, W. W. 2004. Gardens of New Spain: How Mediterranean Plants and Foods Changed America. University of Texas, Austin, TX. Goeden, R. D., and L. A. Andres. 1999. Biological control of weeds in terrestrial and aquatic environments, pp. 871-890. In T. S. Bellows and T.W. Fisher [eds.], Handbook of Biological Control. Academic Press, New York. Goolsby, J. A., and P. J. Moran. 2009. Host range of Tetramesa romana Walker (Hymenoptera: Eurytomidae), a potential biological control of giant reed, Arundo donax L. in North America. Biological Control 49: 160-168. Goolsby, J. A., R. D. Jesusdasan-Alexander, H. Jourdan, B. Mutharaj, A. S. Bourne, and R. W. Pemberton. 2005. Continental comparisons of the interaction between climate and the herbivorous mite Floracarus perrepae (Acari: Eriophyidae). Florida Entomol. 88: 129-134. Goolsby, J. A., P. J. Moran, J. Faulk, and L. Gilbert. 2009a. Distribution and spread of an adventive population of the biological control agent, Tetramesa romana, in Austin, TX. Southwest. Entomol. 34: 329-330. Goolsby, J. A., D. Spencer, and L. Whitehand. 2009b. Pre-release assessment of impact on Arundo donax by the candidate biological control agents, Tetramesa romana (Hymenoptera: Eurytomidae) and Rhizaspidiotus donacis (Homoptera: Diaspididae) under quarantine conditions. Southwest. Entomol. 34. Herrera, A. M., D. D. Dahlsten, N. Tomic-Carruthers, and R. I. Carruthers. 2005. Estimating temperature-dependant developmental rates of Diorhabda elongata (Coleoptera: Chrysomelidae), a biological control agent of saltcedar (Tamarix spp.). Environ. Entomol. 34: 775-784. Moran, P. J., and J. A. Goolsby. 2009. Biology of the galling wasp Tetramesa romana, a biological control agent of giant reed. Biological Control 49:169- 179. Peacock, L., S. Worner, and R. Sedcole. 2006. Climate variables and their role in site discrimination of invasive insect species distributions. Environ. Entomol. 35: 958-963. Sims-Chilton, N., M. P. Zalucki, and Y. M. Buckley. 2009. Patchy herbivore and pathogen damage throughout the introduced Australian range of groundsel bush, Baccharis halimifolia, is influenced by rainfall, elevation, temperature, plant density and size. Biological Control 50: 13-20. Seawright, E. K., M. E Rister, R. D. Lacewell, D. A .McCorkle, A. W. Sturdivant, C. Yang, and J. A. Goolsby. 2009. Economic implications for the biological control of Arundo donax in the Rio Grande Basin. Southwest. Entomol. 34. Spencer, D. F., R. K. Stocker, P. S. Liow, L. C. Whitehand, G. Ksander, A. M. Fox, J. H. Everitt, and L. D. Quinn. 2008. Comparative growth of giant reed (Arundo donax L.) from Florida, Texas, and California. J. Aquatic Plant Manage. 46: 89-96.

356

Sutherst, R. W., and A. S. Bourne. 2009. Modelling non-equilibrium distributions of invasive species: a tale of two modelling paradigms. Biological Invasions 11: 1231-1237. SYSTAT. 2002. SYSTAT for Windows, Version 10.2. SYSTAT Software Inc., Richmond, CA. Thornby, D., D. Spencer, J. Hanan, and A. Sher. 2007. L-DONAX, a growth model of the invasive weed species, Arundo donax L. Aquatic Botany 87: 275-284. Ulrichs, C., and K. Hopper. 2008. Predicting insect distributions from climate and habitat data. BioControl 53: 881-894. Williams, D. W., and A. M. Liebhold. 2002. Climate change and the outbreak ranges of two North American bark beetles. Agric. Forest Entomol. 4: 87-99. Yang, C., J. A. Goolsby, and J. H. Everitt. 2009. Using QuickBird satellite imagery to estimate giant reed infestations in the Rio Grande Basin of Mexico. J. Applied Remote Sensing 3: 033530.

357 VOL. 34, NO. 4 SOUTHWESTERN ENTOMOLOGIST DEC. 2009

Pre-release Assessment of Impact on Arundo donax by the Candidate Biological Control Agents Tetramesa romana (Hymenoptera: Eurytomidae) and Rhizaspidiotus donacis (Hemiptera: Diaspididae) under Quarantine Conditions

John A. Goolsby1,4, David Spencer2, and Linda Whitehand3

Abstract. Impact by two potential biological control agents, Tetramesa romana Walker and Rhizaspidiotus donacis (Leonardi), on the invasive weed, giant reed, Arundo donax L., was assessed in a quarantine greenhouse before release. Tetramesa romana alone and T. romana plus R. donacis significantly damaged A. donax by suppressing leaf and stem lengths and stimulated production of side branches during a 12-week period. R. donacis plus T. romana only slightly more impacted the plant than did T. romana alone, most likely because of the longer life cycle of R. donacis that may require a longer period of time to cause measurable damage. No negative interactions were observed between the two candidate biological control agents. Therefore, based on their potential to significantly damage A. donax under greenhouse conditions and their narrow host ranges, T. romana and R. donacis are suitable candidates for biological control of this invasive reed grass in North America.

Introduction

Pre-release assessment of impact caused by candidate agents is one of the tools that may be useful in the prioritization and selection process in a biological control program. Prioritizing the agents that show the greatest potential to control the target organism could potentially reduce the likelihood that ineffective agents are released (Hoddle and Syrett 2002, Balciunas 2004, McClay and Balciunas 2005). Limiting the number of species released may reduce risk to non-target species and improve success (McEvoy and Coombs 1999, Strong and Pemberton 2001). This paradigm is now part of the North American Technical Advisory Group for Biological Control of Weeds (TAG) and the International Code of Ethics for Biological Control of Weeds Practitioners (Balciunas and Coombs 2004). However, effective agents could be overlooked if the testing process is not realistic. Few biological control programs have attempted to quantitatively assess the impact of a candidate biological control agent on a weed target before release. Goolsby et al. (2004) reviewed these studies that primarily rely on comparisons of ______1USDA-ARS, Beneficial Insects Research Unit, Weslaco, TX. 2USDA-ARS, Exotic and Invasive Weeds Research Unit, Davis, CA. 3USDA-ARS Biometrical Services, Albany, CA. 4Corresponding author: John A. Goolsby, Ph.D., United States Department of Agriculture, Agricultural Research Service, Kika de la Garza Subtropical Agricultural Research Center, Beneficial Insects Research Unit. email: [email protected].

359 standing biomass, seed production, and/or changes in plant architecture. Most of the studies were done under field conditions in the native range (Waloff and Richards 1977, Balciunas and Burrows 1993, Sheppard et al. 1995, Briese 1996, Conrad and Dhileepan 2007). Few attempted to measure impact under quarantine conditions. Yet in most cases, this is the setting in which most biological control practitioners must conduct their research (Kleinjan et al. 2003, Baliciunas and Smith 2006, Williams et al. 2008). Giant reed, Arundo donax L., is the target of a biological control program in the southwestern U.S. and Mexico (Goolsby et al. 2008). As part of the biological control program, data on the impact of the agents was gathered from field observations in the native range (A. Kirk, unpublished) and from quarantine studies in a greenhouse. The target, a reed grass, is novel in that few grasses have been the target of biological control programs. In addition, preliminary observations of the primary herbivores from the native range -- a stem-galling eurytomid wasp, Tetramesa romana Walker; a rhizome and stem-feeding armored scale, Rhizaspidiotus donacis (Leonardi); shoot-feeding chloropid flies, Cryptonevra spp.; and a leafmining cecidomyid, Lasioptera donacis (Gagne), were not definitive as to which agent or combination would be most effective. Two of the agents, T. romana and R. donacis, were selected for further evaluation based on studies of their distributions, climatic adaptation to the Rio Grande Basin, seasonality, and field impact in their native range (Kirk, unpublished data). Adventive populations of T. romana were discovered in Texas and California (Goolsby 2008b, Goolsby et al. 2009a). The Texas population was brought into quarantine facilities for comparison to the imported European populations. The biology and host range of T. romana and R. donacis were documented in quarantine studies (Goolsby 2008a,b; Goolsby and Moran 2009; Moran and Goolsby 2009, Goolsby et al. 2009b). T. romana oviposits into growing stems and lateral shoots. Larvae feed on gall tissue that forms near the site of oviposition. The life cycle takes approximately 1 month at 27°C. The scale life cycle takes 5-6 months from egg to reproductive adult. Adult females produce live crawlers that disperse short distances to settle on the rhizome, leaf collars, or stem nodes, becoming sessile nymphs covered with characteristic armor. Adult males emerge to mate after 2 months. Females continue to feed and produce live crawlers at approximately 5-6 months. In these studies, the impacts of the two candidate agents were evaluated to: 1) confirm that abundant agents could significantly damage the target in a short period of time under ideal conditions and 2) determine if the combination of the two agents would result in negative interactions between them.

Materials and Methods

The studies in the greenhouse were at the USDA-APHIS, Arthropod Biological Control Containment Facility, Moore Airbase, Edinburg, TX, between January and July 2008. Arundo donax plants were freshly dug from sites near Laredo, TX. Robust growth from freshly dug rhizomes is common, and therefore no fertilizer is needed. Rhizomes of the same size were planted in play sand in 25-cm diameter pots and held until emergence of new shoots. Ten plants of similar size were selected for each treatment and held for 12 weeks at 25.1 ± 4.2 SD ºC and 47.2 ± 11.8 SD relative humidity with natural day-lengths. Most plants had three stems growing from the rhizome. Plants were watered with a drip system, with 2 liters of water three times per week. Plants were measured by hand with a 3-m

360 retractable measuring tape at the start of the test and at 2-week intervals. Plant attributes measured were: stem length; leaf length and SPAD (measures green light reflectance, which is an approximation of chlorophyll content or photosynthetic rate); number of lateral shoots; and number of galls per exit hole. Temperature and relative humidity were measured daily. Plants were placed into a cage with two compartments (with and without insects) at the end of the quarantine greenhouse, which was screened with black silk organza. Plants were exposed to full sunlight in the greenhouse on all but the two sides partitioned by black organza. The test plants were exposed to three treatments: T. romana alone, T. romana + R. donacis, and no insects (check). Tetramesa romana were able to freely disperse among the 20 plants in the cage with only insects. Rhizaspidiotus donacis were released as crawlers and were not able to disperse from the potted plants. Tetramesa romana used in the tests were mostly (3,669 or 85%) from field collections of an adventive population near Austin, TX, with 10% from quarantine colonies of T. romana collected in Spain (Goolsby and Moran 2009). Adventive populations of T. romana were discovered in 2006 (Dudley et al. 2006) and 2008 (Goolsby and Moran 2009) in California and Texas, respectively. Individuals from both adventive populations were compared using custom microsatellite markers to European populations collected as part of the biological control program. The adventive populations differed from any of the populations collected in Europe or imported into quarantine for evaluation (J. Manhart, A. Pepper, and D. Tarin, Texas A&M University, College Station, TX, unpublished data). Therefore, the adventive populations do not represent a quarantine breach, but separate introduction(s) of unknown origin. T. romana adults were released daily, with a total of 4,302 females released onto the 20 A. donax plants during the 12-week study. This is a rate of approximately 18 wasps per pot or six wasps per stem per week. Numbers of R. donacis crawlers were not sufficient to study R. donacis alone. R. donacis were from field populations collected near Alicante, Spain. First-instar crawlers were collected from mature scales held in gelatin capsules. Mature females were removed each week from the capsules, leaving only active crawlers. The number of crawlers was counted in 30 capsules from each weekly cohort to estimate the number of crawlers per capsule. The number of capsules was multiplied by the mean number of crawlers per capsule to estimate the total number of crawlers released. From 11 February to 11 April, 57,125 crawlers were released onto the 10 plants in the T. romana + R. donacis treatment. This is a release rate of approximately 158 crawlers per week per stem. For both T. romana and R. donacis, the release rates were intended to evaluate the potential impact of high numbers of the agents under ideal conditions. Three leaves from each plant (top, middle, and bottom) were measured with a Minolta SPAD 502 Chlorophyll Meter (Konica Minolta Photo Imaging (HK), Ltd, Hong Kong) as described by Spencer et al. (2007). SPAD readings from A. donax leaves are strongly correlated with leaf chlorophyll content (Spencer et al. 2008) and leaf nitrogen content (Spencer et al. 2007). This information was collected because both characteristics contribute to leaf photosynthetic capacity, which has been shown to be influenced by herbivory for some species (Novak and Caldwell 1984, Reutuerto et al. 2004). Data were checked for homogeneity of variances and normality of error distributions before further analysis to identify appropriate methods of incorporating heterogeneity into the model. A mixed model analysis was fitted using SAS software, PROC MIXED (Litell et al. 2006), considering the agent and date as fixed

361 effects and plant and stem as random effects. Significance was tested for fixed effects and interactions between them. Measurements on a whole plant basis used “between” and “within plants” variance. When multiple stems were measured, an additional variance of stem within plant was needed. Count variables (number of branches per stem and number of galls per stem), that did not meet the normal distribution assumption, required a Generalized Linear Mixed Model to use a Poisson distribution and were fitted with SAS PROC GLIMMIX (SAS Institute Inc. 2004). Because the same plants were measured across time, a repeated measures covariance structure was used. Tests were considered significant at a probability less than 0.05; exact probability levels for fixed effect tests are shown in the results. When agent or agent x date was significant less than P = 0.05, mean comparisons were computed using Sidek’s adjusted t-tests for the means as estimated by least squares.

Results

The length of the A. donax stem was significantly shorter in the presence of T. romana alone and combination of T. romana plus R. donacis (Fig. 1, Table 1). There was no significant difference between the effect of either of the insect treatments, i.e., either T. romana alone or the combination of T. romana plus R. donacis (Sidek’s adjusted t). The significant interaction term was because stem lengths were similar at the start of the experiment but changed different amounts over time, with the increase for check plants being greater and not leveling off until near the end of the experiment. The number of branches increased over time when A. donax was exposed to the insects (Fig. 2, Table 1). Lengths of the top, middle, and bottom leaves remained short and constant in the presence of the insects, but increased for the checks throughout or at some time period during the experiment (Figs. 3-5, Table 1). Also, in these cases there were no significant differences between the effects of either of the levels of the insect treatments (Sidek’s adjusted t). Arundo donax grown with the insects initially produced more stems than did the check plants (Fig. 6, Table 1), but the number of stems did not change after 6 weeks. This explains the nearly significant (p = 0.054) interaction term. The effects of either level of insect treatment were not significantly different. Leaf SPAD reflectance decreased over time for plants exposed to the insects (Fig. 7, Table 1). However, the decrease was greater for check plants, which were significantly less than either treatment group from the 4th week on, causing the significant interaction term. The number of galls also increased initially for plants exposed to insects and then leveled off (Fig. 8, Table 1). The two insect treatments were not significantly different from each other, but both were significantly different from the check group. One gall was detected in a check plant near the end of the experiment, suggesting a wasp was accidentally allowed into the section of the greenhouse where the check plants were. However, the impacts of the wasps were apparent before this date.

Discussion

Tetramesa romana alone and the combination of T. romana plus R. donacis significantly reduced leaf and stem lengths and caused leaves with greater SPAD values. Both insect treatments reduced stem lengths by 92% by the end of the study. Greater SPAD (photosynthetic output) values imply that the leaves contained

362 363 and Fig. 1.Stemlengths for Stem Length (cm) Rhizaspidiotus donacis 120 150 30 60 90 6E 1A 6A 1P 6P 01MAY 16APR 01APR 16MAR 01MAR 16FEB 0 Mean Control T. romana T. romana + R. donacis romana R. T. + romana T. Control Arundo donax ) inaquarantine greenhouseexperiment. SE grown with and without potential biological controlinsects( grownwithand withoutpotential Tetramesa romana

1.2 Mean SE 1.0

0.8

0.6 364 0.4

0.2 Number of Branches

0.0 16/02 01/03 16/03 01/04 16/04 01/05

Control T. romana T. romana + R. donacis

Fig. 2. Number of branches for Arundo donax grown with and without potential biological control insects (Tetramesa romana and Rhizaspidiotus donacis) in a quarantine greenhouse experiment.

70

60 Mean SE

50

40

30 365 20 Top Leaf Length 10

0 16/02 01/03 16/03 01/04 16/04 01/05

Control T. romana T. romana + R. donacis

Fig. 3. Top leaf length for Arundo donax grown with and without potential biological control insects (Tetramesa romana and Rhizaspidiotus donacis) in a quarantine greenhouse experiment.

70

60 Mean SE

50

40

30 366 20

Middle Leaf Length 10

0 16FEB 01MAR 16MAR 01APR 16APR 01MAY

Control T. romana T. romana + R. donacis

Fig. 4. Middle leaf length for Arundo donax grown with and without potential biological control insects (Tetramesa romana and Rhizaspidiotus donacis) in a quarantine greenhouse experiment.

20 Mean SE 15

10 367

5 Bottom Leaf Length

0 16/02 01/03 16/03 01/04 16/04 01/05

Control T. romana T. romana + R. donacis

Fig. 5. Bottom leaf length for Arundo donax grown with and without potential biological control insects (Tetramesa romana and Rhizaspidiotus donacis) in a quarantine greenhouse experiment.

4.0 3.5 Mean SE 3.0 2.5 2.0

368 1.5 1.0 Number of Stems 0.5 0.0 16FEB 01MAR 16MAR 01APR 16APR 01MAY

Control T. romana T. romana + R. donacis

Fig. 6. Number of stems for Arundo donax grown with and without potential biological control insects (Tetramesa romana and Rhizaspidiotus donacis) in a quarantine greenhouse experiment.

60 Mean SE 50

40

30 369 20

10 Leaf Chlorophyll (SPAD) Leaf Chlorophyll 0 16/02 01/03 16/03 01/04 16/04 01/05

Control T. romana T. romana + R. donacis

Fig. 7. Leaf SPAD index for Arundo donax grown with and without potential biological control insects (Tetramesa romana and Rhizaspidiotus donacis) in a quarantine greenhouse experiment.

grown with and without potential biological control insects ) in a quarantine greenhouse experiment. Arundo donax Rhizaspidiotus donacis SE and Control romana T. donacis R. + romana T. Mean Mean

4 3 2 1 0 02/16 02/23 03/01 03/08 03/15 03/22 03/29 04/05 04/12 04/19 04/26 05/03 Number of Galls of Number Tetramesa romana Fig. 8. Number of galls per plant for (

370

Table 1. Results of Repeated Measures Analysis of Variance for Characteristics of Arundo donax Subjected to Two Levels of Potential Biological Control Agents (Agent) and Measured over Time (Date) Effect NUM df DEN df F-value P > F Stem length Agent 2 35.2 12.82 < 0.0001 Date 5 112 18.17 < 0.0001 Agent x Date 10 225 6.56 < 0.0001 Number of branches1 Agent 2 26.4 3.57 0.043 Date 5 114.6 9.28 < 0.0001 Agent x Date 10 117.8 3.83 0.0002 Top leaf length Agent 2 65 32.27 < 0.0001 Date 5 87.9 19.35 < 0.0001 Agent x Date 10 117 7.13 < 0.0001 Middle leaf length Agent 2 81.1 50.44 < 0.0001 Date 5 155 19.60 < 0.0001 Agent x Date 10 190 9.10 < 0.0001 Bottom leaf length Agent 2 43.9 8.85 0.0006 Date 5 58.5 5.56 0.0003 Agent x Date 10 78.4 2.46 0.013 Number of stems1 Agent 2 25.2 1.11 0.34 Date 5 120.3 4.63 0.0007 Agent x Date 10 123.1 1.88 0.054 Leaf SPAD Agent 2 27.1 12.85 0.0001 Date 5 101 44.44 < 0.0001 Agent x Date 10 108 2.82 0.0037 Number of galls1 Agent 2 27.5 10.29 0.0005 Date 5 96.5 9.87 < 0.0001 Agent x Date 10 80.4 3.78 0.0003 PROC MIXED was used for all calculations except for Branches, Stems, and Galls, which were evaluated using PROC GLMMIX. Values are numerator (NUM) and denominator (DEN) degrees of freedom (df), the F-value, and the probability of a greater F-value (P > F). Degrees of freedom estimated using the Kenward-Roger method. 1Coded by adding 0.5.

more chlorophyll and greater amounts of nitrogen (Spencer et al. 2007, 2008), which likely impacts photosynthetic capacity of remaining leaves and is consistent with previous reports (Retuerto et al. 2004, Dorchin et al. 2006). Damage by T. romana producing galls in shoot meristems stimulated branch production during this 12-week study. This agrees with previously observed response of A. donax to the cutting off of shoots, which resulted in increased branching (Sher et al. 2003). Production of side shoots is desirable because this reduces the height of the stem and provides additional suitable oviposition and feeding sites for T. romana and R. donacis, thus potentially increasing the impact of the agents on the plant. The number of galls in the insect treatments was low. It appeared that successive bouts of oviposition from the newly released wasps may have harmed developing larvae. However, the impact of repeated oviposition appears to have dramatically altered

371 plant growth. Some shoots died after the conclusion of the experiment, which demonstrates the potential for control of A. donax from inundative releases of the agents. Ongoing rearing studies of the wasp have found that two wasps per stem in a one-time release produced large tip galls, which dramatically alter stem elongation and force lateral shoot formation (Goolsby, unpublished data). Therefore, the experimental release rate of 72 wasps per stem during 12 weeks may be much more than needed to cause field-level impacts to A. donax in an inundative release program. Other studies with gall-forming insects have also shown alterations in plant size. Russian knapweed, Acroptilon repens (L.) DC, had reduced shoot length (12- 21%), aboveground biomass (24-25%) and seed production (75-92%) when grown in the field and exposed to either a gall-forming midge, Jaapiella ivannikovi Fedotova, or the wasp, Aulacidea acroptilonica Tyurebaev (Djamankulova et al 2008). The sawfly Euura lasiolepis Smith reduced the number of reproductive buds on shoots of Salix lasiolepis Benth. by 34 to 56% when the number of galls per shoot varied from 0.56 to 2.7 (Sacchi et al. 1988). Goolsby et al. (2004) showed that the leaf-galling mite Floracarus perrepae Knihinicki and Boczek suppressed growth of the climbing fern Lygodium microphyllum (Cav.) R. Br. There is also evidence that the effects of galls on plants may be influenced by environmental conditions. The response of rosin weed, Silphium integrifolium Michx., to apical meristem galls produced by the wasp Antistrophus silphii Gillette differed with growing conditions (Fay et al. 1996). Field-grown plants had reduced height, leaf area, and inflorescence production. They also did not regrow from axillary meristems. Similar plants grown with additional moisture and nutrients in a garden initially showed similar responses but were able to regrow from axillary meristems, resulting in little loss of total biomass or reproductive output. This suggests the possibility that the impacts observed on A. donax in this study may differ with habitat or environmental conditions. A. donax responded to shoot damage by producing additional stems. Successive releases of T. romana adults during the experiment also suppressed regrowth of stems. This is in contrast with two previous reports on the impacts of gall-forming insects. McCrea et al. (1985) noted that the presence of galls caused by the fly Eurosta solidaginis (Fitch) on Solidago altissima L. did not significantly affect final stem height but did slow ramet growth by altering allocation of photosynthate. Djamankulova et al. (2008) reported they did not observe increased formation of new shoots on infested Russian knapweed plants when compared to check plants. The R. donacis plus T. romana combination produced slightly greater plant impacts when compared to T. romana alone. This may be because of few R. donacis or its longer developmental times (6-month life cycle). It is now apparent that a 12-week experiment is not sufficient to fully demonstrate R. donacis. The scale may have also significantly impacted the health of the underground rhizome, which was not examined in this study. It is also possible that the lack of impact was because A. donax photosynthetic rates were elevated for plants exposed to R. donacis. Although there is little information on the effects of scale insects on photosynthesis (Vranjic 1997), they seem to act as photosynthetic sinks and increase rates of photosynthesis. In a recent report, Retuerto et al (2004) demonstrated increased photosynthetic rates by leaves of European holly, Ilex aquifolium, infested by the scale insect Coccus sp. when compared to noninfested leaves.

372 The effects of T. romana and R. donacis in this experiment on immature A. donax (i.e., plants less than 1-year old) were necessarily observed under quarantine conditions (i.e., a sealed, environmentally controlled greenhouse) with an abundant supply of insects. This contrasts with anecdotal reports of limited field impacts from adventive populations of T. romana. While these results in the laboratory clearly demonstrate the damage that can be caused by these potential biological control agents, response of A. donax may differ under field conditions. Plant response to damage by herbivores is variable. Results may also vary between the different populations of T. romana. The number and timing of herbivore-damaging events as well as plant-growth stage, stored nutrients, and soil conditions influence plant response to damage by herbivores (Crawley 1997). Additional study is needed of release rates and interactions among the different populations of T. romana, R. donacis, and A. donax under field conditions in a biological control program. However, timing and numbers released will likely play significant roles in the impacts of these agents in a classical or inundative release program. These studies in a greenhouse showed the utility of measuring plant impacts of candidate biological control agents before release. However, measuring impacts of agents with long life cycles is difficult, especially with a rapidly growing plant such as A. donax. Although the combination of agents did not cause significantly greater impact, it is noteworthy that no negative interactions were observed between the two candidate biological control agents. Further research is needed to document the impact of R. donacis alone. Preliminary observations of attack by R. donacis in quarantine found damage by the scale was significant 4 to 6 months after release. Large colonies of the scale formed at the base of the rhizome, which appeared to further suppress any regrowth, even 8-10 months after release (Goolsby, unpublished data). This is consistent with observations of the impact of R. donacis in the native range. Because of the apparent low dispersal ability of R. donacis, stands with and without the scale can be compared in the native range. Stands with R. donacis have stems that are shorter and smaller in diameter (A. Kirk, unpublished data). Further studies in Europe are measuring the dry weights of 1-m lengths of A. donax rhizomes collected from stands with and without R. donacis. Additional studies are needed to measure the impact of the other candidate agents including the shoot-feeding flies, Cryptonevra spp. and the leafmining fly, L. donacis. The methods presented here should be suitable for measuring the impacts of these candidate agents for A. donax and other large-stemmed grasses, such as the exotic European common river reed, Phragmites australis (Cav.) Trin. ex Steud, that are targets of biological control programs. In conclusion, based on their potential to cause significant damage to A. donax under greenhouse conditions and their narrow host ranges, T. romana and R. donacis are suitable candidates for classical and potentially inundative biological control of the invasive giant reed, Arundo donax, in North America.

Acknowledgment

The authors thank Sarah Snavely, Ann Vacek, and Crystal Salinas for technical support. Quarantine facilities for the research were provided by USDA- APHIS-PPQ-CPHST-PDDML, Edinburg, TX. The biological control agents evaluated in the study were originally discovered and collected by Alan Kirk, USDA- ARS, European Biological Control Laboratory, Montpellier, France. Additional funding for the research was provided by the US Department of Homeland Security,

373 Science and Technology Directorate and the Lower Rio Grande Valley Development Council, Weslaco, TX. Plant material was provided by Hidalgo County Irrigation District #2 and the City of Laredo, TX. Helpful reviews of the manuscript were received from Lincoln Smith and Joe Balciunas, USDA-ARS, Albany, CA.

References Cited

Balciunas, J. K. 2004. Are mono-specific agents necessarily safe? The need for pre-release assessment of probable impact of candidate biocontrol agents, with some examples. In Jim Cullen [ed.], Proceedings, XI International Symposium Biological Control of Weeds, 27 April-2 May 2003, Australian National University, Canberra, ACT, Australia. CSIRO Entomology, Canberra, Australia. Balciunas, J. K., and D. W. Burrows. 1993. The rapid suppression of the growth of Melaleuca quinquenervia saplings in Australia by insects. J. Aquatic Plant Management 31: 265-270. Balciunas, J. K., and E. Coombs. 2004. In E. M. Coombs, J. K. Clark, G. L. Piper, and A. F. Cofrancesco, Jr. [eds.], Biological Control of Invasive Plants in the United States. Oregon State University Press, Corvalis, OR. Balciunas, J., and L. Smith. 2006. Pre-release efficacy assessment, in quarantine, of a tephritid gall fly being considered as a biological control agent for Cape- ivy (Delairea odorata). Biological Control 39: 516-524. Briese, D. T. 1996. Potential impact of the stem-boring weevil Lixus cardui on the growth and reproductive capacity of Onopordum thistles. Biocontrol Science and Technology 6: 251-261. Conrad, K. A., and K. Dhileepan. 2007. Pre-release evaluation of the efficacy of the leaf-sucking bug Carvalhotingis visenda (Heteroptera: Tingidae) as a biological control agent for cat’s claw creeper Macfadyena unguis-cati (Bignoniaceae). Biocontrol Science and Technology 17: 303-311. Crawley, M. J. 1997. Plant herbivore dynamics, pp. 401-474. In M. J. Crawley [ed.], Plant Ecology. Second Edition. Blackwell, Ltd., Osney Mead, Oxford, UK. Djamankulova, G., A. Khamraev, and U. Schaffner. 2008. Impact of two shoot- galling biological control candidates on Russian knapweed, Acroptilon repens. Biological Control 46: 101-106. Dorchin, N., M. D. Cramer, and J. H. Hoffmann. 2006. Photosynthesis and sink activity of wasp-induced galls in Acacia pycnantha. Ecology 87: 1781-1791. Dudley, T.L., Lambert, A., Kirk, A., 2006. Augmentation biological control of Arundo donax, pp. 141-144. In M. S. Hoddle and M. W. Johnson [eds., Proceedings, V California Conference on Biological Control, 25-27 July 2006, Riverside, California, USA. University of California, Riverside, CA. Fay, P. A., D. C. Hartnett, and A. K. Knapp. 1996. Plant tolerance of gall-insect attack and gall-insect performance. Ecology 77: 521-534. Goolsby, J. A. 2008a. Proposed field release of the European stem galler Tetramesa romana Walker (Hymenoptera: Eurytomidae) for biological control of giant reed, Arundo donax L. (Poales: Poaceae) in North America. Release Petition submitted to the USDA-APHIS-TAG, 18 March 2008.

374 Goolsby, J. A. 2008b. Proposed field release of the European rhizome-stem feeding scale Rhizaspidiotus donacis (Leonardi) (Homoptera: Diaspididae), for biological control of giant reed, Arundo donax L. (Poales: Poaceae) in North America. Release Petition submitted to the USDA-APHIS-TAG, 30 March 2008. Goolsby, J. A., and P. J. Moran. 2009. Host range of Tetramesa romana Walker (Hymenoptera: Eurytomidae), a potential biological control of giant reed, Arundo donax L. in North America. Biological Control 49: 160-168. Goolsby, J., P. Moran, A. Kirk, W. Jones, J. Everitt, C. Yang, P. Parker, D. Flores, D. Spencer, A. Pepper, J. Manhart, D. Tarin, G. Moore, D. Watts, and F. Nibling. 2008. Arundo donax – giant reed; an invasive weed of the Rio Grande Basin. Western Weed Science Society Annual Meeting, 12-14 March 2008, Anaheim, CA. Goolsby, J. A., P. J. Moran, J. Falk, and L. Gilbert. 2009a. Distribution and spread of an adventive population of the biological control agent, Tetramesa romana in Austin, Texas. Southwest. Entomol. 34: 329-330. Goolsby, J. A., P. J. Moran, J. A. Adamczyk, A. A. Kirk, W. A. Jones, M. A. Marcos, and E. Cortés. 2009b. Host range of the European, rhizome-stem feeding scale Rhizaspidiotus donacis (Leonardi) (Hemiptera: Diaspididae), a candidate biological control agent for giant reed, Arundo donax L. (Poales: Poaceae) in North America. BioControl Science and Technology (in press). Goolsby, J. A., A. S. Bourne, and R. Zonneveld. 2004. Pre-release assessment of impact on biomass production of an invasive weed, Lygodium microphyllum (Lygodiaceae: Pteridophyta), by a potential biological control agent, Floracarus perrepae (Acariformes: Eriophyidae). Environ. Entomol. 33: 998- 1002. Hoddle, M., and P. Syrett. 2002. Realizing the potential of classical biological control, pp. 395-424. In G. Hallman and C. Schwalbe [eds.], Invasive Arthropods in Agriculture: Problems and Solutions. Science Publishers, Enfield, NH. Kleinjan, C. A., P. B. Edwards, and J. H. Hoffmann. 2003. Impact of foliage feeding by Zygina sp. on tuber biomass and reproduction of Asparagus asparagoides (L): relevance to biological control in Australia. Biological Control 30: 36-41. Litell, R., G. Milliken, W. Stroup, R. L. Wolfinger, and O. Schabenberger. 2006. SAS for Mixed Models, 2nd edition. SAS Institute, Inc., Cary, NC. McClay, A., and J. K. Balciunas. 2005. The role of pre-release efficacy assessment in selecting classical biological control agents for weeds - applying the Anna Karenina principle. Biological Control 35: 197-207. McCrea, K. D., W. E. Abrahamson, and A. E. Weis. 1985. Goldenrod ball gall effects on Solidago altissima: 14C translocation and growth. Ecology 66: 1902-1907. McEvoy, P. B., and E. M. Coombs. 1999. Why things bite back: unintended consequence of biological control of weeds, pp. 167-195. In P. A. Follett and J. J. Duan [eds.], Non-target Effects of Biological Control. Kluwer Academic Publishers, Dordrecht, The Netherlands. Moran, P. J., and J. A. Goolsby. 2009. Biology of the galling wasp Tetramesa romana, a biological control agent of giant reed. Biological Control 49: 169- 179. Novak, R. S., and M. M. Caldwell. 1984. A test of compensatory photosynthesis in the field: implications for herbivory tolerance. Oecologia 61: 311-318.

375 Reutuerto, R., B. Fernandez-Lema, R. Roiloa, and J. R. Obeso. 2004. Increased photosynthetic performance in holly trees infested by scale insects. Functional Ecology 18: 664-669. Sacchi, C. F., P. W. Price, T. P. Craig, and J. K. Itami. 1988. Impact of shoot galler attack on sexual reproduction in the arroyo willow. Ecology 69: 2021-2030. SAS Institute, Inc. 2004. SAS OnlineDoc® 9.1.3. Cary, NC. Sheppard, A. W., J. P. Aeschilimann, J. L. Sagliocco, and J. Vitou. 1995. Below- ground herbivory in Carduus nutans (Asteraceae) and the potential for biological control. Biocontrol Science and Technology 5: 261-270. Sher, A., D. Spencer, S. Sweet, and J. DiTomaso. 2003. Compensatory growth response across two nutrient regimes by invasive Arundo donax. Ecological Society of America Annual Meeting Abstracts 88, 306, Savannah, GA. Spencer, D. F., A. A. Sher, D. Thornby, P.-S. Liow, G. G. Ksander, and W. Tan. 2007. Non-destructive assessment of Arundo donax (Poaceae) leaf quality. J. Freshwater Ecol. 22: 277-285. Spencer, D. F., W. Tan, P.-S. Liow, G. G. Ksander, L. C. Whitehand, S. Weaver, J. Olson, and M. Newhouser. 2008. Evaluation of glyphosate for managing giant reed (Arundo donax) Invasive Plant Science and Management 1: 248- 254. Strong, D. R., and R. W. Pemberton. 2001. Food webs, risks of alien enemies and reform of biological control, pp. 57-74. In E. Wajnberg, J. K. Scott, and P. C. Quimby [eds.], Evaluating Indirect Ecological Effects of Biological Control. CAB International, Wallingford, UK. Vranjic, J. A. 1997. Ecology: effects on host plants, pp. 323-336. In Y. Ben-Dov and C. J. Hodgson [eds.], Soft Scale Insects – Their Biology, Natural Enemies and Control. Elsevier Science, Amsterdam, The Netherlands. Waloff, N., and O. W. Richards. 1977. The effect of insect fauna on growth mortality and natality of broom, Sarothamnus scoparius. J. Applied Ecol. 14: 787-798. Williams, H. E., S. Nesser, and L. G. Madire. 2008. Candidates for biocontrol of Macfadyena unguis-cati in South Africa: biology, host ranges and potential impact of Carvalhotingis visenda and Carvalhotingis hollandi under quarantine conditions. BioControl 53: 945-956.

376 VOL. 34, NO. 4 SOUTHWESTERN ENTOMOLOGIST DEC. 2009

Economic Implications for the Biological Control of Arundo donax: Rio Grande Basin

Emily K. Seawright1, M. Edward Rister1, Ronald D. Lacewell1, Dean A. McCorkle1, Allen W. Sturdivant2, Chenghai Yang3, and John A. Goolsby3

Abstract. Giant reed, Arundo donax L., is a large, bamboo-like plant native to the Mediterranean region. It has invaded several thousand hectares of the Rio Grande riparian habitat in Texas and Mexico. The United States Department of Agriculture- Agricultural Research Service (USDA-ARS) is investigating four herbivore insects as potential biological control agents for giant reed. One of the most important reasons for targeting this invasive weed is to reduce its impact on available water supplies, especially in the Rio Grande Basin. This study examined selected economic implications of this program for agricultural water users in the U.S. The research included (a) estimating the value of the water saved (to agricultural purposes) by reduction of giant reed, (b) benefit-cost analyses, (c) regional economic impact analyses, and (d) an estimate of the per-unit life-cycle cost of water saved during a 50-year planning horizon (2009 through 2058). Positive results related to the benefit-cost ratio, economic impact analyses, and competitive results for the per-unit life-cycle cost of saving water are associated with the biological control project for giant reed. The benefit-cost ratio, calculated with normalized prices, indicates $4.38 of benefits for every dollar of public investment. According to 2009 results for the economic impact analyses, economic output is $22,000, value-added is $11,000, and no employment is supported by the water savings from giant reed. Additionally, the per-unit cost of water saved is $44.08, a value comparable to other projects designed to increase water supply for the region. These results indicate this program will have positive net economic implications for the U.S. and the Lower Rio Grande Valley of Texas.

Introduction

Water supply in the Lower Rio Grande Valley of Texas is an acute issue because the regional economy and population continue to expand at a rapid rate (U.S. Census Bureau 2000). The main source of water for this region is the Rio Grande along the Texas-Mexico border, using two reservoirs -- Amistad near Del Rio and Falcon south of Laredo (Rubinstein 2008). Local water resource managers and community leaders are considering alternative methods to enhance the current water supply for the region. One area of interest is control of the invasive plant species Arundo donax L., commonly called Arundo, or giant reed. ______1Texas A&M University, 2124 TAMU, College Station, TX 77845. 2Texas AgriLife Extension Service, Weslaco, TX 78596. 3United States Department of Agriculture-Agricultural Research Service, Weslaco, TX 78596. This article is from a Master of Science thesis (Seawright 2009).

377 Giant reed is a large, aquatic plant invading the riparian areas of the southwestern U.S., particularly the Rio Grande Basin and California (Tracy and DeLoach 1999, Goolsby and Moran 2009). Giant reed is native to the Mediterranean, Arabian Peninsula, and south Asia (Perdue 1958). The Rio Grande Basin is ideal for establishment and invasion by the weed because of the climatic similarities with its native range and lack of specialist herbivores (Tracy and DeLoach 1999, Goolsby and Moran 2009). Giant reed often grows in dense stands in riparian zones and reproduces solely through vegetative propagation (Decruyenaere and Holt 2001). The plant can grow 5-8 m tall (Bell 1997), exhibits a growth rate approaching 10 cm per day (Dudley 1998), and consumes large quantities of water to support its rapid growth. In Mexico, it is called ‘el ladron de agua,’ the water thief. The data obtained from literature on water use of giant reed varies greatly. The “Arundo Removal Protocol” (Jackson et al. 2002) states the plant consumes 3,800 acre-feet of water per 1,000 acres per year, (i.e., 3.8 acre-feet of water per acre per year). Iverson (1994) compared the water consumption of giant reed to rice at 5.62 acre-feet of water per acre per year. Studies by Zembal and Hoffman (2000), Oakins (2001), and Jackson et al. (2002) indicated that giant reed consumes three times more water than typical native vegetation. In addition to the large amount of water consumed, giant reed is responsible for changing the landscape of the riparian areas. The growth of the plant causes a faster, narrower stream flow, reducing water recreation, and ultimately, undercutting the banks of the river (Oakins 2001). When undercutting occurs, large stands of giant reed break away from the bank and float to infrastructure downstream, often damaging bridges, roads, and water-intake facilities (Dudley et al. 2007). The reduction in native vegetation causes the canopy structure to diminish around the stream, because overhanging trees no longer exist to shade the water. The reduced canopy exposes the river to more sunlight and creates a higher pH of the water, affecting fish and other wildlife native to the area (McGaugh et al. 2006). Because of the high cost of mechanical and chemical control over large areas such as the Rio Grande Basin (Jackson et al. 2002), the USDA-ARS in Weslaco, TX, and cooperators at the Instituto Mexicano del Tecnologia del Aguas in Juitepec, Mexico, are investigating biological control measures for giant reed (Goolsby and Moran 2009). The goals of the project are to mass rear and release the insects in areas along the Rio Grande, as well as its tributaries. The long-term goal is to use biological control agents alone or in combination with selective mechanical/chemical controls to reduce the competitiveness of giant reed, thereby allowing the riparian zones to transition back to native vegetation. A return to native vegetation is predicted to use less water, thus conserving more water in the reservoirs of the Rio Grande Basin. Four insects are being evaluated by the USDA-ARS: Arundo wasp, Tetramesa romana Walker (Hymenoptera: Eurytomidae); Arundo scale, Rhizaspidiotus donacis (Leonardi) (Hemiptera: Diaspididae); Arundo fly, Cryptonevra sp. (Diptera: Chloropidae); and Arundo leafminer, Lasioptera donacis Coutin & Faivre-Amiot (Diptera: Cecidomyiidae). The wasp, T. romana, was permitted and released in April 2009 and a second agent R. donacis was recommended by the USDA-APHIS Technical Advisory Group for Biological Control of Weeds in June 2009.

378 Economic Literature

Measuring the value of water is a key issue in determining the economic implications of saved water. Kaiser and Roumasset (1999) stated that water is usually undervalued and underpriced. In agriculture, many variables influence crop yields (e.g., changes in technology, inputs, weather, etc.). The value estimated for water may also include other exogenous variables (Ward and Michelsen 2002). Additionally, water is a public good, used by the entire population; therefore, the value must include social aspects to account for the impact to the public. Agriculture Composite Acre. Water valuation methods using crop budgets were outlined by Gibbons (1986) and are commonly used in agricultural economic analyses for the U.S. Army Corps of Engineers (Lacewell 2008). In a study by Sturdivant et al. (2004), a composite acre was developed and applied to calculate the benefits to agriculture of flood-control infrastructure along the Rio Grande. In this study, the composite acre is a reflection of the irrigated and dryland cropping patterns in the Lower Rio Grande Valley of Texas. Returns to land are estimated for a composite dryland acre and returns to land and water are identified for an irrigated composite acre, with the difference representing returns to water. Economic Impact Analysis. Economic impact analysis is performed as a method to determine how changes in demand for one industry or economic sector affect the economy (Jenson 2001). The analyses are based on input-output models, or models that create a framework into which data can be collected, categorized, and analyzed (Shaffer et al. 2004). The input-output model is based on the supply and demand relationship for a particular commodity (Deller 2004). IMPLAN is a widely-used input-output model for assessing the economic impacts associated with a change in contributing activity, or shock scenario. The model utilizes data for 509 North American Industry Classification System (NAICS) sectors. IMPLAN estimates multipliers for economic output, value-added, and employment for a designated county, region, or state. These multipliers only capture the backward linkages (i.e., sectors up to and including the farm level in this case) and do not include forward linkages (i.e., further processing) (Minnesota IMPLAN Group Inc. 2004). Per-Unit Life-Cycle Cost of Water Conserved. In the “Economic Methodology for South Texas Irrigation Projects-RGIDECON©,” Rister et al. (2008) documented the methodology used to determine the life-cycle cost per acre-foot of water saved. To determine the life-cycle cost per acre-foot, annuity equivalents are estimated for both the cost stream and the acre-feet of water saved by the project. Dividing the annuity equivalent of the cost stream by the annuity equivalent of the water saved from the construction and implementation of the project results in the estimated life-cycle cost per acre-foot of water saved. The analysis for the giant reed project was based on the water savings being used for agricultural purposes via the conversion of dryland to irrigated crop acreage. Irrigation increases crop yields and contributes to planting additional acreage of higher-value crops. A primary purpose of this research was to estimate the economic benefits of the water saved from the reduction in the size, density, and area infested by giant reed during a 50-year period (2009 through 2058) caused by the release of two biological agents (i.e., wasp and scale). In addition to the estimation of benefits, a comprehensive economic impact analysis for the Texas Lower Rio Grande Valley was conducted for the same time period. Lastly, the per- unit life-cycle cost of water saved (Rister et al. 2008) via the biological control

379 project was derived to facilitate comparisons with other study estimates of life-cycle costs of water saved through Valley irrigation district rehabilitation projects (e.g., Rister et al. 2007). The economic and financial results derived in this research provide the USDA-ARS, local community leaders, U.S. and Mexico government officials, and others with information regarding the expected economic implications of pursuing the release of the biological control agents.

Methods

The number of acres infested with giant reed on both the U.S. and Mexico sides of the River along the 530 miles between San Ignacio and Lajitas, TX, was estimated as 15,715 acres for 2002 and 18,072 acres for 2008, with a total expansion rate of 15% during the 6-year time period (Yang 2008) (Fig. 1). The growth was equally distributed among the years, suggesting an annual growth rate of 2.36% (15% = (1 + 0.0236)6 - 1.0, with 6 representing the number of years of growth between 2002 and 2008). This yearly rate was adopted and used to linearly forecast expected annual growth for each of the 50 years in the planning horizon (2009 through 2058). It was

Fig. 1. Map of the Rio Grande [River] showing the study area of the USDA-ARS, Weslaco, Texas, Arundo donax biological control program. Modified from Everitt et al. (2004).

380 estimated that 80% of the infestation by giant reed occurs between San Ignacio and Del Rio, while the remaining 20% of the infestation occurs between Del Rio and Lajitas (Yang 2008). Recognizing the targeted study area of the biological control agents for the ARS project occurs solely in the 170 river miles between San Ignacio and Del Rio, TX, this analysis was limited to the riparian area of these 170 miles of the Rio Grande. Any incidental control and benefits realized in the 360 miles between Del Rio and Lajitas, TX, or in Mexico were not included. In 2007, an adventive population of T. romana (the wasp, one of the four insects selected for biological control) was discovered near Laredo, TX (Goolsby and Moran 2009). It was estimated that the adventive population of the wasp provided 5% control of giant reed in a restricted section approximately 1 mile long (Goolsby 2008b). This estimate of biological control impact (prior to the release of agents from the biological control project) was divided by the 170 river miles and multiplied by the number of acres infested by giant reed (Arundo acres were assumed to be homogeneous) between San Ignacio and Del Rio to obtain the adjusted baseline acres used for the economic analyses (of the planned introduction of mass-rearing/releasing). Although the results in this analysis identified water saved from the expected reduction in the number of acres infested by giant reed, actual reduction of giant reed from release of the biological agents will likely occur in the form of fewer acres compared to the baseline, as well as in a reduction in the density and height of the existing plants at the time the insects are released. To account for the reduction in abundance of giant reed, a proxy for the reduction in the biomass of giant reed was based on an estimated reduction in acreage. Control Effectiveness. After estimating the area of control, the efficacy of the two insects (i.e., wasp and scale) was estimated. Based on observed success in quarantine facilities in Texas and measured control in Spain, ARS scientists estimated the treated acres within the specified zone would experience 45% control during the first year of treatment, followed by 22% residual control from the original release in the subsequent year, for a total of 67% control during 2 years. Thereafter, steady state conditions were assumed, i.e., residual control will maintain subsequent biomass of giant reed at end-of-Year 2 controlled levels. Results of several sensitivity analyses are included to examine the effects of deviations from the control assumptions within the modeling framework used. The assumption of the 2-year period for the realization of the efficacy of the wasp and scale effects on giant reed follows the life cycle of the plant, because shoots from the plant are perennial, reaching mature height within the first year of growth (Rieger and Kreager 1989) and becoming lignified as the first growing season ends and fall begins (Decruyenaere and Holt 2001). In other words, the shoot reaches maturity in 1 to 2 years. The assumed total 67% control rate also relates to regions of the world where giant reed is native and herbivores maintain the plant at about one-third (33%) of the density seen in its invasive range (Goolsby 2008a). Biological Control Protocol. All costs, past and expected, for the biological control program were estimated by ARS scientists at Weslaco, TX. The expected amount of biological control of giant reed because of release of T. romana (wasp) and R. donacis (scale) along the Rio Grande is directly related to available funds. Release of the biological control agents is expected to begin in 2009 (Year 1 of treatment/control) and continue through 2014 (Year 6 of treatment/control), with residual effects of the 2014 treatment occurring in 2015. In 2009, scientists

381 anticipate treating 1 mile of the riparian zone at Laredo, TX. An arithmetic progression was used to calculate the miles treated in the remaining years of the program. Using this technique, the program is projected to treat 11.27, 22.53, 33.80, 45.07, and 56.33 miles in Years 2, 3, 4, 5, and 6, respectively; completing treatment over the 170-mile-long area. Potential Water Saved. Water is stored at Amistad Reservoir and only released to Falcon Reservoir when required to meet a water request from downstream. Thus, any added water from control of giant reed downstream from Amistad Reservoir allows for water to remain in Amistad Reservoir longer, reducing the Falcon Reservoir losses (occurring via evaporation and seepage). This deduction suggests all "saved" water as a result of control of giant reed is available and will not be lost to conveyance or percolation because these losses are already accounted for (Rubinstein 2008). The annual difference between the nontreated baseline acreage scenario and the reduced treatment acres was calculated to obtain the number of acres of giant reed prevented through the use of biological control agents. The cumulative number of acres controlled and prevented each year was multiplied by the amount of water used per acre (4.37 acre-feet) by giant reed, resulting in the annual amount of water saved. Consumption by giant reed of 4.37 acre-feet of water per acre of infestation provides the basis for estimating the water savings associated with control. The considered assumption of 67% control of giant reed leads to a water savings of 2.93 acre-feet (67% of 4.37). The revised use of this 67% water savings is effectively a re-distribution from giant reed to (a) replacement, native vegetation, (b) Mexico, and (c) U.S. (Texas) irrigated agriculture. Under the assumption of this study, replacement native vegetation was assumed to emerge in the acres cleared of giant reed and use 1/3 of the original water uptake by giant reed for the area; the remaining water saved was divided equally between the U.S. and Mexico. Consequently, the U.S. realized a water savings of only 2/9 of the original 4.37 acre-feet consumed per acre of giant reed. According to Leidner (2009), an average of 577,888 acre-feet of water are diverted each year to irrigation districts for Cameron, Hidalgo, Starr, Willacy, and Zapata counties. The current 14,453 acres of giant reed in the 170-mile reach of the Rio Grande between San Ignacio and Del Rio, TX, consumes an amount of water equivalent to 10.93% of the irrigation water used by Valley irrigation districts, assuming the annual 4.37 acre-feet per acre of water consumption by giant reed. However, as giant reed continues to expand to more acres over time, the water losses become even greater in the absence of control measures. Economic Analyses. Because Rio Grande Valley Basin municipalities have a legal first priority for water and receive sufficient water to meet their needs (Griffin 2006), any increase in Rio Grande water is logically used for irrigation; i.e., agriculture is the residual beneficiary of any increases in water supplies. Because the net water saved is assumed to be used to convert some dryland agricultural acreage to irrigated agricultural crops, a composite acre was developed to reflect the average aggregate effects of additional irrigated acreage, accounting for variations in water intake and profitability across the different crops. The composite acre was developed for both dryland and irrigated cropping patterns with the most current data available on the number of planted acres from the National Agriculture Statistics Service (NASS-USDA 2008a,b). To determine the direct impact of the saved water from the control of giant reed, the value of water in irrigation was used

382 as the appropriate measure of benefits. Returns to crops that comprise the composite acre were based on normalized prices, which smooth seasonal price variation for each commodity (USDA 2009), remove any price impact due to government farm programs/subsidies, and are typically used in determining the social benefits for agricultural projects (Miller 1980, USDA 2009). Market prices are determined by voluntary trading in a market economy (Tietenberg 2006). The weighted dollar amounts, derived in part from the 2007 Texas AgriLife Extension Service crop enterprise budgets, were summed to obtain the returns to land for the dryland composite acre. Additionally, the irrigated crop budgets were used to calculate returns to land and water, because only water delivery costs (not the cost of water itself) were subtracted from the gross revenue in the Texas AgriLife Extension Service budgets. For this study, irrigated crops were limited to maize, Zea mays L.; cotton, Gossypium hirsutum L.; and sorghum, Sorghum bicolor (L.) Moench, because higher value crops such as vegetables, citrus, and sugarcane, Saccharum spp., are considered to already receive the desired amount of water. Returns to water were obtained by subtracting the returns to land (for dryland crops) from the returns to land and water (for irrigated crops). The resulting value was used in conjunction with the baseline model developed for expansion of giant reed to estimate the market benefits at the farm-level, benefits to society, and benefit-cost, sensitivity, and economic impact analyses. The initial estimate of the value of control of giant reed was based on the increase in returns due to the increased availability of irrigation water and conversion of dryland crops to irrigated crop acres, increasing yields and values of production during a 50-year planning horizon (i.e., 2009 through 2058). This net value was estimated annually, accounting for the increasing degree of giant reed acreage mitigation through time as a result of the biological control program. Direct Economic Impact. The composite per acre-foot value for irrigated crops was multiplied by the number of acre-feet of water saved from the use of the biological control agents to determine the value to society from the saved water. The annual costs of the insect control program were inflated at 2.043% for 2007- 2008 costs, and deflated at 6.125% for 2010-2016 costs. Annual benefits were inflated at 2.043% and then discounted at 6.125% to calculate the present value of benefits and costs (Rister et al. 2008). Benefit-Cost Ratio. The present value of benefits to society during 50 years was divided by the present value of the social costs over 50 years to calculate the benefit-cost ratio. This ratio reflects the benefits per dollar of public expenditure. A benefit-cost ratio exceeding a value of one indicates benefits exceed costs to society (Griffin 2006). Sensitivity Analyses. Sensitivity analyses of the benefits were performed to account for uncertainty related to key variables used in the analyses. Sensitivity analyses were conducted for benefit-cost ratios in which water use by giant reed was varied while the (a) control effectiveness of the program, (b) value of water, or (c) costs of the program were simultaneously varied as the second variable, respectively. Economic Impact Analyses. Economic impacts on the Texas Lower Rio Grande Valley associated with the projected water savings were estimated using the IMPLAN model, Version 2.0 (2006 data). The change in gross revenue between dryland and irrigated crops served as the basis for estimating the broader impacts - economic output, value added, and employment.

383 Dividing the total volume of water saved by the composite acre water use resulted in the number of converted acres from dryland to irrigated agriculture. The change in the number of acres for the respective crops -- according to their proportional representation in the composite acre -- resulted in a change in gross revenue for each crop. These net changes in gross revenues were deflated to 2006 dollars to be consistent with the 2006 data in the IMPLAN model. Using the change in gross revenues (deflated) as the direct impact, the IMPLAN model was used to assess the economic impacts associated with the biological control program for giant reed. Per Unit Life-Cycle Costs of Saved Water. The calculated per-unit life- cycle costs of saved water were comparable to life-cycle costs for other programs that add water to the supply for the region. The annuity equivalent of program costs was divided by the annuity equivalent of the water saved. To obtain this value, the total nominal cost of the program was discounted by 6.125% to 2009 dollars (Rister et al. 2008). Additionally, cumulative water (acre-feet) was discounted at the social discount rate of 4.00%. The annuity equivalent (value per year) for both dollars and water were calculated over the 50-year planning horizon. The values were then divided, obtaining the per-unit life-cycle cost of saving water via the biological control program. The water saved is raw water and is the net of the cost of water delivery for irrigation at the farm level or water processing.

Results

The estimated results indicated positive returns and a positive economic impact to the Texas Lower Rio Grande Valley in association with biologically controlling giant reed. The results can be refined with the developed model, ArundoEcon©, as improved input data become available. Economic analyses require estimates of control of giant reed and associated new water savings. The total acres with giant reed controlled by River segment are: 57 acres in the first segment (treated in 2009), 657 in the second segment (treated in 2010), 1,344 in the third segment (treated in 2011), 2,063 in the fourth segment (treated in 2012), 2,814 in the fifth segment (treated in 2013), and 3,600 acres in the sixth segment (treated in 2015). The total acreage controlled is 38 acres in 2009 (Year 1), 460 in 2010 (Year 2), 1,118 in 2011 (Year 3), 1,827 in 2012 (Year 4), 2,568 in 2013 (Year 5), 3,342 in 2014 (Year 6), and 1,182 acres in 2015 (Year 7) (Table 1). Not only is there control of existing giant reed by the biological control agents, but expected future expansion is curtailed. The total number of acres with giant reed remaining at the end of 2009 (first year of treatment) is 14,749. The anticipated 67% control of the entire study area will be reached at the end of 2015, with 5,189 acres remaining at that time. This number of acres is projected to hold constant over the 50-year planning horizon, because equilibrium between the biological control insects and giant reed is projected. The reduced amount of acreage infested by giant reed (resulting from the biological control program) was the difference in the projected acres in the baseline scenario (without control of giant reed) and the estimated acres with control. The reduced acres with giant reed were multiplied by the amount of water used by giant reed (4.37 acre-feet), resulting in the expected gross amount of water saved. After accounting for water uptake from natural vegetation regrowth and Mexico's allotment of the water, the amount of U.S. water saved in Year 1 totals 59 acre-feet.

384 Table 1. Rio Grande Miles Treated and Arundo Acres Controlled with the USDA-ARS Arundo donax Biological Control Program Between San Ignacio and Del Rio, Texas, 2009-2015a

Arundo acres Arundo acres

Residual Beginning Density Miles Acres Controlled controlled Total Cumulative Remaining Year of year per mile treated treated Year 1 Year 2 controlled controlled after control

2009 14,453.3 85.0 1.0 85.0 38.3 --- 38.3 38.3 14,749.4 2010 14,702.6 87.0 11.3 980.2 441.1 18.7 459.8 498.0 14,608.8 385 2011 14,041.6 89.0 22.5 2,006.0 902.7 215.6 1,118.3 1,616.4 13,770.5 2012 12,315.7 91.1 33.8 3,078.9 1,385.5 441.3 1,826.8 3,443.2 12,158.5

2013 9,451.7 93.2 45.1 4,200.7 1,890.3 677.4 2,567.7 6,010.9 9,713.0 2014 5,373.1 95.4 56.3 5,373.1 2,417.9 924.2 3,342.0 9,352.9 6,371.0

2015 0.0 0.0 0.0 0.0 0.0 1,182.1 1,182.1 10,535.0 5,188.9 PROJECT TOTAL 170.0 15,724.0 10,535.0 aIt is anticipated there will be 45% control in the first year (Arundo Acres Controlled Year 1), and another 22% control in the second year (Residual Arundo acres controlled Year 2) for a total of 67% control. This process of 2-year treatment stages continues along the Rio Grande for each segment treated. The amount of water saved continues to increase throughout the 50-year study horizon as the acres treated and controlled increase, with 765 acre-feet saved in 2010, 2,499 acre-feet saved in 2011, 5,371 acre-feet saved in 2012, 9,471 acre- feet saved in 2013, 14,888 acre-feet saved in 2014, and 17,173 acre-feet saved in 2015 (Table 2). The overall control of giant reed in the 170-mile stretch of the Rio Grande during 50 years amounts to more than 58,000 acre-feet of water saved in Year 2058. The net annual water savings for the U.S. amounts to approximately one acre-foot for each acre of giant reed controlled, i.e., 2/9 * 4.37 = 0.97. Estimated returns to water are $139.22 per acre-foot using normalized prices. The water use per irrigated acre (0.54 acre-feet per acre) impacts the number of acres converted from dryland to irrigated crops using the water saved from the control of giant reed. For each acre-foot of water saved, 1.85 dryland acres can be converted to irrigated crops. Total Value of Water Saved to the Rio Grande Valley. The value for water saved and used for irrigation across the Valley was estimated by multiplying the amount of water saved by the returns to water on an annual basis. The estimated value of water saved to the Rio Grande Valley using the irrigated composite acre and normalized prices of crops is more than $8,160 for 2009, $2.39 million in 2015, $3.28 million for 2025, $4.40 million in 2035, $5.81 million in 2045, $7.59 million in 2055, and $8.20 million in 2058. Inflated at an annual rate of 2.043% and discounted at a rate of 6.125%, the present value over 50 years in 2009 dollars is $72.43 million.

Table 2. Annual Acre-Feet of Water Saved and Accruing to the United States with Arundo Control in the Rio Grande Basin, San Ignacio to Del Rio, Texas, 2009 through 2058 Acre-feet of water saved After subtracting consumption After subtracting Year Gross amount by native vegetation Mexico’s sharea 2009 176 117 59 2010 2,294 1,529 765 2011 7,496 4,997 2,499 2012 16,114 10,743 5,371 2013 28,412 18,941 9,471 2014 44,665 29,777 14,888 2015 51,518 34,345 17,173 2025 70,701 47,134 23,567 2035 94,845 63,230 31,615 2045 125,232 83,488 41,744 2055 163,475 108,984 54,492 2058 176,772 117,848 58,924 aAmount of water “saved” and available for use by U.S. (Texas) agriculture for irrigation.

386 Benefit-Cost Analysis. Normalized prices were used in the benefit-cost analyses to reflect the total social benefits of the saved water. The irrigated crop mix has a present value of $72.43 million. The costs incurred and projected throughout the program have a present value of $16.54 million. Dividing the present value of benefits by the present value of costs suggests the returns to irrigated crops mix has a benefit-cost ratio of 4.38:1. That is, society is projected to experience benefits of $4.38 for every $1 of project costs. Because the present value of the benefits are greater than the present value of the costs (i.e., the benefit- cost ratios are greater than one), these results suggest the Arundo biological control project is economically viable. Sensitivity Analyses. Sensitivity analyses were performed to account for uncertainty in selected variables, and provide a range of values encompassing the baseline deterministic results. These sensitivity analyses include varying the assumptions for (a) percent control from beneficial insects, (b) value of water, (c) costs of the program, and for all cases (d) water use of giant reed. These sensitivity results are presented in a pair-way fashion (i.e., with only two variables varying at a time): (a) water use of giant reed (varied from 2.00 to 7.00 acre-feet) and (b) one of the other variables noted previously. In Table 3, the amount of water consumed by giant reed was varied in the baseline scenario, 4.37 acre-feet per year (across the top row). There are three sections of the table, with the first addressing the efficacy of the biological control agents varied about the expected 67% total control from the release of the biological agents (down the left column). The second is the value of water varied about the estimated $139.22 per acre-foot value, and the third section is the cost of the program varied about the baseline estimated amount of $16.54 million. The baseline deterministic values calculated in the model are bold and located in the shaded cells. The sensitivity of the benefit-cost ratios to varying water use rates by giant reed and the efficacy of the biological control agents are shown in the top third of Table 3. The ratio ranges from 1.56:1 at 40% control efficacy from the beneficial insects with water use by giant reed at 2.00 acre-feet of water per year to a ratio of 7.76:1 at 80% control efficacy from the beneficial insects and water use by giant reed of 7.00 acre-feet per year. At the lowest, most conservative set of assumptions examined in this analysis, the return on the project would be $1.56 for every $1.00 of resources invested by the public sector, indicating the project is feasible. Overall, the benefits of the program range from $1.56 to 7.76 for every $1 of public funds spent, depending on the water consumption rate of giant reed, efficacy of the insects, and adopted crop mix (acres converted from dryland to irrigated). This range indicates a positive net outcome in all scenarios indicated. Actual realized benefits are expected to fall within this range. As expected, less water consumed by giant reed and decreased efficacy of the biological control agents produces a smaller return to the costs of the control program; conversely, more water consumed by giant reed and increased efficacy of the biological control agents produces a larger return to the costs of the control program. As demonstrated in the middle-third of Table 3, less water consumed by giant reed and a lower value of water results in lower returns to the control program. At the extreme point, the benefit-cost ratio is not economically justified (0.72), where the value of water is $50.00 and the consumption of water by giant reed is 2.00 acre-feet. The project becomes economical at 2.00 acre-feet when the value of water

387 Table 3. Sensitivity Analysis, Benefit-Cost Ratio of Benefits with Variations in Annual Water Consumption of Arundo and Control Rate from Beneficial Insects (Total %) using Normalized Prices, in the Texas Lower Rio Grande Valley, 2009 Variation in annual water consumption (acre-feet) -2.37 -1.37 -0.37 0 0.63 1.63 2.63 Low-marginal-value crops Annual water consumption by Arundo (acre-feet/year) 2.00 3.00 4.00 4.37 5.00 6.00 7.00 40.0% 1.56 2.34 3.12 3.41 3.90 4.68 5.47 Control rate from 50.0% 1.73 2.59 3.45 3.77 4.31 5.18 6.04 beneficial insects (total 60.0% 1.89 2.83 3.78 4.13 4.72 5.67 6.61 %) 67.0% 2.00 3.01 4.01 4.38 5.01 6.01 7.02 70.0% 2.05 3.08 4.11 4.49 5.13 6.16 7.19 75.0% 2.14 3.20 4.27 4.67 5.34 6.41 7.48 80.0% 2.22 3.33 4.44 4.85 5.54 6.65 7.76

388

$50.00 0.72 1.08 1.44 1.57 1.80 2.16 2.52 $100.00 1.44 2.16 2.88 3.15 3.60 4.32 5.04 $125.00 1.80 2.70 3.60 3.93 4.50 5.40 6.30 Value of water $139.22 2.00 3.01 4.01 4.38 5.01 6.01 7.02 $150.00 2.16 3.24 4.32 4.72 5.40 6.48 7.56 $175.00 2.52 3.78 5.04 5.51 6.30 7.56 8.82 $200.00 3.24 4.86 6.48 7.08 8.10 9.72 11.34

-30.0% 2.86 4.30 5.73 6.26 7.16 8.59 10.02 -20.0% 2.51 3.76 5.01 5.47 6.26 7.52 8.77 -10.0% 2.23 3.34 4.45 4.87 5.57 6.68 7.80 Cost of program 0.0% 2.00 3.01 4.01 4.38 5.01 6.01 7.02 10.0% 1.82 2.73 3.64 3.98 4.56 5.47 6.38 20.0% 1.67 2.51 3.34 3.65 4.18 5.01 5.85 30.0% 1.54 2.31 3.08 3.37 3.85 4.63 5.40 increases to $100 or when the use of water by giant reed increases to 3.00 acre- feet at $50.00 (i.e., more water would be saved from the reduction of giant reed). If more water is consumed by giant reed and greater values of water exist, larger returns to the costs of the control program are realized. Thus, the project will generate more value in benefits than the value spent in cost (i.e., economically feasible) in all scenarios, except for the most conservative scenario presented. The 2009 benefit-cost ratio results from varying the water use by giant reed and the cost of the Arundo donax biological control program (lower third of Table 3) range from 1.54:1 with the cost of the program at 30% greater than the baseline calculations and water use by giant reed at 2.00 acre-feet of water per year, to a ratio of 10.02:1 with the cost of the program at 30% less than the baseline calculations and water use by giant reed of 7.00 acre-feet per year. At the most conservative set of assumptions examined in this analysis, the return on the project would be $1.54 for every $1.00 of public investment, indicating the project is economically feasible. Economic Impacts. Multipliers for economic output, value-added, and employment from the IMPLAN model are used in conjunction with the changes in gross revenue attributable to increased irrigated acres in the Lower Rio Grande Valley to assess the economic impacts associated with the irrigation use of the water saved. Changes in the baseline gross revenues (as a result of the Arundo biological control program) and the associated economic impact occur due to conversion in acreage from dryland to irrigated, as farmers utilize more water. The impacts are estimated based on deflated increases in gross returns to crops for the Texas Lower Rio Grande Valley (i.e., the Texas southernmost four counties of Cameron, Hidalgo, Starr, and Willacy). The increases in gross returns were calculated using 2007 dollars. The deflation of the 2007 dollars to 2006 dollars was necessary, because the IMPLAN model uses 2006 data to project the multipliers for each sector. Impact analysis was conducted for this four-county region during the 50-year planning horizon. State impacts were not analyzed in this study because 100% of the direct impacts were assumed to be spent within the four-county region of the Texas Lower Rio Grande Valley. Additionally, the outcome of state impacts is similar to regional impacts. Although the sector mix of the economy is not likely to remain unchanged, this study assumed the structure of the economy remains constant over the 50-year planning horizon. In 2009, when dividing the 59 acre-feet of potential net water saved by 0.54 acre-feet of water use for the irrigated composite acre, a total of 108 acres could be converted from dryland to irrigated. A similar procedure was applied to project the change in irrigated acres over time. The respective crop acres were calculated for conversion in 2015, 2025, 2035, 2045, 2055, and 2058, indicating 31,516, 43,252, 58,022, 76,611, 100,006, and 108,140 acres converted from dryland (rain-fed) to irrigated for the respective years as compared to the uncontrolled giant reed baseline. For each added irrigated acre, there is a reduction of a dryland acre. The additional irrigated acres were added to the current acreage amount and multiplied by the noninflated gross revenue per acre, by crop, to obtain the increased gross revenue by year. Base year prices for 2007 were used in all future revenue estimation, thus reflecting the increasing amount of giant reed controlled through time and in the multipliers. These net increases in the gross revenues due to the additional irrigated acres were deflated to 2006 dollars. Using IMPLAN, multipliers for economic output, value added, and employment were used to estimate the annual impact for each year of the 50-year planning horizon.

389 Using the irrigated crop mix for the four counties in the Texas Lower Rio Grande Valley, economic output results for 2009 and 2015 were estimated at $22,140 and 6.56 million, respectively (Table 4). As the direct impact (change in gross revenue) increases over time -- following the increased conversion of dryland to irrigated acres -- economic output climbs from $8.9 million in 2025 to $22.26 million in 2058. Value-added, which is also the contribution to the gross domestic product of Texas, is estimated at $11,015 in 2009. By 2058, value-added is estimated at $11.07 million. No additional employment is associated with the change in gross revenues for 2009. In 2015, this level of direct impact ($4.58 million) supports an additional 143 jobs. By 2035, the additional jobs supported by these impacts is estimated at 264, and by 2058, an estimated 492 jobs are supported by these impacts. These estimated employment impacts are not additive, but are annual for each year for the regional economy. This analysis indicates the impact of the saved water increases economic output, value-added, and the number of jobs in the region, and is a positive impact to the Lower Rio Grande Valley of Texas.

Table 4. Regional Economic Impact to the Texas Lower Rio Grande Valley in 2006 Dollars from the USDA-ARS, Weslaco, Texas, Arundo donax Biological Control Program, 2009-2058a Direct impactb Economic output Value-added Year ($ million, 2006) ($ million) ($ million) Employment 2009 0.02 0.02 0.01 0 2015 4.58 6.56 3.23 143 2025 6.28 8.90 4.43 197 2035 8.43 11.94 5.94 264 2045 11.13 15.77 7.84 349 2055 14.53 20.58 10.24 455 2058 15.71 22.26 11.07 492 aRegion includes the lower four counties of the state of Texas: Cameron, Hidalgo, Starr, and Willacy. bDeflated change in gross revenue.

Per-Unit Life-Cycle Costs of Saved Water. Annuity equivalents of the respective present values for the cost of the program and the acre-feet of water saved were obtained for the 50-year planning horizon (i.e., 24.2 thousand acre-feet or 7.9 million gallons of water saved per year). Dividing the annuity equivalent of costs ($1.07 million) by the annuity equivalent of water saved (24.2 thousand acre- feet) resulted in a program cost of $44.08 per acre-foot of raw water, or $0.1353 per 1,000 gallons. The per-unit life-cycle cost of water saved due to the biological control program for giant reed is comparable to the average cost of $45 per acre- foot for several of the on-going rehabilitation projects in the Rio Grande Valley designed to conserve raw water (prevent water loss) (Sturdivant et al. 2007).

390

Discussion

While the preliminary results indicated positive net economic benefits of the biological control program for giant reed, several of the critical data-input variable values are uncertain. This research used sensitivity analyses to account for variations in certain variables that could influence the outcome of the analysis. Further sensitivity analyses (e.g., water use by native vegetation) can be found in “Select Economic Implications for the Biological Control of Arundo donax along the Rio Grande” by Seawright (2009). Additional research is currently being conducted and, as the program progresses, input data and model results can be refined. While many issues were addressed in this research, certain areas were not considered, including potential benefits to the Department of Homeland Security and the Border Patrol, recreational activities, environmental values, and benefits to Mexico as a result of the program. Only the program and U.S. benefits received from the control of giant reed by the release of two of four insects, T. romana and R. donacis (the wasp and scale, respectively), in the limited project study area (i.e., 170 river miles between San Ignacio and Del Rio, TX) were considered. Further, the effectiveness of these agents cannot be entirely predicted. The assumptions for control were based on field observations from Europe and laboratory studies in the U.S.

Conclusions

The baseline benefit-cost analysis suggests returns of $4.38 for every public dollar invested. This result indicates net positive returns for the giant reed biological control project. Additionally, the results indicate a positive impact to the regional economy by (a) increasing economic output by $22,140 in 2009, $11.94 million in 2035, and $22.26 million in 2058, (b) increasing value-added by $11,015 in 2009, $5.94 million in 2035, and $11.07 million in 2058, and (c) supporting an additional 264 jobs in 2035 and 492 jobs in 2058. Additionally, the per-unit life-cycle cost of water saved as a result of the ARS biological control program is $44.08 per acre- foot and is comparable to the per-acre-foot costs of current programs in use or being considered for increasing water supply in the Valley. These results indicate a positive economic impact to the Lower Rio Grande Valley of Texas. Because the evaluation, release, and effectiveness of the biological control agents remain under investigation, the results presented are considered preliminary. As of May 2009, the data results for different aspects of this project continue to be observed and collected. It is expected that more accurate data will be identified as the project continues. Based on the current available data and the results of the economic research reported, however, the release of the two biological control agents, T. romana (wasp) and R. donacis (scale), to control giant reed in the Rio Grande Basin (a) increases water availability to the Rio Grande Valley and (b) creates a positive impact both at the farm level and for the regional economy.

Acknowledgment

Support is gratefully acknowledged to the USDA-ARS (contract 58-6204-8- 051), USGS Student Research Grant, TWRI, and the Rio Grande Basin Initiative

391 (contracts 2006-503772-93041 and 2007-628460-99022) for funding and supporting this research. Additionally, the authors thank the collaborators for providing support through additional research for the project. Any copyrights, trademarks, or registered names used in this article are not an endorsement by the authors. Rather, these copyrights, trademarks, and registered names are used for clarification and better understanding of the material for the reader.

References Cited

Bell, G. P. 1997. Ecology and management of Arundo donax, and approaches to riparian habitat restoration in Southern California, pp. 103-113. In J. H. Brock, M. Wade, P. Pysek, and D. Green [eds.], Plant Invasions: Studies from North America and Europe. Blackhuys Publishers, Leiden, Netherlands. Decruyenaere, J. G., and J. S. Holt. 2001. Seasonality of clonal propagation in giant reed. Weed Sci. 49: 760-767. Deller, S. C. 2004. Wisconsin and the Agricultural Economy. Agricultural & Applied Economics Staff Paper Series. Department of Agricultural & Applied Economics, University of Wisconsin, Madison, WI. Dudley, T. 1998. Noxious wildland weeds of California: Arundo donax. Noxious Wildland Weeds of California. (accessed 9 November 2006) http://ceres.ca. gov/tadn/ecology_impacts/Arundo_ww.html. Dudley, T., A. Kirk, J. Goolsby, G. Coffman, and D. Kazmer. 2007. Riparian InVasives Research Laboratory Arundo donax. (accessed 3 September 2007) http://rivrlab.msi.ucsb.edu/Arundo.htm. Everitt, J. H., C. Yang, M. A. Alaniz, M. R. Davis, F. L. Nibling, and C. J. DeLoach. 2004. Canopy spectra of giant reed and associated vegetation. J. Range Manage. 57: 561-569. Gibbons, D. 1986. The economic value of water. Resources for the Future, Washington, DC, and Baltimore, MD. Goolsby, J. A. 2008a. Personal communication. Entomologist. USDA-ARS, Weslaco, TX. 18 April 2008. Goolsby, J. A. 2008b. Personal communication. Entomologist. USDA-ARS, Weslaco, TX. 30 July 2008. Goolsby, J. A., and P. Moran. 2009. Host range of Tetramesa romana Walker (Hymenoptera: Eurytomidae), a potential biological control of giant reed, Arundo donax L. in North America. Biological Control (2009), doi: 10.1016/j.biocontrol.2009.01.019 (accepted). Griffin, R. C. 2006. Water Resource Economics. MIT Press, Cambridge, MA. Iverson, M. E. 1994. The impact of Arundo donax on water resources, pp. 19-25. Arundo donax Workshop Proceedings, November 1993, Ontario, CA. Jackson, N. E., W. Katagi, and C. Loper. 2002. Southern California Integrated Watershed Program Arundo Removal Protocol. Santa Ana Watershed Project Authority. (accessed 1 March 2007) http://www.sawpa. org.Arundo.protocol.htm. Jensen, I. 2001. The Leontief open production model or input-output analysis.” (accessed 20 January 2009) http://online.redwoods.cc.ca.us/instruct/darnold/ laproj/Fall2001/Iris/lapaper.pdf. Kaiser, B., and J. Roumasset. 1999. Water management and the valuation of indirect environmental services. Working Paper 99-5. Department of Economics, University of Hawaii, Manoa, HI.

392

Lacewell, R. D. 2008. Personal communication. Professor of Agricultural Economics. Department of Agricultural Economics, Texas A&M University. College Station, TX, 15 February 2008. Leidner, A. 2009. Personal communication. Texas A&M University, College Station, TX. E-mail: 21 April 2009. McGaugh, S., D. Hendrickson, G. Bell, H. Cabral, K. Lyons, L. Mceachron, and O. Muñoz J. 2006. Fighting an aggressive wetlands invader: a case study of giant reed (Arundo donax) and its threat to Cuatro Ciénegas, Coahuila, Mexico, pp. 100-115. In M. de Lourdes Lozano-Vilano and A. J. Contreras- Balderas [eds.], Studies of North American Desert Fishes: In Honor of E.P. (Phil) Pister, Conservationist. Universidad Autónoma de Nuevo León, Monterrey, Nuevo León, México. Miller, J. R. 1980. Normalized agricultural prices and water resources planning. Water Resources Bulletin 16: 471-476. American Water Resources Association, Middleburg, VA. Minnesota IMPLAN Group, Inc. 2004. IMPLAN Professional Version 2.0. User’s Guide, Analysis Guide, Data Guide. Minnesota IMPLAN Group, Inc., Stillwater, MN Moran, P. J., and J. A. Goolsby. 2009. Biology of the galling wasp Tetramesa romana, a biological control agent of giant reed. Biological Control (2009), doi: 10.1016/j.biocontrol.2009.01.017 (accepted). National Agricultural Statistics Service (NASS). 2008a. USDA National Agricultural Statistics Service-Quick Stats US & All States County Data-Crops. USDA. (accessed 12 May 2008) www.nass.usda.gov/quickstats/create_county_all. jsp. National Agricultural Statistics Service (NASS). 2008b. USDA National Agricultural Statistics Service-Quick Stats US & All States Data-Vegetables. USDA. (accessed 12 May 2008) www.nass.usda.gov/quickstats/create_county_all. jsp. Oakins, A. 2001. An assessment and management protocol for Arundo donax in the Salinas Valley Watershed. Capstone Project. Presented to Faculty of Earth Systems Science and Policy in the Center for Science, Technology, and Information Resources. California State University, Monterey Bay, Seaside, CA. Perdue, R. E., Jr. 1958. Arundo donax – source of musical reeds and industrial cellulose. Econ. Bot. 12: 368-404. Rieger, J. P., and D. A. Kreager. 1989. Giant reed (Arundo donax): a climax community of the riparian zone, pp. 222-225. In Dana L. Abell [Technical Coordinator], Proceedings of the California Riparian Systems Conference: Protection, Management, and Restoration for the 1990s; 22-24 September 1988; Davis, CA. Gen. Tech. Rep. PSW-GTR-110. Berkeley, CA: Pacific Southwest Forest and Range Experiment Station, Forest Service, U.S. Department of Agriculture. Rister, M. E., R. D. Lacewell, and A. W. Sturdivant. 2007. Economic and financial costs of saving water and energy: preliminary analysis for Hidalgo County irrigation district No. 2 (San Juan) – Replacement of pipeline units I-7A, I-18, and I-22. Texas Water Resources Institute. TR-303. College Station, TX. June 2007.

393

Rister, M. E., C. S. Rogers, R. D. Lacewell, J. R. Robinson, J. R. Ellis, and A. W. Sturdivant. 2008. Economic methodology for South Texas irrigation projects - RGIDECON©. Texas Water Resources Institute TR-203 revised, College Station, TX. Roberts, M. 2007. Data on normalized prices for evaluating alternative development and management plans for water and related land resources. USDA-ERS. (accessed 6 June 2008) http://www.ers.usda.gov/Data/ Normalized Prices/. Rubinstein, C. 2008. Personal communication. Executive Director, Texas Commission on Environmental Quality, Austin, Texas. 19 August 2008. Seawright, E. 2009. Select Economic Implications for the Biological Control of Arundo donax along the Rio Grande. M.S. thesis, Department of Agricultural Economics, Texas A&M University, College Station, TX. Shaffer, R., S. Deller, and D. Marcouiller. 2004. Community Economics: Linking Theory and Practice. 2nd edition. Blackwell Publishing, Ames, IA. Sturdivant, A. W., R. D. Lacewell, A. M. Michelsen, M. E. Rister, N. Assadian, M. Eriksson, R. Freeman, J. H. Jacobs, W. T. Madison, J. T. McGuckin, W. Morrison, J. R. C. Robinson, C. Staats, Z. Sheng, R. Srinivasan, and J. I. Villalobos. 2004. Estimated benefits of IBWC Rio Grande flood-control projects in the United States. Contract report provided to the International Boundary and Water Commission (IBWC). TR-275. Texas Water Resources Institute, College Station, TX. September 2004. Sturdivant, A. W., M. E. Rister, R. D. Lacewell, C. S. Rogers, E. K. Seawright, C. N. Boyer, and A. J. Leidner. 2007. Economists: Task 1. Presentation. Joint Rio Grande Basin Initiatives Annual Conference. South Padre Island, TX, 16 May 2007. Texas AgriLife Extension Service. 2007. Crop and livestock enterprise budgets for South Texas. Texas AgriLife Extension Service website. (accessed 24 April 2007) http://agecoext.tamu.edu. Tietenberg, T. 2006. Environmental and Natural Resource Economics. 7th edition. Pearson Education, Inc., New York. Tracy, J. L., and J. C. DeLoach. 1999. Suitability of classical biological control for giant reed (Arundo donax) in the United States. Arundo and Saltcedar Management Workshop Proceedings, 17 June, 1998, Ontario, CA. University of California Cooperative Extension, Holtville, CA. U.S. Census Bureau. 2000. Census 2000 PHC-T-3. Ranking Tables for Metropolitan Area: 1990 & 2000. (accessed 18 January 2008) http://www.census.gov/population/cen2000/phc-t3/tab05.pdf. United States Department of Agriculture (USDA). 2009. Normalized prices. (accessed 11 March 2009) http://www.ers.usda.gov/Data/NormalizedPrices/. Ward, F., and A. Michelsen. 2002. The economic value of water in agriculture: concepts and policy applications. Water Policy 4: 423-46. Yang, C. 2008. Personal communication. Agricultural Engineer. United States Department of Agriculture-Agricultural Research Service, Weslaco, TX. E- mail: 7 October 2008. Zembal R., and S. Hoffman. 2000. Environmental Assessment of the Santa Ana Watershed Program, Fountain Valley, CA.

394 VOL. 34, NO. 4 SOUTHWESTERN ENTOMOLOGIST DEC. 2009

Cost-benefit Analysis of Sorghum Midge, Stenodiplosis sorghicola1 (Coquillett)-Resistant Sorghum Hybrid Research and Development in Texas

Tebkew Damte2, Bonnie B. Pendleton, and Lal K. Almas

West Texas A&M University, P. O. Box 60998, Canyon, TX 79016-0001

Abstract. Sorghum, Sorghum bicolor (L.) Moench, is important to the economies of Texas and the U.S. Sorghum midge, Stenodiplosis sorghicola (Coquillett), attacks flowering sorghum, causing substantial loss of grain every year. Sorghum midge- resistant genotypes are available for use in breeding programs to develop agronomically superior sorghum hybrids. Estimated yield loss caused by sorghum midge varied between 94.2 and 890.8 and 12.9 and 188.8 kg/ha if sorghum midge- susceptible and resistant hybrids, respectively, were grown. Yield saved by using a resistant hybrid was 48.1 (45.3 kg/ha) to 96.7% (702.0 kg/ha). Benefits accrued during a 25-year period were 66.7, 99.9, and 319.8 million (in 1984 constant dollar) at 1, 5, and 15% compounding/discounting rates, respectively. The corresponding accruable costs were $4.3, 6.7, and 23.4 million. Net present value was positive and cost-benefit ratio was less than one, indicating investment in developing a sorghum midge-resistant hybrid was profitable. Depending on the interest rate, a dollar invested in research and development of a sorghum midge-resistant hybrid would generate $13.6 to 15.7.

Introduction

Sorghum midge, Stenodiplosis sorghicola (Coquillett), is the most widespread and damaging insect pest of sorghum, Sorghum bicolor (L.) Moench, worldwide (Teetes and Pendleton 2000) and is a key pest of sorghum in Texas (Cronholm et al. 2007). This tiny fly during her adult life of less than 1 day oviposits approximately 50 eggs between the glumes of flowering spikelets of sorghum. The larva feeds on and prevents the ovary from developing into a kernel. Sorghum midge can cause complete loss of grain (Teetes 1985). Cultural and chemical controls are effective in managing sorghum midge, but are not free of problems. Planting may be delayed or extended because of drought or frequent spring rains, and the short life span and overlapping generations of sorghum midge, time of infestation, and price of sorghum make insecticide expensive to use (Cronholm et al. 2007). The economic thresholds for applying insecticide to susceptible and resistant hybrids are one and five ovipositing sorghum midge, respectively, per flowering panicle of sorghum (Hallman et al. 1984). The first sorghum midge-resistant sorghum line was released in 1971 (Wiseman et al. 1973). In Texas, eight sorghum midge-resistant lines derived from

1Diptera: Cecidomyiidae. 2Debre Zeit Agricultural Research Center, P. O. Box 32, Debre Zeit, Ethiopia; e-mail: [email protected].

395 TAM2566 were released in 1973 (Johnson et al. 1973). Resistant lines released by the Texas Agricultural Experiment Station were used by commercial seed companies to develop hybrids (Peterson 1985). However, hybrids developed from resistant lines yield less grain than susceptible hybrids in the absence of sorghum midge (Peterson et al. 1992, Peterson 2001). Cost of development and distribution of a sorghum midge-resistant hybrid varies in time, and the search for new and greater resistance is continuous. When offered new technology developed through research, farmers will adjust their usage to exploit opportunities for additional profit, such as use of resistant hybrids (Wise 1986). The objective of this study was to compare the cost of research and development to anticipated benefits from yield saved by using a resistant sorghum hybrid produced commercially in sorghum midge-infested areas of Texas.

Methods

Estimating Yield Saved. The year 1982 was used as the year in which benefit from growing sorghum midge-resistant hybrids began to accrue, because commercial production of resistant hybrids began in 1982 (Hallman et al. 1984). The benefit in year t (Bt) was estimated as the value of crop saved using: Bt = YSt * Pt * At……………………………………………………..……………….(1) where Bt = benefit ($) in year t, YSt = yield saved (kg/ha) in year t, Pt = average marketing year price ($/kg) of sorghum, and At = area infested by sorghum midge (ha). Abundance of sorghum midge varies by year, location, planting time, and hybrid (Teetes et al. 1986, 1990; Parker et al. 2003). When sorghum midge is abundant, 100% of grain can be destroyed (Teetes 1985, Parker et al. 2004). Yield loss caused by sorghum midge on either kind of hybrid was estimated by using: EYLt = D * NSFt * TPN ………………………..…....………………………………..(2) where EYLt = estimated yield loss (kg/ha) in year t, D = damage per ovipositing sorghum midge (kg), NSFt = number of successful ovipositing sorghum midges in year t, and TPN = number of panicles infested per hectare. The percentage of yield saved (PYSt) was calculated using:

§ EYLt SH − EYLt RH · PYSt = ¨ ¸100 …………..………………………………………..(3) ¨ EYL SH ¸ © t ¹ where EYLtSH = yield loss (kg/ha) if a susceptible hybrid was grown in year t, and EYLtRH = yield loss (kg/ha) if a resistant hybrid was grown in year t. Data on the number of ovipositing sorghum midges per panicle were obtained from research in different parts of Texas and published in various publications (Table 1). For years in which data were lacking, the maximum number of ovipositing sorghum midges -- 50 for a susceptible hybrid and 200 for a resistant hybrid (Teetes 1985) was used as the “carrying capacity” (K) of a panicle. The number of ovipositing sorghum midges in year t was estimated using the logistic equation: K − N § t · N t+1 = N t + rN t ¨ ¸ ……………………………………….…………………….(4) © K ¹ where Nt = the number of ovipositing sorghum midges per panicle in year t, K = carrying capacity of a panicle, and r = rate of increase. Abundance of sorghum midges increases 2.5 to 3.7 times on Johnsongrass and 24.6 and 12.6 times on sorghum (Pendleton et al. 1994). The number of times

396 Table 1. Number of Ovipositing Sorghum Midges per Sorghum Panicle in Texas Sorghum hybrid Year Reference Susceptible Resistant 1981 8 11 Hallman et al. (1984) 1984 11 50 Teetes et al. (1986) 1989 4.5 5.2 Teetes et al. (1990) 1995 5.3 2.3 Diarisso et al. (1998a) 2003 6.3 2.0 Parker et al. (2003) 2004 23.6 24.5 Parker et al. (2004)

abundance of overwintering sorghum midges increased (2.5) during one sorghum- growing season was considered a finite rate of increase (λ). Using the relationship λ = er, r can be estimated as lnλ. However, the management practices farmers use, natural enemies, and other mortality factors reduce the finite rate of increase and number of individuals added per generation per season. This reduction was assumed to be 50%. Thus, r = 0.5lnλ was used to estimate the number of ovipositing sorghum midges per panicle of a susceptible hybrid. Sorghum midges produce 50 to 60% fewer progeny on a resistant hybrid (Melton and Teetes 1984), so 50% of the rate of increase of sorghum midges on a susceptible hybrid was used to estimate the rate of increase on a resistant hybrid. Equation 4 becomes: 50 − N ª § t ·º a. N = N + 0.4582N for a susceptible hybrid……………….…. (5) t+1 t « t ¨ 50 ¸» ¬ © ¹¼ 200 − N ª § t ·º b. N = N + 0.2291N for a resistant hybrid……………..………(6) t+1 t « t ¨ 200 ¸» ¬ © ¹¼ The number of progeny produced (amount of damage) per female sorghum midge decreased as the number of ovipositing sorghum midges increased (Melton and Teetes 1984). Successful ovipositing sorghum midges were defined as females competent in laying eggs and able to produce progeny. Even if multiple eggs are laid in a spikelet, only one sorghum midge develops. To account for damage prevented by competition (DPC), the amount of damage by successful females (NSF) was estimated by using equation (7) (R2 = 0.8758, n = 8, F = 17.64, P < 0.01) based on data by Karanjkar and Chundurwar (1978): 2 Dt = 3.8815 − 0.4114X t + 0.0124X t …………………………………………..……(7) where Dt = percentage of damage per ovipositing sorghum midge per panicle in year t and Xt = number of ovipositing sorghum midges per panicle in year t. With no competition, one sorghum midge per panicle damages 4.25% of spikelets (Karanjkar and Chundurwar 1978) and DPCt = 4.25 – Dt …………………………………………………………………..….(8) NSFt = TOt – (DPCt * TOt)/100……………………………………….………..…….(9) where NSFt = number of successful sorghum midge females per panicle in year t, TOt = total number of ovipositing sorghum midges per panicle in year t and DPCt = percentage of damage prevented per female because of competition in year t. With natural infestation, one ovipositing sorghum midge damaged 0.0015 kg (42 to 48 spikelets) of a susceptible hybrid and 0.00032 kg (nine spikelets) of a resistant hybrid (Hallman et al. 1984). Equation 2 becomes:

397 a. EYLt = 0.0015 kg/female * NSFt * TPN for a susceptible hybrid…….……..(10) b. EYLt = 0.00032 kg/female * NSF t* TPN for a resistant hybrid.……….……(11) Depending on location and irrigation available in the sorghum midge-infested part of Texas, the recommended number of plants per hectare ranged from 74,100 to 197,600 (Stichler et al. 1997). An average of 135,850 plants per hectare was used. Fifteen percent of the sorghum was assumed to be damaged by sorghum midge (Khalema 1993, Ervin et al. 1996), and the number of panicles infested was estimated to be 20,377 per hectare. Equations 10 and 11 become: a. EYLt = 0.0015 kg/female * NSFt * 20,377 for a susceptible hybrid………...(12) b. EYLt = 0.00032 kg/female * NSFt * 20,377 for a resistant hybrid…………..(13) Yield saved (kg/ha) was calculated as a product of percentage of yield saved (Equation 3) and estimated loss if a susceptible hybrid was grown (Equation 12). Price of Sorghum. The average marketing-year price of sorghum was obtained from Texas Agricultural Statistics Service annual bulletins. Area Infested. Sorghum midges usually are not found in the Texas Panhandle. The area of sorghum in the rest of the state was obtained from Texas Agricultural Statistics Service annual bulletins (1980 to 2005). The Cross Timbers, Northeast, Southeast, and Trans-Pecos were excluded because of forests, cities, or deserts. Based on a survey of the importance of sorghum midge in the region of the state excluding the Panhandle, only dryland sorghum was considered when estimating the area in which yield loss was avoided by using a resistant hybrid. Cost of Hybrid Development. Costs incurred in 1980 through 1993 for developing a sorghum midge-resistant hybrid were obtained from Khalema (1993). Operational costs between 1994 and 2000 were obtained from International Sorghum and Millet Collaborative Research Support Program documents (Table 2). The scientist-year cost for 1994 through 2000 was estimated based on the salary scale of The Texas A&M University System. Project Period. Sorghum midge-resistant hybrids are needed in most areas of Texas but yield 12.7% less than susceptible hybrids in the absence of sorghum midge (Peterson 2001). It was assumed that sorghum midge-resistant hybrids released at the end of the project period would be produced for an additional 5 years, until 2005. All costs and benefits between 1980 and 2000 were compounded, and benefits between 2001 and 2005 were discounted. There was no outlaid cost between 2001 and 2005. The total project period was 25 years. Evaluation Criteria. Net present value (NPV) and cost-benefit ratio (CBR) methods were used to evaluate the returns on investment in research. The NPV defined as the compounded/discounted sum of the projected series of net cash flows of breeding a sorghum midge-resistant hybrid was estimated by: T T T T −t Bt j T −t NPV = ( Bt *(1+ i) j + ) − ( Ct *(1+ i) j ) j t j ¦ ¦ j ¦ j=2 j=1 (1+ i) j=0 where NPV = net present value, Bt = benefit (value saved) in year t, i = interest rate, Ct = cost incurred in year t, T = project period (1980-2000, 2001-2005), and j = 0, 1, 2,….,t. The CBR was calculated by dividing the total compounded cost by the compounded/discounted benefit. The interest rate is composed of a real risk-free rate, a risk premium, and an inflation premium (Barry et al. 1995). To eliminate the inflation problem, all cash flows were converted using the consumer price index (CPI) for all U.S. items (U.S. Department of Labor 2007) to 1984 constant dollars.

398 Table 2. Costs of Developing a Sorghum Midge-resistant Sorghum Hybrid in Texas Project Texas Scientist- Total cost TAM 125 and Sorghum year Total cost (constant Year TAM 225a TAM 123b Producers cost (nominal $) C. P. I.c 1984 dollars) 1980 45,416.4 42,600 103,898.0 191,914.4 0.824 232,905.8 1981 45,416.4 42,600 109,626.0 197,642.4 0.909 217,428.4 1982 45,416.4 42,600 113,759.0 201,775.4 0.965 209,093.7 1983 45,416.4 42,600 116,183.0 204,199.4 0.996 205,019.5 1984 45,416.4 42,600 119,722.0 207,738.4 1.039 199,940.7 1985 45,416.4 42,600 122,945.0 210,961.4 1.076 196,060.8 1986 40,237.5 22,920 124,760.0 187,917.5 1.096 171,457.6 1987 39,000.0 24,000 128,556.0 191,556.0 1.136 168,623.2 1988 39,000.0 24,000 133,323.0 196,323.0 1.186 165,533.7 1989 39,000.0 24,000 139,600.0 202,600.0 1.240 163,387.1 1990 37,500.0 24,000 50,000 147,779.0 259,279.0 1.307 198,377.2 1991 39,000.0 24000 50,000 155,245.0 268,245.0 1.362 196,949.3 1992 39,000.0 36,000 50,000 161,831.0 286,831.0 1.403 204,441.2 1993 24,000.0 36,000 50,000 169,000.0 279,000.0 1.445 193,079.6 1994 24,000.0 36,000 167,303.8 227,303.8 1.482 153,376.4 1995 22,560.0 48,000 172,027.9 242,587.9 1.524 159,178.4 1996 24,000.0 48,000 176,751.9 248,751.9 1.569 158,541.7 1997 25,500.0 48,000 181,476.0 254,976.0 1.605 158,863.5 1998 25,500.0 48,000 186,200.0 259,700.0 1.630 159,325.2 1999 25,500.0 48,000 190,924.1 264,424.1 1.666 158,718.0 2000 25,500.0 48,000 195,648.2 269,148.2 1.722 156,299.7 2001 0 0 0 0 1.771 0 2002 0 0 0 0 1.799 0 2003 0 0 0 0 1.840 0 2004 0 0 0 0 1.889 0 2005 0 0 0 0 1.953 0 Total 741,795.9 794,520 200,000 3,116,558.9 4,852,874.8 3,826,600.7 aTAM125 (1980 to 1997) and TAM225 (1998 to 2000) represents funds allocated by an entomologist for research on sorghum resistance to sorghum midge. bTAM123 (1980 to 2000) represents funds allocated by a sorghum breeder for development of a sorghum midge-resistant hybrid. c C. P. I. = consumer price index for all commodities in the U.S. (U.S. Department of Labor 2007).

Results and Discussion

The year 1982 was used to illustrate how parameters to estimate benefit (yield saved) were calculated. Hallman et al. (1984) reported that during 1981, eight and 11 sorghum midges oviposited per panicle of susceptible and resistant hybrids, respectively (Table 1). To determine the number of sorghum midges ovipositing on susceptible hybrids in 1982, eight was substituted for the variable Nt in Equation 5, as 8 + ((0.4582 * 8)(50 – 8)/50) = 11.1 (Table 3). The number of ovipositing sorghum midges per panicle of a resistant hybrid was estimated using Equation 6, as 11 + ((0.2291 * 11)(200 – 11)/200) = 13.4. Using these estimates, the percentage of damage Dt (Equation 7) per ovipositing sorghum midge per panicle of susceptible hybrid in year 1982 was estimated as (3.8815 – (0.4114 *

399

Table 3. Estimated Benefit from Developing a Sorghum Midge-resistant Sorghum Hybrid in Texas Estimated Estimated Estimated number ovipositing damage (%) of successful Estimated yield Average Estimated Benefit (in females per prevented by ovipositing females loss (kg/ha) per Yield marketing area infested Value saved 1984 panicle competition per panicle year saved year price by sorghum (nominal constant Year S R S R S R S R (kg/ha) ($/kg) midge (ha) dollars) C.P.I. dollars) 1980 5.7 9.1 2.3 3.1 5.6 8.8 169.9 57.5 112.5 0.117 168,157.9 2,216,715.2 0.824 2,690,188.4 1981 8.0 11.0 2.9 3.4 7.8 10.6 237.5 69.3 168.2 0.101 186,340.1 3,165,070.2 0.909 3,481,925.4 1982 11.1 13.4 3.4 3.7 10.7 12.9 326.8 84.1 242.8 0.099 230,247.0 5,534,775.8 0.965 5,735,519.0 1983 15.0 16.2 3.7 3.8 14.4 15.6 439.8 101.9 337.8 0.115 139,178.1 5,402,206.0 0.996 5,423,901.6 1984 11.0 19.7 3.4 3.7 10.6 18.9 324.8 123.4 201.4 0.102 172,597.2 3,532,390.4 1.039 3,399,798.3 1985 14.9 23.7 3.7 3.2 14.4 23.0 439.3 149.7 289.5 0.087 174,510.1 4,411,750.9 1.076 4,100,140.2 1986 18.1 2.8 3.8 1.4 17.4 2.8 533.4 18.2 515.3 0.063 159,352.2 5,176,143.5 1.096 4,722,758.6 1987 23.0 3.5 3.3 1.6 22.2 3.4 680.0 22.3 657.6 0.066 110,556.7 4,777,418.8 1.136 4,205,474.3 1988 3.2 4.3 1.6 1.9 3.1 4.2 96.0 27.2 68.8 0.100 89,149.8 615,243.2 1.186 518,754.8 400 1989 4.5 5.2 2.0 2.2 4.4 5.1 134.8 33.2 101.6 0.086 118,560.7 1,033,698.9 1.240 833,628.1 1990 6.3 6.4 2.5 2.5 6.2 6.2 189.0 40.4 148.6 0.092 116,817.8 1,591,389.1 1.307 1,217,589.2 1991 8.9 7.8 3.0 2.8 8.6 7.6 262.9 49.2 213.7 0.091 125,896.8 2,442,875.3 1.362 1,793,594.2 1992 12.2 9.5 3.5 3.2 11.8 9.2 360.0 59.7 300.2 0.079 180,686.2 4,303,717.1 1.403 3,067,510.4 1993 16.4 11.5 3.8 3.5 15.8 11.1 483.2 72.6 410.7 0.092 122,732.8 4,643,502.0 1.445 3,213,496.2 1994 21.5 14.0 3.5 3.7 20.7 13.5 633.7 88.1 545.6 0.086 111,740.9 5,214,622.4 1.482 3,518,638.6 1995 5.3 2.3 2.2 1.2 5.2 2.3 158.4 14.8 143.7 0.114 94,008.1 1,538,877.4 1.524 1,009,762.1 1996 7.5 2.8 2.7 1.4 7.3 2.8 222.1 18.1 203.9 0.130 158,319.8 4,199,288.5 1.569 2,676,410.8 1997 10.4 3.5 3.3 1.6 10.0 3.4 306.8 22.2 284.7 0.090 126,194.3 3,246,282.3 1.605 2,022,605.8 1998 14.1 4.2 3.7 1.9 13.6 4.2 416.2 27.1 389.2 0.083 99,898.8 3,222,205.2 1.630 1,976,813.0 1999 18.8 5.2 3.7 2.2 18.1 5.1 552.7 33.0 519.5 0.067 117,327.9 4,099,792.2 1.666 2,460,859.6 2000 24.2 6.3 3.1 2.5 23.4 6.2 715.5 40.2 675.4 0.072 99,686.2 4,867,732.0 1.722 2,826,789.8 2001 3.1 7.7 1.5 2.8 3.1 7.5 94.2 48.9 45. 3 0.083 101,508.1 379,939.5 1.771 214,533.9 2002 4.5 9.4 2.0 3.1 4.4 9.1 134.0 59.5 74.5 0.093 94,797.6 653,548.3 1.799 363,284.2 2003 6.3 2.0 2.5 1.1 6.2 2.0 188.7 12.9 175.9 0.090 104,271.3 1,656,864.5 1.840 900,469.8 2004 23.6 24.5 3.2 3.0 22.9 23.8 698.5 155.0 543.4 0.089 86,028.3 4,174,791.9 1.889 2,210,053.9 2005 29.6 29.4 1.7 1.6 29.1 28.9 890.8 188.8 702.0 .0.090 75,358.3 4,752,842.4 1.953 2,433,611.1 S = susceptible hybrid, R = resistant hybrid, C.P.I. = consumer price index for all items (U.S. Department of Labor 2007). 11.08) + (0.0124 * 11.082)) = 0.8455, and the corresponding value for panicles of a resistant hybrid was (3.8815 – (0.4114 * 13.38) + (0.0124 * 13.382)) = 0.5969. Thus, the percentage of damage prevented by competition (Equation 8) was calculated as 4.25 – 0.8455 = 3.4045 and 4.25 – 0.5969 = 3.6531 for the susceptible and resistant hybrids, respectively. The number of sorghum midges that oviposited (NSFt) per panicle (Equation 9) of a susceptible hybrid in year 1982 was calculated as (11.08 – (11.08 * 0.034045)) = 10.7 and the value for a panicle of resistant hybrid was calculated as (13.38 – (13.38 * 0.0365) = 12.9. If a susceptible hybrid was grown, the estimated yield loss (Equation 12) in 1982 would be (0.0015 kg/female * 10.69 females/panicle * 20,377 panicles/ha) = 326.8 kg/ha. If a resistant hybrid was grown, the yield loss (Equation 13) would be (0.00032 kg/female * 12.89 females/panicle * 20377 panicles/ha) = 84.1 kg/ha. On the bases of the assumptions, the estimated yield loss ranged between 94.2 and 890.8 and between 12.9 and 188.8 kg/ha if susceptible and resistant hybrids, respectively, were grown. The percentage of yield saved (Equation 3) in 1982 was ((326.85 – 84.04)/326.85) * 100 = 74.3%. Yield saved in 1982 was 326.85 * 0.743 = 242.9 kg/ha, and ranged from 45.3 to 702.0 kg/ha. Yield loss is a function of the number of sorghum midges. According to the 1982 Texas Agricultural Statistics Service report, 1,534,979.8 ha of sorghum were harvested in the region of Texas analyzed. In this area, 15% was considered infested by sorghum midge. Thus, the estimated area in which yield was saved was (0.15 * 1,534,979.76 ha) = 230,247.0 ha. The area analyzed varied between 75,385.3 and 230,247.0 ha. The Texas Agricultural Statistics Service also reported that in 1982, the average marketing year price for sorghum was $0.099/kg. Thus, the benefit (value saved in dollars by using Equation 1) in 1982 was estimated as (242.85 kg/ha * $0.099/kg * 230,247.0 ha) = $5,535,632.9. The estimated nominal benefit ranged between $379,939.5 and 5,534,775.8. This benefit was from grain saved from damage by sorghum midge and did not include benefit from less use of insecticide. Benefit from less insecticide was not included because reduced insecticide might be a cost to another sector of society involved in production, transportation, and distribution of insecticides, or to pesticide applicators. Although recent estimates of value of yield lost are not available, sorghum midge caused $10 million in damage per year between 1950 and 1970 (Harris 1976), and Peterson et al. (1994) estimated annual loss at $28 million. Compounded/discounted benefits and costs are given in Table 4. The choice of interest rate affects the outcome of the analysis and ultimately the decision on whether to fund or implement a project. Thus, long-term ‘social time preference’ (1%), 5, and 15% interest rates were used to incorporate sensitivity analysis. Benefit accrued in 1984 constant dollars during the 25-year project period increased from 66.7 million at 1% compounding/discounting rate to 99.9 and 319.8 million at 5 and 15% rates, respectively. During the same time, accruable cost increased from 4.2 million at 1% to 6.7 and 23.4 million at 5 and 15% interest rates, respectively. The number of sorghum midges on a panicle of a susceptible hybrid was significantly (r = 0.78, n = 26, P < 0.0001) correlated with benefit, whereas correlation between benefit and number of sorghum midges on spikelets of a resistant hybrid was weak (r = 0.39, n = 26, P = 0.05). Association between price, area, and benefit also was weak. Most year-to-year variation (61%) in benefit was explained by the number of sorghum midges on a panicle of a susceptible hybrid.

401

Table 4. Compounded/discounted Benefit and Cost (in 1984 Constant Dollars) of Developing a Sorghum Midge- resistant Hybrid in Texas

Compounded/discounted (1%) Compounded/discounted (5%) Compounded/discounted (15%) Year Benefit Total cost Benefit Cost Benefit Cost Benefit Cost 1980 2,690,188.4 232,905.8 0.0 284,189.4 0.0 617,968.5 0.0 3,811,861.9 1981 3,481,925.4 217,428.4 0.0 262,677.2 0.0 549,430.7 0.0 3,094,391.1 1982 5,772,395.8 209,093.7 6,904,636.7 250,106.9 13,891,958.8 503,208.9 71,436,016.4 2,587,629.1 1983 5,423,901.6 205,019.5 6,423,550.7 242,805.5 12,431,681.7 469,908.4 58,368,036.7 2,206,268.7 1984 3,399,798.3 199,940.7 3,986,530.8 234,446.2 7,421,333.2 436,445.5 31,814,023.2 1,870,969.4 1985 4,100,140.2 196,060.8 4,760,135.5 227,620.5 8,523,897.1 407,596.3 33,363,093.7 1,595,358.7 1986 4,722,758.6 171,457.6 5,428,689.3 197,086.1 9,350,739.1 339,474.3 33,416,850.6 1,213,183.3 1987 4,205,474.3 168,623.2 4,786,222.0 191,909.0 7,930,049.0 317,964.3 25,875,390.2 1,037,503.0 1988 518,754.8 165,533.7 584,545.9 186,527.5 931,609.0 297,274.8 2,775,467.8 885,646.8 402 1989 833,628.1 163,387.1 930,052.5 182,285.8 1,425,787.0 279,447.4 3,878,364.4 760,140.7 1990 1,217,589.2 198,377.2 1,344,976.0 219,131.8 1,983,324.5 323,135.6 4,925,827.4 802,546.4 1991 1,793,594.2 196,949.3 1,961,627.6 215,400.6 2,782,453.3 305,533.1 6,309,642.6 692,843.4 1992 3,067,510.4 204,441.2 3,321,674.2 221,380.5 4,532,109.9 302,052.8 9,383,584.4 625,390.3 1993 3,213,496.2 193,079.6 3,445,302.8 207,007.4 4,521,711.8 271,682.4 8,547,963.7 513,595.5 1994 3,518,638.6 153,376.4 3,735,105.8 162,812.1 4,715,312.3 205,539.0 8,138,824.9 354,768.9 1995 1,009,762.1 159,178.4 1,061,270.1 167,298.1 1,288,740.7 203,156.4 2,030,992.2 320,164.6 1996 2,676,410.8 158,541.7 2,785,083.8 164,979.1 3,253,194.0 192,708.4 4,681,059.1 277,290.4 1997 2,022,605.8 158,863.5 2,083,892.8 163,677.3 2,341,419.0 183,904.4 3,076,130.6 241,611.6 1998 1,976,813.0 159,325.2 2,016,546.9 162,527.6 2,179,436.3 175,656.0 2,614,335.2 210,707.5 1999 2,460,859.6 158,717.9 2,485,468.2 160,305.1 2,583,902.6 166,653.8 2,829,988.6 182,525.6 2000 2,826,789.8 156,299.7 2,826,789.8 156,299.7 2,826,789.8 156,299.7 2,826,789.8 156,299.7 2001 214,533.9 0.0 212,409.8 0.0 204,318.0 0.0 186,551.2 0.0 2002 363,284.2 0.0 356,126.1 0.0 329,509.5 0.0 274,695.1 0.0 2003 900,469.8 0.0 873,987.1 0.0 777,859.7 0.0 592,073.5 0.0 2004 2,210,053.9 0.0 2,123,818.4 0.0 1,818,216.8 0.0 1,263,605.5 0.0 2005 2,433,611.1 0.0 2,315,497.4 0.0 1,906,797.9 0.0 1,209,934.8 0.0 Total 66,753,940.1 4,260,473.5 99,952,151.2 6,705,040.6 319,819,241.523,440,696.8 The net present values (NVP) were 62.4, 93.2, and 296.4 million in 1984 constant dollars (Table 5) when 1, 5, and 15% interest rates, respectively, were used to compound/discount, suggesting research and development of sorghum midge-resistant hybrids were economically justified. Another way of incorporating sensitivity analysis was changing infested area or number of infested plants per hectare. Decreasing infested area or number of infested plants per hectare by 5% decreased NPV at all interest rates. A 5% increase in infested area or number of infested plants per hectare increased NPV at all compounding/discounting rates. The estimated cost-benefit ratios were 1:15.7, 1:14.9, and 1:13.6 when 1, 5, and 15% interest rates, respectively, were used. A dollar invested in research and development of a sorghum midge-resistant hybrid would generate $15.7, 14.9, and 13.6 at 1, 5, and 15% rates, respectively. The cost-benefit ratio is less than one, which means public money expended for developing and distributing a sorghum midge-resistant hybrid is justified. A 5% decrease in infested area or number of infested plants per hectare also resulted in a cost-benefit ratio of less than one regardless of interest rate. Increasing the infested area or number of infested plants per hectare by 5% had similar effect. By considering only the benefit of increased crop return, Khalema (1993) estimated benefit-cost ratios of 246.4, 246.4, and 250.2 at 5, 10, and 20% interest, respectively. Ervin et al. (1996) using the same data estimated benefit-cost ratios at 24.2:1, 14.0:1, and 8.2:1 at 5, 10, and 15% interest rates, respectively. Benefits from development and use of resistant hybrids are classified into quantifiable economic benefit and intangible benefits (Ervin et al. 1996). Intangible benefits include sorghum midge-resistant hybrids being compatible with insecticides (Teetes et al. 1986, Peterson et al. 1994) and natural enemies (Kausalya et al. 1997), and allowing more flexible planting time. Although research and development of a sorghum midge-resistant hybrid were estimated to be cost-effective, realizing benefit is difficult. Resistance to sorghum midge is inherited as a quantitatively recessive trait (Peterson 1985). Because they flower in the middle of the night and are difficult to pollinate, resistant hybrids yield 12.7% less than susceptible hybrids in the absence of sorghum midge (Peterson 2001, Mutaliano 2005). Such characteristics as short glumes that contribute resistance might confer susceptibility to other pests (Ratnadass et al.

Table 5. Sensitivity of Net Present Value and Cost-benefit Ratio to Change in Percent Area or Plant Infested per Hectare % area or plant Compounding/discounting rate (%) Evaluation criteria infested/ha 1 5 15 10 25.3 37.6 118.7 Net present value (million) 15 62.4 93.2 296.4 20 114.0 170.7 544.9 10 6.9 6.6 6.1 Benefit-cost ratio 15 15.7 14.9 13.6 20 27.8 26.5 24.3

403 2002). Parker and Livingston (1996) found panicles of resistant sorghum more difficult to thresh. The methodology in this study is dynamic in that it accounts for changes in insect abundance, area, price, and yield saved over years. The limitation of this method is that sorghum midge abundance estimated by the logistic equation always increases in succeeding years, which might not occur under natural conditions.

Acknowledgment

This study was supported in part by the International Sorghum and Millet Collaborative Research Support Program and by the Ethiopian Institute of Agricultural Research.

References Cited

Barry, P. J., P. N. Ellinger, C. B. Baker, and J. A. Hopkin. 1995. Financial Management in Agriculture. 5th edition. Interstate Publishers, Danville, IL. Cronholm, G., K. Knutson, R. Parker, and B. Pendleton. 2007. Managing Insect and Mite Pests of Texas Sorghum. Texas Agricultural Extension Service B- 1220. Diarisso, N. Y., B. B. Pendleton, G. L. Teetes, G. C. Peterson, and R. M. Anderson. 1998. Floret morphology of sorghum midge-resistant sorghum. Southwest. Entomol. 23: 67-75. Ervin, R. T., T. M. Khalema, G. C. Peterson, and G. L. Teetes. 1996. Cost/benefit analysis of a sorghum hybrid resistant to sorghum midge (Diptera: Cecidomyiidae). Southwest. Entomol. 21: 105-115. Hallman, G. J., G. L. Teetes, and J. W. Johnson. 1984. Relationship of sorghum midge (Diptera: Cecidomyiidae) density to damage to resistant and susceptible sorghum hybrids. J. Econ. Entomol. 77: 83-87. Harris, K. M. 1976. The sorghum midge. Ann. Appl. Biology 84: 114-118. Johnson, J. W., D. T. Rosenow, and G. L. Teetes. 1973. Resistance to the sorghum midge on converted exotic sorghum cultivars. Crop Sci. 13: 754- 755. Karanjkar, R. R., and R. D. Chundurwar. 1978. Losses to jowar cob in relation to adult midge population. Sorghum Newsl. 21: 55-56. Kausalya, K. G., K. F. Nwanze, Y. V. R. Reddy, F. E. Nwilene, and D. D. R. Reddy. 1997. Emergence pattern of sorghum midge and its major parasitoids on midge resistant and susceptible genotypes. Biocontrol Sci. Tech. 7: 259-269. Khalema, T. M. 1993. An estimation of the costs and benefits of developing a midge resistant sorghum hybrid. M.S. thesis, Texas Tech University, Lubbock, TX. Melton, K. D., and G. L. Teetes. 1984. Effects of resistant sorghum hybrids on midge (Diptera: Cecidomyiidae) biology. J. Econ. Entomol. 77: 626-631. Mutaliano, J. A. 2005. Evaluation of the value of sorghum midge resistant hybrids in the USA. M.S. thesis, Texas A&M University, College Station, TX. Parker, R. D., and S. D. Livingston. 1996. Performance of sorghum midge resistant hybrids in South Texas. Results of Insect Control Evaluations on Corn, Sorghum, and Cotton in Texas Coastal Bend Counties. Texas Agricultural Extension.

404 Parker, R. D., L. L. Falconer, and S. D. Livingston. 2003. Comparison of midge damage on resistant and susceptible sorghum hybrids and impact on yield with and without insecticides treatment. Texas Cooperative Extension. Parker, R. D., L. L. Falconer, and S. D. Livingston. 2004. Sorghum midge damage on resistant and susceptible hybrids and impact on yield with and without insecticide treatment. Texas Cooperative Extension. Pendleton, B. B., G. L. Teetes, and M. E. Makela. 1994. Predicting sorghum midge (Diptera: Cecidomyiidae) generations and abundance. J. Econ. Entomol. 87: 993-998. Peterson, G. C. 1985. Breeding sorghum for midge and greenbug resistance in the USA, pp. 361-370. In Proceedings of the International Sorghum Entomology Workshop, 15-21 July 1984, Texas A&M University, College Station, TX. ICRISAT, Patancheru, India. Peterson, G. C. 2001. Development of sorghum midge resistant hybrids with increased grain yield potential. Productive Rotation On Farms In Texas (PROFIT) Annual Report (WWW document). URL http://sorghum.tamu.edu/report_database/ files/ sub37/midgeannualreport- 00-0137.pdf. Peterson, G. C., G. L. Teetes, J. W. Jones, and R. M. Anderson. 1992. Yield and selected traits of sorghum midge resistant and hybrids in the presence and absence of the pest insect. Texas Agricultural Experiment Station PR-4922. Peterson, G. C., G. L. Teetes, and B. B. Pendleton. 1994. Resistance of sorghum to sorghum midge in the United States. Inter. Sorghum Millets Newsl. 35: 48- 63. Ratnadass, A., J. Chantereau, M. F. Coulibaly, and C. Cilas. 2002. Inheritance of resistance to panicle feeding bug, Eurystylus oldi, and sorghum midge, Stenodiplosis sorghicola, in sorghum. Euphytica 123: 131-138. Stichler, C., M. McFarland, and C. Coffman. 1997. Irrigated and dryland grain sorghum in South and Southwest Texas. Texas Agricultural Extension Service AGR-14. Teetes, G. L. 1985. Sorghum midge biology, population dynamics, and integrated pest management, pp. 233-245. In Proceedings of the International Sorghum Entomology Workshop, 15-21 July 1984, Texas A&M University, College Station, TX. ICRISAT, Patancheru, India. Teetes, G. L., and B. B. Pendleton. 2000. Insect pests of sorghum, pp. 443-495. In C. W. Smith and R. A. Frederiksen [eds.], Sorghum: Origin, History, Technology, and Production. John Wiley and Sons, New York. Teetes, G. L., M. I. Becerra, and G. C. Peterson. 1986. Sorghum midge (Diptera: Cecidomyiidae) management with resistant sorghum and insecticide. J. Econ. Entomol. 79: 1091-1095. Teetes, G. L., G. C. Peterson, and R. M. Anderson. 1990. Exploitation of the sorghum midge non-preference resistance mechanism by mixed plantings of resistant and susceptible sorghum cultivars. Texas Agricultural Experiment Station PR-4732. United States Department of Labor/U.S. Bureau of Labor Statistics. 2007. C. P. I. Detailed Report. Data for February 2007. XXXIII (2). Wise, W. S. 1986. The calculation of rate of return on agricultural research from production functions. J. Agri. Econ. 37: 151-161. Wiseman, B. R., W. W. McMillan, and N. W. Widstrom. 1973. Registration of SGIRL-MR-1 sorghum germplasm. Crop Sci. 13: 398.

405 VOL. 34, NO. 4 SOUTHWESTERN ENTOMOLOGIST DEC. 2009

Effectiveness of Spring Burning as a Physical Management Tactic for Thrips in Phleum pratense L. (Poales: Poaceae)

Dominic D. Reisig1,2, Larry D. Godfrey3, and Daniel D. Marcum4

Abstract. Timothy, Phleum pratense L., is an important forage crop in many western U.S. states. Thrips are an important pest of this crop. The effectiveness of field burning as an alternative management tactic for thrips is undocumented. Small-plot burn experiments were initiated with a hand-held torch in two fields in late winter 2007 and in one field on 5 March 2008. Thrips numbers assessed 2 weeks after burning. Grower-burned and nonburned fields were sampled post-burning for thrips on 20 March 2007 and 2008. In 2007, one experiment had relatively low overall abundance of thrips, and adults and larva were found only in nonburned plots. In the other 2007 experiment, more larvae were found in nonburned than burned plots, although adult numbers were similar. In 2008, the number of adults was greater in nonburned plots, while numbers of larvae were relatively low and were similar between burned and nonburned plots. There were no significant differences in thrips abundance between burned or nonburned grower fields during either year. Primarily brachypterous adults were found but both brachypterous adults and larva were found when sampling occurred later or in warmer areas. Timing of burning, in correlation with thrips phenology, was the most likely explanation for differences in population structure between years in burned plots. Population distributions were not aggregated over the spatial scales of these plot experiments. Burning was demonstrated as a possible short-term management tool, but other factors, in addition to burning, probably are important for regulating abundance in the long term.

Introduction

Timothy, Phleum pratense L., is an important cool-season (C3) grass grown as a high-value forage crop in California and other western states. In California, it is grown above 900 m elevation in the northern intermountain areas and, because of its value, is grown with agrichemical inputs, including irrigation during xeric summer conditions. For this region, in 2008, retail prices for supreme hay from alfalfa, Medicago sativa L., for domestic cattle averaged $232.85 per metric ton, while prices for premium retail timothy (<50 kg bales) averaged $368.89 per metric ton (USDA AMS 2008). Timothy hay is purchased mostly on aesthetic appearance. Visual ______1Department of Entomology, North Carolina State University, The Vernon James Research and Extension Center, 207 Research Station Road, Plymouth, NC 27962. 2Research was done through the Department of Entomology, University of California, Davis, to fulfill Ph.D. requirements. 3Department of Entomology, University of California, Davis, One Shields Avenue, Davis, CA 95616. 4University of California Cooperative Extension, Shasta-Lassen County, P.O. Box 9, McArthur, CA 96056.

407 appearance of the hay is considered the most important attribute for producers, followed by hay price (Curtis et al. 2007). “Brown leaf” is a condition that refers to dead leaves, usually in the lower canopy of timothy stands. These brown leaves are obvious compared to the rest of the green foliage in a bale of hay and cause a significant loss in marketability. Thrips abundance is related to the amount of brown leaves, although rarely related to reduced yield of timothy hay (Reisig 2009). The grass thrips, Anaphothrips obscurus Müller (Thysanoptera: Thripidae), is cosmopolitan in distribution (Mound 1997). It is also the main thrips species in California timothy, with western flower thrips, Frankliniella occidentalis Pergrande (Thysanoptera: Thripidae), composing 1-5% of the species richness (unpublished data). Other factors that may interact to cause brown leaf include mites (Acari: Tetranychidae and Eriophyidae), nutrient deficiencies -- especially nitrogen and potassium deficiencies -- seeding rates, plant senescence, and disease. Scientific data on foliar burning as an alternative to chemical management is sparse in IPM, with a few notable exceptions. Burning dormant alfalfa to manage alfalfa weevil, Hypera postica Gyllenhal (Coleoptera: Curculionidae), is well- documented (Tippens 1964; Harris et al. 1971; Schaber and Entz 1988, 1991) although not implemented because of cost (Harris et al. 1971). In this system, the effect of burning on insect pests is time, species, and location dependent (Schaber and Entz 1991). In potato, Solanum tuberosum L., Rifai et al. (2004) showed that open flaming with a tractor-mounted propane flamer provided ~50% control of Colorado potato beetle, Leptinotarsa decemlineata Say (Coleoptera: Chrysomelidae), and was compatible with other IPM strategies. In cool-season grasses, abundance of Eripohyid mites can be reduced though burning Kentucky bluegrass, Poa pratensis L. (Smilanick and Zalom 1983), and timothy (Pepper 1942). However, in these systems, as with any other system, the thermosensitivity of the crop and pest must be known and must be divergent enough for burning to be compatible with agronomic practices (Vincent et al. 2003). To truly be compatible with IPM, the environmental impacts of the management effort must be assessed. The effects of burning on the environment are mixed, usually with positive impacts on soil and water when compared to pesticides, but negative impacts on air and energy (Laguë et al. 2001). For example, in the Sacramento Valley of California, burning was once a major recommended alternative management tool to remove straw residue. This practice mitigates problems such as stem rot and aggregate sheath spot diseases, toxic organic acids and gases, weeds, and algae (Hill et al. 1983). However, because of meteorological conditions when rice, Oryza sativa L., straw is burned during the fall, smoke is not diffused away. Because exposure to rice straw smoke is associated with respiratory illnesses, the California legislature voted to reduce the maximum amount of allowable burning to 25% of the planted fields by 2001 (Lindberg 2003). However, meteorological conditions are very different in the Fall River Valley and Big Valley, which are above 900 m elevation. Agricultural burning permits can be obtained easily in Shasta County and are not necessary in Lassen County, because air quality is not a major concern. Burning as a management tactic has been pursued by growers to manage pests in timothy fields in California, but the effectiveness of this tool has not been scientifically measured. Burning may be a viable management tactic for areas throughout the West and Southwest where the benefits of pesticide reduction outweigh the detriments of air-quality reduction. Physical tactics such as burning may manage thrips (Hewitt 1914), but may decrease plant vigor if the timothy is not dormant (Wasser 1982), damage the root

408 system leading to lower yields (Pepper 1942), and even kill the shallow haplocorms (Powell and Hanson 1973). The thermal tolerances for survival between crop plants and pests can be narrow (Laguë et al. 2001). In pest management, thermosensitivity is used to describe the tolerance of burning temperatures of both the plant and pest (Vincent et al. 2003). The effect of burning on timothy has been studied in domestic settings, such as by Pepper (1942), and a natural setting. Pennsylvania pasture, composed of various species that included timothy, was burned on 22 April. Yield of timothy was less following the first year of burning, although there was no change in species composition (Hughes 1985). Anderson and Romme (1991) surveyed areas of “severe”, “moderate”, and nonburned forest after the 1988 fire at Yellowstone. The severe sites were areas of hot crown fire, while the moderate sites were areas of surface fire. Timothy was in the more moderately burned areas, but not the severely burned areas. However, these studies did not quantitatively account for the thermosensitivity of timothy. Studies presented herein do not address questions of thermosensitivity and their impact on the plant, but were designed to test the hypothesis that burning in the spring could reduce abundance of thrips in timothy.

Materials and Methods

Plot Studies. Small-plot experiments were established in a timothy field on 19 March 2007 and 5 March 2008 at ~1,000 m elevation in the Fall River Valley near Fall River Mills, CA (Field 1). Another small-plot experiment was established on 19 March 2007 in a timothy field at ~1,250 m elevation in Big Valley, near Nubieber, CA (Field 2). In both years, fields were dry enough to successfully burn, and timothy had just broken dormancy in Big Valley. Timothy in the Fall River Valley had broken dormancy for several weeks in 2007 and had just broken dormancy in 2008. Plots were 1.3 x 1.4 m in 2007, with 35 replications per treatment in Field 1, 36 replications in Field 2, and an equal number of nontreated checks. Plots were 1.4 x 2 m in 2008, with 32 replications per treatment and an equal number of nontreated checks. All plots were surrounded with a 2-m buffer of nontreated timothy and the relative spatial location of each plot was measured. Plots in 2007 were completely randomized, while 2008 treated plots were alternated with nontreated plots. As the single treatment in this study, timothy was burned using a hand-held propane torch. The fire was contained with sheet metal within the prescribed plot area. To avoid harming the timothy, the torch was held only long enough to ignite dead stubble near the edges of the sheet metal and was allowed to burn until it extinguished naturally. This method is similar to how producers in the area burn timothy. If the burn was incomplete, the timothy was ignited until all available dead material had burned within the plot. After the burn, the sheet metal was removed from the experimental arena. A single tiller was sampled from the middle of each plot on 5 April 2007 and 20 March 2008. All tillers were collected by carefully grasping the halpocorm and gently pulling the tiller from the ground to avoid dislodging organisms; they were not selected by sight, to avoid sample bias. The vertical distribution of adult and larval thrips on each tiller was recorded by inspecting the leaf blade. Ligules of each leaf were disarticulated from the stem to reveal thrips that may have been concealed. Thrips were found in nonburned plots in Field 2, but not in burned plots. As a result, statistical analysis was not needed for this experiment. In contrast, analyses of variance (ANOVA) were used to compare the number of adult or larval thrips

409 between burned and nonburned plots for Field 1, 2007, as well as Field 1, 2008. Because spatial autocorrelation was expected, data were initially analyzed using a mixed models ANOVA and the covariance was structured with a power correlation, using the REPEATED statement (P < 0.05, PROC MIXED, SAS Institute 2003). Violations of model assumptions were addressed through transformations. Results from this model were compared to a generalized linear mixed model ANOVA (PROC GLIMMIX, SAS) fitted with a Poisson distribution, using Akaike’s Information Criterion (AIC). Plot location was the only random factor in both analyses and the fixed factor was either adult or larval thrips per tiller. Because the AIC was lower in all cases using the generalized linear mixed model and because there was no overdispersion in this model, the GLIMMIX procedure was used for analysis. Field Sampling. Ten tillers were collected from various burned and nonburned fields within the Fall River and Big valleys on 20 March. Tiller samples were collected at least 40 m from a field edge, depending on field size, and were collected at approximately 5-m intervals. The tillers were stored in plastic bags and transported in a cooler to Davis, CA, where they were stored at 4°C before processing the following day. Tillers were washed according to the procedure described by Reisig and Godfrey (2006), but organisms were backwashed into vials with alcohol for later quantification. In 2007, five burned and seven nonburned fields were sampled, while in 2008, seven burned and four nonburned fields were sampled. Because timothy producers prefer to burn when temperatures are cool and fields are dry, calendar date for field burning varies among years. Fields sampled in 2007 were burned at various times from 15 November 2006 to 22 January 2007, while most fields sampled in 2008 were burned in March 2008. Recovered thrips were quantified and separated with the aid of a stereoscope into larva and two adult phenotypes, macropterous or brachypterous. A mixed model ANOVA (PROC MIXED, SAS) was used to analyze the data, with a fixed factor specified as total thrips per tiller and a random factor specified as field number. Total thrips were used because in both years, only one larva was recovered from the washing. Violations of model assumptions were addressed through square-root transformations.

Results

Plot Studies. In Field 1, 2007, adult thrips numbers were statistically similar between burned and nonburned plots (F = 0.03; df = 1, 67; P = 0.8530), but larval thrips were significantly fewer in burned than nonburned plots (F = 48.19; df = 1, 67; P < 0.0001) (Fig. 1). Thrips numbers were relatively low in Field 2, 2007, and no thrips were found in the 36 burned plots. Of 36 nonburned plots sampled, only four adult and seven larval thrips were found, averaging three thrips for every 10 tillers. In Field 1, 2008, adult thrips numbers were significantly less in burned than nonburned plots (F = 5.69; df = 1, 61; P = 0.0201), and larval thrips numbers were similar among burned and nonburned plots (F = 0.06; df = 1, 61; P = 0.8142). Field Sampling. There were no significant differences in thrips numbers between burned and nonburned fields sampled in 2007 (F = 0.94; df = 1, 9; P = 0.3581) or 2008 (F = 0.58; df = 1, 8; P = 0.4667). Of five burned fields sampled in 2007, no macropterous thrips and 65 brachypterous thrips were found. Of seven nonburned fields sampled in 2007, one macropterous thrips and 123 brachypterous thrips were found. Fewer thrips were found in 2008, with 56 brachypterous thrips in the seven burned fields and 27 in the four nonburned fields sampled. No

410 macropterous thrips or larvae were found in 2008, but one larva was recovered in 2007 from a nonburned field.

6 Nonburned 5 Burned

4 Field 1, 2007 1

0 6

5

Field 1, 2008 4 1

0 Larvae Adults

Fig. 1. Mean total number of thrips per tiller in each plot in Field 1, 2007 and 2008. Vertical bars represent standard error of the mean.

411

Discussion

Total thrips numbers were reduced with spring burning in 2007 and 2008. However, some larval and adult thrips remained in burned plots in two out of three experiments. Although plot size was relatively small, migration from nonburned areas into burned areas is an unlikely explanation. Of 180 adult thrips recovered from the field sampling experiment, only one was macropterous. Macroptery is a generalized phenotype adapted for dispersal (Harrison 1980), and macropterous thrips should be more prone to long-range movement. Macropterous thrips are less prevalent than are the brachypterous phenotype in the fall and winter (Reisig 2009), which corroborated similar observations by Kamm (1972) that macropterous phenotypes are more prevalent in the spring and summer. Because sampling in the grower fields and small plots was done at the same time or within a few weeks of each other, thrips in the small-plot experiments were most likely brachypterous adults and larvae. Additionally, because tillers were collected from the plot centers, a thrips would have had to walk at least 65-70 cm from an area of nonburned timothy to the area of burned timothy that was sampled. No thrips were found in burned plots in Field 2, 2007, although thrips numbers were low in nonburned plots in this field. A more likely explanation for why thrips were found in plots after burning is that some thrips were able to escape the potential noxious effects of heat and smoke by harboring in protective microhabitats. Although every effort was made to obtain a complete burn, it is possible that the burn did not cover the entire area of the plot, which allowed thrips to escape mortality. This effect of burning on thrips was significant regardless of the field and the relative abundance of thrips within the specific field. However, results were different between years when the population was separated based on developmental stage. For example, in Field 1, adult numbers were similar between both years in the nontreated plots (three thrips per 10 tillers in both years). In this field in 2007, adult numbers were not significantly affected by the burn treatment, whereas 2008 adult numbers were significantly less in the burn-treated than nonburned plots. In Field 2, 2007, adult numbers were less, but comparable to Field 1, 2007, with an average of one adult thrips per 10 tillers in nontreated plots. In this field, no thrips were found in burned plots. As a result, burning had a directly measureable effect on adult thrips numbers in Field 2, 2007, and Field 1, 2008. In contrast, larval numbers were noticeably different between both years in the nontreated plots of Field 1. In 2007, there were five larvae per tiller and in 2008 there were three larvae per 100 tillers. Nymph abundance in Field 2, 2007, was between these extremes, with an average of two thrips per 10 tillers in the nontreated plots. The difference of larval thrips numbers in Field 1, 2007, in comparison to the other experiments may help explain why adult numbers were similar between burned and nonburned plots. Nymph thrips were comparatively abundant before burning in Field 1 in 2007, and larval thrips were not completely controlled by the burn. As a result, larval thrips in later instars could have survived the burning in 2007 and developed into adults by the time of sampling. In contrast, so few larvae were found in 2008 that the development into adulthood from the time of the burn to sampling could not have had a large effect on the number of adults sampled. Moreover, the difference between adult thrips numbers found was small between burned and nonburned plots. The 2007 burning probably reduced adult numbers, but by the time of sampling, this effect could not be measured.

412 The large larval numbers in Field 1, 2007, could be reflective of the phenology and overwintering stage of this thrips. Most thrips in timothy before April were adults. This included the study where burned and nonburned fields were sampled on 20 March in 2007 and 2008, as well as the 2008 small-plot study. The large nonburned portion of Field 1 was included as one of the sampling points for a nonburned field in 2007, when burned and nonburned fields were compared, and only adult thrips were found. Adult thrips in this field on 20 March (20 larvae per 10 tillers) were almost seven-fold greater than in nonburned plots of the small-plot study, which was sampled 2 weeks later in the same field (three larvae per 10 tillers). The small plots were sampled in early April 2007. At this time, thrips overwintering as adults in Field 1 could have laid eggs that eclosed into larvae and died. Although Field 2 was sampled on the same date in 2007, larval as well as adult numbers were relatively low. Field 2 was in Big Valley, which is ~250 m higher than the Fall River Valley, where Field 1 was located. Topography and climate varies between these valleys and they are separated by the Big Valley Mountain volcanic series mountain range. Although the Fall River Valley has areas of vernal pools and several rivers and creeks, numerous tributaries and drainages within Big Valley create an area called Big Swamp adjacent to Field 2, which created a mesic environment in comparison to that of Field 1. Average temperatures are cooler in Big Valley than those in the Fall River Valley, and timothy in Big Valley breaks dormancy at a later date than that in the Fall River Valley. Hence, average temperatures in Field 2 were probably cooler than those in Field 1. If this were the case, adult thrips in Field 2 would not have accumulated enough degree days to oviposit eggs in timothy that had broken dormancy, and/or the first yearly generation of larvae would not have eclosed from eggs by the time of sampling in early April 2008. Many thrips species have an aggregated population structure (Mound 1997, Funderburk 2002, Seal et al. 2006). When analyzing the small-plot experimental data, a model with spatial autocorrelation structure imposed on the covariance was not as appropriate as a model without this structure. Plots in all experiments were separated by small borders, and total experimental size ranged from ~540 to 640 m2. The closest tillers from the same treatment sampled were ~2.6 to 4 m apart. Because spatial autocorrelation was not detectable within these distances, populations measured in these experiments may have been aggregated over distances greater than those in these experiments (<640 m2) or at smaller distances than those in these experiments (>2.6 m). It is also feasible that the thrips in these experiments were not aggregated or that the population structure changes as the season progresses. For example, the population distribution could be close to even, or random, as adult thrips are quiescent during fall and winter, but change as the plants break dormancy. Populations would become more mobile with warmer temperatures and more macropterous in spring and summer. Resources could also become more or less limiting as the growing season changed, which might affect population structure. Results from the small-plot experiments demonstrate that spring burning can reduce thrips abundance in the short-term, but results from the field sampling observational experiment suggest that other factors influence thrips abundance in burned fields. Although in winter 2006-2007, fields were burned in late-fall and throughout the winter, most fields in winter 2007-2008 were burned in March 2008. No thrips population trends were observed among fields burned at various dates. In 2007 and 2008, abundance of thrips did not differ significantly among burned or

413 nonburned fields sampled on the first day of spring. However, in 2007, one of the five burned fields had an average of ~3 thrips per tiller, whereas the other burned fields averaged ~1 thrips per tiller. This burned field with relatively abundant thrips was not unusual from other burned fields and was one of three, of five burned fields sampled, that was burned on 20 November 2006 on the same ranch, by the same grower. Abundance in nonburned fields sampled in 2007 ranged widely, from six to 34 thrips per 10 tillers. In 2008, abundance ranged from 0 to ~20 thrips per 10 tillers in burned fields to 0 to 24 thrips per 10 tillers in nonburned fields. Because field burn status was not predictive of thrips abundance, other factors such as timothy variety, insecticide treatment history, moisture, quality of the burn (i.e., temperature, coverage, duration), etc., must have influenced abundance. Furthermore, because grass thrips is parthenogenetic, with a relatively short generation time (Köppä 1970), abundance could, presumably, increase rapidly from individuals surviving the burn. Because timothy can be damaged, or even killed by burning (Hughes 1985, Anderson and Romme 1991), this should be approached with caution as a management tool for thrips. Although these studies did not evaluate the impact of burning on timothy, plots in 2008 were visited just before the first yearly harvest; plots that had been burned had visibly shorter timothy than nonburned areas. As a result, the effect of burning on timothy plants could hypothetically offset any yield increase or damage reduction as a result of thrips abundance reduced through burning. The effect of burning on long-term reduction of thrips abundance is questionable. If a market for organic timothy expands in the future, these studies will be an important foundation and reference for producers wishing to manage thrips. If the prescribed burns could be timed with favorable weather conditions and the first generation of larvae presumed to eclose in the late winter or early spring, burning might be a viable management tool for producers of organic and conventional timothy in California.

References Cited

Anderson, J. E., and W. H. Romme. 1991. Initial floristics in lodgepole pine (Pinus contorta) forests following the 1988 Yellowstone fires. International J. Wildland Fire 1: 119-124. Curtis, K. R., M. W. Cowee, W. W. Riggs, and A. A. Cirelli, Jr. 2007. Do producers and horse owners agree on important characteristics of cool season hays? University of Nevada Cooperative Extension Special Publication SP-07-02. Funderburk, J. 2002. Ecology of thrips, pp. 121-128. In R. Marullo and L. Mound [eds.], Thrips and Tospoviruses: Proceedings of the 7th International Symposium on Thysanoptera, Reggio Calabria, Italy, 2-7 July 2001, Australian National Insect Collection, Canberra, Australia. Harris, W. L., C. C. Blickenstaff, N. A. Clark, A. L. Steinhauer, J. W. Neal, R. M. Corn, and R. H. Ratcliffe. 1971. Flaming for control of alfalfa weevil in Maryland. University of Maryland Agricultural Experiment Station Bulletin A- 171. Harrison, R. G. 1980. Dispersal polymorphisms in insects. Ann. Rev. Ecolog. Syst. 11: 95-118. Hewitt, G. C. 1914. Sterility in oats caused by thrips. J. Econ. Entomol. 7: 211-218.

414

Hill, J. E., S. C. Scardaci, J. F. Williams, C. M. Wick, D. M. Brandon, D. S. Mikkelsen, and M. L. Petersen. 1983. Managing pests in rice, pp. 17-31. In M. L. Flint and B. L. P. Ohlendorf [eds.], Integrated Pest Management for Rice: Second Edition. Publication 3280. University of California Statewide Integrated Pest Management Project. Division of Agriculture and Natural Resources, University of California, Oakland, CA. Hughes, H. G. 1985. Vegetation responses to spring burning in an improved pasture in central Pennsylvania, pp. 3-9. In J. N. Long [ed.], Proceedings, Symposium: Fire Management: the Challenge of Protection and Use, 17-19 April 1985, Utah State University, Logan, UT. Kamm, J. A. 1972. Environmental influence on reproduction, diapause, and morph determination of Anaphothrips obscurus (Thysanoptera: Thripidae). Environ. Entomol. 1: 16-19. Köppä, P. 1970. Studies on thrips (Thysanoptem) species most commonly occurring on cereals in Finland. Ann. Agric. Fenn. 9: 191-265. Laguë, C., J. Gill, and G. Péloquin. 2001. Thermal control in plant protection, pp. 35-46. In C. Vincent, B. Panneton, and F. Fleurat-Lessard [eds.], Physical Control Methods in Plant Protection. Springer-Verlag, Berlin, Germany. Lindberg, J. 2003. 2003 progress report on the phase-down burning of rice straw in the Sacramento Valley air basin. California Air Resources Board, California Department of Food and Agriculture, December 2003. Sacramento, CA. Mound, L. A. 1997. Biological diversity, pp. 107-216. In T. Lewis [ed.], Thrips as Crop Pests. CAB International, Wallingford, UK. Mound, L. A., and D. A. J. Teulon. 1995. Thysanoptera as phytophagous opportunists, pp. 3-19. In B. L. Parker, M. Skinner. and T. Lewis [eds.], Thrips Biology and Management. Plenum Press, New York. Pepper, B. B. 1942. Studies of the mite Epitrimerus hystrix Nalepa on timothy. J. Econ. Entomol. 64: 247-255. Powell, J. B., and A. A. Hanson. 1973. Timothy, pp. 277-284. In M. E. Heath, D. S. Metcalfe, and R. F. Barnes [eds.], Forages. Iowa State University Press, Ames, IA. Reisig, D. R. 2009. IPM of thrips in timothy and a density-independent effect on wing diphenism of grass thrips (Anaphothrips obscurus Müller). Ph.D. dissertation, University of California, Davis. Reisig, D. R., and L. D. Godfrey. 2006. Remote sensing for detection of cotton aphid- (Homoptera: Aphididae) and spider mite- (Acari: Tetranychidae) infested cotton in the San Joaquin Valley. Environ. Entomol. 35: 1635-1646. Rifai, N. M., T. Astatkie, M. Lacko-Bartosova, and P. Otepka. 2004. Evaluation of thermal, pneumatic and biological methods for controlling Colorado potato beetles (Leptinotarsa decemlineata Say). Potato Res. 47: 1-9. SAS Institute. 2003. SAS/STAT User's Guide, version 9.1. SAS Institute, Inc. Cary, NC. Schaber, B. D., and T. Entz. 1988. Effect of spring burning on insects in seed alfalfa fields. J. Econ. Entomol. 81: 668-672. Schaber, B. D., and T. Entz. 1991. Effect of annual and/or biennial burning of seed alfalfa stubble on populations of alfalfa weevil and pea aphid. Ann. Appl. Biol. 119: 425-431.

415

Seal, D. R., M. A. Ciomperlik, M. L. Richards, and W. Klassen. 2006. Distribution of chilli thrips Scirtothrips dorsalis (Thysanoptera: Thripidae), in pepper fields and pepper plants on St. Vincent. Fla. Entomol. 89: 311-320. Smilanick, J. M., and F. G. Zalom. 1983. Eriophyid mites in relation to Kentucky bluegrass seed production. Entomol. Exp. Appl. 33: 31-34. Tippens, H. H. 1964. Effect of winter burning on some pests of alfalfa. J. Econ. Entomol. 57: 1003-1004. USDA AMS. 2008. Annual CA hay market summary cumulative. (http://www.ams.usda. gov/mnreports/lsamlgr311.pdf). Vincent, C., G. Hallman, B. Panneton, and F. Fleurat-Lessard. 2003. Management of agricultural insects with physical control methods. Annu. Rev. Entomol. 48: 261-281. Wasser, C. H. 1982. Ecology and culture of selected species useful in revegetating disturbed lands in the West. FWS/OBS-82/56. U.S. Department of the Interior, Fish and Wildlife Service, Office of Biological Services, Western Energy and Land Use Team, Washington, DC.

416 VOL. 34, NO. 4 SOUTHWESTERN ENTOMOLOGIST DEC. 2009

Thrips (Thysanoptera: Thripidae) on Cotton in the Lower Rio Grande Valley of Texas: Species Composition, Seasonal Abundance, Damage, and Control

S. M. Greenberg1, Tong-Xian Liu2*, and J. J. Adamczyk1

Abstract. Species, seasonal abundance, damage, control, and predaceous natural enemies of thrips (Thysanoptera: Thripidae) on cotton, Gossypium hirsutum L., were determined at two sites from 2005 to 2007 in the Lower Rio Grande Valley of Texas. Thrips can stunt growth and reduce yield potential. Seven species of thrips were found. When the cotton field was 0.5 km from commercial onion, Allium cepa L., the predominant species were western flower thrips, Frankliniella occidentalis (Pergande) (61.7%) and onion thrips, Thrips tabaci Lindeman (27.2%). A cotton field 50 km from a commercial onion-growing area was infested with western flower thrips (68.5%) and bean thrips, Caliothrips fasciatus (Pergande) (29.2%). At both fields, cotton seedlings from the cotyledon to 3-4-true leaf stages were most susceptible to onion thrips. Western flower thrips and bean thrips were found predominantly on 5- to 6-true-leaf stage and older cotton. Thrips were found in cotton at the beginning of the growing season and increased gradually in abundance, peaking in mid-May to late June. Predators on cotton were: minute pirate bugs, Orius spp.; lady beetles, Hippodamia spp.; green lacewing, Chrysopa rufilabris (Burmeister); bigeyed bugs, Geocoris spp.; spiders, Argiope spp., and syrphid flies, Syrphus spp. Orius spp. were the most abundant predators (69.5% on nontreated cotton). Heavy rainfall temporarily reduced abundance of thrips on cotton. The systemic insecticides thiamethoxam (Cruiser™) and imidacloprid (Gaucho Grande™) applied to seeds protected cotton from thrips for as long as 30 days after planting.

Introduction

During 2004-2007, Texas ranked first in production of cotton, Gossypium hirsutum L., in the U.S. (average of 2.4 million hectares planted, with mean yield of 748.8 kg lint per hectare). In the Lower Rio Grande Valley of Texas, an average of 89,100 hectares of cotton was planted per year, with a mean yield of 668.6 kg lint per hectare (Williams 2005, 2006, 2007). In Texas, arthropod pests reduced yield of cotton by 78.2 million kg of lint per year. The main insect pests were thrips, Thrips spp.; bollworm, Helicoverpa zea (Boddie),/budworm, Heliothis virescens (F.), complex; fleahopper, Pseudatomoscelis seriatus (Reuter); cotton aphid, Aphis gossypii Glover; sweetpotato whitefly, Bemisia tabaci (Gennadius); lygus bug, Lygus spp., and green ______1BIRU-KSARC-ARS-USDA. 2Department of Entomology, Texas AgriLife Research, Texas A&M University System, Weslaco, TX 78596. *For correspondence: [email protected].

417 stink bug, Acrosternum hilare (Say). In the Lower Rio Grande Valley, arthropod pests reduced cotton yield by 4.4 million kg of lint per year. The main insect pests of cotton in the Lower Rio Grande Valley are: boll weevil, Anthonomus grandis grandis Boheman; bollworm/budworm complex; fleahopper; beet armyworm, Spodoptera exigua (Hübner); cotton aphid; sweetpotato whitefly; spider mites, Tetranychus spp.; and thrips. Reduction of yield potential by thrips in cotton in Texas was estimated at 1.11% (second by percentage of loss) in 2004, 0.61% (second by percentage of loss) in 2005, and 0.39% (first by percentage of loss) in 2006 (Williams 2005, 2006, 2007). In the Lower Rio Grande Valley, significant cotton losses from thrips were not reported. We assumed information about cotton losses from thrips in the Lower Rio Grande Valley was underestimated for several reasons. First, cotton in the Lower Rio Grande Valley was 100% infested with thrips during four of the last five years. Second, the subtropical climate aids in growing agricultural crops year around and some of these can be sources from which thrips inundate cotton. For example, onion is a major vegetable crop in South Texas and cotton usually germinates when onions are harvested; onion thrips, Thrips tabaci Lindeman, an important pest of onions in South Texas (Sparks et al. 1998, Liu and Chu 2004) migrates to cotton fields after onions are harvested. Distribution and abundance of thrips during the cotton-growing season are normally based on observations in the field. Monitoring thrips is difficult because they are almost microscopic and are cryptic feeders, which make thrips difficult to control with insecticide. Also, thrips migrating from other sites to cotton are frequently resistant to insecticide because of prior exposure at other sites. The potential for thrips to develop resistance is great (Immaraju et al. 1992) because only a few insecticides are effective against each pest species (pyrethroids controlled onion thrips and organophosphates controlled western flower thrips, Frankliniella occidentalis [Pergande]) and because thrips that survive an application of insecticide produce thrips that can survive exposure to the same kind of insecticide (Sparks and Liu 1999). Natural enemies, especially predators, may play a significant role in suppressing thrips (Sabelis and van Rijn 1997, Liu 2004). Thrips species composition and the importance of natural enemies on cotton in the Lower Rio Grande Valley of Texas have not been investigated. Identification of species is important in selecting a control strategy, particularly when an insecticide will be used. The objectives of this study were to determine seasonal abundance, species, and damage of thrips, and to evaluate the efficacies of foliar insecticides and seeds treated with insecticide, as well as the abundance of Orius spp. on cotton associated with cropping systems in the Lower Rio Grande Valley.

Materials and Methods

Studies of changes in seasonal abundance of thrips, damage, abundance of Orius spp., and efficacies of foliar insecticides were conducted at the North and South Farms of Kika de la Garza Subtropical Agricultural Research Center-ARS- USDA, Weslaco, TX, during the 2005-2007 cotton-growing seasons. Experimental plots had six rows 200 m long, and each treatment was replicated four times. Cotton was planted with 10 cm between seeds in rows 76.2 (North Farm) and 101.6 cm apart (South Farm). Variety DPL 5415 RR was planted on 7 March 2005, DPL 5415 RR was planted on 11 March 2006, and AMX 262R was planted on 4 March

418 2007. The plants were maintained by using standard cultural practices for the Lower Rio Grande Valley. Seasonal Abundance and Species of Thrips. Species, seasonal abundance, damage, and predators of thrips were assessed each week from the 1- to 2-true-leaf to young boll stages of cotton. Thrips were sampled by visual examination, a modified beat-bucket method, and blue sticky card traps. For visual examination, each plot was crossed diagonally from one corner to another (five sites on each diagonal). Leaves from terminal, blooms, and squares were observed from the bottom to the top of a plant (five plants were examined per sample). The data for all samples were averaged for the plots. We modified the beat-bucket method developed by Knutson (Muegge et al. 2003). The bottom was cut from a white 18.9-liter plastic bucket (35.6 cm deep and 25.4 cm in diameter) and a zip- lock bag was attached to the open bottom. While holding the bucket at a 45-degree angle to the ground, we bent the plant into the bucket so the terminal and as much of the plant as possible were inside the bucket. While still holding the stem near the base of the plant, we rapidly beat the plant against the side of the bucket for 3 to 4 seconds (five plants per bucket in a single sample and five samples from each plot diagonal). Dislodged thrips and predators fell into the zip-lock bag. The bag was removed from the bucket and replaced with a new bag. The bag was closed and stored in a refrigerator. Fifty plants were sampled in each plot. In addition, thrips were sampled using blue sticky card traps (10 x 16 cm) purchased from Oecos, Ltd., Kimpton, UK. One blue sticky card trap was set in the center and two others were 10 m from the ends of each plot. Sticky traps were individually mounted vertically on wooden stakes; the bottom edges of the traps were 3 cm above the top of the plant canopy. The sticky traps were collected and replaced each week. Adult thrips were identified based on criteria of Moritz et al. (2001) in two fields at Weslaco (where commercial onions were grown and within 0.5 km of cotton fields), and in two fields at Raymondville, TX (locations where onions were not grown and at least 50 km from a commercial onion-growing area to cotton) the growing seasons of 2005 and 2007. About 100,000 thrips were identified to species. Foliar Insecticides to Manage Thrips. We compared acephate (Orthene™ 90S; 0.6 kg/ha), Neemix™ 4.5 (0.4 kg/ha), and insecticidal soap Safer™ (3.5 kg/ha) on seedling cotton to a nontreated check during 2006 and 2007. Each plot was 20 m2, and each treatment was replicated three times. Foliar spray was applied on 1st-2nd and 3rd-4th true-leaf stages. Treatments with a volume of 93.5 liters per hectare were applied with a carbon dioxide-pressurized (2.76 MPa) backpack sprayer with 3 TX10 hollow-cone nozzles. The number of thrips and damage were recorded one week after spraying (two rows were buffer rows and not sampled). Damage was estimated using the scale in the Scouting Handbook (2006): none – no thrips or damage (scale = 0); light – only occasional thrips, newest unfolding leaves without brown edges and crinkling, no silvering of the bottom of leaves (scale = 1); medium – thrips found readily, most of the newest leaves showing browning of edges and crinkling, many of the leaf bottoms silvery (scale = 2); and heavy – numerous thrips, terminal of plant showing injury, silvering of leaves very noticeable, plant general appearance ragged, deformed, and stunted (scale = 3). Insecticide-treated Seed to Manage Thrips. Seeds treated with systemic insecticides thiomethoxam (Cruiser™ 5FS at 0.320 mg/seed) and imidacloprid (Gaucho Grande™ at 0.375 mg/seed) were supplied by Syngenta Crop Protection (Greensboro, NC). The six treatments were: 1) American 262 R and 2) DPL 555

419 BG RR not treated, 3) American 262 R and 4) DPL 555 BG RR treated with Cruiser™ 5FS, and 5) American 262 R and 6) DPL 555 BG RR treated with Gaucho Grande™. For this study in 2006, cotton was planted in a greenhouse (five pots per treatment and five plants per pot) 40, 30, 20, and 10 days before cotton was infested in the field with thrips. When cotton reached these ages, the pots with plants were moved from the greenhouse to a harvested onion field with a history of infestation by onion thrips and kept there for one week (Fig. 1). Damage to cotton was recorded by counting the number of scars from 6.4 cm2 of leaf area of 20 leaves per treatment. Cotton Losses from Thrips. Cotton yield was estimated from different plant damage ratings (Scouting Handbook 2006) in a field cage. Samples were harvested by hand and processed on an Eagle Laboratory gin. Data Analysis. Data were analyzed using analysis of variance (ANOVA) (SAS Institute 2007). Means were separated using Tukey’s honestly significant difference test at Į ” 0.05.

Fig. 1. Different ages of cotton grown in a greenhouse (seeds were treated with systemic insecticide) and moved for infesting by thrips after onions had been harvested in an onion field with a history of infestation by onion thrips.

Results and Discussion

Damage. Thrips damaged leaves, leaf buds, and flowers and continued to feed on cotton plants throughout the growing season. In early spring as soon as cotton germinated, thrips used their mouthparts to rupture plant cells and suck the released contents. Injury by thrips to young seedlings can delay growth and retard

420 maturity. The leaves may turn brown on the edges, develop a silvery color, or become distorted and curl upward (Fig. 2A). In 2005 and 2007, on some cotton (North Farm, Weslaco), spots were severe where thrips destroyed the growing point and deformed and stunted developing leaves of cotton. In 2007, we estimated yield from plant damage ratings (Scouting Handbook 2006) of 1-2 to 3-4 true-leaf stages of cotton. When cotton damage was ranked between 2nd (medium) and 3rd (heavy) scales, the yield losses were 1.5% compared with nondamaged cotton (Table 1). In late season, thrips continued feeding on leaves, while adults fed on pollen among the blooms. Damage by thrips to cotton can be divided by visually observable (in early season from cotyledon stage to 3-4-true-leaf stage) and visually unobservable (in late season). Unobservable damage (feeding and ovipositing scars) on plants

A

B

Fig. 2. Damage by thrips on young seedling cotton plants (A), and oviposition and feeding scars on 5th or 6th true cotton leaves (B).

421 Table 1. Cotton Yield Losses from Plots with Different Plant Damage Ratings Damage Scale Score Yield, kg/ha Loss, kg/ha / % 0.48 No damage-light 685.7 ± 4.4a 0 2.6 Medium-heavy 675.7 ± 3.6c 10.1 / 1.5 1.4 Light-medium 680.9 ± 4.5b 4.9 / 0.7 1.8 Light-medium 679.8 ± 2.3b 6.0 / 0.9 Means (± SE) followed by the same letter in a column are not significantly different (Tukey’s honestly significant difference, P < 0.05).

with 5-6 true leaves was seen clearly with the aid of a microscope (Fig. 2B). Description of the thrips damage on cotton is the first step in assessing yield loss. We assumed the different color of spots from feeding and oviposition scars on green leaves of late cotton reduced photosynthesis and transpiration, eventually reducing cotton yield. We did not measure or find published references, but this needs to be studied. Species Composition. In the Lower Rio Grande Valley, we found seven species of thrips (Table 2). Based on the percentage of adults collected, the main species were onion thrips, western flower thrips, and bean thrips, Caliothrips fasciatus (Pergande), at non-onion- and onion-growing sites. At onion-growing areas, the predominant species were western flower thrips (61.7%) and onion thrips (27.2%), while at non-onion growing areas, western flower thrips (68.3%) and bean thrips (29.2%) predominated. At both sites, young cotton plants (from cotyledon to 3-4 true-leaf stages) were most frequently infested by onion thrips (11.3-29.2% at North Farm and 10.2-30.8% at Raymondville Farm), while western flower thrips were found at a frequency of only 1.6-8.7 and 0.2-7.0%, respectively. Western flower thrips and bean thrips were found predominantly on older cotton, with 10.7- 22.2% western flower thrips and 4.4-14.5% bean thrips at North Farm and 5.2- 29.0% western flower thrips and 27.3-57.0% bean thrips at Raymondville Farm,

Table 2. Species of Adult Thrips on Cotton in the Lower Rio Grande Valley of Texas During the 2005-2006 Growing Seasons Percentage in Onion- Non-onion- Common name Scientific name planted area planted area Aelothrips Aelothrips bicolor Hinds 0.3 0.5 Bean thrips Caliothrips fasciatus (Pergande) 7.5 29.2 Citrus thrips Scirtothrips citri (Moulton) 2.6 0.6 Composite thrips Microcephalothrips abdominalis 0.5 0.6 (D. L. Crawford) Onion thrips Thrips tabaci Lindeman 27.2 0.4 Tubulifera Tubulifera: Phlaeothripidae 0.2 0.4 Western flower thrips Frankliniella occidentalis 61.7 68.3 (Pergande) Total 100 100

422 while onion thrips were 0.1-3.5 and 0-5.1%, respectively (2005 and 2007). Fig. 3 shows the percentage of distribution of the main thrips species on the sample date in 2007. Seasonal Abundance. The same trends for thrips during the 2005-2007 cotton-growing seasons were found by visual examination or using a beat bucket (Fig. 4). In early spring as soon as cotton emerged, thrips migrated from a multitude of wild hosts (such as weeds and grass), onion, and other vegetables. At this time, thrips can be a problem when weather is cool and wet and plant growth is slow (spring 2005 [average soil surface temperature from 25 February to 15 March was 15.4ºC, soil moisture was 14.3%, rainfall was 3.7 cm, and from 1 to 20 March 2007 [16.5ºC, 17.1%, and 9.4 cm, respectively]). Cotton development from germination of seed to emergence of a young plant with two cotyledon leaves occurred in about 21 days, while under more favorable conditions, it took only 10 days. Plants were infested with numerous thrips (two to five per plant), silvering of leaves was noticeable, plants appeared ragged and deformed, and growth was stunted. Plants were not tolerant of even a few thrips and were not able to outgrow damage. Damage by thrips on cotton in the 2006 growing season was medium -- thrips found frequently, most of the newest leaves with brown edges and crinkling, and the lower surface of many of the leaves were silvery despite a density of 0.5 thrips per plant. When hot and dry during late spring and summer, thrips migrated from other hosts to succulent, irrigated crops like cotton (peaks of 1.5 thrips per plant in 2005 and 4.0-4.5 per plant in 2007). Heavy rainfall (about 9.6 cm May-June 2007) temporarily reduced the abundance of thrips (0.5 per plant on 17 May to zero per plant on 12 June 2007). Shifts in abundance during the season probably also were affected by frequency of encounter among species and their preference for different cotton stages. Insect management decisions are crucial to the success or failure of the overall cotton crop. By using the beat-bucket method, we can obtain easier and faster data about thrips infestation and abundance (22 seconds to collect insects from five plants in the field and 120 seconds for analyzing samples in a laboratory

A B 100 100

80 80

Onion thrips 60 60

Percent Western Flower 40 thrips Percent 40 Bean thrips 20 20

0 0

7 4 1 2 .2 .0 .1 .0 .07 4.13 4.21 4 5 5 5.18 5.26 6 .09 .15 .23 7.14 7.21 26 02 15 23 29 06 4 6 6 6 4.074.13 4.214.275.04 5.11 5.185. 6. 6.096. 6. 6. 7. 7.21 Date Date Fig. 3. Percentage distribution of the main species of thrips on sample date, A = non-onion-growing sites, B = onion-growing sites.

423

2005

5 5

4 4

3 3

2 2

1 1

0 0 5.03 5.09 5.16 5.23 6.01 6.07 6.14 6.21 6.29 7.06 5.03 5.09 5.16 5.23 6.01 6.07 6.14 6.21 6.29 7.06

Beat bucket Visual observation

2006 5 5

4 4

3 3

2 2 Thrips per plant Thrips per plant 1 1

0 0

9 6 2 9 6 9 6 3 1 9 6 2 6 .0 .2 .0 1 0 0 .1 .2 5 5.1 5.2 5 6 6.12 6. 6.2 7.0 7.1 5. 5.16 5.22 5.29 6. 6 6.19 6 7.03 7.11 Beat bucket Visual observation

2007

5 5

4 4

3 3

2 2

1 1

0 0 4.06 4.19 4.24 5.01 5.09 5.17 5.22 6.12 6.18 6.27 4.06 4.19 4.24 5.01 5.09 5.17 5.22 6.12 6.18 6.27

Beat bucket Visual observation

Fig. 4. Mean (± SE) number of thrips per plant on cotton during 2005 (r = 0.9117; P = 0.0003), 2006 (r = 0.7940; P = 0.036), and 2007 (r = 0.9049; P = 0.0004).

versus 360 seconds by visual observation of five plants in the field). The number of thrips captured by blue sticky card-traps may be more informative (Fig. 5). However, this needs further study to determine effective area of one trap and the relationship of capture to plant density.

424 Predator Abundance. Several species of predators were found on cotton, including minute pirate bug, Orius tristicolor (White), 69.5% of total predators in 2005; bigeyed bug, Geocoris punctipes (Say), 4.2%; lady beetle, Hippodamia convergens (Guerin-Meneville), 10.2%; green lacewing, Chrysoperla rufilabris (Burmeister), 6.2%; several species of spiders and predaceous mites, 9.1%; and syrphids, 0.8%. In 2005, when cotton was not treated with malathion for the Boll Weevil Eradication Program, the number of minute pirate bugs per plant was significantly greater (average per season of 0.595 ± 0.1 [0.549 to 0.684] and 69.5% of the total numbers of predators observed), than when cotton was treated with malathion during 2006 and 2007 (average per season of 0.161 [38.6%] and 0.224 [40.4%], respectively) (Table 3, Fig. 6). Results from 2005 to 2007 showed that insecticides significantly reduced predaceous natural enemies from 0.856 (2005) to 0.417 (2006), and 0.554 (2007) per plant. Only 38.6 and 40.4% of predators were O. tristicolor, which is significantly different from nontreated cotton plants. These low percentages of Orius to total predators in the insecticide-treated plots indicated that Orius might be more susceptible to insecticides than other predators. These results confirmed that predaceous insects are typically more susceptible to insecticides than are phytophagous pest species because of evolution of a mechanism for detoxification of plant secondary compounds (Croft 1990). Natural enemies of thrips can play prominent roles in regulating abundance of thrips on plants under natural conditions (Hoffmann et al. 1996). There is no doubt that

700

600

500

400

300

Thrips per Sticky Blue Card 200

100

0

4.06 4.19 4.28 5.07 5.17 5.29 6.12 6.19 6.27

Observation date

Fig. 5. Mean number of thrips per sticky blue card (2007).

425

Table 3. Mean Number ± SE of Predaceous Arthropods of Thrips per Plant and Percentage of Orius on Insecticide-treated and Nontreated Cotton Plants (Weslaco, TX) Mean number ± SE On treated cotton On nontreated cotton Year Predators Orius per plant Predators Orius per plant 2005 (NF**) 0.811 ± 0.008 0.553 ± 0.011 2005 (NF) 0.784 ± 0.010 0.549 ± 0.010 2005 (SF**) 0.973 ± 0.005 0.684 ± 0.022 Avg. 2005 0.856 ± 0.008 0.595 / 69.5* A 2006 (NF) 0.390 ± 0.01 0.174 ± 0.02 2006 (NF) 0.437 ± 0.02 0.164 ± 0.04 2006 (SF) 0.423 ± 0.04 0.145 ± 0.01 Avg. 2006 0.417 ± 0.02 0.161 / 38.6* B 2007(NF) 0.580 ± 0.01 0.220 ± 0.01 2007 (NF) 0.572 ± 0.01 0.244 ± 0.05 2007 (SF) 0.509 ± 0.003 0.209 ± 0.02 Avg. 2007 0.554 ± 0.007 0.224 / 40.4* C *Number Orius per plant per percentage. **NF = North Farm, SF** = South Farm, SARC-ARS-USDA locations from which data were obtained. All predators: P = 0.001 (between A and B), P = 0.003 (between A and C), and P = 0.1 (between B and C). Orius: P = 0.0001 (between A and B), P = 0.0001 (between A and C), and P = 0.03 (between B and C).

natural enemies can be successfully used to manage thrips in greenhouses (Jacobson 1997). However, the importance of natural enemies in suppressing thrips in the field is controversial. Parrella and Lewis (1997) indicated that natural enemies play an insignificant role in regulating thrips abundance under field conditions. We assumed that predation by natural enemies cannot adequately suppress thrips on cotton. Mean number of thrips per plant during the 2005 growing season was twice greater than O. tristicolor (average 1.054 ± 0.3 versus 0.533 ± 0.1, P = 0.006), and during the year predator numbers peaked and declined not synchronically with thrips. Chemical Control of Thrips. Orthene 90S by foliar spray killed 70.5-76.9%, Neemix 4.5 killed 49.0-50.0%, and insecticidal soap killed only 24.1-29.2% of thrips compared with a nontreated check (P = 0.003) (Table 4). Seed treatments with systemic insecticides protected cotton from damage for 30 days after planting (Fig. 7). Cotton damage for 10 days after planting with seeds treated with systemic insecticides ranged from 6.2 ± 1.2 to 9.8 ± 1.1 scars per 6.5 cm2 leaf area, while damage to the nontreated check was 27.5 ± 3.3 and 32.8 ± 1.6 (P = 0.0001). At 20 days, averages were 6.8 ± 2.8 to 10.0 ± 3.1 compared to a check with 30.0 ± 3.9 and 30.2 ± 10.2 (P = 0.004). At 30 days, averages were 4.2 ± 1.6 to 6.0 ± 2.7, while the check had 21.5 ± 3.2 and 25.8 ± 6.9 scars per 6.5 cm2 of leaf area (P = 0.006). At 40 days, no treatment was significantly different from the check (P = 0.928).

426 1.8 1.6 2005 1.4 1.2 1.0 0.8 0.6 0.4 0.2 0.0 1.8 1.6 2006 1.4 1.2 1.0 0.8 0.6 0.4 No. of Orius ± SE 0.2 0.0 1.8 1.6 2007 1.4 1.2 1.0 0.8 0.6 0.4 0.2 0.0 4/16 4/30 5/14 5/28 6/11 6/25 7/9

Date (m/d)

Fig. 6. Mean (± SE) number of Orius per plant of cotton during 2005, 2006, and 2007 growing seasons.

427 Table 4. Efficacies of Foliar Spray of Selected Insecticides on Young Cotton for Thrips Control 1st-2nd true-leaf stage 3rd-4th true-leaf stage Thrips per plant ± Thrips per plant ± SE / percent Damage SE / percent Damage Treatments survival rating ± SE survival rating ± SE Not treated 2.9 ± 0.7a 1.7 ± 0.3a 2.4±0.4a 2.3 ± 0.3a Orthene 90S, 0.67 ± 0.2c / 23.1 0.8 ± 0.2b 0.7 ± 0.1c / 29.2 0.5 ± 0.2c 8 oz/acre Neemix 4.5, 1.49 ± 0.4b / 51.3 1.5 ± 0.3a 1.2 ± 0.2b / 50.0 1.2 ± 0.2b 6 oz/acre Insecticidal soap, 2.2 ± 0.6a / 75.9 1.8 ± 0.4a 1.7 ± 0.4ab / 70.8 1.9 ± 0.3a 50 oz/acre Means (± SE) followed by the same letter in a column are not significantly different (Tukey’s honestly significant difference, P < 0.05).

Cotton Var.: Americot 262R DP-555 BG RR 50 50 10 d* 20 d 10* d 20 d 30 d 40 d 30 d 40 40 40 d *days after planting

30 30

20 20

10 10

0 Damage (number of thrips punctures) 0 Damage (number of thrips punctures) Control Cruiser Gaucho Gaucho Grande Control Cruiser Grande Treatments Treatments Fig. 7. Mean (± SE) number of thrips scars per 6.4 cm2 (1 square inch) of cotton leaf area with different seed treatments at 10, 20, 30, and 40 days after planting.

Conclusions. Based on our data, we conclude that (1) cotton loss from thrips in the Lower Rio Grande Valley of Texas has been underestimated because damage by thrips is in contradiction with that reported; (2) thrips injury to young seedling plants stunts growth and reduces yield (visible damage). Unobservable damage (feeding and oviposition scars) on plants that reached 5-6 true leaves is assumed to reduce photosynthesis and transpiration, eventually reducing the cotton yield. We did not measure or find any relevant publication on the subject, although these studies deserve to be a priority; (3) in the Lower Rio Grande Valley we collected seven species of thrips on cotton. Where onions were grown, the

428 predominant species were western flower thrips (61.7%) and onion thrips (27.2%), while at the non-onion locations (at least 50 km from a commercial onion-growing area to cotton), western flower thrips (68.3%) and bean thrips (29.2%) predominated; (4) onion thrips occurred with greater frequency from cotyledon to 3rd-4th true-leaf stages of cotton, while western flower thrips and bean thrips were predominantly found on older cotton; (5) an improved beat-bucket method was easier and faster for collecting and evaluating numbers of insects compared with visual observation; (6) mean number of thrips per plant was two times greater than O. tristicolor, and during the year the predator peaked and declined not synchronically with thrips; and (7) foliar spray with Orthene 90S in early stages could control 70-80% of thrips while biopesticides like Neemix 4.5 and insecticidal soap controlled only 24-50% of thrips. Seed treatments with systemic insecticides protected cotton from thrips for 30 days. Because thrips significantly damage early cotton stages, the main goal of studies on thrips control was to show how cotton could be controlled with minimum losses and maximum protection of the environment. Seed treatment with systemic insecticides better satisfied these requirements, but research is needed to select effective biorational and botanical insecticides for sucking insects.

Acknowledgment

We thank Y. M. Zhang (Texas AgriLife Research, Weslaco), J. Alejandro, J. Caballero, F. Garza, L. Leal, and J. Yarrito (KSARC-ARS-USDA) for technical assistance.

References Cited

Croft, B. A. 1990. Arthropod Biological Control Agents and Pesticides. John Wiley and Sons, New York. Hoffmann, M. P., C. H. Petzoldt, and A. C. Frodsham. 1996. Integrated Pest Management for Onion. Cornell University, Cornell, NY. Immaraju, J. A., T. D. Paine, J. A. Bethke, K. L. Robb, and J. P. Newman. 1992. Western flower thrips (Thysanoptera: Thripidae) resistance to insecticides in coastal California greenhouses. J. Econ. Entomol. 85: 9-14. Jacobson, R. J. 1997. Integrated pest management (IPM) in glasshouses, pp. 639- 666. In T. Lewis [ed.], Thrips as Crop Pests. CAB, Wallingford, Oxon, UK. Liu, T.-X. 2004. Seasonal population dynamics, life stage composition of Thrips tabaci (Thysanoptera: Thripidae), and predaceous natural enemies on onions in south Texas. Southwest. Entomol. 29: 127-135. Liu, T.-X., and C.-C. Chu. 2004. Comparison of absolute estimates of Thrips tabaci (Thysanoptera: Thripidae) with field visual counting and sticky traps in onion field in south Texas. Southwest. Entomol. 29: 83-89. Moritz, G., L. A. Mound, and D. C. Morris. 2001. Thrips ID: Pest Thrips of the World. CD-ROM. ACIAR, Canberra, Australia. Muegge, M. A., A. Knutson, B. Baugh, W. Multer, R. Baker, and S. Downing. 2003. Development of a reliable and efficient sampling plan for cotton fleahopper and western tarnished plant bug using the beat bucket sampling method, pp. 978-979. In Proc. Beltwide Cotton Production Conf. National Cotton Council of America, Memphis, TN.

429 Sabelis, M. W., and P. C. J. van Rijn. 1997. Predation by insects and mites, pp. 259-354. In T. Lewis [ed.], Thrips as Crop Pests. CAB, Wallingford, Oxon, UK. SAS Institute. 2007. SAS/STAT Version 9.1. Cary, NC. Scouting Handbook. 2006. ANR-409. Alabama Cooperative Extension System. Sparks, A. N., J. Anciso, D. J. Riley, and C. C. Chambers. 1998. Insecticidal control of thrips on onions in south Texas: insecticide selection and application methodology. Subtrop. Plant Sci. 50: 58-62. Sparks, A. N. Jr., and T.-X. Liu. 1999. Thrips on Onions. Texas Agricultural Extension Service, The Texas A&M University System. Williams, M. R. 2005. Cotton Insect Losses. In Proc. Beltwide Cotton Production Conf. National Cotton Council of America, Memphis, TN, CD-ROM. Williams, M. R. 2006. Cotton Insect Losses. In Proc. Beltwide Cotton Production Conf. National Cotton Council of America, Memphis, TN, CD-ROM. Williams, M. R. 2007. Cotton Insect Losses. In Proc. Beltwide Cotton Production Conf. National Cotton Council of America, Memphis, TN, CD-ROM.

430 VOL. 34, NO. 4 SOUTHWESTERN ENTOMOLOGIST DEC. 2009

Tritrophic Interactions among Host Plants, Whiteflies, and Parasitoids

Shoil M. Greenberg1, Walker A. Jones2, and Tong-Xian Liu3

Abstract. Effects of cotton, Gossypium hirsutum L.; green bean, Phaseolus vulgaris L.; and sweet potato, Ipomoea batatas (L.) Lam.; on mortality and development of sweetpotato whitefly, Bemisia tabaci (Gennadius) biotype B; bandedwinged whitefly, Trialeurodes abutilonea (Haldeman); and greenhouse whitefly, T. vaporariorum (Westwood); and on the key biological parameters of an exotic parasitoid species, Eretmocerus mundus Mercet, and an indigenous parasitoid, Encarsia pergandiella Howard, were compared in the laboratory. Cotton was most suitable for sweetpotato whitefly, and bean was most suitable for greenhouse whitefly. No significant differences were found between these two whitefly species on sweet potato. Preimaginal mortality of sweetpotato whitefly on cotton was 35.2% versus 77.3% of greenhouse whitefly. Developmental time of sweetpotato whitefly was significantly shorter (17.5 days) than that of greenhouse whitefly (23.2 days). The mortality and developmental time of bandedwinged whitefly did not differ on the different host plants. Parasitism by Er. mundus was greatest in sweetpotato whitefly and least in greenhouse whitefly when both whiteflies were reared on cotton. Parasitism of bandedwinged whitefly was intermediate. Parasitism by En. pergandiella was significantly greater than that by Er. mundus attacking the same whitefly species reared on bean or cotton, except parasitism of sweetpotato whitefly. Emergence of Er. mundus was greatest from sweetpotato whitefly on cotton, and least for bandedwinged whitefly on bean. Emergence of En. pergandiella was significantly greater than that of Er. mundus among host plants and whitefly species except sweetpotato whitefly.

Introduction

Whiteflies (Hemiptera: Aleyrodidae) are among the most widespread and economically important insect pests worldwide. The sweetpotato whitefly, Bemisia tabaci (Gennadius), may feed on 506 plants in 74 families (Greathead 1986). Sweetpotato whitefly biotype B ranks among the most noxious insects attacking agronomic and ornamental crops (Cock 1993). The bandedwinged whitefly, Trialeurodes abutilonea (Haldeman), is a polyphagous feeder on ≈140 species of plants, including many important species in the genus Hibiscus (Russell 1963, Liu and Stansly 2000). Similarly, the greenhouse whitefly, Trialeurodes vaporariorum (Westwood), is an important pest of vegetables and ornamental crops in greenhouses. ______1Beneficial Insects Research Unit, Kika de la Garza Subtropical Agricultural Research Center, Agricultural Research Service, U.S. Department of Agriculture, Weslaco, TX. 2European Biological Control Laboratory, ARS-USDA, Montpellier, France. 3Vegetable IPM Laboratory, Texas AgriLife Research, Texas A&M University System, Weslaco, TX. Correspondence: S. M. Greenberg, Research Entomologist, BIRU-SARC-ARS-USDA, Weslaco, TX; E-mail address: [email protected].

431 Although the bioecology of sweetpotato whitefly and greenhouse whitefly is well understood, little information is available on the bioecology of bandedwinged whitefly under different geographical and environmental conditions. During 2008, 30,355 ha of cotton in the Lower Rio Grande Valley of Texas were infested by bandedwinged whitefly while only 4,047 ha were infested by sweetpotato whitefly (Williams 2009). This contrasted with the U.S. trend where sweetpotato whitefly was more damaging than bandedwinged whitefly in cotton during the same year (147,138 versus 109,853 ha infested by sweetpotato whitefly and bandedwinged whitefly, respectively; Williams 2009). Liu and Stansly (2000) speculated that bandedwinged whitefly, originally restricted to Malvaceae had adapted to other host species in areas where malvaceous hosts have become rare. If this is the case, the adaptation process probably will continue and the host range of bandedwinged whitefly may continue to expand. Information on the interactions among plant species, whiteflies, and their natural enemies is important in the development of locally appropriate integrated pest management strategies including biological control. At present, whitefly control programs in the southwestern U.S. are based mostly on the application of insecticide (Palumbo et al. 2001). However, frequent and intensive use of broad-spectrum insecticides increases the risk of resistance to insecticide, environmental contamination, and human exposure. Conservation biological control is one of the most economically feasible alternatives to conventional pest management practices in field crops. Among the most effective natural enemies of Bemisia whiteflies are parasitoids in the family Aphelinidae (Hymenoptera) including Eretmocerus mundus Mercet (exotic) and Encarsia pergandiella Howard (indigenous). Er. mundus is the most common parasitoid of sweetpotato whitefly in southern Europe and has been successfully established in many parts of the US. En. pergandiella is a heteronomous hyperparasitoid (autoparasitoid); females develop as primary parasitoids on immature whiteflies, but males develop as secondary parasitoids on females of their own or on related species (Hunter 1989). En. pergandiella can cause 94% parasitism of sweetpotato whitefly in South Texas (Goolsby et al. 1998). Encarsia formosa Gahan and Er. eremicus Rose and Zolnerowich are perhaps the most effective species for control of greenhouse whitefly (Hoddle et al. 1998, Gelman et al. 2005). However, no information is available on the use of these species to control bandedwinged whitefly (Liu and Stansly 2000) and little information is available on the use of Er. mundus and En. pergandiella to control greenhouse whitefly. This information is important in areas such as the southwestern US where both pests species co-exist in the field and greenhouse. Most investigations tend to focus on direct interactions between whiteflies and their host plants or the insects and their natural enemies. The full set of tritrophic interactions needs to be assessed to provide a basis for effective pest management strategies against whiteflies. Qualities of the host plant may have indirect effects on the fitness of the third trophic level. For example, high host/pest densities leading to abundant parasitoids or low parasitoid:host ratio may result in longer developmental time and smaller size of parasitoid progeny and a greater proportion of male progeny because of possible differences in nutritional quantity. The objectives of this study were to determine the effects of three host plants, bean, cotton, and sweet potato, on key fitness parameters of three whitefly species, sweetpotato whitefly, bandedwinged whitefly, and greenhouse whitefly, and two parasitoid species, Er. mundus and En. pergandiella, developing on these

432 pests and host plants. Parameters measured included survivorship, developmental time, sex ratio, preovipositional period, daily fecundity, and size. Knowledge of the interrelationship among host plants, whiteflies, and their parasitoids is critical to developing rearing techniques, making decisions in augmentative releases, developing predictive models, and understanding the mechanisms involved in competition by parasitoids.

Materials and Methods

Host Plants. Bean, cotton, and sweet potato were used as host plants in these studies. Leaves were excised and each leaf petiole was placed in a floral aquapic filled with a hydroponic solution (Aqua-Ponics International, Los Angeles, CA). Excised leaves readily rooted and did not deteriorate under fluorescent lights (20 watt, Vita-Life©, Duro-Test Lighting, Elk Grove, IL) in an incubator. Whitefly Cultures. A culture of sweetpotato whitefly originated from adults collected from cabbage, Brassica oleracea L., in Hidalgo County, TX, and maintained on sweet potato in a greenhouse of the Subtropical Agricultural Research Center, ARS-USDA, Weslaco, TX. A culture of greenhouse whitefly originated from individuals received from the Department of Entomology, University of Georgia, Griffin, GA, where they were reared on green bean. A culture of bandedwinged whitefly was started from individuals collected from cotton in Hidalgo County, TX, and maintained on cotton in a greenhouse. Before the experiment, the sweetpotato whitefly, bandedwinged whitefly, and greenhouse whitefly were cultured for three generations on the three host plants (cotton, bean, and sweet potato). We and others (Van Boxtel et al. 1978, Dorsman and van de Vire 1987, Liu and Stansly 2000) observed that when adult whiteflies were transferred from one plant species to another, the insects required a period of adjustment for at least three generations on the host plant. Parasitoid Cultures. Er. mundus was originally collected from sweetpotato whitefly on cotton near Murcia, Spain, and provided by USDA, APHIS Mission Plant Protection Center, Mission, TX (MPPC culture # M92014). En. pergandiella was collected from sweetpotato whitefly on cotton at Weslaco, TX. We maintained all the parasitoid cultures on sweetpotato whitefly reared on sweet potato. Host Plant Effects on Whiteflies. We determined mortality and developmental time separately by the whitefly instar. Whiteflies were confined within a 4.5-cm-diameter clip cage to the underside of each excised test leaf. Each rooted leaf with eggs was placed in a 120 x 25-mm dish covered with polyester organdy for ventilation. Hydroponics solution was added to floral aquapic as needed. Dishes were kept in an environmental chamber at 25 ± 1ºC, 60 ± 5% relative humidity, and a photoperiod of 16:8 light:dark hours. The development of each instar was monitored daily with the aid of a dissecting microscope until adult eclosion, using the morphological descriptions by Gill (1990) and Lopez-Avila (1988). Number of female progeny was determined by determining the gender of 100-150 adults from each treatment. The preoviposition period was recorded by individually confining 10-15 newly-emerged females per treatment within a clip cage to the underside of tested plant leaves. The leaves were inspected daily until the whitefly species started to oviposit. Number of eggs deposited per female per day was determined from 10 females (2 days old) per treatment during the first 3 days of the experiment. Size of 100-150 fourth-instars per treatment also was recorded. Survival on different whitefly stages (egg, 1st, 2nd, 3rd, and 4th instars, and adult) was recorded.

433 Interactions among Host Plant, Whitefly Species, and Parasitoids. We determined parasitization, development, and emergence rates, as well as progeny sex ratio, longevity, and female size. The leaves were infested with whitefly species as described previously. Second instars were used for parasitization by Er. mundus and third instars by En. pergandiella (Jones and Greenberg 1998, 1999). When the designated instar was reached, all but 35 nymphs were carefully removed by hand with an entomological needle. Subsequently, two mated female parasitoids (<2 days old) were released and confined with the nymphs in a clip cage. After 3 hours, parasitoids were removed, and the leaf with parasitized nymphs was returned to the environmental chamber. Each treatment was replicated three times (§105 total nymphs per treatment). We found this yielded the most satisfactory amount of parasitism while minimizing any advance in development of the exposed nymphs during exposure to the parasitoid. We confined two wasps for the experiment, because using only one would necessitate a much longer exposure period to achieve the desired amount of parasitism necessary for comparisons. Longer exposure times allow individual hosts to transform into the next instar, particularly among the younger stages. Following an initial 10-day incubation period, test leaves were examined daily for development and emergence of parasitoids. Time of emergence was recorded for each individual. All emerged adult parasitoids were placed individually into glass vials (1-cm diameter by 3 cm long) and fed with droplets of undiluted honey to determine the longevity of progeny. Mortality was checked daily at 1100 hours. The number of female parasitoid progeny was recorded by examining the antennae, which are sexually dimorphic (average 50 individuals per treatment). The number of parasitized nymphs produced per female parasitoid was recorded by exposing one parasitoid female daily (during the first 4 days) to a range of numbers of whitefly nymphs. The size of 50 parasitoid females was determined by measuring body length from the frons to the tip of the abdomen. Effects of whitefly species reared on bean and cotton on key biological parameters of Er. mundus and En. pergandiella were studied. Data Analysis. Data were analyzed by using analysis of variance (ANOVA) and the independent t-test function of SYSTAT (Wilkinson et al. 1992). Means were separated using the Tukey-Kramer test (TUKEY option of the LSMEANS statement, SAS Institute 1999). The percentage of parasitism was arcsine transformed for analyses (Sokal and Rohlf 1981), but untransformed means are shown.

Results

Host Plant Effect on Whiteflies. Total mortality of sweetpotato whitefly on cotton (35.2 ± 2.5%) was significantly less than on bean (59.8 ± 4.9%), and significantly greater than on sweet potato (26.1 ± 1.4%) (P = 0.007) (Fig. 1). Total mortality of greenhouse whitefly on bean (20.6 ± 1.5 %) was least, and significantly less than that on sweet potato (30.5 ± 3.3%) or cotton (57.4 ± 4.4%) (P = 0.003). Mortality of bandedwinged whitefly on cotton (76.1 ± 4.2%), bean (68.5 ± 1.8%), and sweet potato (72.5 ± 5.6%) were not significantly different (P = 0.286). Total mortality of eggs and young nymphs (1st and 2nd) of each whitefly species on each host plant was 2.5-3.5-fold greater than for older instars (Fig. 2). Developmental time from egg to adult of sweetpotato whitefly was significantly faster on cotton (17.5 ± 0.9 days) than on bean leaves (22.1 ± 1.3 days), while that of greenhouse whitefly was opposite, 20.5 ± 1.3 days on bean and 24.6 ± 2.0 days on cotton (P = 0.001) (Fig. 3). Developmental time of bandedwinged

434 whitefly was not dependent on host plant (23.2 ± 3.9 days on cotton and 21.9 ± 3.6 days on bean, P = 0.188). The greatest portion of the total developmental time was during the egg stage (26.7 to 42.9%) and the fourth instar (19.4 to 33.9%) for all whitefly species or host plants. Sweetpotato whitefly began to oviposit during the first 24-30 hours after emergence, while there was a 48-52 hour preovipositional period for bandedwinged whitefly, and 36-40 hours for greenhouse whitefly (Table 1). The preovipositional period was not significantly different among whitefly species on the different host plants (P = 0.3 for sweetpotato whitefly, P = 0.6 for bandedwinged whitefly, and P = 0.2 for greenhouse whitefly). The number of eggs deposited per day by sweetpotato whitefly females on cotton was significantly greater than on the other hosts (P = 0.001), while oviposition by greenhouse whitefly was significantly greater on bean and sweet potato (P = 0.002). The number of eggs deposited per day by bandedwinged whitefly females was not significantly different between host plants (P = 0.41), but was significantly less than that for the other two species of whiteflies (P = 0.001).

100 B. tabaci

T. abutilonea d d T. vaporariorum 75 d e a

50 b f c 25 g Mortality from egg to adult, %

0 Cotton Bean Sweet potato Host Plant

Fig. 1. Total mortality of whitefly species after development on different host plants. Bars of the same whitefly species with different letters show significantly different total mortality on the different host plants at the 5% level as determined by Tukey’s studentized range test.

435 T. abutilonea 40 Cotton Bean 30 Sweet potato

20

10 Stage-specific mortality, % 0 Egg L1 L2 L3 L4

B. tabaci 40

30

20

10 Stage-specific mortality, % 0 Egg L1 L2 L3 L4

T. vaporariorum 40

30

20

10 Stage-specific mortality, % 0 Egg L1 L2 L3 L4 Stage

Fig. 2. Whitefly species (Trialeurodes abutilonea; Bemisia tabaci; and T. vaporariorum) stage-specific mortality on different host plants.

436 30 T. abutiloneus 25 a cotton a 20 a bean 15 sweetpotato 10

Development time, d 5

0 Eggs 1st 2nd 3rd 4th Total

30 B. tabaci 25

20 a b 15 c

10

Development time, d 5

0 Egg 1st 2nd 3rd 4th Total 30 T. vaporariorum 25 a 20 b b 15

10 Development time, d 5

0 Egg 1st 2nd 3rd 4th Total Stage Fig. 3. Whitefly species (Trialeurodes abutilonea, Bemisia tabaci, T. vaporariorum) stage-specific developmental time on different host plants.

437 Table 1. Effect of Host Plant on Biological Parameters of Three Whitefly Species Size of red-eyed Preovipositional Eggs/female/ nymphs Female Host plant period, days day (length x width), mm progeny, %

Bemisia tabaci Cotton 1.2 ± 0.4a 7.6 ± 1.8a 0.701 x 0.423b 67.1 ± 6.2a Bean 1.1 ± 0.3a 4.5 ± 1.6c 0.680 x 0.385c 59.0 ± 7.7b Sweet potato 1.3 ± 0.2a 5.2 ± 0.5b 0.747 x 0.424a 66.2 ± 6.8a

Trialeurodes abutiloneus Cotton 2.1 ± 0.7a 1.9 ± 0.7a 0.655 x 0.347a 61.0 ± 5.5a Bean 2.3 ± 0.7a 1.7 ± 0.4a 0.638 x 0.337a 60.0 ± 3.5a Sweet potato 2.2 ± 0.6a 1.8 ± 0.6a 0.648 x 0.344a 62.0 ± 4.3a

Trialeurodes vaporariorum Cotton 2.1 ± 0.5a 3.9 ± 0.6b 0.804 x 0.295b 58.1 ± 7.6a Bean 1.5 ± 0.5a 4.8 ± 0.5a 0.855 x 0.520a 64.3 ± 6.5a Sweet potato 1.6 ± 0.1a 4.5 ± 0.5a 0.838 x 0.514a 63.0 ± 5.7a Means (± SE) followed by different letters within whitefly species in a column are significantly different at the 5% level as determined by Tukey’s studentized range test.

The largest size of sweetpotato whitefly red-eyed nymphs was on sweet potato, then cotton, with the smallest on bean (P = 0.019), but the reverse was recorded for greenhouse whitefly (P = 0.044). The size of bandedwinged whitefly nymphs was not significantly different among host plants (P = 0.33). The percentage of female progeny was not significantly different among host plants (P = 0.218), except significantly more female sweetpotato whitefly progeny were on cotton and sweet potato than on bean (P = 0.016). Host Plant and Whitefly Effect on Parasitoids. The greatest rate of parasitism by Er. mundus was on sweetpotato whitefly maintained on cotton (88.7%) and bean (79.0%). The least parasitism by this parasitoid was on greenhouse whitefly (35.4% on cotton and 38.4% on bean). Results for bandedwinged whitefly were intermediate (52.7% on cotton and 54.7% on bean) (P = 0.001). Parasitism by En. pergandiella was significantly greater than that by Er. mundus attacking the same whitefly species on bean or cotton, except parasitism in sweetpotato whitefly (P = 0.002). The percentage of parasitism was not significantly different within Er. mundus and En. pergandiella on each of the three whitefly species reared on bean or cotton, except parasitism of En. pergandiella in greenhouse whitefly reared on cotton was significantly less than on bean (Fig. 4A). Percentage of emergence of Er. mundus depended on the host species and host plant. The greatest emergence was from sweetpotato whitefly on cotton (93.5%); least was from bandedwinged whitefly on bean (6.1%). Emergence of Er. mundus from sweetpotato whitefly reared on cotton (93.5%) was significantly greater than when reared on bean (62.1) (P = 0.001). Emergence of En. pergandiella adults was significantly greater than that of Er. mundus, except from sweetpotato whitefly (P = 0.003) (Fig. 4B).

438

Fig. 4. Effects of whitefly species and host plants on parasitism by Eretmocerus mundus and Encarsia pergandiella (A) and on emergence of parasitoids (B). Parasitoid species: Eretmocerus mundus - 1, 3, 5; Encarsia pergandiella - 2, 4, 6.

Developmental time for Er. mundus was significantly longer on sweetpotato whitefly and on greenhouse whitefly than on bandedwinged whitefly reared on either bean (P = 0.0001) or cotton (P = 0.003) (Table 2). Developmental time for En. pergandiella was significantly longer on sweetpotato whitefly than on bandedwinged

439

Table 2. Effects of Host Plant and Whitefly Species on Development, Longevity, and Size of Eretmocerus mundus and Encarsia pergandiella Eretmocerus mundus Encarsia pergandiella Whitefly host Bean Cotton Bean Cotton Developmental time, days ± SE B. tabaci 18.1 ± 0.2Aa 17.0 ± 0.1Ab 17.3 ± 0.2Aa 15.7 ± 0.1Ab T. abutilonea 15.2 ± 0.5Ba 15.6 ± 0.4Ba 13.8 ± 0.1Ba 13.6 ± 0.2Ba T. vaporariorum 17.2 ± 0.3Aa 17.6 ± 0.4Aa 13.0 ± 0.1Bb 15.6 ± 0.3Aa

Longevity, days ± SE B. tabaci 5.7 ± 0.4Bb 7.1 ± 0.3Aa 8.8 ± 0.6Bb 10.3 ± 0.2Aa T. abutilonea 7.0 ± 0.6Ba 6.8 ± 0.7Aa 12.1 ± 0.5Aa 11.3 ±1.0Aa T. vaporariorum 10.8 ± 0.7Aa 7.0 ± 0.9Ab 12.8 ± 1.5Aa 7.9 ± 1.6Bb

Female size, mm ± SE B. tabaci 0.517±0.007Bb 0.556±0.006Aa 0.498±0.005Ab 0.526±0.005Aa T. abutilonea 0.503±0.004Cb 0.520±0.004Ba 0.462±0.007Ba 0.493±0.008Ba T. vaporariorum 0.541±0.01Aa 0.504±0.01Cb 0.446±0.008Ba 0.420±0.004Cb

Female progeny, % ± SE B. tabaci 42.7 ± 9.0Bb 55.0 ± 2.5ABa 100 100 T. abutilonea 60.0 ± 1.6Aa 61.0 ± 2.4Aa 100 100 T. vaporariorum 65.6 ± 2.9Aa 46.7 ± 3.3Bb 100 100 Means (± SE) for each parasitoid species in each whitefly host between the two host plants (sub-row) with the same lower-case letter and those among the three whitefly hosts on the same host (sub-column) followed by the same upper-case letters are not significantly different at the 5% level as determined by Tukey’s studentized range test.

whitefly or greenhouse whitefly when reared on bean (P = 0.0001), but significantly longer on sweetpotato whitefly and greenhouse whitefly than on bandedwinged whitefly reared on cotton (P = 0.0001). Both Er. mundus and En. pergandiella required significantly longer to develop on sweetpotato whitefly reared on bean (P = 0.002), while En. pergandiella developed significantly longer on greenhouse whitefly reared on cotton (P = 0.003). There were no significant differences when parasitoids developed on bandedwinged whitefly on bean or cotton. The longevity of Er. mundus developed on greenhouse whitefly on bean was significantly longer than on sweetpotato whitefly or bandedwinged whitefly (P = 0.001). The longevity of Er. mundus developed in three whitefly species on cotton was not significantly different (P = 0.949) (Table 2). The longevity of En. pergandiella maintained in greenhouse whitefly or bandedwinged whitefly on bean was significantly longer than in sweetpotato whitefly (P = 0.049), while those on sweetpotato whitefly or bandedwinged whitefly on cotton was significantly longer than in greenhouse whitefly (P = 0.049). The longevity of both Er. mundus and En. pergandiella progeny was significantly shorter following development in sweetpotato whitefly maintained on bean (5.7 and 8.8 days, respectively) than on cotton (7.1 and 10.3 days,

440 respectively) (P = 0.001). When these parasitoids developed on greenhouse whitefly maintained on bean, the longevity of progeny of both parasitoid species was significantly longer than that on cotton (P = 0.002). There were no host plant effects on the longevity of parasitoids reared on bandedwinged whitefly (Table 2). Progeny of Er. mundus females developed on bandedwinged whitefly or greenhouse whitefly reared with bean were a significantly greater proportion of females than on sweetpotato whitefly (P = 0.045), while those on cotton were significantly less in greenhouse whitefly (P = 0.017). Er. mundus on sweetpotato whitefly reared with cotton produced a greater proportion of females than when reared on bean (55.0 versus 42.7%). When reared in greenhouse whitefly on bean, Er. mundus produced more female progeny than when reared on cotton (65.6 versus 46.7%) (P = 0.014) (Table 2). The size of Er. mundus females developed in greenhouse whitefly on bean was significantly larger than those in bandedwinged whitefly or sweetpotato whitefly (P = 0.003), while the size of Er. mundus females in sweetpotato whitefly on cotton was significantly larger than in bandedwinged whitefly or greenhouse whitefly (P = 0.001). En. pergandiella females developed in sweetpotato whitefly on cotton or bean were significantly larger than those in bandedwinged whitefly or greenhouse whitefly (P = 0.001). Er..mundus and En. pergandiella in sweetpotato whitefly were significantly larger on cotton than on bean, while those in greenhouse whitefly were larger on bean than on cotton (P = 0.001) (Table 2).

Discussion

Mortality of sweetpotato whitefly was least when reared on cotton, while mortality of greenhouse whitefly was least on bean. Costa et al. (1991) found greater mortality of sweetpotato whitefly on cantaloupe, Cucumis melo L.; cotton; pumpkin, Cucurbita maxima Dene.; lettuce, Lactuca sativa L.; and tomato, Lycopersicum esculentum L., but less mortality on zucchini, Cucurbita pepo L. On cotton, mortality of sweetpotato whitefly from egg to adult of 25% at 25ºC was reported by Horowitz et al. (1984) and on poinsettia, Euphorbia pulcherrima L., 39% was reported by Enkegaard (1993). Powell and Bellows (1992b) observed that natural mortality of sweetpotato whitefly on cotton at 25.5ºC was 55.2%, while on cucumber only 14%. Merendonk and Lenteren (1978) showed that total mortality of greenhouse whitefly increased from eggplant (8.8%), to cucumber (10.6%), tomato (21.2%), and paprika (92.5%). We assumed, based on observation of changing biological parameters of whitefly species on different host-plants, that the nutritional value of the food-plant might be the most important factor controlling the whitefly species. Our data demonstrated that sweetpotato whitefly developed significantly faster on cotton than on bean, while development of greenhouse whitefly was significantly faster on bean than on cotton. Coudriet et al. (1985) reported the developmental time of sweetpotato whitefly on 17 plant species. The minimum developmental time was 16 days on sweet potato, with a maximum of 29.8 days on carrot, Daucus carota L. The mean developmental period for sweetpotato whitefly on cotton was 17.7 days and 20.2 days on cucumber at 25.5ºC. Our results agree with other authors in showing the species of host plant exerts significant effects on whitefly preovipositional period (Butler and Henneberry 1985, Enkegaard 1993), number of eggs deposited per day (Van Lenteren et al. 1977), percentage of female progeny (Gameel 1978, Sharaf and Batta 1985), and adult size (Bethke et al. 1991).

441 Our study demonstrated that parasitism by Er. mundus is independent of host plant species. The greatest parasitization rate was on sweetpotato whitefly and the least on greenhouse whitefly. Emergence of Er. mundus was dependent on both host plant and whitefly species. The survival of parasitoids to emergence was greatest in sweetpotato whitefly and least in bandedwinged whitefly. Emergence of Er. mundus from sweetpotato whitefly reared on cotton was significantly greater than when sweetpotato whitefly was reared on bean. Parasitism and emergence by En. pergandiella was greater than by Er. mundus when sweetpotato whitefly was the host. Er. mundus from sweetpotato whitefly reared on cotton produced more female progeny than those that emerged from sweetpotato whitefly reared on bean. Er. mundus that developed on greenhouse whitefly on bean produced more female progeny than those from greenhouse whitefly reared on cotton. Both Er. mundus and En. pergandiella on sweetpotato whitefly on cotton had significantly shorter developmental time and greater longevity and size than when the parasitoids were reared on sweetpotato whitefly on bean. Developmental time was significantly shorter and longevity and size were greater when the two parasitoids developed from greenhouse whitefly reared on bean than on cotton. Interactions in whitefly parasitoid systems have been reported by others. Powell and Bellows (1992a) found that the mean longevity of females of a Hawaiian population of Eretmocerus sp. from sweetpotato whitefly on cotton or cucumber was greater than that of a California population. Milliron (1940) found that degree of parasitization was affected by the pubescence of the plant and by excretions of the host plant and whitefly nymphs. Vet et al. (1980) reported that parasitization of greenhouse whitefly by En. formosa on cucumber leaves was much less than on tomato, eggplant, or paprika leaves. Kapadia and Puri (1990) demonstrated that developmental time of En. transvena (Timberlake) (= En. sophia Girault & Dodd) on sweetpotato whitefly reared on eggplant was shorter (12.3 days) than on cotton (18.7 days), while development of Er. mundus was shorter on cotton (15.9 days) than on eggplant (20.1 days). Er. mundus produced significantly greater parasitism of sweetpotato whitefly developing on tomato compared with that in greenhouse whitefly (Greenberg et al. 2002). The same trend was observed with survival of Er. mundus on the same whitefly species. Parasitism and emergence rates by Er. eremicus on sweetpotato whitefly and greenhouse whitefly on tomato were not significantly different. Er. eremicus was recommended to producers as a promising species for management of B. argentifolii Bellows and Perring and greenhouse whitefly, especially in greenhouses. Knowledge of interrelationship among host plants, whiteflies, and their parasitoids is necessary for management of whiteflies, to know which plants are affected by them, to understand their damage, and how to monitor whitefly populations (sites, population dynamics, and action thresholds). These are also critical for developing various control tactics, which include cultural control, plant resistance, chemical and natural controls. Evaluating tritrophic interactions will help reduce abundance and damage by pests to manageable levels. These practices can be modified to preserve natural enemies of whiteflies. Crop rotation helps escape most whiteflies. Temporal or spatial separation between fall and spring crops are helpful in reducing whitefly migration and primary inoculation by virus. These are helpful in understanding the mechanisms involved in competition by parasitoids, making decisions in augmentative release experiments, increasing knowledge of whiteflies and the biology of their parasitoids and host-parasitoid

442 relationships, and developing predictive models. They also can be useful for the development of rearing strategies for parasitoids of whiteflies. Based on our data we conclude that cotton is most suitable for sweetpotato whitefly while bean is best for greenhouse whitefly. Sweet potato had no significant impact on sweetpotato or greenhouse whiteflies. The greatest rate of parasitism by Er. mundus was observed in sweetpotato whitefly whether developing on cotton or bean. The least parasitism by this parasitoid was in greenhouse whitefly developing on cotton or bean. Parasitism by Er. mundus in bandedwinged whitefly nymphs developing on cotton or bean was intermediate. Parasitization rates by En. pergandiella also depended on whitefly species and host plants and was similar to that by Er. mundus.

Acknowledgment

We thank Drs. Donald Thomas (SARC-ARS-USDA) and Carlos E. Bográn (Texas AgriLife Extension Service, Texas A&M University System) for reviewing this manuscript, and W. C. Warfield, Y. M. Zhang, and J. Alejandro for technical assistance.

References Cited

Bethke, J. A., T. D. Paine, and G. S. Nuessly. 1991. Comparative biology morphometrics and development of two populations of Bemisia tabaci (Homoptera: Aleyrodidae) on cotton and poinsettia. Ann. Entomol. Soc. Am. 84: 407-411. Butler, G. D., and T. J. Henneberry. 1985. Bemisia tabaci, a pest of cotton in the Southwestern United States. USDA Tech. Bull. 1707. Cock, M. J. W. 1993. Bemisia tabaci: an update 1986-1992 (on the cotton whitefly with an annotated bibliography). CAB, Silwood Park, Ascot, Berks, UK. Coudriet, D. L., N. Prabhaker, A. N. Kishaba, and D. E. Meyerdirk. 1985. Variations in developmental rate on different hosts and overwintering of the sweet potato whitefly, Bemisia tabaci (Homoptera: Aleyrodidae). Environ. Entomol. 14: 516-519. Costa, N. S., J. K. Brown, and D. N. Byrne. 1991. Life history traits of the whitefly, Bemisia tabaci (Homoptera: Aleyrodidae) on six virus-infected or healthy plant species. Environ. Entomol. 20: 1102-1107. Dorsman, R., and W. van de Vire. 1987. Population dynamics of the greenhouse whitefly, Trialeurodes vaporariorum, on different gerbera varieties. IOBC/WPRC Bull. 10: 46-51. Enkegaard, A. 1993. The poinsettia strain of the cotton whitefly, Bemisia tabaci (Homoptera: Aleyrodidae), biological and demographic parameters on poinsettia (Euphorbia pulcherrima) in relation to temperature. Bull. Entomol. Res. 88: 219-229. Gameel, O. I. 1978. The cotton whitefly Bemisia tabaci (Genn.) in the Sudan Gezira, pp. 111-131. Third Ciba-Geigy seminar on the strategy for cotton pest control in the Sudan. 8-10 May 1978. Basel, Switzerland. Gelman, D. B., D. Gerling, M. A. Blackburn. 2005. Host-parasitoid interactions relating to penetration of the whitefly, Bemisia tabaci, by the parasitoid wasp, Eretmocerus mundus. J. Insect Sci. 5: 46-47.

443

Gill, R. J. 1990. The morphology of whiteflies, pp. 13-46. In D. Gerling [ed.], Whiteflies: their Bionomics, Pest Status and Management. Intercept, Andover, UK. Goolsby, J. A., M. A. Ciomperlik, M. C. Legaspi, J. C. Legaspi, and L. E. Wendel. 1998. Laboratory and field evaluation of exotic parasitoids of Bemisia tabaci (Gennadius) (Biotype “B”) (Homoptera: Aleyrodidae) in the Lower Rio Grande Valley of Texas. Biol. Contr. 12: 127-135. Greathead, A. N. 1986. Host plants, pp.17-25. In M. J. W. Cock [ed.], Bemisia tabaci - A Literature Survey on the Cotton Whitefly with Annotated Bibliography. C. A. B. International Institute Biological Control, UK. Greenberg, S. M., W. A. Jones, and T.-X. Liu. 2002. Interaction among two species of Eretmocerus (Hymenoptera: Aphelinidae), two species of whiteflies (Homoptera: Aleyrodidae), and tomato. Environ. Entomol. 31: 397- 402. Hoddle, M. S., R. G. van Driesche, J. S. Elkinton, and J. P. Sanderson. 1998. Discovery and utilization of Bemisia argentifolii patches by Eretmocerus eremicus and Encarsia formosa (Beltsville strain) in greenhouses. Entomol. Exp. Appl. 87: 17-28. Horowitz, A. R., H. Podoler, D. Gerling. 1984. Life table analysis of the tobacco whitefly Bemisia tabaci (Genn.) in cotton field in Israel. Acta Oecologica/ Oecologia Applicata 5: 221-233. Hunter, M. C. 1989. Suitability of stages of female Encarsia pergandiella Howard (Hymenoptera: Aphelinidae) for development of conspecific male hyperparasites. Entomophaga 34: 265-274. Jones, W. A., and S. M. Greenberg. 1998. Suitability of Bemisia argentifolii (Homoptera: Aleyrodidae) instars for the parasitoid Eretmocerus mundus (Hymenoptera: Aphelinidae). Environ. Entomol. 27: 1569-1573. Jones, W. A., and S. M. Greenberg. 1999. Host instar suitability of Bemisia argentifolii (Homoptera: Aleyrodidae) for the parasitoid Encarsia pergandiella (Hymenoptera: Aphelinidae). J. Agric. Urban Entomol. 16: 49-57. Kapadia, M. N., and S. N. Puri. 1990. Development, relative proportions and emergence of Encarsia transvena (Timberlake) and Eretmocerus mundus Mercet, important parasitoids of Bemisia tabaci (Gennadius). Entomon 15: 235-239. Liu, T.-X., and P. A. Stansly. 2000. Response of Trialeurodes abutiloneus (Homoptera: Aleyrodidae) to sweet potato and two species of Hibiscus. Ann. Entomol. Soc. Am. 93: 850-855. Lopez-Avila, A. 1988. A comparative study of four species of Encarsia (Hymenoptera: Aphelinidae) as potential control agents for Bemisia tabaci (Gennadius) (Homoptera: Aleyrodidae). Ph.D. dissertation. Department of Pure and Applied Biology, CAB International Institute of Biological Control, London, UK. Merendonk, S., and J. C. van Lenteren. 1978. Determination of mortality of greenhouse whitefly Trialeurodes vaporariorum (Westwood) (Homoptera: Aleyrodidae) eggs, larvae, and pupae on four host plant species: eggplant (Solanum melongena L.), cucumber (Cucumis sativus L.), tomato (Lycopersicum esculentum L.), and paprika (Capsicum annuum L.). Med. Fac. Landbouww. Rijksuniv. Gent 43: 421-429.

444

Milliron, N. E. 1940. A study of some factors affecting the efficiency of Encarsia formosa Gahan, an aphelinid parasite of the greenhouse whitefly, Trialeurodes vaporariorum (Westw.). Mich. Agric. Exp. Sta. Tech. Bull. 173: 1-23. Palumbo, J. C., A. R. Horowitz, and N. Prabhaker. 2001. Insecticidal control and resistance management for Bemisia tabaci. Crop Protection 20: 739-765. Powell, D. A., and T. S. Bellows. 1992a. Preimaginal development and survival of Bemisia tabaci on cotton and cucumber. Environ. Entomol. 21: 359-363. Powell, D. A., and T. S. Bellows. 1992b. Adult longevity, fertility and population growth rates for Bemisia tabaci (Genn.) (Ho., Aleyrodidae) on two host plant species. J. Appl. Entomol. 113: 68-78. Russell, L. M. 1963. Hosts and distribution of five species of Trialeurodes (Homoptera: Aleyrodidae). Ann. Entomol. Soc. Am. 56: 149-153. SAS Institute. 1999. SAS/STAT User’s Guide, version 8, SAS Institute, Cary, NC. Sharaf, N. S., and Y. Batta. 1985. Effect of some factors on the relationship between the whitefly Bemisia tabaci Genn. (Homoptera: Aleyrodidae) and the parasitoid Eretmocerus mundus Mercet. Z. ang. Entomol. 99: 267-276. Sokal, R. R., and F. J. Rohlf. 1981. Biometry. W. H. Freeman and Co., San Francisco, CA. Van Boxtel, W., J. Woets, and J. C. van Lenteren. 1978. Determination of host- plant quality of eggplant (Solanum melongena L.), cucumber (Cucumis sativus L.), tomato (Lycopersicum esculentum L.), and paprika (Capsicum annuum L.) for the greenhouse whitefly (Trialeurodes vaporariorum [Westwood]) (Homoptera: Aleyrodidae). Meded. Fac. Landbouwwet. Rijksuniv. Gent. 43: 397-408. Van Lenteren, J. C., J. Woets, N. van der Poel, W. van Boxtel, S. van de Merondonk, R. van der Kamp, H. Nell, and L. Sevenster-van der Lelie. 1977. Biological control of the greenhouse whitefly Trialeurodes vaporariorum (Westwood) (Homoptera: Aleyrodidae) by Encarsia formosa Gahan (Hymenoptera: Aphelinidae) in Holland, an example of successful applied ecological research. Med. Fac. Landbouww. Rijksuniv. Gent 42: 1333-1343. Vet, L. E. M., J. C. van Lenteren, and J. Woets. 1980. The parasite-host relationship between Encarsia formosa (Hymenoptera: Aphelinidae) and Trialeurodes vaporariorum (Homoptera: Aleyrodidae). IX. A review of the biological control of the greenhouse whitefly with suggestions for future research. Z. Angew. Entomol. 90: 26-51. Vinson, S. V. and G. F. Iwantsch. 1980. Host regulation by insect parasitoids. Q. Rev. Biol. 55:143-165. Wilkinson, L., M. A. Hill, and E. Vang. 1992. SYSTAT: Statistics, version 5.2 ed. Systat, Evanston, IL. Williams, M. R. 2009. Cotton insect loss estimates. Proceedings of Beltwide Cotton Conference. National Cotton Council of America. Memphis, TN. CD- ROM.

445 VOL. 34, NO. 4 SOUTHWESTERN ENTOMOLOGIST DEC. 2009

Oviposition Characteristics of Pecan Weevil1

Michael W. Smith2 and Phillip G. Mulder3

Oklahoma State University, Stillwater, OK 74078

Abstract. Pecan weevil, Curculio caryae (Horn) (Coleoptera: Curculionidae), is a major fruit pest of pecan, Carya illinoinensis (Wangenh.) K. Koch. The majority of economic damage results from larvae feeding on the kernel. Pecan fruit were collected at three sites to describe oviposition activity related to location on the fruit, cluster size, and determine the likelihood of multiple females ovipositing in the same fruit. Results indicated that oviposition was primarily at the distal (stigma) end of the fruit, although feeding punctures seemed to occur without respect to fruit location. Infestation of a fruit cluster rarely resulted in all fruit in the cluster being infested; in fact, one or two fruit were the most common infestation level, regardless of cluster size. Infestation by larvae typically occurred at greater frequency in larger clusters and less frequency in single fruit clusters than the infestation rate across all cluster sizes, suggesting that more fruit produced a greater concentration of volatile compounds attractive to females. Multiple punctures in the shuck were frequently observed, but multiple punctures of the testa were rare when infestation was moderate but increased when infestation was high. This indicated that feeding by adult weevils was not affected by earlier feeding, but deep penetration of the fruit leading to oviposition was inhibited by a previous penetration of the testa. In orchards with the majority of the crop infested, fruit were identified with larval development up to the third instar, combined with eggs. Evidence suggests an oviposition-marking pheromone produced by the host or female weevil deters other females from using the fruit for reproduction, but will not avert multiple oviposition episodes when non-infested fruit are sparse.

Introduction

Pecan weevil, Curculio caryae (Horn) feed on developing Carya fruit (Gibson 1969). Pecan is a principal host, with economic losses resulting from adults feeding on immature fruit and larvae feeding within the fruit, on the developing cotyledons. Adult pecan weevil feeding or probing fruit causes drop within 15 days if the testa (seed coat) is penetrated while the endosperm is non-cellular (water stage) (Calcote 1975). When the shell (pericarp) lignifies, feeding by male weevils is restricted to the shuck (involucre), but females penetrate the shuck, lignified shell, and testa. Females that probe into the cellular endosperm cause a black spot or pit on the kernel, or mold results destroying the kernel. Deeper penetration of the kernel ______1Coleoptera: Curculionidae. 2Department of Horticulture and Landscape Architecture, Oklahoma State University, Stillwater, OK 74078. 3Department of Entomology and Plant Pathology, Oklahoma State University, Stillwater, OK 74078.

447 through the endosperm, at the onset of cotyledon deposition results in mold and/or collapse of the developing kernel, referred to as a “pop” by producers. Females initiate oviposition during cotyledon deposition through fruit maturity. Mulder and Grantham (2007) provided a thorough description of pecan weevil biology, life cycle, and damage. Female pecan weevils seem to search for suitable sites for oviposition at random within the tree (Hall et al. 1979, Harris and Ring 1979). When females locate multi-fruit clusters, more than one fruit is usually infested. However, previously infested fruit appear not to be infested again, suggesting the existence of a strong oviposition-inhibiting mechanism (Harris and Ring 1979). Oviposition deterrents may originate from the host (Hilker and Meiners 2002) or the pest (Prokopy 1981, Meiners et al. 2005). An oviposition deterrent may be useful in managing pecan weevil to discourage egg laying and allow redirection of female weevils to locations that suggest bountiful oviposition sites. Insecticides could then be deployed to these targeted areas. This strategy could reduce insecticide usage, thus reducing costs and environmental impact. The primary objective of this study was to ascertain if fruit were only infested a single time at various population densities of pecan weevil. In addition, the influence of cluster size on infestation by pecan weevil and the puncture location on the fruit were characterized.

Materials and Methods

Stillwater Research Station. ‘Gormely’ fruit began to form a cellular endosperm (gel stage) on 16 August 2006, followed by cotyledon deposition (dough stage) on 30 August 2006. One fruit per cluster was sampled on 11 September 2006 and examined for the presence and location of punctures in the shuck, shell, and testa (Fig. 1). The kernel was dissected and examined for the presence of eggs and/or larvae (Fig. 2). A total of 400 fruit was sampled from 15 trees. The site had not been managed to control pecan weevil. Feyodi Creek State Park, Cleveland, OK. One hundred fruit clusters of varying sizes were sampled from each of six native pecan trees (600 clusters total) on 2 October 2006. Clusters were individually bagged to determine the number of

Fig. 1. Two pecan weevil punctures in the involucre (A) and the pericarp (B).

448

Fig. 2. Multiple pecan weevil eggs showing the ability of the female to extend the ovipositor into various areas of the pecan kernel.

infested fruit per cluster. Nut weight (shell + kernel) ranged from 1.7 to 3.7 g, with 30.1 to 44.9% kernel. Fruit were examined for punctures of the shuck, shell, and testa, and infestation by pecan weevil. No management practices were used to control pecan weevil at this site. Hoffman Orchard, Stillwater, OK. One hundred fruit clusters each of varying sizes were sampled from several native pecan and ‘Burkett’ trees (200 clusters total) on 5 October 2006. Trees had not been managed for pecan weevil infestation. Clusters were individually bagged to determine the number of infested fruit per cluster. Native pecans averaged 3.2 g per nut, with 30.1% kernel, and ‘Burkett’ nuts weighed 6.4 g per nut, with 36.6% kernel. Fruit were examined for punctures of the shuck, shell, and testa, and infestation by pecan weevil. Data Analysis. Data were analyzed using a chi-square distribution with appropriate hypotheses tested using Proc Freq in SAS (SAS Institute) (Agresti 1990).

Results and Discussion

Stillwater Research Station. In this historically nontreated orchard, infestation by pecan weevil was great, with 97.5% shucks, 92.2% shells, and 85.5% of the testae punctured (Table 1). Sixty-one percent (244 of 400) of the fruit were

449 Table 1. Number of Punctures by Pecan Weevil Penetrating the Shuck, Shell, and Testa of 400 ‘Gormely’ Pecan Fruit Sampled on 11 September 2006 Number of fruit with punctures in selected part Number of punctures Shuck Shell Testa 0 10 31 52 1 129 235 272 2 102 93 65 3 60 23 8 4 58 15 3 5 24 1 0 6 5 1 0 7 2 0 0 10 1 0 0 11 2 1 0

infested with pecan weevil eggs or larvae. When the testa had been punctured, 70% (244 of 348) of the fruit were infested by eggs or larvae. The discrepancy between the testa being punctured and the presence of eggs or larvae may result from females feeding, particularly virgin females, or females that found the fruit unsuitable for reproduction. Several fruit with pecan weevil eggs and first- to third- instar larvae were observed, as well as fruit with eggs on both sides of the middle septum, providing evidence that when noninfested fruit was meager, females oviposit in fruit previously attended by another female. Sixty-five percent of the fruit had multiple punctures in the shuck, 33.5% in the shell, and 19% in the testa (Table 1). The large percentage of multiple punctures in the shuck combined with a decreasing percentage of punctures in deeper tissue within the fruit indicated that feeding by weevils did not seem to affect subsequent feeding activity. Because many of the feeding punctures only penetrated the shuck, it is likely that many of the punctures were by males, although feeding females may only puncture the shuck unless they are probing for suitable reproduction sites. The frequency of multiple punctures of the testa supports the previous observation that females will use a fruit for reproduction that was attended by another female when noninfested fruit are sparse. Pecan weevils punctured the fruit primarily at the distal end (89%) (Table 2). Only about 10% had punctures at the basal (stem) end of the fruit. A preference for the distal end was initially reported by Harris and Ring (1979). They suggested that this preference was likely associated with cotyledon development that proceeds basipetally. Puncturing the shuck at the distal end of the fruit resulted in 93% (331 of 357) shell penetration and 87% (311 of 357) testa penetration; however, when the basal end of the shell and the testa were compromised, 58% (23 of 40) and 30% (12 of 40) penetration rates were recorded, respectively. This indicates that many or perhaps most of the punctures at the basal location were feeding punctures. The number of punctures near the middle of the fruit was intermediate between the distal and basal locations. Feyodi Creek State Park, Cleveland, OK. The number of fruit punctured by pecan weevil at Feyodi Creek State Park was substantially less than at the Stillwater Research Station. At Feyodi Creek, 39% (492 of 1256) of the fruit had punctures in the shuck. Multiple punctures in the shuck occurred in 33% (162 of

450 492) of the fruit. Pecan weevil penetrated the testa 36% (175 of 492) of the time when the shuck was punctured, but only 1.8% (9 of 492) of the fruit had multiple punctures of the testa. The testa was punctured only two times when multiple punctures were found. No eggs were found in samples from Feyodi Creek (probably because of the later sampling date), but larvae infested 12.3% (154 of 1,256) of the fruit sampled. The large number of punctures or infestation of the shuck compared to the testa (492, 175, and 154 of 1,256 for shuck or testa puncture and larval infestation, respectively) suggests much feeding or probing for oviposition suitability. The distribution of punctures was similar to that at the Stillwater Research Station, with preference for the distal end of the fruit (data not shown).

Table 2. Punctures by Pecan Weevil Penetrating Selected Fruit Parts by Location on 400 ‘Gormely’ Pecan Fruit Sampled on 11 September 2006 Number of fruit punctured ȋ2 probability of a equal Fruit location punctured number of punctures in Fruit part Basal Middle Distal each location by fruit part Shuck 40 138 357 0.0001 Shell 23 69 331 0.0001 Testa 12 45 311 0.0001 ȋ2 probability of the same number of punctures in 0.0002 0.0001 0.0001 each fruit part by location

The number of fruit within each cluster size infested with larvae was compared to the total fruit infestation rate to determine if fruit within a particular cluster size was preferred (Table 3). Fruit in clusters with two to four fruit were infested at a rate similar to the overall level. Five fruit clusters were infested at a greater rate than the overall infestation level, and single fruit clusters tended to be infested at a lower level (null hypothesis rejected at the 7% level).

Table 3. Effect of Cluster Size on the Frequency of Fruit Infested by Pecan Weevil on Native Pecan Trees at Cleveland, OK Larval ȋ2 probability of Fruit examined Fruit with larvae infestation 87.7% not infested Fruit/cluster (no.) (no.) (%) 12.3% infested 1 190 15 7.9 0.0645 2 448 43 9.6 0.2646 3 396 56 14.1 0.2646 4 172 23 13.4 0.6686 5 50 17 34.0 0.0001 Total 1,256 154 The chi-square statistic tested frequency of infestation by cluster size compared to overall frequency of infestation of fruit (87.7% not infested, 12.3% infested).

451 The number of clusters with any fruit infested was compared with the overall infestation rate of clusters to determine if a particular cluster size was preferred (Table 4). Single fruit clusters were infested less than the infestation rate across all cluster sizes. Two fruit clusters were similar to the overall infestation rate, and clusters with three or more fruit were infested at a greater rate. Five fruit clusters were especially infested. Harris and Ring (1979) reported that a female weevil encountered a pecan cluster at random, but when the insect located multi-fruit clusters, normally more than one fruit was infested. In contrast, results of this study indicated that pecan weevil specifically sought larger fruit clusters although they seldom infested all the fruit within a cluster (Table 5). A feeding preference study using an olfactometer demonstrated that volatile compounds from pecan and hickory fruit attracted adult pecan weevil (Collins et al. 1997). Larger fruit clusters probably produce more volatile compounds, and thus are more attractive than smaller clusters.

Table 4. Effect of Cluster Size on the Frequency of Any Fruit within a Cluster Being Infested by Pecan Weevil on Native Pecan Trees at Cleveland, OK Clusters Clusters with Larval ȋ2 probability of examined larvae infestation 80.2% not infested Fruit/cluster (no.) (no.) (%) 19.2% infested 1 190 15 7.9 0.0001 2 224 36 16.1 0.2345 3 132 42 31.8 0.0002 4 43 15 34.9 0.0090 5 10 7 70.0 0.0001 Total 599 115 The chi-square statistic tested infestation frequency by cluster size to the overall cluster infestation frequency (80.2% not infested, 19.2% infested).

Table 5. Number of Pecan Fruit per Cluster Infested with Pecan Weevil Larvae for Clusters of Different Sizes on Native Pecan Trees at Cleveland, OK Number Number of clusters clusters Number of infested fruit per cluster Fruit/cluster observed 0 1 2 3 4 5 1 190 175 15 ------2 224 188 29 7 ------3 132 90 29 12 1 ------4 43 28 8 6 1 0 --- 5 10 3 2 2 2 0 1

Pecan weevil larvae usually infested only one or two fruit per cluster, regardless of the cluster size (Table 5), although female weevils were attracted to larger clusters (Tables 3 and 4). Failure to take advantage of fruit once located suggests inefficient reproduction. However, multiple larvae developing from a

452 single oviposition event (Harris and Ring 1979) mitigates the inefficient use of available fruit, contributing to a reproduction efficiency that can result in sufficient abundance of pecan weevil to devastate production in managed orchards producing a consistent supply of fruit. Hoffman Orchard, Stillwater, OK. Fifty-two percent (246 of 474) of the fruit were infested with pecan weevil larvae. Infestation rates of native pecans (126 of 254) and ‘Burkett’ (120 of 220) were similar (ȋ2 probability = 0.28), suggesting little, if any, fruit size preference because trees were close and ‘Burkett’ nuts were almost twice as large as the native pecans. Harris (1976) found the number of pecan weevil larvae per infested nut was independent of fruit size. Because infestation frequency was similar, data were pooled over ‘Burkett’ and native pecans to examine the relationships of cluster size, puncture location on the fruit, and the occurrence of multiple punctures by pecan weevils. The shuck was punctured on 77.2% (366 of 474) of the fruit, indicating abundant feeding by a relatively large population of pecan weevils. The testa was punctured on 56.5% (268 of 474) of the fruit. Multiple punctures of the shuck were recorded in 27% (128 of 474) of the fruit, and 6.5% (31 of 474) of the fruit had the testa punctured more than once. The preference for certain sizes of fruit clusters was tested by comparing the incidence of infestation by pecan weevil for a given cluster size to the frequency over all cluster sizes (Table 6). The overall infestation rate (at least one fruit infested in a cluster) was 68.3%. Clusters with a single fruit were infested less frequently than the overall rate. Clusters with two fruit were infested more often than the overall infestation rate, and all of the five fruit clusters were infested. Although these results are not as definitive as data from Feyodi Creek State Park, they also suggest a cluster size preference. In this case, single fruit clusters were not as desirable as larger clusters and all five fruit clusters were infested, supporting the results from Feyodi Creek State Park.

Table 6. Effect of Cluster Size on the Frequency of Any Fruit within a Cluster Being Infested by Pecan Weevil on Native Pecan Trees at Hoffman Orchard, Stillwater, OK Clusters Clusters with Larval ȋ2 probability of examined larvae infestation 31.7% not infested Fruit/cluster (no.) (no.) (%) 68.3% infested 1 48 17 35.4 0.0001 2 72 57 79.2 0.0475 3 48 34 70.8 0.7060 4 23 20 87.0 0.0545 5 9 9 100 --- Total 200 63 The chi-square statistic tested infestation frequency by cluster size to the overall cluster infestation frequency (31.7% not infested, 68.3% infested).

Pecan weevil clearly preferred puncturing fruit at the distal rather than the basal end (data not shown), as was observed at the other two sites and by Harris

453 and Ring (1979). The testa was punctured 38% (139 of 366 fruit observed) of the time at the distal end compared to only 9% (35 of 366 fruit) (ȋ2 probability for equal puncture frequency = 0.0001) at the basal end. The trend in shuck punctures was similar, but the magnitude of difference was not as great, probably because feeding punctures by males and females do not seem to be site specific on the fruit.

Conclusions

The distal end of the fruit is clearly the preferred location for oviposition. Harris and Ring (1979) said this may be related to the basipetal development of cotyledons within the fruit, thus the distal end is capable of supporting developing larvae before the basal location. Punctures at locations other than the distal end of the fruit were mostly for feeding. Pecan weevil females were attracted to larger fruit clusters, as demonstrated at the Hoffman Orchard and Feyodi Creek State Park. This contrasts with results of Hall et al. (1979) and Harris and Ring (1979) who reported weevils located fruit clusters randomly. The more frequent infestation of at least one fruit in larger clusters may be associated with a greater abundance of volatiles produced by fruit that are attractive to female weevils (Collins et al. 1997). Pecan weevil females seemed averse to ovipositing in a fruit if it was infested previously. This behavior may contribute to the rarity of all fruit within a cluster being infested. Oviposition activity suggests a deterrent was produced by the female or plant, supporting speculation of Harris and Ring (1979). However, at the Stillwater Research Station and at the Hoffman Orchard abundant pecan weevils that used most available resources resulted in multiple penetrations of the testa and various developmental stages of pecan weevil (eggs and larvae in the same fruit), indicating more than one oviposition event occurred in a single fruit. This result indicated that although females avoid fruit with oviposition, they use such fruit if others are in short supply. Pecan weevil management using an oviposition-marking pheromone could be useful to deter females from one area if a suitable substrate was available for reproduction nearby. The attractiveness of this nearby area could be enhanced by incorporating attractive volatiles produced by pecan fruit. Concentrating females in a smaller area might allow for less use of insecticide or other control methods. However, if suitable fruit are unavailable in the immediate area, female weevils may use fruit previously used for oviposition by other females.

Acknowledgment

Approved for publication by the Oklahoma Agricultural Experiment Station. Funding for this study was from the Oklahoma Agricultural Experiment Station and the Oklahoma Pecan Growers’ Association. The authors thank Ms. Becky Cheary and Mr. Kelly Seuhs for their tireless efforts in sorting through many of the fruit from sampled trees.

References Cited

Agresti, A. 1990. Categorical Data Analysis. John Wiley & Sons, New York. Calcote, V. R. 1975. Pecan weevil: feeding and initial oviposition as a related to nut development. J. Econ. Entomol. 68: 4-6.

454

Collins, J. K., P. G. Mulder, R. A. Grantham, W. R. Reid, M. W. Smith, and R. D. Eikenbary. 1997. Assessing feeding preferences of pecan weevil (Coleoptera: Curculionidae) adults using a Hardee olfactometer. J. Kansas Entomol. Soc. 70: 181-188. Gibson, L. P. 1969. Monograph of the genus Curculio in the New World (Coleoptera: Curculionidae). Part I. U.S. and Canada. Entomol. Soc. Am. Misc. Pub. 6: 240-285. Hall, M. J., R. D. Eikenbary, and W. D. Warde. 1979. Effects of pecan nut cluster size on the selection of nuts for feeding and oviposition by the pecan weevil, Curculio caryae (Coleoptera: Curculionidae). Can. Entomol. 111: 1193-1196. Harris, M. K. 1976. Pecan weevil infestation of pecans of various sizes and infestations. Environ. Entomol. 5: 248-250. Harris, M. K., and D. R. Ring. 1979. Biology of pecan weevil from oviposition to larval emergence. Southwest. Entomol. 4: 73-85. Hilker, M., and T. Meiners. 2002. Induction of plant responses to oviposition and feeding by herbivorous arthropods: a comparison. Entomologia Experimentalis Applicata 104: 181-192. Meiners, T., N. K. Hacker, P. Anderson, and M. Hilker. 2005. Response of the elm leaf beetle to host plants induced by oviposition and feeding: the infestation rate matters. Entomologia Experimentalis Applicata 115: 171-177. Mulder, P. G., and R. A. Grantham. 2007. Biology and control of the pecan weevil in Oklahoma. Oklahoma Cooperative Extension Service EPP-7079. Prokopy, R. J. 1981. Oviposition-deterring pheromone system of apple maggot flies, pp. 477-494. In E. R. Mitchell [ed.], Management of Insect Pests with Semiochemicals: Concepts and Practices. U.S. Dep. Agri., Gainesville, FL.

455 VOL. 34, NO. 4 SOUTHWESTERN ENTOMOLOGIST DEC. 2009

Phylogenetic Analysis of Heat Shock Proteins in Glassy-winged Sharpshooter, Homalodisca vitripennis

Henry Schreiber IV, Daymon Hail, Wayne Hunter1, and Blake Bextine

University of Texas at Tyler, 3900 University Boulevard, Tyler, TX 75799

Abstract. The glassy-winged sharpshooter, Homalodisca vitripennis (Germar) (Hemiptera: Cicadellidae), is the major vector of Xylella fastidiosa Wells et al., the causal agent of Pierce’s disease of grapes, Vitis spp. More research on leafhopper stress response is possible as more genomic information becomes available. To gain information on the stress response of the glassy-winged sharpshooter, a cDNA library was constructed from adults and fifth instars, resulting in 5,906 expressed sequence tags (ESTs). After quality scoring, 4,445 sequences underwent assembly, which produced a set of 2,123 sequences that putatively represented distinct transcripts. BLASTX analysis identified four significant homology matches to heat shock proteins (HSP), which are the focus of this study. The overall importance and function of heat shock proteins lie in their ability to maintain protein integrity and activity during stressful conditions, such as extreme heat, cold, or drought. Phylogenetic analyses using these four heat shock protein sequences provided further support of transcript by the identification of specific motifs. This study shows that highly conserved genes such as heat shock proteins are a viable supplement to ribosomal DNA in elucidating phylogenetic relationships.

Introduction

Organisms respond to heat shock or other environmental stress by inducing the synthesis of certain proteins, some known as heat shock proteins (Lindquist 1986, Sorenson 2003). Infections, temperature changes, inflammation, toxins, hypoxia, starvation, and even exercise can result in increased production of these proteins (Sorenson 2003). Heat shock proteins aid in folding, targeting, and tracking of nascent proteins, promote transcription, are involved in cellular division, and can be up regulated via cell signaling in addition to environmental stimuli (Feder and Hofmann 1999). Small heat shock proteins have an approximate molecular weight of less than 30kDa and are molecular chaperones, maintaining proper protein structure by blocking aggregation of denatured proteins, aiding nascent protein folding, and assisting construction of quaternary structure (Bova et al. 2002, Gu et al. 2002, Sobott 2002, Fu and Chang 2004). Among the heat shock protein families is a group of well-conserved proteins with an approximate molecular weight of 70 kDa, known as the HSP70 family. Most species have several proteins belonging to this family. Some of these members are ______1U.S. Horticultural Research Laboratory, Genomics Laboratory, Ft. Pierce, FL.

457 expressed only under stress conditions (strictly inducible), while some are present in cells under normal growth conditions (Craig 1989), are not heat-inducible (Pelham 1986), and are known as heat shock cognates (HSC). In eukaryotes, HSP70 can work with small heat shock proteins to restore functionality to heat- denatured proteins (Lee and Vierling 2000) or co-chaperone with HSP40 to fold nascent proteins into proper tertiary structure by temporarily binding to hydrophobic domains until sequence translation is complete (Douglas et al. 1994) The 90 kDa heat shock protein (HSP90) is one of the most prolific proteins in eukaryotic cells, constituting 12% of cellular protein under baseline conditions (Sreedhar 2004). The functions and morphological evolution have been studied extensively and include signal transduction, protein folding, and degradation of denatured proteins (Nadeau 1993, Jakob 1994). Increased functionality of HSP90 is acquired when associated with its co-chaperones, playing an important role in the folding of newly synthesized proteins. HSP90 binds with an array of substrate proteins and co-chaperones. The interactions of HSP90 with the co-chaperones are determined by the substrate protein being bound (Jakob 1995). Understanding heat shock proteins in insects, especially leafhoppers, will provide insights into the biological adaptive elasticity to stressors such as insecticides, parasitization, and temperature of these important agricultural pests. The glassy-winged sharpshooter, Homalodisca vitripennis Germar (Hemiptera: Cicadellidae), is an insect pest that occurs throughout most of the southern USA and is endemic to most regions of Texas (Young 1958, Turner and Pollard 1959). Without naturally occurring forms of biological control, glassy-winged sharpshooters have established populations in new areas and have negatively affected yields of grape, Vitis spp. (de Leon et al. 2004). Glassy-winged sharpshooter is a voraciously-feeding, xylem-limited pest reported to feed on plants from at least 35 families, including both woody and herbaceous types (Hoddle et al. 2003), and can impact plant health directly by depriving the plant of nutrients and damaging the xylem sufficiently to preclude vascular flow. However, indirect plant damage occurs during feeding and subsequent transmission of the xylem-limited bacterium Xylella fastidiosa Wells et al. (Xanthomonadales: Xanthomonadaceae). The invasion of glassy-winged sharpshooter into grape-growing regions of California, especially the Temecula Valley, produced enormous risk to the wine and table-grape industry by spreading the phytopathogen X. fastidiosa, the causative agent of Pierce’s disease (Purcell 1997, de Leon et al. 2004). Additionally, many other economically important plants including citrus; almond, Prunus dulcis (Mill.) D.A.Webb; and oleander, Nerium oleander L.; are affected by strains of X. fastidiosa, resulting in a multitude of plant diseases including citrus variegated chlorosis (Chang et al. 1993, Pooler and Hartung 1995), almond leaf scorch (Mircetich et al. 1976), and oleander leaf scorch (Purcell 1999). A search of the National Center for Biotechnology Information for glassy- winged sharpshooter genes or protein sequences revealed less than 25 complete, non-mitochondrial genes or complete genomic proteins. Although the complete mitochondrial sequence of glassy-winged sharpshooter has been described (Genbank AY875213), the genomic DNA sequence is incomplete. More than 20,000 expressed sequence tags (ESTs) from glassy-winged sharpshooter have been submitted to the National Center for Biotechnology Information; however, many of these expressed sequence tags are duplicates. This study is an initial step in examining the potential use of this information by examining the utility of these

458 heat shock proteins to describe the phylogeny of leafhoppers in relation to other insects.

Materials and Methods

Leafhopper Collection. Glassy-winged sharpshooters were obtained from Heather Costa, Riverside, CA. Abundant leafhoppers on citrus in fields surrounding Riverside, CA, were collected by using sweep nets. Glassy-winged sharpshooters were immediately macerated in RNAlater® RNA Stabilization Reagent (Ambion, Austin, TX) and shipped on ice to Wayne Hunter, U.S. Horticultural Research Laboratory, Genomics Laboratory, Ft. Pierce, FL, for further processing. cDNA Library Construction. Approximately 140 adult and fifth-instar larval glassy-winged sharpshooters were used to construct the cDNA library. The library was produced as described in Coudron et al. (2006). Whole leafhoppers were ground in liquid nitrogen, and total RNA was extracted by using the guanidinium salt-phenol-chloroform procedure previously described by Strommer et al. (1993). cDNA was synthesized using Stratagene’s ZAP-cDNA Synthesis Kit (Stratagene, La Jolla, CA). Sequencing of Clones. Clone sequencing was processed as in Coudron et al. (2006). Reactions were prepared in 96-well format using the Biomek2000™ liquid-handling robot (Beckman Coulter, Inc.). Sequencing reaction products were precipitated with 70% isopropanol, resuspended in 15 µL of sterile water, and loaded onto an ABI 3730 DNA Analyzer (Applied Biosystems, Foster City, CA) at the Genomics Center, U.S. Horticultural Research Lab, Ft. Pierce, FL. Sequence Analysis. The software TraceTuner™ (Paracel, Pasadena, CA) and Sequencher™ (Gene Codes, Ann Arbor, MI) were used for quality and validation of sequences. Sequencher contig assembly parameters were set using a minimum overlap of 50 bp and 90% identity. Putative sequence identity was determined based on BLAST similarity searches using the NCBI BLAST server (www.ncbi.nlm.nih.gov). Matches with an E-value ”-10 were considered significant and were classified according to the Gene Ontology classification system. Translated proteins were analyzed with National Center for Biotechnology Information’s BLASTp, Pfam (www.pfam.org), InterProScan (www.ebi.ac.uk), and Expert Protein Analysis System (www.expasy.org). Four partial protein sequences were analyzed for phylogenetic relationships with homologous heat shock protein sequences. Phylogenetic Analysis. Four protein sequences identified as heat shock proteins were aligned using T-Coffee (www.tcoffee.org) and ClustalW (www.ebi.ac.uk/Tools/clustalw/) against a variety of homologous heat shock proteins in different taxa. Alignments were retrieved and visualized in Treeview v1.6.6.

Results

Mining of the 5,906 cDNA clones produced from adult and fifth-instar glassy- winged sharpshooters resulted in 4,445 high-quality (i.e., •200 bases with a TraceTuner™ score of 20 or better) expressed sequence tags. Sequence alignment of the expressed sequence tags resulted in a Unigene set of 2,123 total assembled sequences, at Phred 20 score, 40-bp overlap, 100-bp minimum length,

459 using Sequencher™ 8.0 (Gene Codes Corp, Ann Arbor, MI). Translated proteins were analyzed with National Center for Biotechnology Information’s BLASTp, Pfam (www.pfam.org), InterProScan (www.ebi.ac.uk), and Expert Protein Analysis System (www.expasy.org). Four partial protein sequences were analyzed for phylogenetic relationships with homologous heat shock protein sequences. A BLASTp and BLASTn analysis showed that WHWTC-contig[1627] and WHWTC- contig[1325] had significant homology with the small heat shock proteins, while WHHC-contig[1333] displayed homology with HSP70, and WHWTC-contig[1285] was homologous with HSP90 (Table 1). A homology search in the Pfam database identified protein sequences with their respective heat shock protein families (Table 2). A functional analysis and homology search using PantherDB annotated and classified the sequences as belonging to the heat shock protein superfamily (Table 3). Phylogenetic trees illustrated accurate grouping of taxa into clades relative to known heat shock proteins from closely related Hemipteran species (Figs. 1-3). The clades were separated according to taxonomic order.

Table 1. Protein Sequence Similarities from Homalodisca vitripennis Contigs H. vitripennis clones Descriptor E-value WHWTC-Contig[1627] gb|ABC84494.1| heat shock protein 20.7 3.00E-43 694 bases Locusta migratoria WHWTC-Contig[1325] gb|ACH85196.1|heat shock protein 20 2.00E-44 954 bases Bemisia tabaci WHHC-Contig[1333] gb|AAZ17399.2| 70 kDa heat shock protein 2.00E-149 1047 bases Bemisia tabaci WHWTC-Contig[1285] gb|AAZ17403.1| 90 kDa heat shock protein 6.00E-179 1144 bases Bemisia tabaci Nucleotide matches (accesion|protein description organism) and e-values for query contigs using National Center for Biotechnical Information’s BLASTx.

Table 2. Hidden Markov Models Homology Search of In Silico-translated Protein Sequences Using Pfam Protein Database (www.pfam.org) with Protein Family Description, Pfam Identification, Sequence Coverage, and Corresponding E-value Contig Pfam family Sequence HMM number Description identification Start End From To E-value WHWTC- Hsp20/alpha Contig[1627] crystallin family PF00011 86 182 1 109 1.50E-40 WHWTC- Hsp20/alpha Contig[1325] crystallin family PF00011 63 159 1 109 2.30E-38 WHHC- Contig[1333] Hsp70 protein PF00012 1 325 291 619 2.20E-201 WHWTC- Contig[1285] Hsp90 protein PF00183 3 380 101 489 0

460

Table 3. Homalodisca vitripennis In Silico-translated Heat Shock Protein Sequence and Functional Analysis Homology Search Using Hidden Markov Models (HMM) of Panther Protein Database (www.pantherdb.org) with Protein Family Description, PantherDB Identification, Sequence Coverage, Corresponding E-value, Gene Ontology (www.geneontology.org) Code, and Annotation Descriptor Sequence Gene Ontology Annotations

461 Contig number Description Panther ID Start End E-value Code Descriptor

WHWTC-Contig[1627] Small Heat Shock Protein (HSP20) PTHR11527 23 195 5.30E-57 GO:0009408 Response to heat

WHWTC-Contig[1325] Small Heat Shock Protein (HSP20) PTHR11527 1 172 3.60E-56 GO:0009408 Response to heat

WHHC-Contig[1333] Heat Shock Protein 70KDA PTHR19375 1 240 7.70E-143 GO:0005524 ATP binding

WHWTC-Contig[1285] Heat Shock Protein 90 PTHR11528 3 380 1.40E-290 GO:0006950 Response to stress

Fig. 1. Cladogram of small heat shock proteins constructed using glassy-winged sharpshooter sequences WHWTC-Contig[1627] and WHWTC-Contig[1325]. Subject sequences were analyzed using NCBI BLASTp search and aligned using T- Coffee multiple alignment tool (www.tcoffee.org) and visualized using Treeview v1.6.6. (Organism name|accession number| reference number).

462

Fig. 2. Cladogram of heat shock 70 proteins constructed using sequence WHHC- Contig[1333]. Subject sequences were analyzed NCBI BLASTp search and aligned using ClustalW2 multiple alignment tool (www.ebi.ac.uk/Tools/clustalw2/index.html) and visualized using Treeview v1.6.6. (Organism name|accession number| reference number).

463

Fig. 3. Cladogram heat shock 90 proteins constructed using sequence WHWTC- Contig[1285]. Subject sequences were analyzed using NCBI BLASTp search and aligned using ClustalW2 multiple alignment tool (www.ebi.ac.uk/Tools/clustalw2/ index.html) and visualized using Treeview v1.6.6. (Organism name|accession number| reference number).

464

Discussion

The heat shock proteins from glassy-winged sharpshooter had homology to the heat shock proteins from other insects and grouped most closely with other Hemiptera when subjected to phylogenetic analysis. Two small heat shock protein sequences from glassy-winged sharpshooter grouped with that of blue-green sharpshooter, Graphocephala atropunctata (Signoret) (Fig. 1). The HSP70 sequence from glassy-winged sharpshooter grouped with two HSP70 sequences from the pea aphid, Acyrthosiphon pisum (Harris) (Fig. 2). And, the HSP90 sequence from glassy-winged sharpshooter grouped with three sequences from the pea aphid (Fig. 3). These phylogenetic analyses corroborate evidence from Pfam and PantherDB protein databases that describe the glassy-winged sharpshooter partial protein sequences as heat shock proteins. Additionally, the phylogenetic trees created using these protein sequence comparisons show that heat shock proteins can be used to determine phylogenetic and cladistical associations. Heat shock proteins have a variety of functions within the cell, including the prevention of protein aggregation and denaturation by heat, are conserved across all taxa, and are present in every species analyzed (Feder and Hofmann 1999, Sorensen 2003). Heat shock protein families are organized via their level of expression in the cell (i.e., inducible or constitutive expression) as well as the complexes formed by the heat shock proteins; however, the greatest organization criteria is the molecular weight of the heat shock proteins that include families of 20kDa, 40kDa, 60kDa, and 90kDa proteins (Gething 1997). Conserved domains exist in these families, including alpha-crystalline structure in small heat shock proteins, an N-terminal pentapeptide sequence in HSP70, and a highly conserved N-terminal domain in HSP90. The heat shock protein sequences collected from glassy-winged sharpshooter contain these conserved domains, permitting significant in-silico comparisons (data not shown). Phylogenetic analysis and alignment searches of the four heat shock protein sequences were confounded by the overwhelming number of heat shock protein sequence isoforms submitted to the National Center for Biotechnology Information. However, with consideration of heat shock protein isoforms, phylogenetic comparisons showed accurate clades of glassy-winged sharpshooter heat shock protein with those of other closely related insect taxa (Figs. 1-3). Phylogenetic trees formed on heat shock protein comparison verified other phylogenetic analyses based on mitochondrial DNA. Although many expressed sequence tags were found to be heat shock protein homologues, many heat shock proteins in the glassy- winged sharpshooter genome, including members of the small heat shock proteins, HSP60, and HSP70 families, have not been analyzed. Additionally, many members of the HSP90 family and its co-chaperones have not been sequenced. The need for more in-depth sequencing of glassy-winged sharpshooter is evident by the paucity of heat shock proteins currently identified in the glassy-winged sharpshooter genomic database. In fruit fly, Drosophila melanogaster Meigen, whose entire genome is sequenced, more than 200 heat shock proteins have been identified and submitted to the National Center for Biotechnology Information. Glassy-winged sharpshooter has a predicted genome size of ~1.24 pg (Hunter, unpublished), similar to the haploid male whitefly, Bemisia argentifolii Bellows & Perring, at ~1.1pg (Leshkowitz et al. 2006), three times the size of the Asian citrus psyllid, Diaphorina citri Kuwayama, ~0.35pg (Hunter et al. 2009) and approximately five times the size

465 of the fruit fly, which is ~0.18pg (Brown et al. 2005). Thus, we suspect that glassy- winged sharpshooters will have heat shock proteins that approximate the number in genomes of other insects. Systematic biases can distort evidence via improper gene sampling. Therefore, it is necessary to limit the effects of these entanglements by analyzing multiple genes that undergo relatively uniform evolution. Ribosomal DNA is a useful molecule for examining phylogenetic relationships among many eukaryotes, primarily because no other molecule has been sequenced as extensively (Stechmann 2003). However, phylogenetic analysis using ribosomal DNA can cause artefactual groupings of unrelated genera that have undergone rapid rRNA evolution (Philippe and Adoutte 1998, Philippe et al. 2000). Previous studies have used heat shock proteins to elucidate phylogenetic relationships in eukaryotes (Plesofsky-Vig 1992, Stechmann 2003). The ubiquitous and metropolitan prevalence of heat shock proteins allow for comparison of organisms as distantly related as that of bacteria, Escherichia coli (Migula) Castellani and Chalmers, and fruit flies (Lindquist 1986). Additionally, the importance of heat shock proteins in evolution and speciation has been well documented (Sorensen 2003). Finally, the difference between families of heat shock proteins allows researchers many options in determining precision and resolution in describing phylogenetic relationships by using the more conserved HSP90 domain or the more varied small heat shock protein family to define relationships at any level of categorization from kingdom to species (Feder and Hofmann 1999). One of the greatest determining factors in host range for an invasive species is stress tolerance, an attribute directly related to expression of heat shock proteins. As such, the sequence variation of heat shock proteins offers an excellent resource to apply in defining phylogenetic relationships and aid in revealing the ranges of pest species.

Acknowledgment

Funded in part by the USDA-APHIS and the Texas Pierce's Disease Research and Education Program. Special thanks to Christine Lynch for technical advice.

The use or mention of a trademark or proprietary product does not constitute an endorsement, guarantee, or warranty of the product by the U.S. Department of Agriculture and does not imply its approval to the exclusion of other suitable products.

References Cited

Bova, M. P., Q. Huang, L. Ding, and J. Horwitz. 2002. Subunit exchange, conformational stability, and chaperonelike function of the small heat shock protein 16.5 from Methanococcus jannaschii. J. Biol. Chem. 277: 38468- 38475. Brown, J. K., G. M. Lambert, M. Ghanim, H. Czosnek, and D. W. Galbraith. 2005. Nuclear DNA content of the whitefly Bemisia tabaci (Aleyrodidae: Hemiptera) estimated by flow cytometry. Bull. Entomol. Res. 95: 309-312. Craig, E. A. 1989. Essential roles of 70kDa heat inducible proteins. Bioessays 11: 48-52.

466

Chang, C. J., M. Garnier, L. Zreik, V. Rossetti, and J. M. Bove. 1993. Culture and serological detection of xylem-limiting bacterium causing citrus variegated chlorosis and its identification as a strain of Xylella fastidiosa. Curr. Microbiol. 27: 137-142. Cyr, D. M., and M. G. Douglas. 1994. Differential regulation of Hsp70 subfamilies by the eukaryotic DnaJ homologue YDJ1. J. Biol. Chem. 269: 9798-9804. Feder. M. E., and G. E. Hofmann. 1999. Heat-shock proteins, molecular chaperones, and the stress response: evolutionary and ecological physiology. Annu. Rev. Physio. 61: 243-282. Fu, X., and Z. Chang. 2004. Temperature-dependent subunit exchange and chaperone-like activities of Hsp16.3, a small heat shock protein from Mycobacterium tuberculosis. Biochem. Biophys. Res. Commun. 316: 291- 299. Gething, M. J. [ed.]. 1997. Guidebook to Molecular Chaperones and Protein- Folding Catalysts. Oxford University Press, Oxford, UK. Gu, L., A. Abulimiti, W. Li, and Z. Chang. 2002. Monodisperse HSP16.3 nonamer exhibits dynamic dissociation and reassociation, with the nonamer dissociation prerequisite for chaperone-like activity. J. Mol. Biol. 319: 517- 526. Hoddle, M. S., S. V. Triapitsyn, D. J. W. Morgan. 2003. Distribution and plant association records for Homalodisca coagulata (Hemiptera: Cicadellidae) in Florida. Florida Entomol. 86: 89-91. Hunter, W. B., S. E. Dowd, C. S. Katsar, R. G. Shatters, Jr., C. L. McKenzie, and D. G. Hall. 2009. Psyllid biology: expressed genes in adult Asian citrus psyllids, Diaphorina citri Kuwayama. Open Entomol. J. 3: 18-29. Jakob, U., and J. Buchner. 1994. Assisting spontaneity: the role of Hsp90 and small Hsps as molecular chaperones. Trends Biochem. Sci. 19: 205-211. Jakob, U., H. Lilie, I. Meyer, and J. Buchner. 1995. Transient interaction of Hsp90 with early unfolding intermediates of citrate synthase. J. Biol. Chem. 270: 7288-7294. Lee, G. J., and E. Vierling. 2000. A small heat shock protein cooperates with heat shock protein 70 systems to reactivate a heat-denatured protein. Plant Physiol. 122: 189-197. de Leon, J. H., W. A. Jones, and D. J. W. Morgan. 2004. Population genetic structure of Homalodisca coagulata (Homoptera: Cicadellidae), the vector of the bacterium Xylella fastidiosa causing Pierce’s disease in grapevines. Ann. Entomol. Soc. Am. 97: 809-818. Lindquist, S., and E. A. Craig. 1988. The heat-shock proteins. Annu. Rev. Genet. 22: 631-677. Mircetich, S. M., S. K. Lowe, W. J. Moller, and G. Nyland. 1976. Etiology of almond leaf scorch disease and transmission of the causal agent. Phytopath. 66: 1- 24. Nadeau, K., A. Das, and C. T. Walsh. 1993. Hsp90 chaperonins possess ATPase activity and bind heat shock transcription factors and peptidyl prolyl isomerases. J. Biol. Chem. 268: 1479-1487. Phillippe, H., and A. Adoutte. 1998. The molecular phylogeny of Eukaryota: solid facts and uncertainties, pp. 25-56. In G. Coombs, K. Vickerman, M. Sleigh, and A. Warren [eds.], Evolutionary Relationships Among Protozoa. Chapman and Hall, London.

467

Plesofsky-Vig, N., J. Vig, and R. Brambl. 1992. Phylogeny of the alphacrystallin- related heat-shock proteins. J. Mol. Evol. 35: 537-545. Pooler, M. R., and J. S. Hartung. 1995. Specific PCR detection and identification of Xylella fastidiosa strains causing citrus variegated chlorosis. Curr Microbiol. 31: 377-381. Purcell, A. H. 1997. Xylella fastidiosa, a regional problem or global threat? J. Plant Pathol. 79: 99-105. Purcell, A. H., S. Saunders, M. Hendson, M. Grebus, and M. Henry. 1999. Causal role of Xylella fastidiosa in oleander leaf scorch disease. Phytopathol. 89: 53- 58. Sobott, F., J. L. P. Benesch, E. Vierling, and C. V. Robinson. 2002. Subunit exchange of multimeric protein complexes: real-time monitoring of subunit exchange between small heat shock proteins by using electrospray mass spectrometry. J. Biol. Chem. 277: 38921-38929. Sorensen, J. G., G. T. N. Kristensen, and V. Loeschcke. 2003. The evolutionary and ecological role of heat shock proteins. Ecol Lett. 6: 1025-1037. Sreedhar, A. S., E. Kalmar, P. Csermely, and Y. F. Shen. 2004. Hsp90 isoforms: functions, expression and clinical importance. FEBS Lett. 562: 11-15. Stechmann, A., and T. Cavalier-Smith. 2003. Phylogenetic analysis of eukaryotes using heat-shock protein Hsp90. J. Mol. Evol. 57: 408-419. Turner, W., and H. Pollard. 1959. Life histories and behavior of five insect vectors of phony peach disease. Tech. Bull. U.S. Dep. Agric. 28: 1188. Young, D. A. 1958. A synopsis of the species of Homalodisca in the United States. Bull. Brooklyn Entomol. Soc. 53: 7-13.

468 VOL. 34, NO. 4 SOUTHWESTERN ENTOMOLOGIST DEC. 2009

A New Method for Collecting Clean Stable Fly (Diptera: Muscidae) Pupae of Known Age

Dennis R. Berkebile, Anthony P. Weinhold, and David B. Taylor

USDA-ARS-AMRU, 305 Entomology Hall, University of Nebraska, Lincoln, NE 68583-0938

Abstract. Stable flies, Stomoxys calcitrans L., are important pests of confined and pasture cattle. They have been reared in the laboratory to study their biology and to test new methods of control. Research on rearing modifications has concentrated on developing larval diets from materials locally abundant. Under current protocols, pupae form in the medium. Aggregations of pupae were located and removed, often with a considerable amount of extraneous material. Various methods have been developed to separate the pupae from waste material. We describe a method by which wandering larvae are enticed to leave the medium prior to pupariation. The larvae were attracted to a moist cloth on a shelf positioned at the end of the rearing pan. Almost 85% of the wandering larvae were collected on the shelf. This simplifies obtaining clean pupae and allows for collecting pupae of known age for experimental work. We also include data on the rate at which the larvae wandered onto the shelf under conditions we used in the laboratory.

Introduction

Stable flies, Stomoxys calcitrans L., are important pests of livestock throughout the world. Adults feed on the blood of many mammals, and are an economically important pest of livestock, especially cattle (Campbell et al. 1977, 1987). Stable flies have been reared in the laboratory since the early 1900s to study their biology. Early methods allowed generation of small numbers of flies to describe basic life history (Newstead 1906, Mitzmain 1913). Doty (1937) developed a method for rearing large numbers of stable flies and emphasized the importance of adding oat hulls to the larval medium. Materials such as wood shavings (McGregor and Dreiss 1955), vermiculite (Goodhue and Cantrel 1958), straw, bagasse (Bridges and Spates 1983), and peanut hulls (Hogsette 1992a) have been used to reduce the compaction of the medium and increase larval survival. Pupation normally occurs under a crust that forms on the surface of the medium and along the inside perimeter of the rearing container (Campau et al. 1953, McGregor and Dreiss 1955), but occasionally pupae are scattered throughout the medium (Bridges and Spates 1983). Clean pupae can be recovered by water flotation (Bailey et al. 1975) and dried with forced air. This method can be labor intensive, require additional space and equipment, and have the potential to expose workers to air-borne allergens.

469

Larvae of some fly species have a period before pupariation when they wander in search of a suitable habitat in which to pupariate (Frankel and Bhaskaran 1973). Some leave their larval habitats such as wounds (e.g., screwworms, Cochliomyia hominivorax Coquerel), fruit (Tephritidae), or manure (e.g., face flies, Musca autumnalis DeGeer) in search of a suitable place to pupariate in the soil (Graham and Dudley 1959, Arends and Wright 1981, Alyokhin et al. 2001). The site selected may depend on moisture, medium texture, or orientation with light. Other flies, such as the stable fly and house fly, Musca domestica L., may remain in the larval habitat, but go through a wandering stage to select the best microhabitat for development of pupae. Mature stable fly larvae search for a dark habitat with ~68% moisture, high pH, and low osmolality to pupariate (McPheron and Broce 1996). The fact that mature stable fly larvae are attracted to moisture provided an important lead in the development of a simplified method for collecting pupae from laboratory-reared stable flies. The effectiveness and efficiency of the method was evaluated and compared with our previous method. Its use for collection of pupae of known age was explored.

Materials and Methods

Stable Fly Rearing. Stable flies were maintained at the United States Department of Agriculture, Agricultural Research Service, Agroecosystem Management Research Unit, Lincoln, NE. Immatures and adults were maintained at 23 ± 2°C with variable relative humidity (30-50%) and a photoperiod of 12:12 light:dark hours. Adults were fed daily by placing an unscented Stayfree® feminine napkin (McNeil-PPC, Morris Plains, NJ), saturated with citrated bovine blood, on top of the screen cages. Eggs from 7- to 10-day-old adults were collected on moist black cotton cloth and rinsed into a dish. A modified pipette was used to transfer 1 ml of eggs (approximately 8,000) to a rearing medium consisting of wheat bran (500 g), fish meal (115 g), wood chips (200 g), and water (1.6 liters). Immatures were allowed to develop in the medium for 12-14 days before pupae were harvested. Aggregations of pupae were spooned from around the perimeter of the plastic rearing pan (36.7 x 31.9 x 14.4 cm) where they tended to aggregate. The medium in the center of the pan was also examined for additional aggregations of pupae. Pupae isolated in this way were contaminated with considerable amounts of waste medium. Laboratory experiments and quality-control measurements (average pupal weight, total pupae weight, etc.) require clean pupae. Most of the waste medium was removed by placing the pupae in water and skimming the floating pupae from the surface. The washed pupae were allowed to dry and the remaining dry wheat bran was removed by pouring the pupae in front of a fan. Development of Pupal Shelf. Wandering larvae were attracted to moist areas. Therefore, a moist cotton towel (30.5 x 42 cm) was placed on the surface of the medium at one end of the rearing pan. This procedure attracted many larvae, but the technique was messy and the cloth quickly disintegrated. Subsequently, a shelf (10.2 cm tall x 10.2 cm wide x 31.9 cm long), fabricated from the end of a rearing pan, was placed with a moist cotton towel on top of the medium at one end of the rearing pan. This method reduced the amount of rearing medium brought onto the cloth from the migrating larvae but had a problem with drying out. Wrapping a wet sponge in the moist cotton towel increased the length of time the cloth remained moist (Fig. 1).

470

Fig. 1. The shelf showing accumulation of pupae on the sponge wrapped in a towel.

Comparison of Pupae Collected by Two Methods. The number of pupae collected using the shelf was compared with pupae collected using a flotation method. After 14 days, pupae were removed from the shelf, and the dry medium and medium under the shelf were examined for pupae to quantify the efficiency of the method. The total weight of pupae collected, weight of 100 pupae, estimated number of pupae (total weight of pupae/weight of 100 pupae), pupal survival (adult emergence), and sex ratio (percentage female) were compared with pupae collected using flotation. Pupae recovered by each method were placed in separate cages. Adults were fed blood for 9 days. On Day 10, 20 females from each cage were dissected to determine their physiological age (Scholl 1980) and the number of ovarioles in each female. The test was replicated eight times. Rate of Pupation. Rearing pans were set up with a shelf as described previously. Shelves were inspected daily to ensure the sponges were moist and to check for wandering larvae. Larvae and pupae were removed from the sponge daily and placed into separate plastic containers (100-mm diameter by 15 mm) with a moist paper towel until 14 days after the eggs were placed in the rearing medium. All of the pupae in each container were weighed, then 100 were counted, weighed, and placed into emergence containers (100-mm diameter by 25 mm tall) for determination of viability and sex ratio. This procedure was replicated eight times. Variations of the Shelf Technique. Three variations of the shelf compared the influence of different amounts of moisture as an attractant for wandering larvae. Three rearing pans with shelves were set up: one with a moist sponge, one with a dry sponge that was moistened after 5 days, and one with a sponge that remained dry throughout the study. Larvae were allowed to wander until 14 days after the eggs were placed in the medium, and the pupae that formed were collected as described previously. Media in the dry area of the rearing pan, at the edge of the shelf, and under the shelf were sampled for characterization and comparison of the different areas of the rearing pan. Two 16-g samples were weighed in 100-ml disposable polypropylene beakers. One beaker was placed in a 100°C oven for 48 hours and reweighed to

471 determine the moisture content. Distilled water (80 ml) was added to the second sample, stirred on a magnetic stirrer (Velp Scientifica, Milano, Italy) for 1 hour, and allowed to set for 30 minutes. Conductivity and pH were measured with a benchtop pH/conductivity/TDS meter (Oakton Instruments, Vernon Hills, IL). Ionic strength adjuster (5 M sodium chloride, 1.6 ml) was added to the sample, stirred, and the ammonium content was determined using a benchtop ion meter (Oakton Instruments, Vernon Hills, IL) with an ammonium-combination, ion-selective electrode. This was replicated three times. Statistical Analysis. Data were analyzed with Proc GLM (SAS Institute 2004). Least-significant difference tests (LSD) were used to separate the means. Arcsine transformations were performed on percentages before mean separations were made. Unless otherwise stated, Į < 0.05.

Results and Discussion

A new method for recovering stable fly pupae was developed using a plastic shelf and a moist cloth-wrapped sponge. The moist sponge provided the wandering late third instars a suitable habitat for pupation and the shelf provided a clean environment from which they could be observed and recovered. The pupae recovered by this method were compared with those isolated using a flotation method (standard). The number and mean weight of pupae collected using the shelf did not differ significantly from those collected using flotation (Table 1). An average of 85% of the pupae produced in a pan was collected on the shelf. Percentage of viable pupae did not differ between the two methods. Significantly fewer females (54.3 and 46.3%) were collected with the shelf than by the flotation method. Females made up a significantly larger percentage (62.9%) of the stable flies collected in the medium when using the shelf than those collected on the shelf. This may indicate that the medium is a slightly more preferable habitat for pupation of female stable flies. Further investigation is required to determine possible reasons for this. Female stable flies were dissected 10 days after emergence. A significantly greater number of ovarioles (95.0 and 97.4) was counted in females collected on the shelf than by the flotation method. This small difference would have minimal effect on colony production. Most females were gravid (Scholl 1980, Stage 5) at the time of dissection (96.9% shelf, 100% standard). The shelf method was used to determine the timing and rate of migration of the mature larvae. Third instars began dispersing to the shelf 6 days after the medium was seeded with eggs. Larvae collected before Day 10 had a greater survival rate to adult than those collected later (Table 2). A greater portion of the larvae that migrated to the shelf before Day 9 were male and those migrating later were predominantly female. Most larvae were collected on Days 8 and 9 and the sex ratio was ~1:1 (50% female) during this time. The number of larvae decreased rapidly on Days 10-12. This information is important when planning studies that require stable fly pupae of known age. Variations of the shelf method were compared to verify the importance of the moist sponge as a pupation site for the stable fly and to develop options to simplify its use. When a dry sponge was used with the shelf, only 13% of the pupae were collected when compared to the shelf with the moist sponge (Table 3). The dry sponge was inconsistent. On one occasion more than 3,200 pupae were recovered when moist media was on the shelf due to shifting of the pans and the movement of the larvae. When medium was absent, no stable flies pupated on the dry sponge.

472 Table 1. Characteristics of Pupae Collected by the Flotation Method vs. the Shelf Method Shelf Flotation Shelf Mediuma Total pupal weight (g) 54.1 ± 2.80a 53.9 ± 6.46a 9.9 ± 2.51b Weight per 100 pupae (g) 1.1 ± 0.04a 1.1 ± 0.04ab 1.2 ± 0.06b Number of pupae 5128.6 ± 374.78a 4768.3 ± 587.03a 858.1 ± 254.07b % Adult emergence 82.8 ± 4.07a 77.8 ± 2.93a 78.8 ± 4.72a % Female 54.3 ± 1.56a 46.3 ± 1.85b 62.9 ± 1.29c Ovarioles per female 95.0 ± 0.68a 97.4 ± 0.64b -- Means ± SE followed by the same letter in a row are not significantly different (P < 0.05). aPupae recovered from the medium.

Table 2. Characteristics of Pupae Collected from the Shelf on Days 7-12 After Seeding the Medium with Eggs Total pupal weight (g) Weight/100 % Adult and % pupae (g) Number of pupae emergence % Female 4.9 ± 1.71a Day 7 1.08 ± 0.04a 452.9 ± 150.85c 82.6 ± 9.13a 32.8 ± 4.51a (9.2%) 24.2 ± 2.77b Day 8 1.12 ± 0.03a 2,176.3 ± 268.02a 91.8 ± 1.73a 38.9 ± 1.84a (44.8%) 19.3 ± 3.13b Day 9 1.16 ± 0.04a 1,678.9 ± 278.64b 86.8 ± 3.61a 55.3 ± 2.98abc (35.7%) 4.4 ± 1.90a Day 10 1.03 ± 0.15a 388.0 ± 173.76c 32.1 ± 11.50b 64.2 ± 3.73bc (8.2%) 1.0 ± 0.51a Day 11 0.91 ± 0.21a 88.7 ± 49.87c 7.6 ± 3.82c 71.0 ± 7.19c (1.8%) 0.2 ± 0.09a Day 12 0.91 ± 0.21a 15.4 ± 8.17c 7.6 ± 6.64c 64.4 ± 18.39c (0.3%) Means ± SE followed by the same letter in a column are not significantly different (P < 0.05).

Table 3. Comparison of Three Variations of the Shelf Method Where the Sponge Received Moisture on Day 1, Day 5, or None Sponge in pan on Sponge in pan on Day 1 Day 5 Dry sponge Total pupal weight/pan (g) 72.2 ± 8.85a 66.3 ± 9.24a 9.0 ± 6.27b Weight/100 pupae (g) 1.2 ± 0.05a 1.2 ± 0.05a 1.0 ± 0.02a Number of pupae 5,943.1 ± 931.64a 5,553.5 ± 805.02a 874.9 ± 620.09b % Adult emergence 92.8 ± 1.91a 93.4 ± 2.54a 90.0 ± 4.73a % Female 52.4 ± 2.52a 54.8 ± 1.82a 45.5 ± 4.36a % Pupae on shelf 95.8 ± 2.19a 91.5 ± 5.18a 12.9 ± 8.31b Means followed by the same letter in a row are not significantly different (P < 0.05).

473 This verifies the observations by McPheron and Broce (1996) of the importance of moisture for pupation. No significant effect on the number of pupae collected, mean weight of 100 pupae, or adult emergence was observed when the moist sponge was placed on the shelf 5 days after seeding the medium with eggs than when the moist sponge was in place at the time of seeding. The number of pupae collected on the sponge moistened on Day 5 was more variable but some maintenance time was saved because the sponge did not need to be monitored for moisture on Days 1 to 4. More than 90% of the pupae were collected regardless of when the sponge was moistened. This was an improvement over the number of pupae recovered during the initial development of the shelf method. The pH was not affected by the use of the moist sponge on the shelf. Moisture of the medium at the edge of the shelf was greater than that in the general medium for either method using the moist sponge (Table 4). Conductivity was less and moisture greater at the edge of the shelf when the sponge was moistened on Day 1 than when it was moistened on Day 5. Ammonia was significantly greater in the moist zone when the sponge was moistened on Day 1 than when the dry sponge was used. The migrating larvae may be attracted to the increase in moisture and ammonia at the edge of the shelf and wander onto the shelf in search of a suitable pupation habitat.

Table 4. Comparison of Physical Properties of Medium Sampled from 3 Areasa of Pans Where the Shelf Was Installed on Day 1 with the Sponge, Day 5 with the Sponge, and Day 1 without a Sponge Conductivity Ammonia content Area pH (ms) (ppm) % Moisture Day 1 8.8 ± 0.15a 4.7 ± 0.15bc 252.8 ± 23.30ab 48.7 ± 3.34ab Dry Day 5 8.8 ± 0.15a 4.8 ± 0.11bc 228.4 ± 21.22ab 48.4 ± 2.85ab zone Dry sponge 8.8 ± 0.18a 5.0 ± 0.15c 197.6 ± 22.98a 46.3 ± 3.40a Day 1 8.7 ± 0.12a 3.7 ± 0.43a 207.4 ± 18.42a 60.1 ± 5.26cd Moist Day 5 8.7 ± 0.10a 4.2 ± 0.20abc 195.2 ± 25.98a 54.0 ± 6.19bc zone Dry sponge 8.8 ± 0.13a 4.6 ± 0.23bc 177.4 ± 27.23a 49.0 ± 3.11ab Day 1 8.8 ± 0.11ab 4.0 ± 0.22ab 302.4 ± 33.33b 64.7 ± 5.16d Under Day 5 8.9 ± 0.10b 3.7 ± 0.18a 234.2 ± 14.27ab 64.3 ± 5.10d shelf Dry sponge 8.9 ± 0.25b 4.1 ± 0.17abc 193.5 ± 51.87a 56.9 ± 4.71bc Means followed by the same letter in a column are not significantly different (P < 0.05). aDry = center of pan away from shelf, moist = area at edge of shelf, and under = material under the shelf.

The method described for obtaining stable fly pupae is an improvement over the flotation method previously used. Isolating pupae by flotation (Bailey et al. 1975) is time consuming and inefficient. Young pupae (<24 hours old) do not float (Frankel and Bhaskaran 1973) and are lost or must be recovered by hand. We observed that the time required to harvest stable fly pupae from a pan was reduced by about 10 to 15 minutes per pan by using the shelf technique. Much time is saved when multiple pans are processed for a project. This method also reduces the equipment required for processing pupae, freeing space for other activities.

474 The males were found to seek pupation habitats before females. This is important when developing research or control techniques that require the bias of one gender over the other. The Sterile Male Technique has been proposed as a method for controlling stable flies (LaBrecque et al. 1975, 1981). Both adult male and female stable flies are blood feeders and contribute to economic damage. For such a control program to gain the support of livestock producers, mass releases would need to be made when populations are low and at levels where the populations are kept below economic injury levels. The shelf method could be used to select pupae that are predominately males, reducing the number of females released. This would increase the efficiency of such a program and reduce the number of the pest that would need to be released. In the past, much laboratory and field research has relied on the use of laboratory-reared house fly pupae. A number of methods have been devised that make it relatively easy to obtain clean house fly pupae of a known age. To properly monitor pupal parasitoids in the field it is important to use pupae of known age because some parasitoids are more successful on young muscoid pupae (King 1997, 1998). Unlike methods using sand (Hogsette 1992b) to obtain pupae of known age, the shelf makes it easy to observe the larvae/pupae with minimum disturbance to the habitat. The shelf can be observed as frequently as needed to increase the precision of the pupal age. Stable fly pupae are easily isolated using the shelf method. This method can be used to improve rearing and reduce the amount of rearing space required. This technique could also be used to improve the efficiency of control programs, and increase our knowledge of stable fly biology and their parasitoids. The shelf is a simplified method for obtaining pupae for routine rearing. It has potential use for programs in which mass rearing is required and is an equally important tool for obtaining pupae of a known age in order to obtain important information on the effect of abiotic and biotic factors on this life stage.

Acknowledgment

This work was done in cooperation with the Institute of Agriculture and Natural Resources, University of Nebraska, Lincoln, NE. The authors thank Jerry Hogsette, Muhammad Chaudhury, and Shripat Kamble for reviewing earlier versions of this manuscript. Mention of a proprietary product does not constitute endorsement or a recommendation for use by the USDA.

References Cited

Alyokhin, A. V., C. Mille, R. H. Messing, and J. J. Duan. 2001. Selection of pupation habitats by oriental fruit fly larvae in the laboratory. J. Insect Behav. 14: 57-67. Arends, J. J., and R. E. Wright. 1981. Mass rearing of face flies. J. Econ. Entomol. 74: 355-358. Bailey, D. L., T. L. Whitfield, and G. C. Labrecque. 1975. Laboratory biology and techniques for mass producing the stable fly, Stomoxys calcitrans (L.) (Diptera: Muscidae). J. Med. Entomol. 12: 189-193. Bridges, A. C., and G. E. Spates. 1983. Larval medium for the stable fly, Stomoxys calcitrans (L.). Southwest. Entomol. 8: 6-10.

475

Campau E. J., G. J. Baker, and F. D. Morrison. 1953. Rearing stable fly for laboratory tests. J. Econ. Entomol. 46: 524. Campbell, J. B., I. L. Berry, D. J. Boxler, R. L. Davies, D. C. Clanton & G. H. Deutscher. 1987. Effects of stable flies (Diptera: Muscidae) on weight gains. J. Econ. Entomol. 80: 117-119. Campbell, J. B., R. G. White, J. E. Wright, R. Crookshank, and D. C. Clanton. 1977. Effects of stable flies on weight gains and feed efficiency of calves on growing or finishing rations. J. Econ. Entomol. 70: 592-594. Doty, A. E. 1937. Convenient method of rearing the stable fly. J. Econ. Entomol. 30: 367-369. Fraenkel, G., and F. Bhaskaran. 1973. Pupariation and pupation in cyclorrhaphous flies (Diptera): terminology and interpretation. Ann. Entomol. Soc. Am. 66: 418-422. Goodhue, L. D., and K. E. Cantrel. 1958. The use of vermiculite in medium for stable fly larvae. J. Econ. Entomol. 51: 250. Graham, A. J., and F. H. Dudley. 1959. Cultural methods for mass rearing of screw-worm larvae. J. Econ. Entomol. 52: 1006-1008. Hogsette, J. A. 1992a. New diet for production of house flies and stable flies (Diptera: Muscidae) in the laboratory. J. Econ. Entomol. 85: 2291-2294. Hogsette, J. A. 1992b. Autoseparation method for harvesting house fly (Diptera: Muscidae) pupae of known age. J. Econ. Entomol. 85: 2295-2297. King. B. H. 1997. Effects of age and burial of house fly (Diptera:Muscidae) pupae on parasitism by Spalangia cameroni and Muscidifurax raptor (Hymenoptera: Pteromalidae). Environ. Entomol. 26: 410-415. King. B. H. 1998.Host age response in the parasitoid wasp Spalangia cameroni (Hymenoptera: Pteromalidae). J. Insect Behav. 11: 103-117. LaBrecque, G. C., D .W. Meiffert, and D. E. Weidhaas. 1975. Potential of the sterile-male technique for the control or eradication of stable flies, Stomoxys calcitrans Linnaeus. P. 449-60. In Proc. IAEA-FAO Symp. Sterility Principle for Insect Control, Innsbruck, Austria, 22-26 July 1974. LaBrecque,G. C., R. S. Patterson, D. F. Williams, and D. E. Weidhaas. 1981. Control of the stable fly, Stomoxys Calcitrans (Diptera: Muscidae), on St. Croix, U. S. Virgin Islands, using integrated pest management measures I. Feasibility of sterile male releases. J. Med. Entomol. 18: 194-196. McGregor, W. S., and J. M. Dreiss. 1955. Rearing stable flies in the laboratory. J. Econ. Entomol. 48: 327-328. McPheron, L. J., and A. B. Broce. 1996. Environmental components of pupariation-site selection by the stable fly (Diptera: Muscidae). Environ. Entomol. 25: 665-671. Mitzmain, M. B. 1913. The bionomics of Stomoxys calcitrans Linnaeus; a preliminary account. Phillip. J. Sci. 8: 29-48. Newstead, R. 1906. On the life-history of Stomoxys calcitrans, Linn. J. Econ. Biol. 1: 157-166. SAS Institute. 2004. SAS 9.1.3 Help and Documentation. SAS Institute, Inc. Cary, NC. Scholl, P. J. 1980. A technique for physiologically age grading female stable fly, Stomoxys calcitrans (L.). Neb. Agric. Exper. Stat. Res. Bull. 298, Lincoln, NE.

476 VOL. 34, NO. 4 SOUTHWESTERN ENTOMOLOGIST DEC. 2009

Esterases in Aedes albopictus (Skuse) from Northeastern Mexico

Gustavo Ponce-Garcia, Mohammad Badii, Mercado Roberto, and Adriana E. Flores

Laboratorio de Entomología Medica, Facultad de Ciencias Biológicas, Universidad Autonoma de Nuevo León, Mexico. Ap. Postal 109-F. San Nicolás de los Garza, NL, 66450 Mexico. Telephone (52) 818 3321453, e-mail: [email protected].

Abstract. A microassay was used to determine the activities of Į- and ȕ-esterases and acetyl cholinesterase in larvae and adults of populations of Asian tiger mosquito, Aedes albopictus (Skuse) (Diptera: Culicidae), in northeastern Mexico. Main collection sites were cemeteries and tire dumps. Three mechanisms of resistance were determined; however, elevated ȕ-esterase activity was prominent, with frequencies greater than 92% in adult Asian tiger mosquitoes in three of the populations. Based on these results, we concluded that the mechanisms of resistance in the populations studied are caused mainly by esterase activity. This study can serve as a reference for future programs aimed at controlling this species.

Introduction

Alpha- and ȕ-esterases are responsible for resistance through detoxification to organophosphates and carbamates in mosquitoes (Lee et al. 1990, Ketterman et al. 1992, Cuany et al. 1993, Georghiou 1994, Pasteur and Raymond 1996). Esterase activity also is associated with resistance to pyrethroids (Brogdon and Barber 1990; Flores et al. 2005, 2006). Until the early 1980s, it was believed that the Asian tiger mosquito, Aedes albopictus (Skuse) (Diptera: Culicidae), lived exclusively on a few islands in the Indian Ocean, in the eastern part of Asia, and on the Hawaiian Islands (Huang 1972). The first known infestation on the American continent was reported in 1985 (CDC 1986), and currently, Asian tiger mosquito is distributed between parallels 40º N and 40º S throughout the American continent. The potential threat posed by the presence of this exotic and efficient vector of dengue and possibly yellow fever and other arboviruses has caused concern in the Americas (PAHO 1987). Few campaigns by cities to control the Asian tiger mosquito have been successful. The mosquito was eradicated in Indianapolis, IN (Jardina 1990). Currently in Singapore, programs are in place to control this species. However, there still are reports of resistance to some insecticides such as organophosphates in Houston and New Orleans (New Orleans Mosquito Report 1987, Khon et al. 1988, Robert and Olson 1989). In addition, resistance to organochlorides and chlorinated hydrocarbons has been registered in China, India, Japan, Malaysia, and the Philippines (Neng et al. 1993). The Asian tiger mosquito has been affected in an indirect way by campaigns aimed at controlling yellowfever mosquito, Aedes aegypti (L.). These species often share the same habitat and seem to show similar resistance patterns to some

477 insecticides. Therefore, we have urged identification of mechanisms of resistance for Asian tiger mosquito in populations from northeastern Mexico, a geographical area through which this species was introduced into Mexico for the first time, in addition to being the principal distribution point in the country (CDC 1989).

Materials and Methods

Mosquito Collections. Enzyme activity was determined in four populations of Asian tiger mosquito distributed in northeastern Mexico: Piedras Negras, Coahuila (N 29º 39´, 211´´); Reynosa (N 26º 03´ 37.9´´ W 98º 14´37.1´´) and Río Bravo, Tamaulipas (N 26º 01´ 12.5´´ W 98º 07´ 21.4´´); and Allende, Nuevo Leon (N 25º 17.19´ W 100º 0.405´) (Fig. 1). The mosquitoes were collected as larvae from 200-liter drums, flower vases in cemeteries, and used tires and transported to an insectarium where they were maintained under controlled conditions (13:11 light:dark hours, 27ºC, and 70% relative humidity) to establish colonies. The larvae were maintained in 35 x 25 cm plastic trays and fed balanced dog food. The pupae were transferred to 250-ml beakers and placed in 30 x 30 cm breeding cages. Adult mosquitoes were fed with cotton soaked with honey (10%) water. The Aedes females were offered mice blood for the production of eggs. The mosquitoes used for biochemical assays belonged to the F2 generation. We used as a susceptible strain of Asian tiger mosquito an insectarium colony maintained for at least 2 years without exposure to insecticide.

E.U.A

1 1.- Piedra Negras, Coah. 3 2.- Allende, N. L 2 4 3.- Reynosa, Tamps. 4.- Río Bravo, Tamps.

Fig. 1. Geographical distribution of Aedes albopictus (Skuse) and area of study in northeastern Mexico.

478 Biochemical Assay. The microplate assay method was used (Brogdon 1984, 1988; Brogdon and Barber 1990; Brogdon et al. 1997). Fourth-instar larvae and 1- to 2-day-old adult females not fed blood were used for biochemical analysis. Enzymes were determined for each mosquito: Į- and ȕ-esterases, acetylcholinesterase (AChE), and insensitive acetylcholinesterase (iAChE). A minimum of two positive and negative checks was used per plate, depending on the test. For Į- ҏand ȕ-esterases, individual larvae and adults were homogenized in 100 µl potassium phosphate buffer (pH 7.2), and brought to 1 ml. The assay used 100 µl of diluted homogenate to which 100 µl Į- or ȕ-naphthyl acetate was added. The reaction mixture was incubated for 10 minutes at room temperature and the absorbance was read at 540 nm in a microplate reader (Benchmark, Biorad Laboratories, Philadelphia, PA). To determine insensitive acetylcholinesterase, extracts of individual mosquito larvae and adults were prepared and diluted as described previously. The assay was done with 100 µl of diluted homogenate to which was added 100 µl ATCH reagent mix (ATCH, propoxur, acetone, and potassium phosphate buffer) and 100 µl of DTNB. Absorbance was read immediately at 415 nm in a microplate reader (Benchmark, Bio Rad) and again at 10 minutes. The To reading was subtracted from the T10 reading for statistical analysis. Frequency of Resistance. The percentage of resistance was estimated by obtaining the number of individuals of each population that exceeded a greater absorbance value for each enzyme obtained in the susceptible strain. Results from different enzyme activity of the same mosquito population were compared (Table 2) with the susceptible strain by analysis of variance (ANOVA) and Tukey mean multiple comparison (P < 0.05). Enzyme activity was classified as “unaltered”, “incipiently altered”, or “altered” if the rate was <15, between 15 and 50, or >50%, respectively (Montella et al. 2007).

Results

Approximately 1,200 mosquitoes, including larvae and adults, were used to determine the mechanisms of resistance involving Į- and ȕ-esterases and insensitive acetylcholinesterase. Three mechanisms of resistance were found for larvae as well as adults; elevated insensitive acetylcholinesterase activity usually associated with carbamate use was not observed, while ȕ-esterase showed the greatest values in adults (Table 1). A mechanism of resistance involving Į-esterase was observed in larvae from the four populations (Tables 1 and 2), whereas the population from Piedras Negras showed enzyme activity incipiently altered (25%), followed by mosquitoes from Reynosa (14.5%), Rio Bravo (6.7%), and Allende (1.7%) that showed unaltered enzyme activity. A mechanism of resistance involving ȕ-esterase (Table 1) was noted. Unaltered enzyme activity was found in larvae from Allende and Rio Bravo (8.5% each) and incipiently altered enzyme activity was found in larvae from Reynosa (17.6%) and Piedras Negras (35%). An iAChE-related mechanism of resistance was not apparent in the larvae, which indicated that this mechanism did not occur in these populations. Only adults from Piedras Negras showed Į-esterases incipiently altered (22.9%) (Table 1). Flores et al. (2006) determined the presence of this mechanism of resistance in populations of yellowfever mosquito from the municipalities of Benito Juarez and Cozumel in the state of Quintana Roo, southern Mexico.

479 Table 1. Resistant Frequencies of Aedes albopictus Larvae and Adults for Each Enzyme Mechanism per Location Based on the Threshold Established for the Susceptible Strain Larvae Adults

Population Į-esterases ß-esterases AChE iAChE Į-esterases ß-esterases AChE iAChE Allende 1.7 8.5 31.7 0 0 0 28.6 0 Río Bravo 6.7 8.5 1.7 0 0 94.9 0 0 Piedras Negra 25.0 35.0 14.6 0 22.9 100 15.0 0 Reynosa 14.5 17.6 68.3 0 0 92.0 83.6 0

480 Table 2. Mean (± SD) Values of Enzyme Activities of Larvae and Adult Aedes albopitus Mosquito Populations Collected in Northeastern México Į-esterase ȕ-esterase Acetylcholinesterase Insensible acetylcholinesterase Population Larvae Adults Larvae Adults Larvae Adults Larvae Adults Susceptible 0.522 ± 0.13 0.612 ± 0.07 0.854 ± 0.09 0.527 ± 0.06 0.089 ± 0.03 0.171 ± 0.042 0.017 ± 0.01 0.072 ± 0.007 Nuevo León State: Allende 0.545 ± 0.12 0.569 ± 0.03 0.830 ± 0.26 0.557 ± 0.06 0.102 ± 0.04 0.235 ± 0.04 0.014 ± 0.008 0.009 ± 0.007 Tamaulipas State: Rio Bravo 0.612 ± 0.13 0.551 ± 0.02 0.933 ± 0.19 0.936 ± 0.08 0.075 ± 0.02 0.136 ± 0.02 0.014 ± 0.006 0.011 ± 0.006 Reynosa 0.703 ± 0.10 0.603 ± 0.04 1.09 ± 0.14 0.934 ± 0.11 0.151 ± 0.05 0.303 ± 0.047 0.02 ± 0.009 0.007 ± 0.007 Coahuila State: Piedras Negras 0.696 ± 0.16 0.757 ± 0.10 1.14 ± 0.19 1.24 ± 0.06 0.08 ± 0.03 0.183 ± 0.06 0.011 ± 0.007 0.012 ± 0.007

Except for adult mosquitoes from Allende, ȕ-esterase activity showed altered enzyme activity compared with the susceptible strain (Table. 1). One-hundred percent of the population from Piedras Negras showed altered enzyme activity, followed by populations from Rio Bravo (94.4%) and Reynosa (92%). However the population from Allende showed unaltered enzyme activity of 0%. As with larvae, an insensitive AChE-related mechanism of resistance was not apparent for adults. Table 2 shows the mean enzymatic activity of Į-, ȕ-, and insensitive acetylcholinesterase. Alpha-esterases in larvae from Rio Bravo, Piedras Negras, and Reynosa were significantly different from those in the susceptible strain, whereas for adults, only the strain from Piedras Negras was significantly different from the susceptible strain. The strains of larvae from Piedras Negras and Reynosa showed significant difference in ȕ-esterases compared with the susceptible strain. For the adults, the populations from Rio Bravo, Piedras Negras, and Reynosa showed significant difference compared with the susceptible strain. The iAChE did not show significant difference in larvae or adults of these three strains.

Discussion

Asian tiger mosquito is resistant to insecticide in many countries such as China, Japan, Malaysia, Thailand, and Vietnam in the Pacific region of Asia, Venezuela in South America, and those in the Caribbean Islands (WHO 1986, Wu et al. 1992, Mazarri and Georghiou 1995). In Singapore, Rawlins and Joseph (1995) reported Asian tiger mosquito was resistant to DDT, dieldrin, and propoxur. However, just Pethuan et al. (2007) determined activities of enzymes mixed function oixdases (MFO), nonspecific esterases (Į and ȕ), glutathione-S-transferases (GST), and insensitive acetylcholinesterase (AChE) for this species from Thailand. In larval populations, it was possible to determine mechanisms of resistance involving Į-and ȕ-esterases, albeit at low enzyme activity, but greater activity for ȕ- esterase was observed mainly in adults in populations from Piedras Negras, Rio Bravo, and Reynosa. In adults, elevated Į-esterase activity was determined only in the population from Piedras Negras. Flores et al. (2006) determined frequencies of resistance of 100% in populations of yellowfever mosquito from Benito Juarez and Cozumel, Quintana Roo, Mexico, although absorbance values were not greater than those of the population from Piedras Negras. High ȕ-esterase activity was found in populations from Río Bravo, Piedra Negras, and Reynosa, with values similar to those of the populations from Benito Juarez and Cozumel. Both mechanisms have been determined also with high activity levels for Į2- and ȕ2- esterases in southern house mosquito, Culex quinquefasciatus Say (Wirth et al. 1990), Į3- and ȕ3-esterases in C. tarsalis Coquillett (Prabheker et al. 1987), and Į4-, Į5-, ȕ4- and ȕ5-esterases in northern house mosquito, C. pipiens L. (Poierié et al. 1992). There was no indication of involvement of insensitive AChE in resistance in adults or larvae. Yellowfever mosquito from Baja California Sur and Norte, Veracruz, and Quintana Roo, Mexico (Medina 2003; Flores et al. 2005, 2006) also showed the absence of this mechanism, similar to that reported by Mazarri and Georghiou (1995) for populations from Coro and Maracay, Venezuela. However, this mechanism has been reported for other species such as Anopheles albimanus Wiedmann in populations from Izabal and Jutiapa, Guatemala (Brogdon et al. 1989)

481 and northern house mosquito from Padova (Severini et al. 1993) and Cyprus, Italy (Wirth 1998). Based on the results of this study, we determined that Asian tiger mosquito has enzyme activity levels similar to those of yellowfever mosquito in Mexico (Medina 2003; Flores et al. 2005, 2006), as well as An. albimanus (Say) (Brogdon et al. 1989) and northern house mosquito (Raymond et al. 1987, Bisset et al. 1990, Severini et al. 1993, Rivet et al. 1994), species resistant to distinct groups of insecticides currently used throughout the world.

Acknowledgment

This study was supported by the research program PAICYT, Universidad Autonoma de Nuevo Leon, Monterrey, Mexico.

References Cited

Bisset, J., M. Rodríguez, C. Díaz, E. Ortiz, M. Marquetti, and J. Hemingway. 1990. The mechanisms of organophosphate and carbamate resistance in Culex quinquefasciatus (Diptera.Culicidae) from Cuba. Bull. Entomol. Res. 80: 245- 250. Brogdon, W. G. 1984. Mosquito protein microassay. I. protein determinations from small portions of single mosquito homogenates. Comp. Biochem. Physiol. B: 79: 457-459. Brogdon, W. G. 1988. Microassay of acetylcholinesterase activity in small portions of single mosquito homogenates. Comp. Biochem. Physiol. 90: 145-150. Brogdon, W. G., and A. M. Barber. 1990. Fenitrothion-deltamethrin cross resistance conferred by esterases in Guatemala Anopheles albimanus. Pestic. Biochem. Physiol. 37: 130-139. Brogdon, W. G., R. F. Beach, J. M. Stewart, and L. Castanaza. 1989. Analisis por ensayo de microplaca de la distribución de la residencia de Anopheles albimanus a organofosforados y carbamatos en Guatemala. Bol. Of Sanit Panam. 106: 139-151. Brogdon, W. G., J. C. McAllister, and J. Vulule. 1997. Heme peroxidase activity in single mosquitoes identifies individuals expressing an elevated oxidase for insecticide resistance. J. Am. Mosq. Control Assoc. 13: 233-237. Centers for Disease Control and Prevention (CDC). 1986. Aedes albopictus (Say) introduction in Texas. MMWR. 35: 141-142. Centers for Disease Control and Prevention (CDC). 1989. Update: Aedes albopictus (Say) infestation United States, México. Morb. Mort. Week. Rpt. 38: 440, 445-446. Cuany A., J. Handani, J. Berge, D. Fournier, M. Raymond, G. P. Georghiou, and N. Pasteur. 1993. Action of esterase B1 on clorpyrifos in organophosphate resistant Culex mosquitoes. Pestic. Biochem. Physiol. 45: 1-6. Flores, A. E., W. V. Albeldaño, I. S. Fernandez, M. H. Badii, M. H. B. Loaiza, G. G. Ponce, S. F. Lozano, W. G. Brogdon, W. C. Black IV, and B. Beaty. 2005. Elevated Į esterase levels associated with permethrin tolerance in Aedes aegypti (L.) from Baja California, Mexico. Pestic. Biochem. Physiol. 82: 66- 78.

482 Flores, A. E., J. S. Grajales, I. S. Fernandez, G. G. Ponce, M. H. B. Loaiza, S. Lozano, W. G. Brogdon, W. C. Black IV, and B. Beaty. 2006. Mechanisms of insecticide resistance in field populations of Aedes aegypti (L.) from Quintana Roo, Southern Mexico. J. Am. Mosq. Control Assoc. 22: 672-677. Georghiou, G. P. 1994. Principles of insecticide resistance management. Phytoprotection 75 Suppl. 51-59. Huang, H. M. 1972. The subgenus Stegomyia of Aedes in Southeast Asia. The Scutellaris group of species. Contrib. Am. Entomol. Inst. 9: 108. Jardina, B. 1990. The eradication of Aedes albopictus (Say) in Indianapolis, Indiana. J. Am. Mosq. Control. Assoc. 6: 310-311. Ketterman, A. J., K. G. I. Jayawardena, and J. Hemingway. 1992. Purification and characterization of a carboxy-esterase involved in insecticide resistance from the mosquito Culex quinquefasciatus. Biochem. J. 287: 355-360. Khon, B., D. Shuterland, D. Sprenger, D. Dickerson, and H. Nguyen. 1988. Susceptibility status of Ae. Albopictus (Say) to three topically applied adulticides. J. Am. Mosq. Control Assoc. 4: 310-313. Lee, J. G., N. Pasteur, and P. Georghiou. 1990. Preliminary results of an Elisa test for detections of organophosphate-resistance in Culex populations due to increased detoxification by esterase. 58th Annual Conference of the California Mosquito and Vector Control Association.111-115. Mazarri, M. B., and G. P. Georghiou. 1995. Characterization of resistance to organophosphate, carbamate and pyrethroid insecticides in field populations of Aedes aegypti (L.) from Venezuela. J. Am. Mosq. Control Assoc. 11: 315- 322. Medina, S. A. 2003. Determinación de Mecanismos de Resistencia en Aedes aegypti (L.) a un Insecticida Piretroide en el Estado de Veracruz, México. Tesis (inédita),.Instituto Tecnológico de Sonora. Montella I. R, A. J. Martins, V. Fernandez, L. B. Pereira, I. A. Braga, and D. Valle. 2007. Insecticide Resistance Mechanisms of Brazilian Aedes aegypti Populations from 2001 to 2004. Am. J. Trop. Med. Hyg. 77: 467-477. Neng Yan, W. X., H. Fuming, and D. Dazong. 1993. Susceptibility of Aedes albopictus (Say) from China and mechanism of DDT resistance. J. Am. Mosq. Control. Assoc. 9: 394-397. New Orleans Mosquito Control Board. 1987. Annual Report. City of New Orleans. /NOMCB PAHO. 1987. Aedes albopictus (Say) en la Americas: plan de accion. Washington DC. Documento CE99. Pasteur, N., and M. Raymond. 1996. Insecticide resistance genes in mosquitoes: their mutations, migration, and selection in field populations. J. Heredity 87: 444-449. Pethuan, S., N. Jirakanjanakit, S. Saengtharatip, T. Chareonviriyaphap, D. Kaewpa, and P. Rongnoparut. 2007. Biochemical studies of insecticide resistance in Aedes (Stegomyia) aegypti and Aedes (Stegomyia) albopictus (Diptera: Culicidae) in Thailand. Tropical Biomedicine 24: 7-15. Poierié, M., M. Raymond, and N. Pasteur. 1992. Identification of two distinct amplifications of the esterase B locus in Culex pipiens (L.) mosquitoes from Mediterranean countries. Biochem. Genet. 30: 13-26. Prabhaker, N., G. P. Georghiou, and N. Pasteur. 1987. Genetic association between highly-active esterases and organophoshpate resistance in Culex tarsalis. J. Am. Mosq. Control Assoc. 3: 473–475. Medline, CSA.

483 Rawlins, S. C., and W. O. U. Joseph. 1995. Resistance in some Caribbean populations of Aedes aegypti (L.) to several insecticides. J. Am. Mosq. Control Assoc. 11: 59-65. Raymond, M., N. Pasteur, G. P. Georghiou, R. B. Mellon, M. C. Wirth, and M. K. Hawley. 1987. Detoxification esterases new to California, USA, in organophosphate-resistant Culex quinquefasciatus (Diptera: Culicidae). J. Med. Entomol. 24: 24-27. Rivet, Y., M. Raymond, J. A. Rioux, A. Delalbre, and N. Pasteur. 1994. Resistance monitoring in Culex pipiens (Diptera: Culicidae) from central-eastern France. J. Med. Entomol. 31: 231-239. Robert, L., and J. Olson. 1989. Susceptibility of female Ae. albopictus (Say) from Texas to commonly used adulticides. J. Am. Mosq. Control Assoc. 5: 251- 253. Severini, C., R. Romi, M. Marinucci, and M. Raymond. 1993. Mechanisms of insecticide resistance in field populations of Culex pipiens from Italy. J. Am. Mosq. Control Assoc. 9: 164-168. WHO. 1986. Resistance of vectors and reservoirs of disease to pesticide. Tenth report of the WHO Expert Committee on Vector Biology and Control. Technical Report Series 737. World Hearth Organization, Geneva, Switzerland. Wirth, M. C. 1998. Isolation and characterization of two novel organophosphate resistance mechanisms in Culex pipiens from Cyprus. J. Am. Mosq. Control Assoc. 14: 397-405. Wirth, M. C., M. Marnique, G. P. Georghiou, and N. Pasteur. 1990. Esterases A2 and B2 in Culex quinquefasciatus (Diptera: Culicidae): role in organophosphate resistance and linkage. J. Medical Entomol. 27: 202-206. Wu, N., Y. Xiao, F. M. Huang, and D. Z. Chen. 1992. Susceptibility of Aedes albopictus (Say) from China to insecticides and mechanism of DDT resistance. J. Am. Mosq. Control Assoc. 8: 394-397.

484 VOL. 34, NO. 4 SOUTHWESTERN ENTOMOLOGIST DEC. 2009 SCIENTIFIC NOTE

A Visual Guide for Identification of Euschistus spp. (Hemiptera: Pentatomidae) in Central Texas

Jesus F. Esquivel1, Roger M. Anderson1, and Robert E. Droleskey2

Stink bugs have become problematic in cotton, Gossypium hirsutum L., following efforts by Boll Weevil Eradication Programs. These programs have led to a reduced number of insecticide applications that normally suppressed stink bugs. Several phytophagous species of Euschistus in cotton and other crops are often grouped into the “brown stink bug complex.” However, this does not enable accurate identification of individual Euschistus species; unlike the southern green stink bug, Nezara viridula L., and green stink bug, Acrosternum hilare Say, which are relatively easy to identify in the field. In addition to their similarities in coloration causing potential problems with identification, brown stink bugs are less susceptible to certain classes of insecticides (Snodgrass et al. 2005). Dichotomous keys are available for identifying Euschistus spp. (Rolston 1974, McPherson and McPherson 2000) but are not helpful for accurate and rapid identification of stink bugs in the field. Because of the advanced stage of the Boll Weevil Eradication Program in the Blacklands region, known tolerances to certain classes of insecticides, and seemingly apparent similarities between species, this study was undertaken to identify Euschistus spp. in Central Texas and provide a visual guide for accurate identification in the field. From January 2007 through January 2009, five black-light and two pheromone-baited traps were used to collect adult brown stink bugs in Central Texas (Burleson County). Traps were placed in or near pecan, Carya illinoinensis (Wangenh.) K. Koch, orchards adjacent to fields previously planted with maize, Zea mays L.; cotton; or soybeans, Glycine max (L.) Merr. Traps were serviced three times a week, and captured adults were taken to a laboratory where dichotomous keys of Rolston (1974) and McPherson and McPherson (2000) were used for identification. Voucher specimens (Accession Number 672) were confirmed and cataloged by Ed G. Riley, Museum Curator, Department of Entomology, Texas A&M University, College Station, TX. Six Euschistus species were identified in the region (Fig. 1). The absence of spots in the membranous area of the hemelytra (Fig. 1A, inset upper right) is the key characteristic for distinguishing E. quadrator Rolston from other species. Spots in the membranous area (Fig. 1B, inset upper right) or in confluence with (i.e., in line with) venation of the hemelytra are present in the other five species. Similar to E. quadrator, E. tristigmus (Say) has a unique character in the presence of large median black spot(s) on the abdominal venter (Fig. 1B, inset lower right) that can be used to distinguish this species from all others in the region.

1USDA, ARS, Southern Plains Agricultural Research Center, Areawide Pest Management Research Unit, College Station, TX. 2USDA, ARS, Southern Plains Agricultural Research Center, Food and Feed Safety Research Unit, College Station, TX.

485 A B

C D

E F

Fig. 1. Euschistus spp. in Central Texas: A) E. quadrator (inset upper right, membranous wing area without spots [7X]); B) E. tristigmus (inset upper right, membranous wing area with spots [7X]; inset lower right, spot ventrally on abdomen [7X]); C) E. obscurus (inset lower right, evaporative area adjacent to scent gland not punctuate with black markings [3X]); D) E. crassus (inset lower right, evaporative area adjacent to scent gland punctuate with black markings [3X]); E) E. ictericus (inset lower right, ventral abdominal spiracles narrowly ringed with black [5X]); and, F) E. servus (inset lower right, ventral abdominal spiracles narrowly ringed with pale color [5X]). In all figures, the scale bar unit = 1 mm.

486 The number of spots can vary from one to three, and insects with three spots have been observed in southern Brazos County (R. Anderson, personal observation). Additionally, the shape of the spots varies between males and females; females have more elongated spots. Although not shown, the predatory spined soldier bug, Podisus maculiventris (Say), also has markings on the venter but can be differentiated by the more acute and lateral orientation of the humeri, or ‘shoulders’, on the pronotum than those shown in Fig. 1B. The remaining four species (Fig. 1C-F) do not have median black spots on the abdominal venter. The four species can be partly differentiated based on the presence or absence of a callous ridge, or fascia, located dorsally between the humeri. The callous ridge is formed by the absence of punctate black markings (i.e., spots) across the dorsal surface of the pronotum between the humeri. Euschistus obscurus (Palisot) (Fig. 1C), E. crassus Dallas (Fig. 1D), and E. ictericus L. have a callous ridge between the humeri but each has additional criteria to help in differentiation. To differentiate between E. obscurus and E. crassus, the ventral evaporative areas adjacent to the scent glands should be examined. In E. obscurus, the evaporative area is not punctate with black markings (Fig. 1C, inset). Conversely, the evaporative area is punctate with black markings on E. crassus (Fig. 1D, inset lower right), and this is the only species of the six identified that has such markings on the evaporative area. Also, E. obscurus has more densely punctate markings on the anterior half of the pronotum than do other species, making the anterior half of the E. obscurus pronotum appear darker than the posterior half. Euschistus ictericus (Fig. 1E) can be differentiated from E. obscurus and E. crassus based on the acute humeri and presence of black rings around the abdominal spiracles (Fig. 1E, inset lower right). However, while E. ictericus does have a callous ridge between humeri, the degree of punctate markings can vary such that the callous ridge is not clearly defined. Unlike E. ictericus, E. servus (Say) (Fig. 1F) does not have ventral abdominal spiracles ringed with black, acute humeri, or a callous ridge on the pronotum. The spiracles are narrowly ringed with pale color on E. servus (Fig. 1F, inset lower right) and has variably shaped humeri. E. servus is the largest of the six species. In conjunction with the aforementioned species-specific characteristics, coloration can be used to assist in separating some of the species. E. tristigmus and E. obscurus are typically darker than the other four species. E. ictericus occasionally has russet (or red) coloration at the humeri, and this coloration can extend dorsally. Coloration should not be a sole determining factor for species identification, however. Size of the insect can be helpful to differentiate species. Morphometrics or repeated measurements of adults were not done but the specimens shown in Fig. 1 are representative of each species. At the two extremes, E. quadrator (Fig. 1A) was the smallest (in length and width), approximately 9 mm long and 5 mm wide along the abdomen (excluding protrusion of humeri); E. servus (Fig. 1F) was the most robust and largest of all species, approximately 14 mm long and 8 mm wide along the abdomen (excluding protrusion of humeri). Specimens of E. tristigmus, E. crassus, and E. ictericus are of similar size but each has the unique aforementioned characteristics to accurately identify the species. Observations and character designations presented here are in agreement with terminologies by Rolston (1974) and McPherson and McPherson (2000), but a more precise description is needed to accurately identify E. obscurus in Central

487 Texas. Rolston (1974) and McPherson and McPherson (2000) described the membrane of the hemelytra as “unspotted” or “not marked with brownish spots” for E. obscurus, suggesting absence of spots. Examination of voucher specimens reveals that E. obscurus does not have spots in the membranous area of the hemelytra per se, but the spots are in confluence with the venation of the membranous area. Museum specimens from 12 localities indicate the numbers of spots vary, but personal communications (S. Bundy, New Mexico State University) suggest the presence and localization of the spots within the venation may be a regional occurrence. Regardless, all voucher specimens from Central Texas exhibited the markings within the wing venation. The species identified were found in or near corn, cotton, soybeans, and pecan orchards in Central Texas and have been reported as economically important pests in row, forage, and tree fruit crops (McPherson and McPherson 2000, and references therein). This visual guide allows simplified and accurate identification of Euschistus spp. by row, forage, and fruit crop producers and researchers, as well as assists in selection of appropriate insecticides to control these insects.

Mention of trade names or commercial products in this report is solely for the purpose of providing specific information and does not imply recommendation or endorsement by the U.S. Department of Agriculture.

References Cited

McPherson, J. E., and R. M. McPherson. 2000. Stink Bugs of Economic Importance in America North of Mexico. CRC Press, Boca Raton, FL. Rolston, L. H. 1974. Revision of the genus Euschistus in Middle America (Hemiptera, Pentatomidae, Pentatomini). Entomologica Americana 48: 1- 102. Snodgrass, G. L., J. J. Adamczyk, Jr., and J. Gore. 2005. Toxicity of insecticides in a glass-vial bioassay to adult brown, green, and southern green stink bugs (Heteroptera: Pentatomidae). J. Econ. Entomol. 98: 177-181.

488 VOL. 34, NO. 4 SOUTHWESTERN ENTOMOLOGIST DEC. 2009 SCIENTIFIC NOTE

Fortuitous Establishment of Rhyzobius lophanthae (Coleoptera: Coccinellidae) and Aphytis lingnanesis (Hymenoptera: Encyrtidae) in South Texas on the Cycad Aulacaspis Scale, Aulacaspis yasumatsui (Hemiptera: Diaspididae)

Daniel Flores and Jason Carlson

USDA, APHIS, PPQ, CPHST, Mission Laboratory, Edinburg, TX

The cycad aulacaspis scale, Aulacaspis yasumatsui Takagi, is currently found in China, Singapore, Hong Kong, Cayman Islands, Puerto Rico, U.S. Virgin Islands, Hawaii, and Florida (Germain and Hodges 2007). It was originally described from specimens collected on a Cycas sp., in Bangkok, Thailand, in 1972 (Takagi 1977). In recent years, finds have also been reported in California, Georgia, and Nevada (IUCN/SSC Cycad Specialist Group 2009). In 2006, severe outbreaks of cycad aulacaspis scale were reported in South Texas where sago palms, Cycas revoluta Thunb., adorn landscapes and are important ornamental plants for commercial nursery growers (Bográn et al. 2006). Of the more than 20 species of scale insects that occur on cycads in Florida, the most damaging species is cycad aulacaspis scale (Hodges et al. 2003). In Thailand, this armored scale is considered a pest of cycads and is usually maintained in low densities by parasitoids (Tang et al. 1997). In 1998, two natural enemies of cycad aulacaspis scale, a predaceous beetle, Cybocephalus nipponicus Endroudy-Younga (misidentified as Cybocephalus binotatus Grouvelle - Smith and Cave 2006), and a parasitic wasp, Coccobius fulvus (Compere and Annecke), were released in Florida by the Tropical Research and Education Center, Homestead, FL. These natural enemies seem to control cycad aulacaspis scale effectively, but the scale undergoes outbreaks that are more severe when natural enemies are absent (Hodges et al. 2003). The predator Rhyzobius lophanthae Blaisdell 1892 (Coleoptera: Coccinellidae) is a natural enemy of many species of diaspidid scale (Hodek 1973, Stathas 2000). Because of its prey specificity, high fecundity, adult longevity, absence of diapause, good mobility, rapid development (five to seven generations per year), and lack of parasitism, R. lophanthae is considered an important natural enemy of many species of armored scale (Smirnoff 1950, Rubstov 1952, Katsoyannos 1996). It is known for its effectiveness in integrated pest management of scale insects (Hodek 1973), via inundative releases (Katsoyannos 1996) and classical biological control programs in several countries including the USA (Yus 1973), Italy (Bouvier 1913, Smirnoff 1950), Argentina (Salvadores 1913), Bermuda (Bennet and Hughes 1959), Algeria, Tunisia, Morocco (Rungs 1950, Smirnoff 1950) and Georgia (Rubstov 1952). It is also reported from Greece, where it may have spread from neighboring countries (Argyriou and Katsoyannos 1977, Katsoyannos 1996). There are no records of this predator species being released in Texas. However, since October 2006, R. lophanthae has been found associated with

489 infestations of cycad aulacaspis scale in South Texas. On 9 July 2008, insects collected in South Texas were positively identified by Dr. Natalia J. Vandenberg (USDA / APHIS / PPQ / Systematic Entomology Laboratory, Beltsville, MD) as Rhizobius lophanthae. All developmental stages have been observed on sago palms, with larvae and adults being most prevalent. During visual inspections, adults were found feeding on cycad aulacaspis scale or in copulo on cycad foliage heavily infested by cycad aulacaspis scale. Adults were found on all parts of the plants; larvae were found mostly along the midrib (top and bottom) of the fronds. Few larvae were observed on the stem of the plants. Eggs were found mostly at the base of the frond where it meets the stem; a few times eggs were observed close to the midrib on the underside of a frond. Pupae were found along the bases of the fronds. To determine the distribution and abundance of cycad aulacaspis scale in South Texas, Brownsville, Harlingen, McAllen, and the town of South Padre Island were surveyed between October 2006 and May 2009. Yellow sticky traps (6.4 x 7.6 cm) were placed on 10 mature sago palm trees (1 to 2 m in height) at each location and monitored monthly to determine the number of winged-males of cycad aulacaspis scale and seasonal variation between locations. While monitoring populations of cycad aulacaspis scale in South Texas, large numbers of a parasitoid were observed on the yellow sticky traps at all locations. Infested fronds were collected and placed in plexi-glass emergence cages (40.6 x 45.7 x 61 cm). We found cycad aulacaspis scale being parasitized by a wasp identified by Dr. Gregory A. Evans (USDA / APHIS / PPQ / Systematic Entomology Laboratory, Beltsville, MD) as belonging to the Aphytis lingnanensis group. Interestingly, Woolley et al. (1994) reported A. lingnanensis attacking the California red scale, Aonidiella aurantii (Maskell) (Hemiptera: Diaspididae) in Texas. This is the first report of an Aphytis species attacking A. yasumatsui. Several Aphytis species have been used successfully in biological control programs against armored scales (Rosen and DeBach 1990, Van Driesche and Bellows 1996). For several years, many scientists have searched for natural enemies of cycad aulacaspis scale, but to our knowledge an Aphytis species was not previously found. Species in the A. lingnanensis group are taxonomically similar and difficult to separate (G.A. Evans, personal communication). The insects in South Texas may be A. lingnanensis Compere, which attacks several species of scale and is widespread. DNA analysis may perhaps offer the best solution to identification of species. Field data suggest that cycad aulacaspis scale populations are variable throughout the year and establishment and abundance of A. lingnanensis and R. lophanthae on cycad aulacaspis scale are contributing to this variability. These beneficial insects have not been released in South Texas and no one knows how they managed to colonize the area or if they changed hosts. However, we believe they fortuitously entered South Texas along with cycad aulacaspis scale on sago palms and proceeded to spread. Since the initial discovery near Miami, cycad aulacaspis scale has become widely distributed in Florida (Hamon 2000). In 2006, cycad aulacaspis scale was detected in South Texas, making it a new state record and suggesting that it was an incipient infestation. Our data also indicate that Aphytis lingnanesis and Rhyzobious lophanthae have become fortuitously established on cycad aulacaspis scale. Although abundance of cycad aulacaspis scale varies throughout the year, to date the scale has not reached problematic population levels in South Texas.

490

References Cited

Argyriou, L. C., and P. Katsoyannos. 1977. Coccinellidae species found in the olive groves of Greece. Ann. Inst. Phytopathol. Benaki (N.S.) 11: 331-345. Bennet, F. D., and I. W. Hughes. 1959. Biological control of insect pests in Bermuda. Bull. Entomol. Res. 50: 424-428. Blaisdell, F. E. 1892. A new species of Coleoptera from California. Entomol. News 3: 51. Bogran, C. E., B. A. Castro, and S. Ludwig. 2006. The cycad aulacaspis scale, a pest of sago palms in Texas. University of Texas Cooperative Extension Pub. EEE-00038. Bouvier, E. L. 1913. Coccinelles contre Cochenilles. Rev. Sci. Paris 20: 673-677. Germain J. F., and G. S. Hodges. 2007. First report of Aulacaspis yasumatsui (Hemiptera: Diaspididae) in Africa (Ivory Coast), and update on distribution, Florida Entomol. 90: 4. Hamon, A. 2000. Cycad aulacaspis scale, Aulacaspis yasumatsui. http://doacs.state.fl.us/~pi/enpp/ento/aulacaspis.html (currently unavailable). Hodek, I. 1973. Biology of Coccinellidae. Czechoslovak Academy of Sciences. Prague, Czechoslovakia. Hodges, G. S., F. W. Howard, and E. A. Buss. 2003. Update on management methods for cycad aulacaspis scale. Fla. Coop. Ext. Ser. Bul. ENY-680. IUCN/SSC Cycad Specialist Group. 2009. Cycad Aulacaspis Scale Information Page. http://www.cycadsg.org/pages/CAS.htm. Katsoyannos, P. 1996. Integrated Insect Pest Management for Citrus in Northern Mediterranean Countries. Benaki Phytopathological Institute, Kifissia, Athens, Greece. Rosen, D., and P. DeBach. 1990. Conservation of natural enemies, pp. 461-472. In D. Rosen [ed.], Armored Scale Insects: Their Biology, Natural Enemies and Control. Elsevier, Amsterdam. Rubstov, I. A. 1952. Lindorus – an effective predator of diaspine scales. Entomol. Obozr. 32: 96-106. Rungs, C. 1950. Sur l’extension spontanee au Maroc du Rhyzobius (Lindorus) lophanthae Blaisd. (Col.:Coccinellidae). Bull. Soc. Entomol. Fr. 55: 9-11. Salvadores, A. Z. 1913. El Durazno (The Peach). Reprint from Bol. Minist. Agric. Buenos Aires (in Rev. Appl. Entomol. 1914) II. Smirnoff, W. 1950. Sur la biologie au Maroc de Rhyzobius (Lindorus) lophanthae Blaisd. (Col.: Coccinellidae). Rev. Pathol. Veg. Entomol. Agric. Fr. 29: 190- 194. Smith, T. R., and R. D. Cave. 2006. Pesticide susceptibility of Cybocephalus nipponicus and Rhyzobius lophanthae (Coleoptera: Cybocephalidae, Coccinellidae). Florida Entomol. 89: 502-507. Stathas, G. J. 2000. Rhyzobius lophanthae prey consumption and fecundity. Phytoparasitica 28: 203-211. Takagi, S. 1977. A New Species of Aulacaspis associated with cycad in Thailand. Insecta Matsumurana, New Series 11: 63-72. Tang W., S. L. Yang, and P. Vatcharakorn. 1997. Cycads of Thailand. Nong Nooch Tropical Garden and the Cycad Conservation Company, Bangkok. Van Driesche, R., and T. S. Bellows, Jr. 1996. Biological Control. Chapman and Hall, New York.

491

Woolley, J. B., M. Rose, and P. C. Krauter. 1994. Morphometric comparisons of Aphytis species in the lingnanensis group (Hymenoptera: Aphelinidae), pp. 223-244. In D. Rosen [ed.], Advances in the Study of Aphytis (Hymenoptera: Aphelinidae). Intercept Ltd., Andover, UK. Yus, R. 1973. On the presence in the Iberian Peninsula of Rhizobius lophanthae (Blaisdell, 1892) (Col.: Coccinellidae). Graellsia 29: 111-115.

492 VOL. 34, NO. 4 SOUTHWESTERN ENTOMOLOGIST DEC. 2009

SUBJECT INDEX TO VOLUME 34

Aedes aegypti Asian citrus psyllid resistance to permethrin in abundance in Florida, Puerto Rico, northern Mexico, 167 and Texas citrus groves, 1

Aedes albopictus Asian tiger mosquito, 477 esterases in from northeastern Mexico, 477 Aulacaspis yasumatsui fortuitous establishment of Aleyrodidae, 189, 431 Rhyzobius lophanthae and Aphytis lingnanesis in South Allium cepa, 13, 219, 417 Texas on the cycad Aulacaspis scale, 489 Anthonomus grandis grandis, 31 Bacillus thuringiensis, 61, 239 Anyphaenidae, 103 Bandedwinged whitefly, 431 Aphelinidae, 327, 431 Bean thrips, 417 Aphididae, 121, 131, 205 Beauveria bassiana Aphytis lingnanesis effects of relative humidity on fortuitous establishment of efficacy on nymphs of Rhyzobius lophanthae and in sweetpotato whitefly, Bemisia South Texas on the cycad tabaci on hibiscus in Aulacaspis scale, Aulacaspis greenhouses, 189 yasumatsui, 489 Bemisia tabaci, 327, 431 Arctiidae, 239 effects of relative humidity on efficacy of BotaniGard™ on Arundo donax, 329, 347, 359 nymphs on hibiscus in economic implications for the greenhouses, 189 biological control of: Rio Grande Basin, 377 Blacklegged tick, 273

Arundo wasp Boll weevil seasonality and movement of effect of hexaflumuron on gestation adventive populations of, a and reproduction of adult, 31 biological control agent of giant reed in the Lower Rio Grande Braconidae, 213 Basin in South Texas, 347 Brassica oleracea, 193

493 Brevycorine brassicae, 193 Clover mite effects of supplemental irrigation Bryobia praetiosa on populations of and other effects of supplemental irrigation arthropods in a Kentucky on populations of and other bluegrass lawn, 69 arthropods in a Kentucky bluegrass lawn, 69 Coccinellidae, 179, 489

Caliothrips fasciatus, 417 Coreidae, 305

Canine, 273 Corn leafhopper a newly discovered parasitoid of in Carabidae, 43 California, 99

Carabids Cotton, 31, 61, 103, 239, 417, 431, impact of land management 485 practices on and other arthropods on the western High Crambidae, 213 Plains of North America, 43 Crane fly, 85 Carya illinoinensis, 111, 227, 305, 319, two autumnal from Michigan: 447, 485 comments on the difficulties of identification and a review of Casearia corymbosa, 75 the Tipula subgenus Platytipula in the United States and Cattle, 263, 469 Canada, 85 Tipula (Yamatotipula) jacobus: Cecidomyiidae, 395 description of the female and a Midwestern record for a long Chinches perceived as Eastern in especies, fluctuación poblacional y distribution, 151 enemigos naturales de asociadas a nogal pecanero, Cryptolaemus montrouzieri 305 potencial de depredacion de hacia Planococcus citri, 179 Chironomidae, 245 Culicidae, 167, 477 Chlorpyrifos, 227 Curculio caryae Cicadellidae, 99 efficacy of entomopathogenic fungi in suppressing in commercial Cicer arietinum, 61 pecan orchards, 111 oviposition characteristics, 447 Chrysomelidae, 289 Curculionidae, 31, 111, 159, 447 Cicadellidae, 457 Cycas revoluta, 489 Citrus, 1 Cydia caryana, 227

494 Cydia pomonella Eurytomidae, 329, 347, 359 protección de frutos de manzano con el virus de la granulosis de, Euschistus 331 visual guide for identification in Central Texas, 485 Dalbulus maidis newly discovered parasitoid of in Euschistus crassus, 485 California, 99 Euschistus ictericus, 485 Diaphorina citri, 1 Euschistus obscurus, 485 Diatraea considerata impact of sugarcane burning on Euschistus quadrator, 485 the stalkborer and its parasitoid Macrocentrus prolificus in Euschistus servus, 485 western Mexico, 213 Euschistus tristigmus, 485 Diabrotica virgifera zeae DNA isolation from and Diabrotica Flower flies virgifera virgifera by a CTAB sympatry of Milesia scutellata and simplified procedure, 289 Milesia virginiensis at their western range limits in North Diabrotica virgifera virgifera America, and the previous DNA Isolation from Diabrotica unknown juvenile stages of M. virgifera zeae and by a CTAB scutellata, 141 simplified procedure, 289 Frankliniella occidentalis, 417 Diaspididae, 359, 489 Garbanzo, 61 Diuraphis noxia, 121 Giant reed, 329, 347, 359, 377 Dryinidae, 99 Glassy-winged sharpshooter Encarsia pergandiella, 431 phylogenetic analysis of heat shock proteins in Homalodisca Encarsia sophia vitripennis, 457 new state record for the silverleaf whitefly parasitoid in Texas, 327 Glycine max, 485

Encyrtidae, 489 Gonatopus agropyrus a newly discovered parasitoid of Eretmocerus mundus, 431 the corn leafhopper, Dalbulus maidis, in California, 99 Erythroxylum havanense, 75 Gossypium hirsutum, 31, 239, 417, Estigmene acrea 431, 485 evaluation of seedling transgenic cotton containing Bacillus Green bean, 431 thuringiensis toxins to saltmarsh caterpillar, 239

495 Greenbug Ixodidae, 273 airborne remote sensing to detect stress in wheat, 205 Kentucky bluegrass, 69

Greenhouse whitefly, 431 Largidae, 305

Gryllidae, 43 Leaffooted bugs species, seasonal occurrence, and Haematobia irritans, 263 natural enemies of stink bugs in pecans, 305 Heat shock proteins, 457 Lycosidae, 43 Heliothis virescens incidence on garbanzo varieties in Maconellicoccus hirsutus northwestern Mississippi, 61 isolation of microsatellites in a Mexican population and Hexaflumuron, 31 amplification in populations from different geographical Hibana futilis origins, 295 extreme susceptibility of spiderlings to selected insecticides in a Macrocentrus prolificus laboratory bioassay, 103 impact of sugarcane burning on the stalkborer Diatraea Hickory shuckworm considerata and its parasitoid in effect of two insecticides on and western Mexico, 213 predators of pecan pests, 227 Maize, 99, 159, 485 Homalodisca vitripennis phylogenetic analysis of heat shock Maize weevil, 159 proteins in glassy-winged sharpshooter, 457 Mexican corn rootworm, 289

Horn fly Milesia scutellata nutritional limitation on growth and sympatry of and Milesia virginiensis development of larvae, 263 flower flies at their western range limits in North America, Hylesia lineata and the previous unknown growth and survival of a tropical juvenile stages, 141 polyphagous caterpillar: effects of host and group size, 75 Milesia virginiensis sympatry of and Milesia scutellata Ipomoea batatas, 431 flower flies at their western range limits in North America, Iris yellow spot virus, 13 and the previously unknown juvenile stages of M. scutellata, Ixodes scapularis 141 population genetics and phylogeography of from canines Muscidae, 263, 469 and deer in Arkansas, 273 Noctuidae, 61

496 Onion, 13, 219, 417 Rhizaspidiotus donacis pre-release assessment of impact Onion thrips, 417 on Arundo donax by the effect of different nitrogen regimes candidate biological control on Thrips tabaci on onions, agents Tetramesa romana and Allium cepa, 219 under quarantine conditions, 359 Pecan, 111, 227, 305, 319, 447, 485 Rhyzobius lophanthae Pecan weevil fortuitous establishment of and efficacy of entomopathogenic fungi Aphytis lingnanesis in South in suppressing in commercial Texas on the cycad Aulacaspis pecan orchards, 111 scale, Aulacaspis yasumatsui, oviposition characteristics, 447 489

Pentatomidae, 305, 485 Russian wheat aphid seasonal presence on alternate Permethrin, 167 hosts in Colorado, 121 grain and vegetative biomass Phaseolus vulgaris, 431 reduction by in winter wheat, 131 Phleum pratense, 407 Saccharum, 213 Pieris rapae, 193 Sago palm, 489 Pink hibiscus mealybug, 295 Saltmarsh caterpillar Pistachio, 305 evaluation of seedling transgenic cotton containing Bacillus Pistacia vera, 305 thuringiensis toxins to Estigmene acrea, 239 Planococcus citri potencial de depredacion de Saturniidae, 75 Cryptolaemus montrouzieri hacia, 179 Schizaphis graminum, 205

Plutella xylostella, 193 Silverleaf whitefly new state record for the parasitoid Polypedilum (Polypedilum) newaygo, Encarsia sophia in Texas, 327 a new species of midge, 245 Sitophilus zeamais Pseudoccocidae, 179, 295 relationship between chemical and physical parameters of maize Psyllidae, 1 varieties and susceptibility, 159

Sorghum, 395

Sorghum bicolor, 395

497 Sorghum midge pre-release assessment of impact cost-benefit analysis of on Arundo donax by the Stenodiplosis sorghicola- candidate biological control resistant sorghum hybrid agents and Rhizaspidiotus research and development in donacis under quarantine Texas, 395 conditions, 359

Soybean, 485 Tetranychidae, 69

Spiderlings, 103 Thouinia paucidentata, 75

Stable fly Thripidae, 13, 219, 407, 417 new method for collecting clean pupae of known age, 469 Thrips straw mulch and reduced-risk Stenodiplosis sorghicola pesticide impacts on and Iris cost-benefit analysis of sorghum yellow spot virus on western- midge-resistant sorghum hybrid grown onions, 13 research and development in effectiveness of spring burning as Texas, 395 a physical management tactic in Phleum pratense, 407 Stink bugs, 485 on cotton in the Lower Rio Grande species, seasonal occurrence, and Valley of Texas: species natural enemies of and composition, seasonal leaffooted bugs in pecans, 305 abundance, damage, and control, 417 Stomoxys calcitrans, 469 Thrips tabaci, 13, 417 Straw mulch, 13 effect of different nitrogen regimes on onion thrips on onions, Sweet potato, 431 Allium cepa, 219

Sugarcane, 213 Timothy, 407

Sweetpotato whitefly, 431 Tipula effects of relative humidity on two autumnal crane flies from efficacy of BotaniGard™ on Michigan: comments on the nymphs on hibiscus in difficulties of identification and greenhouses, 189 a review of the subgenus Platytipula in the United States Syrphidae, 141 and Canada, 85

Tebufenozide, 227 Tipula longiventris note on the identification, 95 Tetramesa romana, 347 distribution and spread of an Tipula (Platytipula) spenceriana, 85 adventive population of the biological control agent in Tipula (Platytipula) ultima, 85 Austin, Texas, 329

498 Tipula (Yamatotipula) jacobus Trichogramma platneri, 319 description of the female and a Midwestern record for a crane Trichogramma pretiosum, 319 fly long perceived as Eastern in distribution, 151 Trichogrammatidae, 319

Tipulidae, 85, 95, 151 Trichoplusia ni, 193

Tobacco budworm, 61 Triticum aestivum, 43, 131, 205

Tortricidae, 227, 331 Western corn rootworm, 289

Trialeurodes abutilonea, 431 Western flower thrips, 417

Trialeurodes vaporariorum, 431 Wheat, 43, 131, 205

Trichogramma Whiteflies dispersal on pecan trees and its tritrophic interactions among host susceptibility to selective plants and parasitoids, 431 insecticides, 319 dispersión de en arboles de nogal White-tailed deer, 273 y susceptibilidad a insecticidas selectivos, 319 Yellowfever mosquito, 167

Trichogramma exiguum, 319 Zea mays, 99, 159, 485

499 VOL. 34, NO. 4 SOUTHWESTERN ENTOMOLOGIST DEC. 2009

AUTHOR INDEX TO VOLUME 34

Abel, Craig A., 61 Fernandez Salas, Ildefonso, 167 Adamczyk, J. J., 417 Fichtner, Scott M., 13 Almaraz-Abarca, Norma, 289 Fleenor, Scott B., 141 Almas, Lal K., 395 Flores, Adriana E., 167, 477 Álvarez-Zagoya, Rebeca, 289 Flores, Daniel, 1, 489 Anderson, Roger M., 485 French, B. Wade, 43 Ansley, Jim, 131 Gallegos-Morales, Gabriel, 331 Badii, Mohammad, 477 García-Gutiérrez, Cipriano, 193 Barajas Ontiveros, C. G., 319 García Nevárez, Gerardo, 319 Barragán-Valencia, Gabriela, 289 Gardner, Wayne A., 111 Bautista-Martínez, Néstor, 193 Gent, David H., 13 Blanco, Carlos A., 61, 227 Gilbert, Lawrence, 329 Behle, Robert W., 111 Giles, Kris, 205 Berkebile, Dennis R., 469 Godfrey, Larry D., 407 Bextine, Blake, 457 González Hernández, Alejandro, 305 Borboa-Flores, Jesús, 159 González-Maldonado, María Berenice, Burgos-Hernández, Armando, 159 193 Buschman, Lawrent L., 43 Goolsby, John A., 327, 329, 347, 359, Carlson, Jason, 489 377 Camper, Matt, 13 Greenberg, S. M., 417 Caro, A., 213 Greenberg, Shoil M., 431 Catana, Vasile, 205 Guerrero-Rodríguez, Eugenio, 331 Chaírez-Hernández, Isaías, 193 Hail, Daymon, 457 Chávez Sánchez, N., 319 Haines, R. D., 99 Cortez-Rocha, Mario Onofre, 159 Hall, David G., 1 Cottrell, Ted E., 111 Hammon, Robert, 13 Cranshaw, Whitney S., 13, 69 Harris, Marvin K., 111 Currie, Randall S., 43 Hoffmann, Wesley C., 31 Damte, Tebkew, 395 Hunter, Wayne, 457 Davis, Holly N., 43 Jenkins, David A., 1 de Jesús de Luna-Santillana, Erick, Jones, Doug, 205 179 Jones, Walker A., 431 Delgado-Alvarado, Amanda Elí, 289 Kerns, D. L., 239 Droleskey, Robert E., 485 Kerzicnik, Lauren M., 121 Durán-Martínez, Erika Patricia, 179 Kesey, B. J., 239 El-Heneidy, A. H., 219 Kramer, Karen, 69 Ellington, J., 219 Lacewell, Ronald D., 377 Elliott, Norman, 131, 205 Latheef, Mohamed A., 31 Elzen, G. W., 103 Leos-Martínez, Josue, 159 Esquivel, Jesus F., 485 Liu, Tong-Xian, 189, 417, 431 Evans, Gregory A., 327 López, Juan D., Jr., 31, 61 Falk, Jay, 329 Mahaffey, Linda, 13

501 Malik, M. F., 219 Rosas-García, Ninfa M., 179, 295 Marcum, Daniel D., 407 Sánchez-Mariñez, Reyna Isabel, 159 Martínez-Montoya, Humberto, 295 Sánchez-Peña, Sergio R., 331 McCorkle, Dean A., 377 Sánchez-Pérez, Félix de J., 331 McMillan, Mark, 13 Sanchez Ramos, Francisco J., 167 Michels, G. J., Jr. 205 Sánchez-Valdez, Víctor M., 331 Michels, Jerry, Jr., 131 Sanderson, R., 219 Mirik, Mustafa, 131, 205 Schreiber, Henry, IV, 457 Moran, Patrick, 329, 347 Schwartz, Howard F., 13 Mulder, Phillip G., 447 Seawright, Emily K., 377 Nawaz, M., 219 Setamou, Mamoudou, 1 Newton, A. S., 99 Shapiro-Ilan, David I., 111 Otto, Kristen, 13 Smith, Michael W., 447 Peairs, Frank B., 121 Spencer, David, 359 Pendleton, Bonnie B., 395 Stansly, Philip A., 189 Pérez-Domínguez, Juan Francisco, Steelman, C. D., 273 289 Sturdivant, Allen W., 377 Pescador-Rubio, Alfonso, 75 Summers, C. G., 99 Pfannenstiel, R. S., 103 Szalanski, A. L., 273 Pfannenstiel, Robert S., 327 Taber, Stephen W., 95, 141, 151, 245 Phoofolo, Mpho, 205 Tarango Rivero, Socoro Héctor, 227, Ponce Garcia, Gustavo, 167, 477 305 Quiñones Pando, F. J., 319 Taylor, David B., 469 Quiñones-Pando, Francisco Javier, Temeyer, Kevin B., 263 227 Terán-Vargas, Antonio P., 61 Racelis, Alexis E., 347 Trout, R. T., 273 Randolph, Terri L., 121 Vejar-Cota, G., 213 Ree, Bill, 111 Villegas-Mendoza, Jesús Manuel, 179 Reisig, Dominic D., 407 Weiland, Aubrey A., 121 Reyes Solis, Guadalupe, 167 Weinhold, Anthony P., 469 Ríos-Soto, José Luis, 159 Whitehand, Linda, 359 Rios-Velasco, Claudio, 331 Wong-Corral, Francisco Javier, 159 Rister, M. Edward, 377 Yang, Chenghai, 377 Roberto, Mercado, 477 Yang, Zhiming, 205 Rodríguez-del-Bosque, L. A., 213

502 THE SOCIETY OF SOUTHWESTERN ENTOMOLOGISTS

Application for Membership

Name: ______Date______Mailing Address: ______City: ______State: ______Country: ______Zip Code: ______Telephone No.: ______Business Affiliation or Specialty: ______

Membership in The Society of Southwestern Entomologists includes a subscription to the Southwestern Entomologist.

Return this application with a check in the amount of $20.00 (U.S. Dollars, drawn on an American Bank) to cover membership dues for one year, to:

Allen E. Knutson, Secretary-Treasurer Southwestern Entomological Society 17360 Coit Road Dallas, Texas 75252-6599

Membership in The Society of Southwestern Entomologists is open to all persons interested in entomology.

503 VOL. 34, NO. 4 SOUTHWESTERN ENTOMOLOGIST DEC. 2009

505

506

View publication stats