Hoddle & Van Driesche: Bemisia control with Encarsia formosa 1

Transcription

1 Hoddle & Van Driesche: Bemisia control with Encarsia formosa 1 EVALUATION OF ENCARSIA FORMOSA (HYMENOPTERA: APHELINIDAE) TO CONTROL BEMISIA ARGENTIFOLII (HOMOPTERA: ALEYRODIDAE) ON POINSETTIA (EUPHORBIA PULCHERRIMA): A LIFETABLE ANALYSIS MARK S. HODDLE AND ROY VAN DRIESCHE Dept. of Entomology, University of Massachusetts, Amherst, MA, ABSTRACT Weekly releases of the parasitoid Encarsia formosa Gahan failed to control a low density population (initially, 0.51 nymphs and pupae per plant) of the whitefly Bemisia argentifolii Bellows & Perring on greenhouse grown poinsettia plants in Massachusetts when released at the rate of 4-7 adult females per plant. A lifetable constructed for uncaged B. argentifolii in the presence of E. formosa indicated that survivorship from the first/second instar to adult emergence was 14%. In contrast, in a lifetable constructed for B. argentifolii on caged poinsettia from which E. formosa was excluded, survivorship was 67%. Release of E. formosa reduced the number of insecticide applications on poinsettia by 75%, but the cost of using E. formosa (on a per m 2 basis) was 9.5 times that of insecticides alone. Key Words: Augmentation, biological control, integrated pest management, greenhouse, whitefly, parasitoid. RESUMEN Las liberaciones semanales de 4-7 hembras adultas del parasitoide Encarsia formosa Gahan por planta no pudieron controlar una baja densidad poblacional (inicialente, 0.51 ninfas y pupas por planta) de la mosca blanca Bemisia argentifolii Bellows & Perring en plantas de flor de pascua en Massachusetts. Una tabla de vida construida para B. argentifolii no mantenida en jaulas y en presencia de E. fomosa indicó que la sobrevivencia desde el primero/segundo instar hasta la emergencia del adulto fue del 14%. En contraste, en una tabla de vida construida para B. argentifolii sobre plantas de flor de pascua mantenidas en jaulas de las cuales E. formosa fue excluida, la supervivencia fue del 67%. La liberación de E. formosa redujo el número de aplicaciones de insecticida en las plantas de flor de pascua en un 75%, pero el costo del uso de E. formosa (por metro cuadrado) fue 9.5 veces el del insecticida solo. The primary phytophagous pest affecting poinsettia (Euphorbia pulcherrima Willd. ex Klotzsch) is the silverleaf whitefly, Bemisia argentifolii Bellows & Perring [= the B strain of Bemisia tabaci (Gennadius)] (Homoptera: Aleyrodidae) (Perring et al. 1993, Bellows et al. 1994). In Massachusetts, 200 greenhouses produce approximately one million individually potted poinsettias annually, which have a wholesale value of $5 million. Poinsettia is the highest ranked ornamental sold in the last quarter of the year in this state [unpublished University of Massachusetts Integrated Pest Management Program Annual Report (1994)]. Because poinsettia is grown for its aesthetic qualities, growers have an extremely low tolerance for the presence of whitefly nymphs, adults, or honeydew. Insecticides are often applied on a calendar schedule, with applications being made every 3-5 days to reduce B. argentifolii populations to

2 2 Florida Entomologist 79(1) March, 1996 acceptably low levels. Adverse effects of such intensive pesticide use against this pest have been documented (Parrella et al. 1992; Heinz & Parrella 1994 a,b). The University of Massachusetts Cooperative Extension System for floricultural crops has initiated an integrated pest management program (IPM) to design a more effective management program for B. argentifolii on poinsettia. In the first phase of this program, scouts were employed to recommend pesticide application when susceptible whitefly stages exceeded tolerable densities. Targeted spraying in this manner has reduced insecticide use by 18-50% with most poinsettia growers in the program [unpublished University of Massachusetts Integrated Pest Management Program Annual Report (1994)]. The second objective of the IPM program is to reduce insecticide use even further with biological control agents (in particular parasitic wasps) for suppression of B. argentifolii. One of the parasitoids that has been considered for use in the IPM program is Encarsia formosa Gahan. Encarsia formosa is a commercially available parasitoid that is used to control the greenhouse whitefly, Trialeurodes vaporariorum (Westwood), a serious pest of greenhouse vegetable crops. Encarsia formosa is used worldwide to control T. vaporariorum on vegetable crops in greenhouses (van Lenteren & Woets 1988). Encarsia formosa tested under laboratory conditions with B. tabaci (strain B) on poinsettia developed more slowly, exhibited higher mortality with reduced longevity, and was less fecund than parasitoids that developed on T. vaporariorum (Boisclair et al. 1990, Szabo et al. 1993). Several studies on the use of E. formosa to control B. tabaci on poinsettia in greenhouses suggest this parasitoid is effective, contrary to the laboratory findings [the strain was not identified in these studies and others, but is assumed to be strain B which became problematic on poinsettia in Europe after 1987 (see Boisclair et al. 1990, Szabo et al. 1993)]. Investigations by Benuzzi et al. (1990) in Italy, Albert & Schneller (1989) and Albert & Sautter (1989) in Germany, and Stenseth (1993) in Norway concluded that E. formosa successfully suppresses B. tabaci (unidentified strain) on poinsettia grown for the Christmas market when T. vaporariorum is present. Work by Parrella et al. (1991) in California, USA, reports that E. formosa is an ineffective control agent for populations of B. argentifolii on poinsettia plants grown for cutting production in the spring. Because poinsettia growing conditions in the northeastern USA in the fall are more similar to those in Europe in the fall than to spring growing conditions in California, there was a need to evaluate the ability of E. formosa to control B. argentifolii populations on poinsettia in Massachusetts. To measure the efficacy of E. formosa, we constructed lifetables for B. argentifolii in both the presence and absence of the parasitoid. MATERIALS AND METHODS Greenhouses and Cultivars Two greenhouses at one commercial poinsettia producer in Massachusetts were monitored to determine whether insecticides or E. formosa provided better control of B. argentifolii. One greenhouse received E. formosa as a control measure and is designated here as the biological control greenhouse. The second greenhouse was managed using synthetic pesticides and is designated as the insecticide greenhouse. Each was a 170 m 2 A frame greenhouse (glass construction) with three benches in Cambridge, Massachusetts. The two side benches (1 m 27 m) each held 156 pots (18 cm diam) with 5 single-stem unpinched plants per pot. The middle bench (2 m 26 m) held 264 pots (19 cm diam) with 6 single-stem unpinched plants per pot for a total of

3 Hoddle & Van Driesche: Bemisia control with Encarsia formosa pots and 3144 plants per greenhouse. The poinsettia cultivars were white and marble Angelika ; red, pink and white Celebrate 2 ; and pink Gutbier V-14. The study started immediately after both greenhouses were filled with potted cuttings in August, Some plants were removed during the test from both houses to satisfy spacing requirements. Population Density Estimation and Lifetable Construction To estimate whitefly population densities, six leaves (2 from the bottom of the plant, 2 middle, and 2 top) of 30 plants in each greenhouse were inspected weekly for B. argentifolii. The number of eggs, first and second instars, third instars, fourth instars, red eyed pupae, and adults were recorded. T. vaporariorum was not observed in either greenhouse. Three treatments were established in the biological control greenhouse: uncaged plants (Treatment 1), cages without E. formosa (Treatment 2), and cages with E. formosa (Treatment 3). Treatment 2 acted as the control, and Treatment 3 was a check for a cage effect on whitefly development in the presence of the parasitoid. In addition to estimating whitefly densities on randomly selected leaves in Treatment 1, the fate of marked cohorts of nymphs was determined. Cohorts were established by tagging and numbering naturally-infested plants bearing first or second instar B. argentifolii. Numbers were written on tagged leaves with an indelible marker beside young nymphs. Numbered nymphs were examined weekly, and their developmental stage recorded. Young nymphs (approximately 1-30 nymphs per leaf) found on 1-3 leaves of each of 3-5 plants were recruited every week for the lifetable study. Observations were continued until the nymphs had either died of unknown causes, disappeared, been parasitized, or emerged as adult whiteflies. Parasitism was noted when the whitefly pupa turned brown or a parasite had emerged. The recorded fates of all nymphs (204) in Treatment 1 were used to create a partial lifetable for B. argentifolii. For Treatments 2 and 3, nine pots (19 cm diam) with single-stem unpinched poinsettia plants (cultivars used; white Angelica, red and pink Celebrate 2 ) were selected at random and enclosed by fine mesh bags (28 cm 28 cm 36 cm; mesh threads per cm 2 ). Each bag was supported by four 50 cm stakes that were driven into the potting medium. A rubber band was used to seal the bottom of the bag against the exterior of the pot. One male and one female adult B. argentifolii were released into each bag. In Treatment 2, the resulting whitefly population was allowed to develop on poinsettia in the absence of E. formosa. In Treatment 3, three E. formosa were introduced into each of the nine cages weekly. For Treatments 2 and 3, whitefly population density estimates were made, and cohorts of nymphs established in the same manner as Treatment 1. Numbered nymphs on tagged leaves were observed weekly for survivorship and parasitism. Mortality data for numbered nymphs in Treatments 2 and 3 were used to create partial lifetables for B. argentifolii populations on caged poinsettia plants in the presence and absence of E. formosa. Calculating Marginal Probabilities of Mortality and k-values Marginal attack rates were calculated to separate mortality from each observed source (unknown death, disappearance, and parasitism). The marginal probability of attack is the number of pests that would be attacked by an agent in the absence of all other contemporaneous mortality agents. It is the net probability of dying (as opposed to the crude probability of dying, which is the apparent mortality calculated from

4 4 Florida Entomologist 79(1) March, 1996 numbers observed to die from a cause) (Royama 1981, Elkinton et al. 1992). The marginal probability of attack was calculated for each factor (Table 1) as: m i = 1-(1-d) di/d where m i = marginal probability of attack from the ith cause, d i = death rate from the ith cause and d= death rate from all causes combined (Elkinton et al. 1992). Killing powers or k-values (the negative logarithm of the proportion surviving in each stage) for each mortality factor were determined as: k i = -log 10 (1-marginal probability of attack by the ith cause), where k i = the k-value for the ith cause of mortality. Wasp Releases, Percent Emergence and Emergence Pattern Parasitoid releases in both the cages and open greenhouse began immediately after the biological control greenhouse was filled with poinsettias. Bunting Biological North America supplied release cards, each bearing 100 parasitized T. vaporariorum pupae. The number of cards put into the biological control greenhouse each week ranged from 140 to 268. Cards were hung on strings stretched between the pots and tied at the same height as the pot rims. In this position, wasps emerged below the foliage and were assumed to move upwards through the canopy searching for B. argentifolii nymphs. Every week, all cards were removed from the biological control greenhouse before new cards were put out. Ten cards were randomly selected from those recovered and soaked in water and detergent for 30 min in the laboratory. Parasitized greenhouse whitefly and exuviae were rubbed off the card with a size 2 insect pin, and the mean number of parasitized greenhouse whitefly per card and the percent emergence of wasps were determined for each weekly release. These values were used to calculate the mean number of wasps released per plant per week. On two occasions, the emergence pattern of the parasitoid was determined by counting the number of wasps that emerged from the cards each day in the laboratory. Cost Analysis The cost of biological control vs. the cost of insecticides was determined by analyzing insecticide application records for both the insecticide and biological control greenhouses. The price of purchasing the required number of parasitoids was based on an averaged estimate from suppliers of beneficial insects. Labor costs associated with releasing parasitoids and applying insecticides were not included in the analysis. Sales Inspection At week 16 of the growing period, the finished plants were shipped to retailers. To determine the final whitefly density, six leaves on 15 plants from both the biological control and the insecticide greenhouses were inspected. The number of live nymphs and pupae on each leaf was recorded.

5 Hoddle & Van Driesche: Bemisia control with Encarsia formosa 5 TABLE 1. PARTIAL LIFETABLES FOR B. ARGENTIFOLII FOR EACH OF THE THREE TREATMENTS IN THE BIOLOGICAL CONTROL GREENHOUSE. Stage l x 1 d x f dx q x Marginal Probability of Attack k-value T 1 2 T 2 T 3 T 1 T 2 T 3 T 1 T 2 T 3 T 1 T 2 T 3 T 1 T 2 T 3 T 1 T 2 T 3 3 I 1 /I unknown death: disappeared: I unknown death: disappeared: I unknown death: disappeared: P unknown death: disappeared: parasitized: A lx = Number entering the stage. d x = Number dying in the stage. f dx = Factor responsible for observed mortality. q x = Proportion dying in that stage (see Southwood 1978 for more information on parts of lifetables). 2 T1 = Lifetable for B. argentifolii on naturally infested poinsettia plants in the biological control greenhouse (Treatment one). Mortality from 1st/2nd instar to adult was 85.78%. T 2 = Lifetable for B. argentifolii on poinsettia plants inside mesh bags which excluded E. formosa in the biological control house (Treatment 2). Mortality from 1st/2nd instar to adult was 33.33%. T 3 = Lifetable for B. argentifolii on poinsettia plants inside mesh bags into which three E. formosa were introduced each week in the biological control greenhouse (Treatment 3). Mortality from 1st/2nd instar to adult was 97.90%. 3 I1 /I 2 = B. argentifolii first and second instar, I 3 = third instar, I 4 = fourth instar, P= pupae, A= adults.

6 6 Florida Entomologist 79(1) March, 1996 RESULTS Population Density Estimates and Lifetable Construction Whitefly densities on caged poinsettia plants onto which one female whitefly had been introduced at the start of the experiment (Treatment 2, Fig. 1A) increased steadily, reaching 246 nymphs and pupae per plant by week 10. In contrast, whitefly densities on caged plants inoculated with three adult E. formosa per week (Treatment 3) were substantially lower by week six (Fig. 1A) and averaged only 20 live nymphs and pupae per plant in week ten, 8% of the recorded density on the control plants (Treatment 2) (Fig. 1A). Whitefly populations on uncaged plants in the biological control greenhouse remained below six nymphs and pupae per plant until week 7 of the experiment. The number of immature whiteflies on these plants increased to approximately 39 nymphs and pupae per plant by week 10 (Fig. 1B). At that time parasitoid releases were terminated, and two insecticide applications were made. In contrast, whitefly densities in the insecticide greenhouse increased to 32 nymphs and pupae per plant by week 4, but declined to low densities (< 5 nymphs and pupae per plant) by week six (Fig. 1B). This low level of infestation was maintained in the insecticide greenhouse through week 10 because of regular insecticide applications (see Fig. 1C). The percentage of plants that were infested reached 100% in weeks 4 and 6 for the insecticide greenhouse and the biological control greenhouse respectively. The percentage of B. argentifolii-infested plants steadily declined in the insecticide greenhouse after week six. This trend was not observed in the biological control greenhouse (Fig. 1C). In the biological control greenhouse, 86% of the whitefly nymphs that were followed individually died prior to adult emergence (see footnote Table 1); however, nymphal densities ultimately exceeded the grower s damage threshold for the crop. In Treatment 2, in which a B. argentifolii population developed on caged poinsettia plants in the absence of E. formosa, 33% of the nymphs died of natural causes (see footnote Table 1). In Treatment 3, where three E. formosa per week were released into identical cages, the level of nymphal mortality was 98% (see footnote Table 1). Marginal Probabilities of Mortality and k-values Mortality from three factors (parasitism, unknown death, and disappearance) occurred contemporaneously. The marginal probability of attack and k-values for these factors in Treatments 1-3 are presented in Table 1. Treatment 3 consistently exhibited the highest levels of mortality for each of the immature lifestages (Table 1). The highest observed k-values were those for pupae which exhibited high levels of parasitism in Treatments 1 and 3 (Table 1). Wasp Releases, Percent Emergence and Emergence Pattern Number of parasitized pupae per card, percentage of wasps emerging, number of release cards put into the greenhouse each week, number of wasps released per plant, and number of wasps released per m 2 are shown in Table 2. Two shipments of E. formosa exhibited different emergence patterns in the laboratory. Group one exhibited a unimodal emergence pattern with wasp numbers peaking 5 days after receipt (mean daily maximum temperature= 24.7 C ± 0.7; mean daily minimum temperature= 23 C ±0.7) (Fig. 2A). Group two exhibited a bimodal emergence pattern, with wasp num-

7 Hoddle & Van Driesche: Bemisia control with Encarsia formosa 7 Figure 1. (A) The mean number of Bemisia argentifolii nymphs and pupae (± S.E.M.) on caged poinsettia plants in the absence (Treatment 2) and presence of E. formosa (Treatment 3). (B) The mean number of Bemisia argentifolii nymphs and pupae (± S.E.M.) per plant in the insecticide house and the biological control house (Treatment 1). (C) Percentage of plants infested with adult or immature stages of Bemisia argentifolii in the biological control and insecticide greenhouse; asterisk indicates dates of insecticide applications in the insecticide greenhouse.

8 8 Florida Entomologist 79(1) March, 1996 TABLE 2. WEEKLY PERCENTAGE EMERGENCE OF E. FORMOSA, MEAN NUMBER OF PARASITIZED PUPAE PER CARD, TOTAL NUMBER OF CARDS WITH PARASITIZED T. VAPORARIORUM PUT INTO GREENHOUSE EACH WEEK, NUMBER OF ADULT PARASITOIDS EMERGING PER PLANT AND PER M 2, AND NUMBER OF PLANTS IN THE BIOLOGICAL CONTROL GREENHOUSE FOR EACH WEEK OF THE TRIAL. Week# Mean No. Parasitized Pupae/Card % Wasp Emergence No. Cards/ Week Wasp No./ Plant/Week No. Wasps/ No. Plants/ M 2 Greenhouse % % % % % % % % %

9 Hoddle & Van Driesche: Bemisia control with Encarsia formosa 9 Figure 2. (A) Daily emergence of Encarsia formosa in the laboratory and cumulative percent (B) emergence of Encarsia formosa in the laboratory. bers peaking on days 2 and 5 after receipt (mean daily maximum temperature= 24.4 C ± 0.7; mean daily minimum temperature= 23.3 C ± 0.3) (Fig. 2A). Over 97% of E. formosa had emerged after 7 days (Fig. 2B). In the biological control greenhouse,

10 10 Florida Entomologist 79(1) March, 1996 the mean daily maximum temperature was 22.8 C ± 0.6; mean daily minimum temperature= 17 C ± 0.5. Cost Analysis The costs of controlling B. argentifolii with E. formosa or insecticides are presented in Table 3. Weekly releases of 4-7 E. formosa per plant for nine weeks, followed by two insecticide applications, were 9.5 times more expensive than using insecticides alone. Sales Inspection At week 16 of the growing period, the numbers of live nymphs on plants from both the biological control and insecticide greenhouse were low at the time of shipment. The mean numbers of nymphs per leaf were 0.01 ± 0.11 in the insecticide greenhouse and 0.02 ± 0.21 in the biological control greenhouse. There was no obvious difference in foliage quality between the biological control and the insecticide greenhouses. DISCUSSION Release of high numbers of E. formosa (4-7 wasps per plant per week) did not successfully control a population of B. argentifolii on poinsettia, even though parasitoid releases were initiated at the beginning of the growing period when the infestation of B. argentifolii nymphs was very low (0.09 per leaf or 0.51 per plant). The high mortality (98%) observed in Treatment 3 (caged plants with wasps) (Table 1) may have occurred because cages prevented wasps from abandoning plants, thereby increasing residence and searching time. The differences between the observed k-values (Table 1) of Treatment 2 (caged plants with no wasps) and those of Treatment 1 (uncaged plants) and Treatment 3 (caged plants with wasps), with respect to unknown death for all nymphal stages, may be due to aborted parasitism (in older nymphs) or host feeding by E. formosa. In addition, some whitefly death observed in Treatment 3 may have resulted from superparasitism. Another problem inherent with the use of cages to enclose single plants is the need to introduce adult whiteflies. An introduction rate of just 1 female and 1 male per TABLE 3. COMPARISON OF THE COSTS OF WHITEFLY CONTROL IN THE INSECTICIDE GREENHOUSE AND THE BIOLOGICAL CONTROL GREENHOUSE 1. Insecticide House Biological Control House Total cost of sprays $ $43.28 Total cost of E. formosa NA $ Total treatment cost $ $ Treatment cost per plant $0.09 $1.02 Cost m 2 $1.58 $ Insecticide costs are based on 1993 catalogue prices. The E. formosa price was based on a rate of $12.00 per 1000 wasps. Insecticide costs per m 2 in Massachusetts range from $ $2.26 in poinsettia crops [unpublished University of Massachusetts Integrated Pest Management Program Annual Report (1994)].

11 Hoddle & Van Driesche: Bemisia control with Encarsia formosa 11 plant to establish an experimental population resulted in the caged plants having an initial adult whitefly density nine times that of the biological control greenhouse; before E. formosa was released, the mean number of adult whiteflies in the biological control house was 0.22 ± 0.45 adults per plant. Consequently, this may have exaggerated the observed densities in the control cages (Treatment 2). However, this would not affect comparisons between Treatments 2 and 3, as both were inoculated with equal numbers of whiteflies. The limited control provided by E. formosa was 9.5 times more expensive than insecticides on a per m 2 basis (Table 3). Albert & Sautter (1989) achieved cheaper control of B. tabaci (the strain of whitefly was not identified) on poinsettia with E. formosa than with chemicals, but T. vaporariorum, a preferred host for E. formosa, was present in the crop. This may have affected parasitoid population levels in the greenhouse. At sale of the crop, whitefly densities in the chemical greenhouse and biological control greenhouse were 0.01 ± 0.11 and 0.02 ± 0.21 nymphs per plant respectively. Plants from both houses were marketed successfully. In the biological control greenhouse, insecticide use was reduced by 75% by release of wasps (from eight to two insecticide applications), but pest control costs were increased from $0.09 to $1.02 per plant. Several agronomic practices associated with poinsettia production should favor biological control of B. argentifolii in Massachusetts. First, poinsettias are grown from June (if cuttings are being produced) until December and entire greenhouses are devoted to poinsettia production. Monocultural production simplifies pest management because B. argentifolii is the only arthropod causing foliar damage, and this nullifies incompatible management programs for pest complexes (Heinz & Parrella 1994a, Parrella et al. 1991). Second, the majority of growers in Massachusetts purchase poinsettia cuttings in July or August from suppliers who typically sell plants with very low densities of adult and immature B. argentifolii. Therefore, initial B argentifolii densities are sufficiently low that a favorable ratio of parasitoids to whiteflies could be established. Third, fungal diseases of poinsettia foliage can be controlled with fungicides that are compatible with biological control agents (Parrella et al. 1991). Fourth, winters in the northeastern USA prevent continual immigration of B. argentifolii into greenhouses from outdoor host plants, and growers need only manage the whitefly population that had established in the greenhouse before the onset of cold weather. In view of these considerations, two major constraints to successful biological control of B. argentifolii on commercially grown poinsettia in Massachusetts are: (1) the commercial non-availability of an effective natural enemy for B. argentifolii, and (2) lack of information as to which release strategies would maximize the impact of a suitable control agent. A suitable release program should span the entire window of pest susceptibility to ensure maximum mortality by host feeding and parasitism. Variable release rates and timings may be necessary to achieve this objective as foliage density, pest density, and levels of parasitism change over the season. Variable release rates of a parasitoid may result in higher levels of parasitoid recycling through reproduction. Natural reproduction in the greenhouse could augment weekly releases and reduce the cost of parasitoid releases. Complete reliance on biological control may not be feasible for this ornamental crop, but incorporation of natural enemies in the context of an IPM program for poinsettia should be an attainable goal. Colonies of aphelinid parasitoids that attack B. argentifolii exist at several research institutes in the United States. Further work in greenhouses is needed to evaluate the efficacy of these parasitoids for B. argentifolii control on poinsettia.

12 12 Florida Entomologist 79(1) March, 1996 ACKNOWLEDGMENTS We thank J. Mason for her assistance in the field and E. Norberg for allowing this trial to be run on his premises. The cooperation of Dr. R. Greatrex CIBA-Bunting Ltd, France, and D. Cahn of Bunting Biological North America is gratefully acknowledged. This research was supported by grants from the Massachusetts IPM Program, and USDA/NRICGP grant # REFERENCES CITED ALBERT, R., AND H. SAUTTER Parasitoids protect Christmas stars from whiteflies. Deutscher Gartenbau 43: ALBERT, R., AND H. SCHNELLER Successful biological control in ornamental plants -1. Poinsettias (Euphorbia pulcherrima Willd. ex Klotzsch). Gesunde Pflanzen 41: BELLOWS, T. S., T. M. PERRING, R. J. GILL, AND D. H. HEADRICK Description of a species of Bemisia (Homoptera: Aleyrodidae). Ann. Entomol. Soc. America. 87: BENUZZI, M., G. NICOLI, AND G. MANZAROLI Biological control of Bemisia tabaci (Genn.) and Trialeurodes vaporariorum (Westw.) by Encarsia formosa Gahan on poinsettia. SROP/WPRS Bull 13: BOISCLAIR, J., G. J. BRUEREN, AND J.C. VAN LENTEREN Can Bemisia tabaci be controlled with Encarsia formosa? SROP/WPRS Bull. 13: ELKINTON, J. S., J. P. BUONACCORSI, T. S. BELLOWS, AND R. G. VAN DRIESCHE Marginal attack rate, k-values and density dependence in the analysis of contemporaneous mortality factors. Res. Pop. Ecol. 34: HEINZ, K. M., AND M. P. PARRELLA. 1994a. For silverleaf whiteflies on poinsettias... bios are in control. Greenhouse Grower 12: HEINZ, K. M., AND M. P. PARRELLA. 1994b. Poinsettia (Euphorbia pulcherrima Willd. ex Koltz.) cultivar mediated differences in performance of five natural enemies of Bemisia argentifolii Bellows and Perring, n. sp. (Homoptera: Aleyrodidae). Biological Control 4: PARRELLA, M. P., T. D. PAINE, J. A. BETHKE, K. L. ROBB, AND J. HALL Evaluation of Encarsia formosa for biological control of sweetpotato whitefly (Homoptera: Aleyrodidae) on poinsettia. Environ. Entomol. 20: PARRELLA, M. P., T. S. BELLOWS, R. J. GILL, J. K. BROWN, AND K. M. HEINZ Sweetpotato whitefly: prospects for biological control. California Agriculture 46: PERRING, T. M., A. D. COOPER, R. J. RODRIGUEZ, C. A. FARRAR, AND T. S. BELLOWS Identification of a whitefly species by genomic and behavioral studies. Science 259: ROYAMA, T Evaluation of mortality factors in insect lifetable analysis. Ecol. Monogr. 5: SOUTHWOOD, T. R. E Ecological methods with particular reference to the study of insect populations. Chapman and Hall London. 524 pp. STENSETH, C Biological control of cotton whitefly Bemisia tabaci (Genn.) (Homoptera: Aleyrodidae) by Encarsia formosa (Hymenoptera: Aphelinidae) on Euphorbia pulcherrima and Hypoestes phyllostachya. WPRS/SROP Bull. 16: SZABO, P., J. C. VAN LENTEREN, AND P. W. T. HUISMAN Development time, survival and fecundity of Encarsia formosa on Bemisia tabaci and Trialeurodes vaporariorum. SROP/WPRS Bull. 16: VAN LENTEREN, J. C., AND J. WOETS Biological and integrated control in greenhouses. Ann. Rev. Entomol. 33:

13 Castineiras et al.: Temperature response of Ceranisus menes 13 TEMPERATURE RESPONSE OF TWO STRAINS OF CERANISUS MENES (HYMENOPTERA: EULOPHIDAE) REARED ON THRIPS PALMI (THYSANOPTERA: THRIPIDAE) ANTONIO CASTINEIRAS, RICHARD M. BARANOWSKI AND HOLLY GLENN University of Florida, IFAS, Tropical Research and Education Center, SW 280th Street, Homestead, FL ABSTRACT The development response to temperature of a Japanese uniparental strain and a Thai biparental strain of Ceranisus menes (Walker) (Hymenoptera: Eulophidae) was studied. The parasitoids were reared on first instar Thrips palmi larvae in incubators at constant temperatures of 21, 23, 25, 27 and 29 C. Total developmental time decreased with the increase of temperature from 35.1 to 21.9 days in females and from 33.4 to 18.8 days in males. Lowest mortality (12%) was recorded in both strains at 23 C and highest (95%) in the Japanese strain at 29 C. Seventy-three percent of the Thai parasitoids survived at 29 C, but 39% mortality was observed at 21 C. Percent parasitism ranged from 23.8 to 28.9% at C, but decreased to 11.5% at 21 C. The sex ratio (male:female) was not affected by temperature and averaged 1:1.9. A thermal constant of 500 degree-days and a developmental threshold of 8 C (from egg to adult emergence) were obtained for both Japanese and Thai females. For the Thai males, the thermal constant was degree-days and the minimum threshold was 13.7 C. Key Words: Ceranisus menes, development, parasitism, mortality, Thrips palmi, biological control. RESUMEN Fue estudiada la respuesta a la temperatura en una cepa uniparental japonesa y en una cepa biparental tailandesa de Ceranisus menes (Walker) (Hymenoptera: Eulophidae), un parasitoide de Thrips palmi Karny (Thysanoptera: Thripidae). Los parasitoides fueron criados en incubadoras a temperatura constante de 21, 23, 25, 27 y 29 C. La duración del desarrollo disminuyó con la temperatura de 35.1 a 21.9 días en las hembras, y de 33.4 a 18.8 días en los machos. La más baja mortalidad (12%) fue registrada en ambas cepas a 23 C y la más alta (95%) en la cepa japonesa a 29 C. El 73% de los parasitoides tailandeses sobrevivió a 29 C, pero a 21 C fue observado un 39% de mortalidad. El porcentaje de parasitismo estuvo en el rango de 23.8 a 28.9%, pero bajó hasta el 11.5% a 21 C. La relación sexual (macho:hembra) no fue afectada por la temperatura y promedió 1:1.9. Se obtuvo una constante térmica de 500 gradosdías y un umbral de desarrollo de 8 C (desde el huevo hasta la emergencia del adulto) en las hembras de ambas cepas. En los machos tailandeses la constante térmica fue de grados-días y el umbral mínimo de 13.7 C. Since its first detection in 1990, Thrips palmi Karny has become an important pest of eggplant, potato, beans, peppers and cucurbits in South Florida (Seal et al. 1992). The eulophid wasp Ceranisus menes (Walker) was selected as a possible biological control agent for suppression of T. palmi. In , C. menes strains were introduced from Japan and Thailand into Dade County, Florida.

14 14 Florida Entomologist 79(1) March, 1996 Ceranisus menes is an endoparasitoid of Thysanoptera with worldwide distribution (Loomans 1995). Life cycle length of C. menes changes with temperature and host (Loomans 1995). The duration of the egg-larval stage of C. menes is 14 days and adult emergence occurs at 37 days at 20 C when reared on T. palmi (Tagashira 1992). When reared on Thrips tabaci Lind. at C, the egg-larval stage takes 11 days and adults emerge at 18 days (Sakimura 1937). Under laboratory conditions, 80% parasitism by C. menes of Frankliniella intonsa (Trybom) has been obtained (Murai 1990). A maximum of 75% field parasitism of T. palmi was reported in Japan (Hirose et al. 1992). Females are more abundant in nature, but sex ratio may change under laboratory rearing conditions due to arrhenotokous parthenogenesis. Uniparental populations have been reported from several countries (Loomans 1995). The existence of biotypes of C. menes with differences in color patterns, temperature response, host acceptance, developmental time, sex ratio and reproduction has been documented (Loomans, 1995). Selection of biotypes adapted to particular conditions of host and climate can improve the efficacy of biological control programs. The objective of this study was to evaluate the effect of temperature on developmental time, percent parasitism and sex ratio of two biotypes of C. menes using T. palmi as the host. Data on temperature response will be used in laboratory rearing of the parasitoid and in selection of strains for the biological control of T. palmi in the field. MATERIALS AND METHODS Two strains of C. menes were studied, a uniparental strain collected in Kyushu, Japan in 1990 (Tagashira 1992) and a biparental strain collected in Chiang-Mai, Thailand, by R. M. Baranowski and F. D. Bennett (University of Florida) in January A laboratory reared colony of the uniparental strain was brought into Florida by Y. Hirose (Kyushu University, Japan) in February Groups of 20 first instar T. palmi were taken from a laboratory colony and placed on 25 mm eggplant leaf discs. The discs with the larvae were held with parafilm at the top of mm plastic tubes as described by Tagashira (1992). One parasitoid female was placed in every tube and the open end of the tube was sealed with parafilm. In the case of the Thai strain, as unmated females produce only males, pairs of C. menes were kept for 6 h at 25 C in daylight in mm vials to ensure mating before females were placed in the plastic tubes. Tubes were maintained in incubators at 21, 23, 25, 27 and 29 C (± 1 C), and a photoperiod of 12:12 h (L:D). After twenty-four hours, tubes were opened and female parasitoids were removed. Leaf discs with thrips were transferred from the tubes into 500 mm petri dishes having a layer of plaster of Paris in the bottom and covered with humid filter paper. Thrips larvae in the petri dishes were kept in incubators at the conditions described above. Last instar thrips were considered to be parasitized only when the parasitoid could be clearly recognized through the body wall. Parasitized larvae were counted daily and transferred to 25 8 mm moist filter paper strips. Filter paper strips with parasitized larvae were placed in mm vials. Vials were sealed with parafilm and returned to the incubators. Percent mortality [(number of dead parasitoid larvae and pupae per total parasitized larvae) 100], percent parasitism [(number of parasitized thrips larvae per 20) 100] and sex ratio (male: female) were calculated. Three replicates of 10 tubes containing a female wasp and 20 thrips larvae were studied at each temperature. Developmental time data were square root transformed. Percent mortality and parasitism data were arcsine transformed. Developmental time data were analyzed using Statistical Analysis System general linear models (SAS 1985) for a balanced ANOVA. Means were separated by Waller-Duncan k-ratio

15 Castineiras et al.: Temperature response of Ceranisus menes 15 t-tests. Developmental thresholds were estimated from linear regression equations of the developmental rates against temperature, as described by Varley et al. (1973), using Sigma Stat statistical software (Kuo et al. 1992). RESULTS Parasitoid developmental time decreased with increased temperature only between 21 and 27 C (Tables 1 and 2). Thai males emerged one day before females at 25 C, but at lower or higher temperatures the difference in the dates of emergence increased (Table 2). Japanese strain percent mortality increased at temperatures above or below 23 C (Table 3). In the Thai strain no significant difference was observed in mortality at 23 and 25 C, but above 25 C and below 23 C mortality increased. Only 4-6% of the Japanese parasitoids survived at 29 C (Table 3). At 21 C and above 25 C pupae were frequently contaminated with fungi. No statistical differences in percent parasitism were observed between 23 and 29 C in either strain, but parasitism decreased about 50% at 21 C compared with temperatures equal to or above 23 C. (Table 3) Sex ratios of the Thai strain reared at different temperatures were not significantly different (F= 0.30; df= 4,10; p= 0.87). The mean sex ratio (male:female) was 1:1.9 (± 0.008). Linear regression models show the effect of temperature on the rates of development of the parasitoids. Coefficients of determination equal to or above 84% were obtained for developmental rates against temperature (Table 4). The Japanese and Thai strains showed different thermal constants in the time periods from egg to pupa, but the number of degree-days needed for pupation and for total development were the same for the females of the two strains. Thai males had lower thermal constants and higher developmental thresholds than the Japanese and Thai females (Table 4). DISCUSSION The data in Tables 1 and 2 show that strains responded to temperature by increasing or decreasing their developmental time. Similar results on the duration of the egglarval and pupal stages at comparable temperatures were reported for a Japanese strain of C. menes reared on T. tabaci (Sakimura 1937). TABLE 1. EFFECT OF TEMPERATURE ON THE DEVELOPMENTAL TIME OF THE JAPANESE STRAIN OF CERANISUS MENES REARED ON THRIPS PALMI. Developmental Time in Days 1 (Mean ± SD) Temp ( C) Egg-pupa Pupa Total ± 0.33 a ± 0.07 a ± 0.39 a ± 0.37 b ± 0.30 b ± 0.09 b ± 0.26 c ± 0.15 c ± 0.11 c ± 0.16 d ± 0.06 d ± 0.08 d ± 0.39 d ± 0.31 d ± 0.33 d 1 Means within a column followed by the same letter are not significantly different according to a Waller-Duncan k-ratio t-test on square root transformed data (p > 0.001, k-ratio = 100). Untransformed means are presented.

16 16 Florida Entomologist 79(1) March, 1996 TABLE 2. EFFECT OF TEMPERATURE ON THE DEVELOPMENTAL TIME OF THE THAI STRAIN OF CERANISUS MENES REARED ON THRIPS PALMI. Developmental Time in Days 1 (Mean ± SD) Temp ( C) Egg-prepupa Pupa (Female) Pupa (Male) Total (Female) Total (Male) ± 0.28 a ± 0.34 a ± 0.26 a ± 0.57 a ± 0.11 a ± 0.04 b ± 0.23 b ± 0.13 b ± 0.18 b ± 0.16 b ± 0.13 c ± 0.06 c ± 0.02 c ± 0.16 c ± 0.13 c ± 0.15 d ± 0.09 d ± 0.50 d ± 0.23 d ± 0.17 d ± 0.16 d ± 0.35 d ± 0.60 d ± 0.39 d ± 0.63 d 1 Means within a column followed by the same letter are not significantly different according to a Waller-Duncan k-ratio t-test on square root transformed data (p > 0.001, k-ratio = 100). Untransformed means are presented.

17 Castineiras et al.: Temperature response of Ceranisus menes 17 TABLE 3. PERCENT MORTALITY AND PERCENT PARASITISM OF TWO STRAINS OF CE- RANISUS MENES REARED AT DIFFERENT TEMPERATURES ON THRIPS PALMI. % Mortality 1 (Mean ± SD) % Parasitism (Mean ± SD) Temp ( C) Japanese Strain Thai Strain Japanese Strain Thai Strain ± 0.41 b ± 0.12 a ± 0.83 b ± 0.53 b ± 0.79 e ± 0.72 d ± 0.72 a ± 0.50 a ± 0.89 d ± 1.35 d ± 1.15 a ± 0.85 a ± 0.96 c ± 0.36 c ± 1.11 a ± 0.89 a ± 0.88 a ± 1.09 b ± 1.58 a ± 0.30 a 1 Means within a column followed by the same letter are not significantly different according to a Waller-Duncan k-ratio t-test on arcsine transformed data (p > 0.001, k-ratio = 100). Untransformed means are presented. Percent mortality was high (Table 3). According to Hirose et al. (1993), a high level of mortality occurs between the larval and pupal stages in the laboratory due to handling. Sakimura (1937) stated that females may kill some of the young host larvae by the insertion of the ovipositor without laying eggs. That could explain the low percent parasitism found in our experiments (Table 3). Temperature did not affect the sex ratio in the biparental strain, but the mean sex ratio of 1:1.9 was higher than other reports for C. menes: 1:1.5 (Sakimura 1937, Daniel 1986) and 1:1 (Murai 1990). Minimum temperature thresholds of the Japanese parasitoids were between 5 and 8 C (Table 4). Japanese parasitoids developed and pupated at 29 C, but 95% of pupae died (Table 3). Tagashira (1992) observed that wasps of the Japanese uniparental strain reared on F. intonsa did not emerge at 30 C. We suggest that the maximum temperature threshold for the Japanese biotype of C. menes is near 30 C. The Thai biotype is probably better adapted to warmer temperatures than the Japanese biotype. Minimum temperature thresholds of this strain ranged from 7.8 to 14.2 C (Table 4). The maximum temperature threshold for the Thai strain seems to be over 29 C. At 29 C mortality was 27.4% and percent parasitism was 28.9%(Table 3). Comparisons of the behavior of the Thai strain with former studies are not possible. These parasitoids were found in Thailand in 1987 and only some field data are available from Hirose et al. (1993). The best temperature for laboratory rearing of C. menes is 25 C. At 25 C development from egg to adult was 26 days for Thai males and 27 days for females of both strains, percent mortality fluctuated between 14-20% and percent parasitism averaged 27%. At 21 C parasitism averaged 12% and above 27 C mortality was generally higher than 18%. In South Florida, where the average temperature during the growing season (September-April) is C, temperature does not appear to be a limiting factor for field establishment of C. menes. Considering their response to temperature, both strains are good candidates for biological control of T. palmi in the field. ACKNOWLEDGMENTS We thank J. Peña and W. Meyer, Tropical Research and Education Center, University of Florida, for reviewing the manuscript, and J. B. Coulliette for technical assis-

18 18 Florida Entomologist 79(1) March, 1996 TABLE 4. PARAMETERS AND R 2 OF THE REGRESSION EQUATIONS FOR DEVELOPMENTAL RATES VS. TEMPERATURE; THERMAL CONSTANTS AND DEVEL- OPMENTAL THRESHOLDS OF TWO STRAINS OF CERANISUS MENES REARED ON THRIPS PALMI. Japanese Strain Thai Strain Egg-pupa Pupa Total Egg-pupa Pupa (F) Pupa (M) Total (F) Total (M) Slope Intercept r Thermal Constant (Deg-Days) Develop. Threshold

19 Castineiras et al.: Temperature response of Ceranisus menes 19 tance. This research was supported by CSRS Grant Development of Management Strategies for the Melon Thrips, Thrips palmi. This is Florida Agricultural Experiment Station Journal Series No R REFERENCES CITED DANIEL, M. A Thrips-parasite interactions in some Panchaetothripine Thysanoptera (Insecta: Arthropoda). Proc. Indian Natl. Sci. Acad. 52: HIROSE, Y., M. TAKAGI, AND H. KAJITA Discovery of an indigenous parasitoid of Thrips palmi Karny (Thysanoptera: Thripidae) in Japan: Ceranisus menes (Walker) (Hymenoptera: Eulophidae) on eggplant in home and truck gardens. Appl. Entomol. Zool. 27: HIROSE, Y., H. KAJITA, M. TAKAGI, S. OKAJIMA, B. NAPOMPETH, AND S. BURANAPAN- ICHPAN Natural enemies of Thrips palmi and their effectiveness in the native habitat, Thailand. Biol. Control 3: 1-5. KUO, J., E. FOX, AND S. MCDONALD Sigma Stat. User s Manual. Jandel Scientific, San Rafael, CA. LOOMANS, A. J. M., AND J. C. VAN LENTEREN Hymenopterous parasitoids of thrips. Wageningen Agricultural University Papers. 95: MURAI, T Rearing method and biology of thrips parasitoid, Ceranisus menes. Bull. IOBC/WPRS. 13: SAKIMURA, K On the bionomics of Thripoctenus brui Vuillet, a parasite of Thrips tabaci Lind., in Japan. Kontyû 11: SAS INSTITUTE SAS User s Guide: Statistics. 5th ed. SAS Institute, Cary, N.C. TAGASHIRA, E Host acceptance by Ceranisus menes (Walker) (Hymenoptera: Eulophidae), a larval parasitoid of thrips, with reference to host suitability for the parasitoid. M. S. Thesis, Fac. of Agric., Kyushu University, Fukuoka, 44 pp. VARLEY, G. C., G. R. GRADWELL, AND M. P. HASSELL Insect Population Ecology, an Analytical Approach. Univ. of California Press. Berkeley and Los Angeles.

20 Tsai: Biology of Peregrinus maidis 19 DEVELOPMENT AND OVIPOSITION OF PEREGRINUS MAIDIS (HOMOPTERA: DELPHACIDAE) ON VARIOUS HOST PLANTS JAMES H. TSAI Fort Lauderdale Research and Education Center University of Florida, IFAS Fort Lauderdale, FL ABSTRACT The development and oviposition of Peregrinus maidis (Ashmead) (Homoptera: Delphacidae), a serious pest and the only known vector of maize stripe tenuivirus and maize mosaic rhabdovirus in tropical and subtropical areas, was studied on the following plants in the laboratory: corn (Zea mays L. var. Saccharata Guardian ), itch grass (Rottboellia exaltata L.), rice (Oryza sativa L. var. Mars, Saturn, Nato, Bellevue, Labelle, Labonnet, and Starbonnet), sorghum (Sorghum bicolor (L.) Moench var. AKS 614), goose grass (Eleusine indica (L.) Gaertn), oats (Avena sativa L.), rye (Secale cereale L.), gama grass (Tripsacum dactyloides L.), barnyard grass (Echinochloa crusgalli L.) and sugarcane (Saccharum officinarum L.). Peregrinus maidis nymphs did

21 20 Florida Entomologist 79(1) March, 1996 not develop on rye, oats, rice and sugarcane, but the adults survived for various lengths of time on these test plants. The average length of nymphal development on corn, itch grass, sorghum, goose grass, barnyard grass and gama grass was 17.20, 17.87, 20.21, 24.97, and days, respectively. Adult longevity (X ± SD) on corn, gama grass, itch grass, sorghum, goose grass, and barnyard grass was 36.1 ± 20.0, 42.7 ± 16.6, 28.3 ± 11.9, 7.6 ± 6.4, 8.1 ± 7.3 and 7.3 ± 6.6 days, respectively. Oviposition rarely occurred on sorghum, goose grass and barnyard grass. The numbers of eggs laid per day per female on corn, itch grass and gama grass was (X ± SD) 21.0 ± 2.0, 6.4 ± 6.6, 3.5 ± 3.0 eggs, respectively; the numbers of eggs per female per life on these respective plants was (X ± SD) 612 ± 170.1, 146 ± and 48 ± 45.6 eggs. Key Words: Corn delphacid, corn, itchgrass, rice, sorghum, goosegrass, barnyard grass, gama grass. RESUMEN El desarrollo y la ovoposición de Peregrinus maidis (Ashmead) (Homoptera: Delphacidae), una seria plaga y el único vector conocido del tenuivirus de la raya del maíz y del rhabdovirus del mosaico del maíz en áreas tropicales y subtropicales, fue estudiado en el laboratorio en las siguientes plantas: arroz (Oryza sativa L. var. Mars, Saturn, Nato, Bellevue, Labele, Labonnet y Starbonnet), avena (Avena sativa L.), caña de azúcar (Saccharum officinarum L.), centeno (Secale cereale L.), maíz (Zea mays L. var. Saccharata Guardian ), sorgo (Sorghum bicolor (L.) Moench var. AKS 614), Echinocloa crusgalli L., Eleusine indica (L.) Gaertn, Rottboellia exaltata L., y Tripsacum dactyloides L. Las ninfas de P. maidis no se desarrollaron en centeno, avena, arroz y caña de azúcar pero los adultos sobrevivieron diferente tiempo en estas plantas. El desarrollo promedio de las ninfas en maíz, R. exaltata, sorgo, E. indica, E. crusgalli y T. dactyloides fue de 17.20, 17.87, 20.21, 24.97, y días, respectivamente. La longevidad de los adultos (X ± SD) en maíz, T. dactyloides, R. exaltata, sorgo, E. inidica y E. crusgalli fue de 36.1 ± 20.0, 42.7 ± 16.6, 28.3 ± 11.9, 7.6 ± 6.4, 8.1 ±7.3 y 7.3 ± 6.6 días, respectivamente. La ovoposición raramente ocurrió en sorgo, E. indica y E. crusgalli. El número de huevos puesto por día por hembra en maíz, R. exaltata y T. dactyloides fué de (X ± SD) 21.0 ± 2.0, 6.4 ± 6.6, y 3.5 ± 3.0, respectivamente. El número de huevos por hembra puestos durante toda su vida en esas plantas fue (X ± SD) 612 ± 170.1, 146 ± y 48 ± The corn delphacid, Peregrinus maidis (Ashmead), is not only a major pest of corn and sorghum (Namba & Higa 1971, Chelliah & Basheer 1965), it is also the only known vector of two important maize viruses [maize mosaic rhabdovirus (MMV) and maize tenuivirus (MStV)] (Nault & Knoke 1981, Tsai 1975), and it is of particular economic importance in the lowland humid tropics. It has even been suggested that its introduction and the spread of two devastating viral diseases into Central America resulted in the collapse of the Mayan civilization (Brewbaker 1979; however, see Nault 1983). Namba & Higa (1971) reported that P. maidis was able to survive for various lengths of time on napiergrass (Pennisetum purpureum Schumach), vaseygrass (Paspalum urvillei Steud.), sugarcane (Saccharum officinarum L.), sorghum (Sorghum vulgare Pers.), sourgrass (Trichachne insularis Nees.), Californiagrass (Brachiaria mutica Stapf), Job s tears (Coix lacryma-jobi L.), pangolagrass (Digitaria decumbens Stent), and nutgrass (Cyperus rotundus L.). Chelliah & Basheer (1965) stated that this insect utilized Pennisetum typhoides (Stapf. and Hubbard), Sorghum

22 Tsai: Biology of Peregrinus maidis 21 halepenise (L.), Setaria italica (Beauv.), Echinochloa colona var. frumentacca (L.) and Paspalum scrobiculatum (L.) as breeding or feeding hosts. Earlier studies have shown that itch grass (Rottboellia exaltata L.) was a common host for two corn viruses, MStV and MMV, and their vector, P. maidis (Tsai 1975, Falk & Tsai 1983, Nault & Knoke 1981). However, little is known of the relationship between P. maidis and its alternate hosts in south Florida. The role of alternate hosts in the epidemiology of these diseases between growing seasons and the extension of the host range of P. maidis is of great significance in studying the ecology of this insect. This paper reports the development and oviposition of P. maidis on six host plants. MATERIALS AND METHODS A stock colony of P. maidis was maintained on corn (Zea mays L. var. Saccharata Guardian ) in an insectary at 27±1 C and a 12:12 (L:D) photoperiod for over 15 years. P. maidis eggs were excised from the midribs of corn plants and placed on cut leaf pieces of corn and allowed to hatch. At least 30 newly hatched nymphs were singly transferred to individual culture tubes (25 mm diam) containing fresh leaf pieces of corn, itch grass (Rottboellia exaltata L.), sorghum (Sorghum bicolor (L.) Moench. var. AKS 614), gama grass (Tripsacum dactyloides L.), sugarcane (Saccharum officinarum L.), and seedlings of goose grass (Eleusine indica (L.) Gaertn.), barnyard grass (Echinochloa crusgalli L.), rye (Secale cereale L.), oats (Avena sativa L.) and rice (Oryza sativa L. var. Mars, Saturn, Nato, Bellevue, Labelle, Labonnet and Starbonnet). The opening of the culture tube was covered with a piece of stretched Parafilm to prevent escape or desiccation. At the same time, a group of newly hatched nymphs was reared on the potted young plants of each test plant species or variety and used as replacements for dead or missing insects. The ages of replacement insects were determined by body sizes and morphological characteristics as described by Tsai & Wilson (1986). All tests were conducted at 27±1 C, 60% RH, and a photoperiod of 12:12 (L:D) in three growth chambers (Percival Scientific Inc., Boone, IA) over a period of nine months. Each insect was checked daily for molting, and plant tissue was replaced every two days or sooner. The dates of molting, duration of stadia, mortality, and adult longevity were recorded. Differences in nymphal development and adult longevity on different hosts were analyzed by Student-Newman-Keul (SNK) multiple range test (Sokal & Rohlf, 1969). Ten to 15 pairs of newly emerged adults were placed on single plants at the 4- to 5-leaf stage and covered by a plastic cage to test for oviposition. Pairs were removed daily to new plants which were then dissected for egg counts. RESULTS Of the 17 species and varieties of plants tested, only corn, itch grass, gama grass, goose grass and barnyard grass were able to support P. maidis development. At least 60 first instar nymphs were singly tested on each of seven varieties of rice and one variety of sugarcane, oats and rye; the survivorships of test insects ranged from 2 to 5 days. A group of 32 adults was also tested individually on these plants. They only survived (X ± SD) 3.0 ± 2.4, 4.0 ± 3.2, 2.0 ± 1.5, 5.0 ± 3.2, 3.0 ± 2.1, 3.0 ± 1.8, 4.0 ± 2.2, 4.0 ± 3.5, 5.0 ± 5.3, and 10.0 ± 9.9 days on rice var. Mars, Saturn, Nato, Bellevue, Labelle, Labonnet, and Starbonnet, sugarcane, oats, and rye, respectively. There were significant (P < 0.05) differences in total nymphal development times among some hosts, with shortest development time on corn (17.20 ± 1.50 days; X ±

23 22 Florida Entomologist 79(1) March, 1996 TABLE 1. DEVELOPMENT (X NO. DAYS ± S.D.) OF P. MAIDIS ON SIX HOST PLANTS. 1 Instar Host 1st (n) 2nd (n) 3rd (n) 4th (n) 5th (n) Total Z. mays 3.8±1.5 (90) cd 3.0±1.5 (59) c 3.1±1.4 (44) d 3.1±1.5 (38) c 4.2±1.4 (21) cd 17.20±1.50 c R. exaltata 3.0±1.0 (31) d 3.1±1.0 (34) c 4.0±1.5 (25) cd 4.3±1.7 (24) bc 3.5±1.1 (14) d 17.87±2.48 c S. bicolor 3.9±0.9 (27) bcd 3.5±0.7 (27) c 4.1±1.3 (22) cd 2.8±0.9 (26) c 4.9±1.7 (20) cd 20.21±1.97 c E. indica 4.4±1.3 (29) bc 5.3±1.5 (24) b 4.8±1.1 (25) bc 5.3±1.4 (32) b 5.3±1.7 (32) bc 24.97±2.81 b E. crusgalli 5.2±0.9 (38) a 5.1±0.9 (20) b 5.7±1.6 (20) b 5.1±1.0 (20) b 6.2±2.0 (22) b 27.24±2.21 b T. dactyloides 4.6±1.9 (64) b 8.6±2.6 (50) a 12.3±2.5 (43) a 16.4±3.2 (35) a 18.6±1.1 (11) a 60.50±10.30 a 1 Means within each column followed by the same letters are not significantly different (P>0.05, SNK multiple range test).

24 Tsai: Biology of Peregrinus maidis 23 SD) and longest on gama grass (60.50 ± 10.30) (Table 1). In general, first instars had higher survival rates, compared with other stages, on most plants tested (Table 2). Cumulative percentage of survival from first instar to adult was higher on sorghum (60.9%) than on other hosts (Fig. 1). The adults lived significantly longer on gama grass (42.70 ± days), corn ± days), and itch grass (28.30 ± days) than on sorghum (7.64 ± 6.44 days), goose grass (8.00 ± 7.32 days) and barnyard grass (7.27 ± 6.55 days; P < 0.05) (Fig. 2). Adults rarely deposited eggs on sorghum, goose grass and barnyard grass. The number of eggs laid per day per female on corn, itch grass, and gama grass was (X ± SD) 21.0 ± 2.0, 6.4 ± 6.6, and 3.5 ± 3.0 eggs, respectively; the number of total eggs per female on these respective hosts was (X ± SD) 612 ± , 146 ± and 48 ± DISCUSSION This study showed that there was no significant difference in length of P. maidis nymphal development on sweet corn, itch grass and sorghum, but development times were significantly shorter on these hosts than on goose grass, barnyard grass and gama grass (Table 1). Although adult longevity and egg production on itch grass are significantly less than on corn, the average total number of eggs produced by each female on itch grass is still considerably higher when compared to other host plants. Itch grass is considered to be one of the world s most serious weed pests and is currently found in Florida, Louisiana, Texas, Arkansas, Alabama and Georgia, and as far north as latitude 23 (Holm et al. 1977). Research has shown that itch grass is a good host for MStV and MMV (Bradfute & Tsai 1983, Falk & Tsai 1983, Herold 1972, Gingery et al. 1981, Lastra 1977). It is the most important alternate host for P. maidis and MStV and MMV in south Florida between corn growing seasons (Tsai, unpublished). Sorghum is grown throughout the eastern United States for grain and forage (Bailey & Bailey 1976). Like itch grass, several sorghum species have been shown to be hosts of MStV and MMV (Greber 1983, Herold, 1972, Gingery et al. 1981). Chelliah & Basheer (1966) reported that sorghum was the preferred host for P. maidis in laboratory tests. Peregrinus maidis has been reported as a serious pest on new sorghum varieties in India (Agriwal et al. 1981). Sorghum species have thus been suggested as the ancestral hosts for P. maidis and MStV and MMV (Nault 1983). Although P. mai- TABLE 2. AGE-SPECIFIC SURVIVORSHIP (%) OF P. MAIDIS ON SIX HOST PLANTS. Instar Host 1st 2nd 3rd 4th 5th Z. mays R. exaltata S. bicolor E. indica E. crusgalli T. dactyloides

25 24 Florida Entomologist 79(1) March, 1996 dis feeding on sorghum was demonstrated in the present study, it may not be suitable for breeding because few eggs were laid on this host. Goose grass is found through the great plains and occasionally in California, Oregon, Utah and Arizona (Bailey & Bailey 1976). Gama grass and barnyard grass are found throughout the eastern United States (Small, 1972). Gama grass, now established in Florida and Louisiana has the potential for spreading as far north as Minnesota as a row crop competitor (Patterson & Quimby 1978; Patterson et al. 1979). This study has demonstrated that P. maidis could also feed and reproduce on gama grass, but it is not as suitable a host as sweet corn and itch grass. Both itch grass and gama grass flower and shed seeds year-round. Germination of seed is staggered, resulting in continuous growth of new seedlings throughout the year. The northward spread of these two weeds and the adaptation of MStV and MMV to a perennial host or to other delphacid species with a broader host range and wider distribution than P. maidis might allow these viruses to become established in temperate regions of the United States. Despite the short adult longevities on sorghum, goose grass and barnyard grass, P. maidis lived long enough to oviposit a few eggs on these plants. Tsai & Zitter (1982) reported transovarial passage of MStV. This eliminates the need for an alternate host for the virus, with newly hatched insects already carrying the disease agent. The long distance spread of P. maidis and two viruses by strong winds has been proposed (Brewbaker 1979, Bradfute et al. 1981). It is possible that viruliferous P. maidis could be transported in a similar manner to the corn belt during the spring and summer. Although the northward distribution of P. maidis is probably prevented by winter temperatures and the apparent absence of an overwintering stage (Tsai & Wilson 1986), the adaptability of the insect and the presence of these alternate hosts as possible overwintering sites may pose a serious threat to northern corn growers. Figure 1. Cumulative percentage survival of Peregrinus maidis nymphs reared on six host plants.

26 Tsai: Biology of Peregrinus maidis 25 Figure 2. Mean adult longevity (in days) of Peregrinus maidis on six host plants (means ± SD). ACKNOWLEDGMENT Appreciation is extended to Ms. Lori A. Bohning for conducting various experiments and to K. H. Wang for statistical analysis. This research was supported in part by grants from the Pioneer Hi-Bred International, Inc. and The American Seed Research Foundation, Washington, D.C. Florida Agricultural Experiment Stations Journal Series # R REFERENCES CITED AGRIWAL, R. K., R. S. VERMA, AND G. S. BHARAJ Screening of sorghum lines for resistance against shoot bug, Peregrinus maidis (Ashmead) (Homoptera: Delphacidae). JNKVV (Jawaharal Nehru Krishi Vishwa Vidyalaya) Res. J. 12: 116. BAILEY, L. H., AND E. Z. BAILEY Hortus Third. Macmillan Publishing Co., Inc pp. BRADFUTE, O. E., J. H. TSAI, AND D. T. GORDON Corn stunt spiroplasma and viruses associated with a maize epidemic in southern Florida. Plant Dis. 65: BRADFUTE, O. E., AND J. H. TSAI Identification of maize mosaic virus in Florida. Plant Dis. 67: BREWBAKER, J. L Diseases of maize in the wet lowland tropics and the collapse of classic Maya civilization. Econ. Bot. 33: CHELLIAH, S., AND M. BASHEER Biological studies of Peregrinus maidis (Ashmead) (Araeopidae: Homoptera) on sorghum. Indian J. Entomol. 27:

27 26 Florida Entomologist 79(1) March, 1996 FALK, B. W., AND J. H. TSAI Physiochemical characterization of maize mosaic virus. Phytopathology 73: HEROLD, F Maize mosaic virus. No. 94, in Descriptions of Plant Viruses. Commonw. Mycol. Inst., Assoc. Appl. Biol. Kew, Surrey, England. 4 pp. HOLM, L. G., D. L. PLUCKNETT, J. V. PANCHO, AND J. P. HERBERGER The world s worst weeds, distribution and biology. The East-West Center. The University Press of Hawaii. 609 pp. GINGERY, R. E., L. R. NAULT, AND O. E. BRADFUTE Maize stripe virus: Characteristics of a member of a new virus class. Virology 112: GREBER, R. S Characteristics of viruses affecting maize in Australia, pp in D. T. Gordon, J. K. Knoke, L. R. Nault, and R. M. Ritter, [eds.]. Proc. Int. Maize Virus Dis. Colloq. and Workshop. Ohio Agric. Res. Dev. Cent., Wooster, Ohio. 266 pp. LASTRA, R. J Maize mosaic and other maize virus and viruslike diseases in Venezuela, pp in L. E. Williams, D. T. Gordon, and L. R. Nault, [eds.], Proc. Int. Maize Virus Dis. Colloq. and Workshop. Ohio Agric. Res. Dev. Cent., Wooster, Ohio, 145 pp. NAMBA, R., AND S. Y. HIGA Host plant studies of the corn planthopper, Peregrinus maidis (Ashmead), in Hawaii. Proc. Hawaii Entomol. Soc. 21: NAULT, L. R., AND J. K. KNOKE Maize vectors, pp in D. T. Gordon, J. K. Knoke, and G. E. Scott [eds.], Virus and viruslike diseases of maize in the United States. Southern Coop. Series Bull p NAULT, L. R Origin of leafhopper vectors of maize pathogens in Mesoamerica, pp in D. T. Gordon, J. K. Knoke, L. R. Nault, and R. M. Ritter, [eds.], Proc. Int. Maize Virus Dis. Colloq. and Workshop. Ohio, Agric. Res. Dev. Cent., Wooster, Ohio, 266 pp. PATTERSON, D. T., AND P. C. QUIMBY, JR Itchgrass: a potential noxious weed in Mississippi. Mississippi Agric. For. Exp. Stn. Res. Rep. 3: 1-3. PATTERSON, D. T., C. R. MEYER, E. P. FLINT, AND P. C. QUIMBY, JR Temperature responses and potential distribution of itch grass (Rottboellia exaltata) in the United States. Mississippi Agric. For. Exp. Stn. Res. Rep. 27: SMALL, J. K Manual of the southeastern flora. Hafner Publishing Co pp. SOKAL, R. R., AND F. J. ROHLF Biometry - the principles and practice of statistics in biological research. W. H. Freeman and Co. 776 pp. TSAI, J. H Occurrence of a corn disease in Florida transmitted by Peregrinus maidis. Plant Dis. Rep. 59: TSAI, J. H., AND S. W. WILSON Biology of Peregrinus maidis with descriptions of immature stages (Homoptera: Delphacidae). Ann. Entomol. Soc. America 79: TSAI, J. H., AND T. A. ZITTER Transmission characteristics of maize stripe virus by the corn delphacid. J. Econ. Entomol. 75:

28 Thorne et al.: Nasutitermes Nests and Nodules 27 ARCHITECTURE AND NUTRIENT ANALYSIS OF ARBOREAL CARTON NESTS OF TWO NEOTROPICAL NASUTITERMES SPECIES (ISOPTERA: TERMITIDAE), WITH NOTES ON EMBEDDED NODULES BARBARA L. THORNE 1, MARGARET S. COLLINS 2, AND KAREN A. BJORNDAL 3 1 Department of Entomology University of Maryland College Park, MD Smithsonian Institution Museum Support Center 4210 Silver Hill Road Suitland, MD Department of Zoology 223 Bartram Hall University of Florida Gainesville, Florida ABSTRACT Nest architecture of the arboreal Neotropical termites Nasutitermes acajutlae (Holmgren) and N. nigriceps (Haldeman) is described, with special reference to carton inclusions or nodules found within the normal gallery matrix of some nests. Nutrient analyses of these nodules show that they have high cellulose and low cutin concentrations in comparison to normal nest carton. These data support the hypothesis that the nodule inclusions serve as a form of facultative food storage in some nests of these termite species. These cases appear to represent a rare situation in which food is not stockpiled or cultured by termites, but rather some partially processed, masticated food is incorporated into the nest matrix for future consumption. Key Words: Termites, Nasutitermitinae, inclusions, food storage. RESUMEN Se describe la arquitectura del nido de las termitas neotropicales Nasutitermes acajutlae (Holmgren) y N. nigriceps (Haldeman), con referencia especial a inclusiones de cartón o nódulos encontrados dentro de la matriz de la galería de algunos nidos. El análisis de nutrientes de los nódulos muestra que estos tienen concentraciones altas de celulosa y bajas de cutina, en comparación con el cartón normal de los nidos. Los datos sostienen la hipótesis de que las inclusiones de los nódulos sirven como una forma facultativa de almacenamiento de alimento en algunos nidos de termitas de esas especies. Estos casos parecen representar una rara situación en la cual el alimento no es almacenado en pilas o cultivado por las termitas, sino masticado y parcialmente procesado e incorporado a la matriz del nido para consumo futuro. The tropicopolitan termite genus Nasutitermes (Termitidae; Nasutitermitinae) is the most speciose of all isopteran genera, containing approximately 75 described species from the Neotropics alone (Araujo 1977). Unlike most termites, many species of Nasutitermes build arboreal carton nests composed of wood and salivary and fecal flu-

29 28 Florida Entomologist 79(1) March, 1996 ids (Light 1933), and occasionally other materials such as sand (Thorne & Haverty, pers. obs.). Most other nest-building termites build mounds on the ground (e.g., Emerson 1938), but nesting in trees has enabled species of Nasutitermes and several other genera to colonize and exploit a new habitat. Nasutitermes nigriceps (Haldeman) is a geographically widespread termite, ranging at least from Panama north throughout the lowland forests of Central America into Mexico. It is also found on Jamaica and on Grand Cayman Island (Araujo 1977, Thorne et al. 1994). N. acajutlae (Holmgren), which is morphologically very similar to N. nigriceps, is found on Puerto Rico, the US and British Virgin Islands (BVI), Trinidad, and Guyana (Emerson 1925, Araujo 1977, Thorne et al. 1994). There are isolated reports of members of the N. nigriceps complex from other parts of South America, but a comprehensive taxonomic analysis of specimens is needed to verify species identity of the South American fauna. Despite the abundance of Nasutitermes arboreal nests, the chemical composition of the carton material has not been examined in detail in any species (but see Becker & Seifert 1962 for data on ash and lignin content). Knowledge of the composition of the nest is fundamental in determining origin of nesting materials, cost of construction, variation among colonies and species, and ability of the termites to allocate components of their diet for nest construction. A distinctive feature of some N. acajutlae and N. nigriceps nests is the presence of rounded carton inclusions or nodules within the typical gallery matrix (Hubbard 1877 pp. 268, 270, Andrews 1911 pp , Emerson 1938 p. 264, Wolcott, cited in Martorell 1945 p. 361). These nodules appear to be of a similar carton construction as the rest of the nest, but they are a lighter brown color, are formed in dense concentric sheaths (Fig. 1), and they may possibly serve as a form of food storage (Hubbard 1877; Andrews 1911). Kemner (1929) interprets the presence of carton nodules in the Javan termite Microcerotermes depokensis Kemner as food storage structures. Noirot (1959) reported compact masses of wood fragments in the central cavity of nests of Globitermes sulphureus (Haviland). Some termite genera do store food as dried vegetative elements in specialized portions of their nests ( attics ) [e.g., Hodotermitinae (Hodotermes, Microhodotermes, Anacanthotermes); Rhinotermitidae (Psammotermes); Termitidae: Amitermitinae (certain Amitermes and Drepanotermes), Nasutitermitinae (certain species of Tumulitermes, Nasutitermes and Trinervitermes)] (Noirot 1970, Bouillon 1970). The fungus growing termites (certain Macrotermitinae) culture fungus within the nest as a supplemental food source. Interestingly, some fungus growing termites store vegetative materials within the nest before they are included in the fungus garden (Pseudacanthotermes, Acanthotermes, some Macrotermes) (Grasse & Noirot 1951). If the Nasutitermes nodules described in this paper are indeed food reserves, they are not simply stored food but rather elements which have already been masticated and partially processed by the termites, then positioned within the nest matrix for future consumption. In this paper we describe the architecture of N. acajutlae and N. nigriceps nests from sites in Panama and the BVI. Observations of the nodule inclusions are reported. Nutrient analyses of two nests without nodules and comparative chemical analyses of nodule material versus the surrounding normal carton matrix of two nests with nodules are presented and reported. MATERIALS AND METHODS Eight N. nigriceps nests were collected within 5 km of the Panama Canal in 1980 and 1981; only one of these contained the distinctive nodules within the carton next matrix. This arboreal nest was collected from the Gigante East Peninsula near Barro

30 Thorne et al.: Nasutitermes Nests and Nodules 29 Colorado Island on 7 April The entire nest was pried from the host tree, placed within a plastic bag, and taken to the laboratory of the Smithsonian Tropical Research Institute on Barro Colorado Island. The nest was dissected by sequential shaving after being refrigerated for 24 hours to inactivate the termites (technique described in Thorne 1984). Nest carton from four colonies (one N. nigriceps nest collected near Barro Colorado Island, Panama in 1981; three N. acajutlae nests collected in 1988 and 1989 on Guana Island, BVI) was analyzed in Two of the nests (the N. nigriceps nest from Panama and a 1988 N. acajutlae nest from the BVI) contained nodules. Chemical composition of both the nodules and a sample of the more typical dark carton material was determined from those two nests, and samples of typical carton matrix from parts of two additional Guana Island N. acajutlae nests (which did not contain nodules) were also analyzed. Type of nest material examined is presented in Table 1. Materials and Methods for Nutritional Analyses of Nest Samples In the laboratory, samples were dried at 60 C to constant mass (approximately 24 h). Dried samples were ground to pass through a 1 mm screen in a Wiley mill. A portion of each sample was dried at 105 C to determine percent dry matter and then placed in a muffle furnace for 3 h at 500 C to determine percent organic matter and ash (an estimate of total mineral content). In vitro organic matter digestibility, or percent fermentable substrate, was determined by the Tilley & Terry (1963) method as modified by Moore & Mott (1974). This analysis consists of a 48 h incubation under CO 2 at 39 C with an inoculant of steer rumen fluid followed by a 48 h acid-pepsin treatment to remove undigested microbes. The percent of organic matter that disappears during the 96 h is the in vitro organic matter digestibility. Percentage of neutral detergent fiber (NDF: cellulose, hemicellulose, lignin and cutin) was measured by the Van Soest technique (Goering & Van Soest 1970) with decalin and sodium sulfite omitted (Golding et al. 1985). Analyses for percentages of acid detergent fiber (ADF: cellulose, lignin and cutin), potassium permanganate lignin, and cutin followed Goering & Van Soest (1970). Percent hemicellulose is estimated by subtracting ADF from NDF. Lipids were extracted with ethyl ether in a Goldfisch apparatus for 8 h. Percent concentrations of total (Kjeldahl) nitrogen and phosphorus were measured with a block digester (Gallaher et al. 1975) and an automated Technicon analyzer (Hambleton 1977). Energy content of food and feces was determined in a bomb calorimeter following standard procedure (Parr Instrument Co. 1960). One sample was analyzed from each source of nest material. In each analysis, two subsamples were evaluated. Values for replicates of each sample were accepted within 1% relative error. Relative error is calculated as (a-b)/(a+b) where a and b are replicate values. Rarely, the values obtained for the duplicates were not within 1 relative percentage, in which case a third subsample was analyzed. Table 1 reports the mean of the analyzed values for each sample. RESULTS Nest Architecture Nests built by N. acajutlae and N. nigriceps can be among the largest of any arboreal nesting Nasutitermes. Maximum dimensions of an ellipsoidal nest can approach 2 m in height and 1 m in girth (Thorne et al. 1994). The exterior of these nests is typically medium to greyish brown in color and irregularly mottled, generally with

31 30 Florida Entomologist 79(1) March, 1996 rather shallow bumps, unlike the dark nests with small bumps characteristic of the exterior of N. corniger (Motschulsky) and N. costalis (Holmgren) nests or the lighter brown, smooth shell typical of N. ephratae (Holmgren) (Thorne 1980; Haverty et al. 1990). Young nests may be difficult to identify to species, but differences in exterior appearance make it possible to visually discriminate most mature nests of N. acajutlae and N. nigriceps from those of N. corniger, N. costalis, or N. ephratae. The outer carton shell of nests of all of these Nasutitermes species has small pinpoint holes, visible if a piece is held up to a light. These holes presumably function in gas exchange. The intercalated matrix of galleries within mature nests of N. acajutlae and N. nigriceps tends to be larger (chamber diam up to 2.1 cm) and with thicker carton walls (up to 0.6 cm near the exterior of a nest; exceeding 1.7 cm near the interior of the nest) than in nests of arboreal Nasutitermes found sympatrically with one or both of these species (N. columbicus, N. corniger, N. costalis, N. ephratae). The royal cell within the nest is often positioned near the central longitudinal axis of the nest, and frequently located in or near a branch fork or knothole of the host tree. In small to medium sized nests (< 60 cm diam) the royal cell is a distinctly thicker sphere or ellipsoid of layered carton (generally up to cm in diam) surrounding the royal chamber. In nests exceeding 1 m on an axis the royal chamber can be embedded in the dense carton center of the nest, with the royal cell becoming indistinct from the remainder of the central, reinforced portion of the nest. We have not found distinguishing characters to differentiate nest architecture of N. acajutlae from N. nigriceps. The carton-covered foraging trails built by large N. acajutlae and N. nigriceps colonies are wider and less regular than in N. corniger, N. costalis, or N. ephratae. Termites occupying small to medium sized N. acajutlae and N. nigriceps nests frequently build simple, linear trails cm wide, thus they are indistinct from trails of N. corniger, N. costalis, or N. ephratae. However, trails leaving large nests of N. acajutlae and N. nigriceps are often broad (up to 14 cm in width) and deep (up to 8 cm from the tree to ceiling of the gallery). Trails from large nests are often highly irregular along the edges. Occasionally a floor is built as well so that the trail becomes a tube that can, for a limited distance, be separate from the tree or branch. As is typical for many arboreal Nasutitermes, tunnels built on the exterior of tree branches are frequently on the underside of the branch. We hypothesize that this minimizes disturbance by hard rainfall or by creatures traveling along the tops of branches. Building galleries in the shade of branches would also minimize desiccation from direct sunlight. A further advantage would be that foraging tunnels built on the undersides of branches would receive maximum moisture from rain running off the branch. This would be beneficial for N. acajutlae or N. nigriceps since they often live in relatively dry habitats (Thorne et al. 1994). Description of Nodules and Nest Population We describe nodule inclusions in three nests: one N. nigriceps nest dissected in April, 1981 in Panama, one N. acajutlae nest dissected in July, 1988 on Guana Island, BVI, and a nest dissected on the island of Tortola, BVI in October, 1994 (nodules from the latter nest were not analyzed for nutritional content). Photographs of the interior of the N. nigriceps nest collected in Panama are shown in Fig. 1. The nest was generally spherical, about 46 cm in diam, which placed it in a medium size category for conspecific nests in that area. Twenty nodule formations, 14 of which measured cm in diam, but some as small as 1 cm diam, were removed from the nest. All nodules were positioned within 4-10 cm of the nest exterior. The nodules were of a uniform light brown color in contrast to the dark brown surrounding

32 Thorne et al.: Nasutitermes Nests and Nodules 31 Figure 1. Photographs of nodules embedded within normal nest carton matrix in the Nasutitermes nigriceps nest collected in Panama in April 1981 (A,B) and in an N. acajutlae nest collected on Tortola, BVI in October 1994 (C,D).

33 32 Florida Entomologist 79(1) March, 1996 nest matrix. Nodule shape was generally spherical although some had distortions or were irregular ellipses. The nest contained an active population of soldiers and workers, as well as a conspicuous brood of wingbud nymphs in the penultimate and ultimate instars. Many of the nymphs occupied the interiors of the nodule spheres. No reproductives, eggs or immatures were found within the nest. The N. acajutlae nest on Guana Island was irregularly ellipsoidal, measuring approximately 1 m in height with a maximum diam of 75 cm. The nest contained a primary queen, developing nymphs of a variety of instars, many eggs and white larvae, and a large population of workers and soldiers. No primary king was retrieved but that is typical during field dissections because mature kings are small enough to retreat quickly and evade capture. The light-colored nodules were located in a zone surrounding the hard, inner core of the nest, all positioned at least 2 cm from the exterior nest wall. Many of the nodules were scalloped, possibly indicating consumption by termites. As with the N. nigriceps nest, immature termites occupied the interior of hollowed-out nodules. An N. acajutlae nest collected on the island of Tortola, BVI was brought to us in several pieces during a field trip in October This large nest, estimated to have been just over a meter in height and about 80 cm in diam, contained eggs, white immature, soldiers, and workers. There were relatively few brachypterous nymphs, but numerous mature alates were present. The primary queen and king were not recovered, but the presence of egg caches suggests that the reproductives were present in the intact nest or in the portion of the host tree surrounded by the nest. This nest had clusters of nodules positioned within the inner perimeter of the nest (Fig.1 C, D) Again, the outer 2 cm or more of nest material was dark, more typical carton matrix with no nodules. Because the nest arrived in pieces, it was impossible to tell whether nodules were built in the center core of the nest. We did not do nutrient analyses of nodule material from the Tortola nest, but we measured each of the 75 nodules that were retrieved. The distribution of nodule sizes recovered from this nest is shown in Fig. 2. Some of the irregularly spherical nodules in this nest were solid, dense material; most were hollowed to some extent as seen in the Panama and Guana Island nests. Hollowed nodules contained large numbers of immatures. Eight of the nodules from this nest were bilobed, as if two units had been constructed or fused together. The nodules from all three of these nests were generally similar in size, shape, color, and position within the nest matrix. In each case immatures occupied excavated nodules. Chemical Analyses of Nest and Nodule Material The most striking aspect of the nutrient composition of nest materials (Table 1) is the consistency among these nests. The only apparent differences are the higher cellulose and lower cutin concentrations in nodule samples than in carton samples and the higher in vitro organic matter digestibility values of the samples from Panama. No statistical analyses were performed because only two samples were available in each category. Hemicellulose was absent or present in only trace amounts in the samples, so was not included in Table 1. Nutrient composition of the nest material apparently does not change with age. The composition of recently constructed normal carton was very similar to that of old carton material from the North Bay and White Bay Beach nests on Guana Island. The high organic matter content indicates that little, if any, soil or sand is incorporated into these samples of nest or nodule material.

34 Thorne et al.: Nasutitermes Nests and Nodules 33 TABLE 1. NUTRIENT COMPOSITION OF NODULE AND NORMAL NEST CARTON SAMPLES 1 Organic Matter Cellulose Lignin Cutin Lipids Nitrogen Phosphorus In Vitro Organic Matter Digestibility Energy Panama-1981 Nodule Normal nest carton BVI-1988 Nodule Normal nest carton BVI-1989 North Bay Beach Interior, dense carton Exterior, thin carton BVI-1989 White Bay Beach New, thin carton Fresh, very thin carton Dry nest carton All values are based on one sample, each analyzed in duplicate. If the values obtained for the duplicates were not within 1 relative percentage, a third replicate was analyzed. Means of the two or, rarely, three values are listed. All values are presented as percent dry matter except in vitro organic matter digestibility is expressed as percent organic matter and energy is expressed a kj/g dry matter. Panama samples are from an Nasutitermes nigriceps colony; BVI (British Virgin Island) samples are from N. acajutlae colonies.

35 34 Florida Entomologist 79(1) March, 1996 Figure 2. Size frequency distribution of nodules removed from the Nasutitermes acajutlae nest collected on the island of Tortola, BVI in October 1994 (N = 75; x = 2.6 ± 1.2 cm). DISCUSSION Nasutitermes acajutlae and N. nigriceps are exceptional among termites in building distinctive inclusions or nodules within the normal carton matrix of their nests. The only other termite reported to build similar structures is the Javan termite Microcerotermes depokensis (Kemner 1929). Two contrasts between the composition of nodules versus normal nest carton analyzed in this study may be biologically significant. First, the nodule samples have lower cutin and higher cellulose percentages than do samples of the surrounding, dark carton matrix. Cutin degradation is not possible for most organisms except some specialized fungi; digestion of cutin by termites is unknown (Breznak, pers. comm.). The differences in cutin and cellulose percentage may indicate that the termites are constructing the nodules from materials with greater digestibility. The relatively high cutin percentages in typical carton probably enhances water-proofing and construction strength. It is unlikely that the difference in cutin abundance is due to transfer of waxes from the insect exocuticle to the nest walls. The percentage of cutin in fresh, newly constructed carton (having minimal opportunity for contact transfer of waxes from passing insects) does not differ markedly from that of old, dense, interior carton

36 Thorne et al.: Nasutitermes Nests and Nodules 35 (see samples from North Bay and White Bay Beach nests, Guana Island, BVI in Table 1). A second distinction is that both the nodule sample and the normal gallery within the nodule sample from Panama have higher in vitro digestibility than do any of the BVI samples. This may reflect species differences in carton processing, or a difference in diet among the two populations (N. acajutlae sampled from Guana Island were feeding substantially on sea grape, Coccoloba uvifera, the diet of the N. nigriceps from Panama is unknown). Clearly, further sampling and geographic variation within each species must be examined before differences of this type can be further evaluated. Hubbard (1877) and Andrews (1911) hypothesized that these Nasutitermes nodules serve as food storage. Kemner (1929) came to a similar conclusion in the case of Microcerotermes depokensis. The food storage hypothesis is supported by the higher cellulose content of nodules in comparison to surrounding nest carton in both N. acajutlae and N. nigriceps. It is difficult to know the conditions under which the nodules might be naturally consumed in a nest, but 0.3 g portions of both N. acajutlae and N. nigriceps nodules offered to 100 workers of N. acajutlae, N. nigriceps, N. costalis, and Zootermopsis nevadensis (Hagen) were consumed in the laboratory within 24 h, and consumed by 100 Reticulitermes flavipes (Kollar) workers within 48 hr (N=3 per species). These species did not consume the normal carton of either N. acajutlae or N. nigriceps nests. Termite nest material can be used as nutritional food reserves in some species. Hegh (1922) commented that termites in mature colonies of Microcerotermes fuscotibialis (Sjostedt) eat the internal walls of their nests during times of food stress. Noirot (1970) reported that central walls of Cephalotermes rectangularis (Sjostedt) nests can be used to culture the termites in the laboratory. The function of nodules and circumstances under which they are constructed are difficult to identify because they are found so rarely. In both Panama and the BVI, examination of nests of approximately the same size, in the same local area, at the same season never revealed another live colony with nodules. Because young were found within the nodules of both N. acajutlae (white immatures instars 1-3) and N. nigriceps (developing alate nymphs) the nodule food reserves may be sequestered for juveniles. Comparable nests with immatures, however, did not have nodules. Nodule construction may be influenced by individual colony health, age, microhabitat, food resources, caste proportions, or population size. Even among colonies producing nodules, they may be ephemeral within a nest. Nodules may only be present seasonally, stockpiled as food reserves and then consumed during times of high demand (as when alate brood is present), when food is scarce, or when travel from the nest is physiologically expensive (as in a drought). It is notable that the only two Nasutitermes species known to construct these nest inclusions are the closely related species N. acajutlae and N. nigriceps, both of which can occupy dry and thus potentially stressful environments (Thorne et al. 1994). The facultative ability to store food in nodules, combined with an exceptional desiccation tolerance of individuals, may contribute to the survival of these two Nasutitermes species in arid or otherwise marginal habitats not colonized by other members of the genus. ACKNOWLEDGMENTS We extend sincere thanks to Dr. and Mrs. Henry Jarecki and the staff of The Guana Island Club for their support and hospitality during the course of this project. John A. Breznak, Michael I. Haverty, and James D. Lazell provided constructive input and productive discussions. Field assistance was contributed by James Egelhoff,

37 36 Florida Entomologist 79(1) March, 1996 Thomas Jarecki, and Ralph Rusher. Portions of this research were funded by The Conservation Agency through a grant from the Falconwood Foundation, The Smithsonian Tropical Research Institute, NSF grant DEB to BLT, and by a cooperative agreement between the University of Maryland and the Pacific Southwest Research Station, U.S. Department of Agriculture. REFERENCES CITED ANDREWS, E. A Observations on termites in Jamaica. J. Anim. Behav. 1: ARAUJO, R. L Catalogo dos Isoptera do Novo Mundo. Academia Brasileira de Ciencias, Rio de Janeiro, 92 pp. BECKER, V. G., AND K. SEIFERT Ueber die chemische zusammensetzung des nest - und galeriematerials von termiten. Ins. Soc. 9: BOUILLON, A Termites of the Ethiopian Region, pp in K. Krishna and F. M. Weesner [eds.] Biology of Termites, Volume II. Academic Press, N.Y. EMERSON, A. E The termites of Kartabo, Bartica District, British Guiana. Zoologica 6: EMERSON, A. E Termite nests. A study of the phylogeny of behavior. Ecol. Monogr. 8: GALLAHER, R. N., C. O. WELDON, AND J. G. FUTRAL An aluminum block digester for plant and soil analysis. Soil Sci. Soc. American Proc. 39: GOERING, H. K., AND P. G. VAN SOEST Forage fiber analyses (Apparatus, reagents, procedures, and some applications). Agric. Handbook 379. U.S.D.A., Washington, D.C. GOLDING, E. J., M. F. CARTER, AND J. E. MOORE Modification of the neutral detergent fiber procedure for hays. J. Dairy Sci. 68: GRASSE, P. P., AND C. NOIROT Nouvelles recherches sur la biologie divers Termites champignonnistes (Macrotermitinae). Ann. Sci. Nat. Zool. Biol. Animale [11] 13: HAMBLETON, L. G Semiautomated method for simultaneous determination of phosphorus, calcium and crude protein in animal feeds. J. Assoc. Off. Agric. Chem. 60: HAVERTY, M. I., B. L. THORNE, AND M. PAGE Surface hydrocarbon components of two species of Nasutitermes from Trinidad. J. Chem. Ecol. 16: HEGH, E Les Termites. Imprimerie Industrielle & Financiere, Bruxelles. HUBBARD, H. G Notes on the tree nests of termites in Jamaica. Proc. Boston Soc. Nat. Hist. 19: KEMNER, N. A Ans der Biologie der Termiten Java. 10th Congr. Int. Zool., Budapest, 2: LIGHT, S. F Termites of Western Mexico. Univ. of California Publ. in Entomol. 6: MARTORELL, L. F A survey of the forest insects of Puerto Rico. Part II. The Journal of Agriculture of the University of Puerto Rico. The Agricultural Experiment Station, Rio Piedras, P.R. 29: MOORE, J. E., AND G. O. MOTT Recovery of residual organic matter from in vitro digestion of forages. J. Dairy Sci. 57: NOIROT, C Le nid de Globitermes sulphureus Haviland au Cambodge. Ins. Soc. 6: NOIROT, C The nests of termites, pp in K. Krishna and F. M. Weesner [eds.] The Biology of Termites, Volume II. Academic Press, N.Y. PARR INSTRUMENT CO Oxygen bomb calorimetry and combustion methods. Technical Manual Parr Instrument Company 130: THORNE, B. L Differences in nest architecture between the Neotropical arboreal termites Nasutitermes corniger and Nasutitermes ephratae (Isoptera: Termitidae). Psyche 87:

38 Thorne et al.: Nasutitermes Nests and Nodules 37 THORNE, B. L Polygyny in the Neotropical termite Nasutitermes corniger: life history consequences of queen mutualism. Behav. Ecol. Sociobiol. 14: THORNE, B. L., M. I. HAVERTY, AND M. S. COLLINS Taxonomy and biogeography of Nasutitermes acajutlae and N. nigriceps (Isoptera: Termitidae) in the Caribbean and Central America. Ann. Entomol. Soc. America 87: TILLEY, J. M. A., AND R. A. TERRY A two-stage technique for the in vitro digestion of forage crops. J. British Grassl. Soc. 18:

39 Heath et al.: Papaya Fruit Fly Trapping System 37 IMPROVED PHEROMONE-BASED TRAPPING SYSTEMS TO MONITOR TOXOTRYPANA CURVICAUDA (DIPTERA: TEPHRITIDAE) R. R. HEATH 1, N. D. EPSKY 1, A. JIMENEZ 2, B. D. DUEBEN, P. J. LANDOLT, W. L. MEYER 3, M. ALUJA 4, J. RIZZO 5, M. CAMINO 2, F. JERONIMO 5 AND R. M. BARANOWSKI 3 1 Insect Attractants, Behavior, and Basic Biology Research Laboratory, Agricultural Research Service, United States Department of Agriculture, Gainesville, FL Centro de Desarrollo de Productos Bióticos, Km. 8.5 Carretera Yautepec-Jojutla, Yautepec, Morelos, Mexico 3 University of Florida, IFAS, Tropical Research and Education, Homestead, Florida 4 Instituto de Ecología, A.C., Apartado Postal 63, Xalapa, Veracruz, Mexico 5 APHIS-PPQ, United States Embassy Guatemala, APO Miami, Florida ABSTRACT A membrane-based formulation method that provided a constant release rate of synthetic pheromone for the papaya fruit fly, Toxotrypana curvicauda Gerstaecker, was developed. Release rate measurements over 23 days indicated that lures loaded with 5, 15, 25, and 50 µl of synthetic pheromone released an average of 120, 360, 580 and 1120 ng per hr and the half-life of the lures was estimated to be 67, 184, 300 and 48 days, respectively. Field tests conducted in Mexico compared efficacy of blank and pheromone-baited sticky green spheres, cylindrical traps made from green opaque plastic that either contained a toxicant or were coated with sticky material, and cylindrical traps prepared from green sticky paper. Green opaque traps containing a toxicant and sticky paper traps captured approximately five times more papaya fruit flies than either the sticky-coated green opaque traps or the sticky-coated green spheres, and the presence of pheromone did not affect numbers of flies captured. Thus, the combination of the green color and the cylindrical shape provided a visual cue sufficient for papaya fruit fly capture. The pheromone lure significantly increased trap capture in similar tests conducted in Guatemala. Capture was highest in the sticky paper traps and in sticky-coated spheres. Use of the membrane-based synthetic pheromone in a cylindrical trap may provide an effective tool for monitoring papaya fruit flies. Key Words: Insecta, papaya fruit fly, pheromone formulation, trap.

40 38 Florida Entomologist 79(1) March, 1996 RESUMEN Fue desarrollado un método de formulación basado en una membrana que provee una velocidad constante de liberación de la feromona sintética para la mosca frutera de la papaya, Toxotrypana curvicauda Gerstaecker. Las medidas de la velocidad de liberación durante 23 días indicaron que los cebos cargados con 5, 15, 25 y 50 µl de la feromona sintética liberaron un promedio de 120, 360, 580 y 1120 ng por hora y la vida media de los cebos fue estimada en 67, 184, 300 y 48 días, respectivamente. En experimentos de campo llevados a cabo en México se comparó la eficacia de esferas verdes adhesivas con y sin feromona, trampas cilíndricas hechas con plástico verde opaco que contenían un tóxico o estaban cubiertas con material adhesivo, y trampas cilíndricas hechas de papel adhesivo verde. Las trampas verdes opacas que contenían un tóxico y las trampas de papel adhesivo verde capturaron aproximadamente cinco veces más moscas de la papaya que las trampas verdes opacas cubiertas de goma o las esferas verdes cubiertas de goma. La presencia de la feromona no afectó el número de moscas capturado. La combinación del color verde y la forma cilíndrica proveyeron una pista visual suficiente para la captura de las moscas fruteras. El cebo de feromona aumentó significavivamente la captura en pruebas similares conducidas en Guatemala. La captura fue más alta en las trampas de papel adhesivo y en las esferas cubiertas de goma. El uso de la feromona sintética con el método de la membrana en la trampa cilíndrica podría ser una herramienta efectiva para el monitoreo de las moscas fruteras de la papaya. Toxotrypana curvicauda Gerstaecker is the principal insect pest of commercial papaya (Carica papaya L.); it occurs throughout the tropical and subtropical areas of the New World (Wolcott 1933). Studies of papaya fruit fly behavior in the field and in the laboratory indicated the existence of a male-produced pheromone that mediates interactions with females (Landolt & Hendrichs 1983, Landolt et al. 1985). This male-produced pheromone was identified by Chuman et al. (1987) as 2-methyl-6-vinylpyrazine (2,6-MVP). Synthetic 2,6-MVP elicited the same behavioral responses from sexually mature unmated female papaya fruit flies as did male-produced pheromone. Field trials demonstrated that the pheromone used with 12.7 cm diam sticky-coated green spheres as visual fruit cues (Landolt et al. 1988) and a release rate of approximately 1 µg per hr (12 male equivalents) was optimal for papaya fruit fly attraction (Landolt & Heath 1990). However, general field use of this trapping system has been limited because of the difficulty in keeping the fragile, open glass capillaries, in which the highly volatile pheromone is formulated, vertical in order to avoid spillage. Furthermore, the short life span of the adhesive necessitates continuous trap maintenance. Cylindrical traps, which were painted to provide a visual cue and were baited with a food-based synthetic attractant, captured males and females of the Mediterranean fruit fly, Ceratitis capitata (Wiedemann), and a number of Anastrepha spp. (Heath et al. 1995). Cylindrical traps with appropriate visual cues may provide an alternative to the spheres currently used for papaya fruit fly monitoring. Herein, we report the application of a membrane-based formulation system that provides a variety of release rates depending on the amount of chemical used. Field tests were conducted to determine if the membrane-release formulation and green cylindrical traps could be used to overcome the shortcomings of the earlier trapping system.

41 Heath et al.: Papaya Fruit Fly Trapping System 39 MATERIALS AND METHODS Membrane-Based Formulation System A 3 by 5 cm lure was prepared by folding a 6 by 5 cm piece of 6 mill impermeable polyethylene (backing) in half. A 1.17 cm diam hole was cut in the center of the front of the lure and a piece of 1 mill high density polyethylene (membrane) film (Consep Inc., Bend, OR) was placed inside the lure. The bottom and sides were heat sealed to form an envelope and to secure the membrane. The release area of the membrane was reduced to a 5 mm diam circle by placing a piece of aluminum tape (United Tape Company, Cumming, GA) over the 1.17 cm hole in the front of the lure. A piece of rectangular filter paper and a plastic grid that were slightly less than 3 by 5 cm were placed in the envelope to provide stability; synthetic pheromone was added and the top of the lure was heat sealed. Lures were loaded with 5, 15, 25 and 50 µl of >99% pure 2,6- MVP (Fuji Flavour Co. LTD., Tokyo, Japan). Lures were placed in a hood with a 0.25 cm per sec air flow and, beginning two days after filling, the release rates from at least three lures of each load were measured every three to four days over a 23-day period. Volatiles were collected and analyzed using collection systems and gas chromatographic conditions described previously (Heath & Manukian 1992, Heath et al. 1993). Mean release rates for each pheromone load were used in linear regression analysis to determine the change in release rate over time and the half-life of each lure. A field-test comparison of the membrane-based lures and the capillary lures used by Landolt & Heath (1990) was conducted in Homestead, Florida, using sticky-coated sphere traps. Both lures released approximately 1 µg 2,6-MVP per h. Solid styrofoam spheres (12.7 cm diam) painted dark green were used (Great Lakes IPM, Vestaburg, MI). Capillary lures were mounted in holes drilled in the spheres and lures were positioned so that lure openings were 2-3 cm from the top of the sphere. A 4 by 7 cm piece of 6 mill plastic (Faulkner Plastics, Gainesville, FL) was folded in half to form a tent. The membrane-based lure was attached to the inside of the tent to protect the lures from rain, and this assembly was placed on the top of the sphere (Fig. 1a). Paintable sticky coating (Tangle Trap, The Tanglefoot Co., Grand Rapids, MI) was applied to the outside of the sphere to retain responding flies. Ten pairs of sticky spheres baited with either a capillary lure or a membrane-based lure were hung by wires from papaya leaf petioles near fruit clusters located about m above ground in trees along the outside edge of a papaya grove. Traps were placed in the border rows because the papaya fruit fly activity tends to be the highest and fruit damage the greatest along the borders of a grove (Landolt & Heath 1988). The experiment was replicated over time; traps were checked every 7-10 days per replicate. There were four consecutive replicates. Total numbers of males and females captured on the ten traps per lure type were summed separately to give the number of each sex captured per replicate. Separate t-test comparisons were conducted for numbers of males and females captured using Proc TTEST (SAS Institute 1985). Cylindrical Traps Cylindrical traps (Fig. 1b,c) consisted of three major components; the main trap body, which was a cylindrical container (9 cm diam by 15 cm long); two removable end caps for quick access into the trap for bait replacement; and a wire hanger for holding the trap together and supporting the complete assembly on a tree (Heath et al. 1995). The trap bodies for the green opaque traps (Fig. 1b) were prepared from a rectangular

42 40 Florida Entomologist 79(1) March, 1996 piece (15.0 cm wide 30.0 cm long) of green opaque plastic (0.025 cm thick, Faulkner Plastics, Gainesville, FL), which was rolled to form a 9.0 cm diam cylinder. The green opaque traps used one of two methods for catching insects that responded to the trap and/or bait. One type used paintable sticky coating, as was used for the sphere traps, applied to the outside of the trap. The second type used internally-placed toxicant panels (Heath et al. 1995), which contained a feeding stimulant and a toxin, to kill insects after they entered the trap. The toxicant panels are coated with a solution (1.0: 0.5: 0.01) of paint [Hunter Green 100% acrylic latex paint (Glidden, Cleveland, OH)], sugar [American Chemical Society grade sucrose (Mallinkrodt, Paris, KY)], and pesticide [technical grade methomyl (DuPont, Newark, DE); 98%(AI)]. Panels were air dried for at least 48 h before use and were placed on the inside of both the top and bottom end caps using double-sided tape. Sticky-paper traps were made from dark green fruit fly adhesive paper (FFAP) supplied by the Atlantic Paste and Glue Co., Inc. (Brooklyn, NY). The efficacy of the adhesive paper was determined in laboratory tests of the retention of alighting flies. Flies were obtained as mature larvae from field-collected papaya fruit in Dade County, Florida. Larvae exited the fruit and pupated in vermiculite (Landolt & Heath 1988). In tests conducted in a greenhouse, a piece of FFAP (12 by 12 cm) was placed in a screen cage (30 by 30 by 30 cm) containing 30 sexually mature females. Males were not tested because females are the primary target of the traps. An observer counted the number of fly landings on the sticky paper over a one-h time period. Total Figure 1. Illustration of three types of papaya fruit fly traps used in field studies conducted in North and Central America. A) Solid styrofoam spheres were painted with dark green paint and coated with paintable adhesive to retain flies. B) Opaque cylindrical traps were constructed using green plastic. This trap either contained toxicant panels or was coated with paintable adhesive to retain attracted flies. C) Stickypaper cylindrical traps were similar to the opaque cylindrical traps, but the trap body was made with dark green fruit fly adhesive paper. The adhesive paper was supplied with protective paper (shown as white overlay) that was removed when traps were placed in the field to expose the sticky trap surface.

43 Heath et al.: Papaya Fruit Fly Trapping System 41 numbers of flies remaining on the paper after the one-h time period were counted, and ratio of landed to captured flies was determined. The test was replicated three times over a period of three days. The sticky-paper traps (Fig. 1c) were prepared similarly to the green opaque traps. A rain guard was used with this trap to protect the paper trap body; it was made from the top half of a mm petri dish (P/N# 1058, Becton Dickinson, Lincoln Park, NJ). A 5 cm length of polyvinyl chloride tubing (9.0 cm outside diam, Hughs Supply, Gainesville, FL) was glued to the center underside of the petri dish, which provided a holder for the trap body. The protective paper supplied with the adhesive paper was removed when traps were placed in the field to expose the sticky trap surface. Field Tests A study was conducted at a papaya plantation on the grounds of the Centro de Desarrollo de Productos Bióticos (CEPROBI) of the Instituto Politécnico Nacional (IPN), Morelos, Mexico, from November 1994 through January 1995, the dry season during the coolest time of the year. The native vegetation adjacent to the study site is classified as selva baja caducifolia or lowland deciduous forest (Soria 1985). Gonolobus sorodius (Asclepiadaceae) and Jacaratia mexicana (Caricaceae), which are native hosts for the papaya fruit fly (Castrejon-Ayala & Camino-Lavin 1991), occur in the area. The papaya plantation had two rows of papaya that serve as a trap crop and are separated by 10 m from the main plot. The experimental plot was 177 by 63 m, with papaya trees planted every 3 m. Treatments consisted of four trap types: 1) sticky-coated green spheres, 2) sticky-coated green opaque cylinders, 3) green opaque cylinders with internally-placed toxicant panels, and 4) sticky-paper cylinders made with the dark green FFAP material. These traps were either baited with membrane-based pheromone lures, which released about 1 µg 2,6-MVP per h, or were unbaited. Lures were taped to the inside of the opaque cylinders or attached to a plastic rain tent of sticky-paper traps using double-sided sticky tape. Traps were placed near the immature fruit within the papaya tree because papaya fruit fly females successfully oviposit in those fruit (Landolt 1985). The eight treatments (four trap types times two bait treatments) were placed randomly within a block and treatments were moved sequentially at time of sampling. There were six blocks placed in the trap crop, i.e., around the periphery of the papaya orchard. Traps were placed in every other tree within a block, and there were 8-19 papaya trees without traps between the experimental blocks. The sticky-paper trap bodies were replaced weekly, the sticky-coated traps were cleaned and recoated as needed. Pheromone lures were replaced after four weeks of field use. The numbers of males and females captured per trap were recorded weekly for six weeks (replicates) and then after two weeks for the final sample, for a total of seven replicates. Similar field tests were conducted in a papaya orchard at finca Cuilapa, Guatemala. However, liquid protein-baited McPhail traps, bell-shaped glass traps with a water reservoir (Newell 1936), were added because of local interest in using this trap for papaya fruit flies. McPhail traps were baited with five torula yeast-borax pellets (ERA Int., Freeport, NY) in 300 ml of water (Gilbert et al. 1984). Traps were placed in papaya trees located around the periphery of the papaya orchard, with individual traps placed near the smaller fruit within the tree, as above. All nine treatments (four trap types, two bait treatments, plus McPhail traps) were placed randomly within a block and treatments were moved sequentially at time of sampling. There were four blocks placed in the periphery of the papaya orchard. There were at least ten papaya trees without traps that separated the experimental blocks, and traps within a block

44 42 Florida Entomologist 79(1) March, 1996 were placed in every third tree. Traps were sampled weekly. Sticky-coated traps were cleaned and recoated as needed, sticky-paper trap bodies were replaced weekly. Pheromone lures were replaced after four weeks of field use. McPhail traps were cleaned and protein solutions were replaced every two weeks. The numbers of males and females captured per trap were recorded weekly for eight weeks, for a total of eight replicates. Statistical Analysis. The total number of trapped flies per treatment was determined from the sum of each sex collected per replicate. Thus, one replicate consisted of the sum total captured in six traps (Mexico) or four traps (Guatemala) per trap type. Sum total was converted to percentage trapped within replicate for statistical analysis. Data were analyzed with two-way analysis of variance (ANOVA) with interaction using Proc GLM (SAS Institute 1985) followed by LSD mean separation tests (P = 0.05). Factors used in the ANOVA were bait (2 levels: pheromone baited or no pheromone; data from McPhail traps were not included in this analysis) and trap type (4 levels: sticky sphere, sticky green opaque, green opaque with toxicant or sticky-paper trap). One-way ANOVAs were conducted to test all nine trap type/bait combinations in the Guatemalan tests. Data were log-transformed (x + 1) to stabilize the variance prior to analysis (Box et al. 1978). Sex of the trapped flies was considered to be a response variable, not a treatment factor, therefore separate analyses were conducted for females, males and total (females plus males) papaya fruit flies captured in each country. RESULTS Lures loaded with 5, 15, 25, and 50 ml released an average (± std) of 0.12 ± 0.01, 0.36 ± 0.02, 0.58 ± 0.02 and 1.12 ± 0.11 µg per h during the 23 days that release rates were obtained (n = 16, 16, 16 and 15, respectively). The projected half-lives of the lures were 67, 184, 300 and 48 days, respectively. Lures prepared with 5, 15, or 25 µl of material showed very little decrease during the time that release rates were measured (Fig. 2). To cover the optimum range of pheromone release, we selected the 50 µl loaded pheromone lure, which released an average of about 1 µg per hr, for subsequent field experiments. There were no significant differences in catch (average ± std) on sticky spheres baited with capillaries or the membrane-based lure of either males (2.0 ± 3.4 versus 2.5 ± 0.6) or females (7.8 ± 4.5 versus 8.3 ± 3.3). Number of papaya fruit flies captured in Mexico was low, with an average of 15 flies captured among all six blocks (48 traps total) per replicate. Trap type significantly affected capture of females (F = 7.41; df = 3, 48; P = ) and total flies (F = 3.79; df = 3, 48; P = ), but not males (F = 1.39; df = 3, 48; P = 0.26). A higher percentage of females was captured on the green opaque traps with toxicant and on the sticky-paper traps than on the traps that used paintable sticky coating (sphere-sticky, opaque-sticky). Presence of the pheromone lure did not affect capture of the papaya fruit flies, although pheromone-baited traps usually captured slightly more flies than their unbaited counterparts (Fig. 3). Capture of papaya fruit flies was also low in the tests conducted in Guatemala, with an average of 17 flies captured among all four blocks (36 traps total) for the first 5 weeks of the study. Numbers trapped dropped to 0-3 for the last 3 weeks, therefore these data were deleted and only the first five replicates were included in the analyses. Presence of pheromone lure and trap type significantly affected capture of females, males and totals, and there was no interaction between these factors. Percentage of papaya fruit flies captured in pheromone-baited traps was always higher than capture in the same traps without pheromone (Fig. 4). Percentage of pa-

45 Heath et al.: Papaya Fruit Fly Trapping System 43 paya fruit flies captured in the liquid protein-baited McPhail traps was no higher than that in the least effective traps tested. No females were captured in the sticky-coated green opaque traps, and percentage captured in these traps was lower than in the other three trap types. For both males and total flies, percentage captures in the sticky-paper traps and the sticky-coated spheres were higher than in either of the green opaque traps. In laboratory trials, the FFAP material used for the sticky-paper traps was highly effective in retaining flies that landed on the surface. In the three tests, nine, six and four flies landed on the paper and all (19 out of 19) were captured on the sticky surface. DISCUSSION These studies demonstrated that a membrane-based lure can be used to formulate synthetic papaya fruit fly sex pheromone for field use, and that cylindrical traps may be used to replace the sphere traps used previously. The cylindrical traps apparently provided a sufficient visual cue because the unbaited cylindrical traps also captured papaya fruit flies. Landolt & Heath (1990) found that although mated females responded to the visual aspects of the spheres alone, virgin females represented most of the females captured in response to pheromone. Number of males also increased directly with pheromone dose, indicating that males responded to the pheromone presence. It is not known why there was no effect of pheromone in the trials in Mexico as Figure 2. Average release rates (µg per h) of papaya fruit fly synthetic pheromone which was formulated in membrane-based lures, over time. Regressions were determined from lures (n=3) containing 5, 15, 25 and 50 µl of 2-methyl-6-vinylpyrazine.

46 44 Florida Entomologist 79(1) March, 1996 Figure 3. Results of trap capture of female (top), male (middle) and total (bottom) papaya fruit flies captured in field trials conducted in Mexico (n =7). Traps tested (left to right) were green sticky spheres, green opaque cylinders with sticky exterior, green opaque cylinders with internally-placed toxicant, and green sticky-paper cylinders. Traps were either baited with papaya fruit fly pheromone membrane-based lures (shaded bars) or left unbaited (open bars). Pairs of bars within a graph headed by the same letter are not significantly different (LSD mean separation test on log (x + 1) transformed data, P = 0.05; non-transformed means presented).

47 Heath et al.: Papaya Fruit Fly Trapping System 45 Figure 4. Results of trap capture of female (top), male (middle) and total (bottom) papaya fruit flies captured in field trials conducted in Guatemala (n = 5). Traps tested (left to right) were green sticky spheres, green opaque cylinders with sticky exterior, McPhail traps, green opaque cylinders with internally-placed toxicant, and green sticky-paper cylinders. Except for the McPhail traps, which were baited with aqueous torula yeast plus borax solution, the traps were either baited with papaya fruit fly pheromone membrane-based lures (shaded bars) or left unbaited (open bars). Bars within a graph headed by the same letter are not significantly different [LSD mean separation test on log (x + 1) transformed data, P = 0.05; non-transformed means presented].

48 46 Florida Entomologist 79(1) March, 1996 was observed in the trials in Guatemala. A number of factors, including the demographics of the fly population (e.g. presence or absence of virgin females in the population) and environmental parameters at each site, may have affected the results. The mated status of the females captured was not determined in our studies. The membrane-based lure can be formulated to provide a multitude of release rates, and release rates were directly related to amount of 2,6-MVP added to the lure. The lure used in trials reported herein was based on optimal release rate for stickycoated spheres (Landolt & Heath 1990). It is not known if this release rate is optimal for traps other than sticky spheres. A lower release rate may improve catch in the opaque cylindrical traps containing the toxicant panel because this trap requires that the flies enter the trap to be captured. Additional studies are needed to examine a range of pheromone doses for use in cylindrical traps that use toxicant or sticky materials. The use of the FFAP material provided a facile method to prepare cylindrical traps. This material was easily fabricated and the sticky material did not adhere to personnel who contacted the adhesive. Exposure over time indicated that it was impervious to rain when the material was used as described. The adhesive material did not drip or run as was observed when paintable sticky coating was used. It should be noted, however, that in experiments conducted with sticky-paper traps for other tephritid pests in southern Florida, we observed that on occasion small birds and lizards were either blown into the trap or came in contact with the FFAP material. While the sticky-paper trap caught non-target insects in studies in both Mexico and Guatemala, we did not observe the capture of small animals such as was observed in Florida. The captured insects and animals can be remove from the FFAP material with the use of mineral oil. This process had no observed detrimental effect on the animals if the animals were removed shortly after capture. Another problem with this, as well as other sticky-coated traps, is that wind-borne dust or debris may coat the trap and limit the longevity of the adhesive in the field. The green opaque trap used with the toxicant provided an alternative to a sticky trap, although it may not be as sensitive under all environmental conditions. Papaya fruit flies were captured in protein-baited McPhail traps in this study in Guatemala, but in low numbers. This suggests that volatiles from protein baits may be attractive to papaya fruit flies. If so, addition of protein bait volatiles may provide synergists that could further improve trap capture of the pheromone-baited traps. It is envisioned that the information presented here may provide a facile method to monitor papaya fruit fly infestations. This would result in decreased pesticide application and, potentially, offer a method of control through trapping alone. ACKNOWLEDGMENTS We thank Charles Powell (USDA/ARS, Gainesville, FL), Sanjay Kulkarni, John Howell, Yasmin Cardoza (Univ. of Florida), Vicot Castrejón and Jaime Piñero (CEP- ROBI, Mexico) for their assistance; D. Martínez (Laboratorio de Botánica, Universidad Autónoma del Estado de Morelos, Cuernavaca, Mexico) for plant identification; and Peter E. A. Teal (USDA/ARS, Gainesville, FL), Jennifer L. Sharp (USDA/ARS, Miami, FL), Timothy C. Holler (USDA/APHIS, Gainesville, FL) and two anonymous reviewers for critical review and helpful comments on the manuscript. The authors thank Michael Bentivegna, Jr. and the Atlantic Paste and Glue Co., Inc. (Brooklyn, NY), for providing the fruit fly adhesive paper; Janice Gillespie, Edie Christensen and Consep Membranes, Inc. (Bend, OR) for providing materials for the lures. We thank

49 Heath et al.: Papaya Fruit Fly Trapping System 47 Gordon Tween and the staff at the USDA, APHIS-International Services, U.S. Embassy, Guatemala and Pedro Rendon and the staff at the USDA, APHIS-PPQ, Guatemala for their support of this research. Partial financial support was provided by the International Foundation for Science (Grant C/1741-1), CEPROBI and Instituto de Ecologia, A. C. (Proyecto Ecologia y Comportamiento Animal). This article reports the results of research only. Mention of a proprietary product does not constitute an endorsement or recommendation for its use by USDA. REFERENCES BOX, G. E. P., W. G. HUNTER, AND J. S. HUNTER Statistics for experimenters. An introduction to design, data analysis, and model building. J. Wiley & Sons, New York, New York. CASTREJON-AYALA, F., AND M. CAMINO-LAVIN New host plant record for Toxotrypana curvicauda (Diptera: Tephritidae). Florida Entomol. 74: 466. CHUMAN, T., P. J. LANDOLT, R. R. HEATH, AND J. H. TUMLINSON Isolation, identification, and synthesis of male-produced sex pheromone of papaya fruit fly, Toxotrypana curvicauda Gerstaecker (Diptera: Tephritidae). J. Chem. Ecol. 13: GILBERT, A. J., R. R. BINGHAM, M. A. NICOLAS, AND R. A. CLARK Insect trapping guide. Pest Detection/Emergency Projects, State of California Department of Food and Agriculture, Sacramento, California. HARRIS, E. J., S. NAKAGAWA, AND T. URAGO Sticky traps for detection and survey of three tephritids. J. Econ. Entomol. 64: HEATH, R. R., AND A. MANUKIAN Development and evaluation of systems to collect volatile semiochemicals from insects and plants using a charcoal-infused medium for air purification. J. Chem. Ecol. 18: HEATH, R. R., N. D. EPSKY, P. J. LANDOLT, AND J. SIVINSKI Development of attractants for monitoring Caribbean fruit flies (Diptera: Tephritidae). Florida Entomol. 76: HEATH, R. R., N. D. EPSKY, A. GUZMAN, B. D. DUEBEN, A. MANUKIAN, AND W. L. MEYER Development of a dry plastic insect trap with food-based synthetic attractant for the Mediterranean and the Mexican fruit fly (Diptera: Tephritidae). J. Econ. Entomol. 88: LANDOLT, P. J Papaya fruit fly eggs and larvae (Diptera: Tephritidae) in fieldcollected papaya fruit. Florida Entomol. 68: LANDOLT, P. J., AND R. R. HEATH Effects of age, mating, and time of day on behavioral responses of female papaya fruit fly, Toxotrypana curvicauda Gerstaecker (Diptera: Tephritidae) on synthetic sex pheromone. Environ. Entomol. 17: LANDOLT, P. J., AND R. R. HEATH Effects of pheromone release rate and time of day on catches of male and female papaya fruit flies (Diptera: Tephritidae) on fruit model traps baited with pheromone. J. Econ. Entomol. 85: LANDOLT, P. J., AND J. HENDRICHS Reproductive behavior of the papaya fruit fly, Toxotrypana curvicauda Gerstaecker (Diptera: Tephritidae). Ann. Entomol. Soc. America 76: LANDOLT, P. J., R. R. HEATH, AND J. R. KING Behavioral responses of female papaya fruit flies, Toxotrypana curvicauda Gerstaecker (Diptera: Tephritidae) to male-produced sex pheromone. Ann. Entomol. Soc. America 78: LANDOLT, P. J., R. R. HEATH, H. R. AGEE, J. H. TUMLINSON, AND C. O. CALKINS A sex pheromone-based trapping system for the papaya fruit fly, Toxotrypana curvicauda Gerstaecker (Diptera: Tephritidae). J. Econ. Entomol. 81: NEWELL, W Progress report on the Key West (Florida) fruit fly eradication project. J. Econ. Entomol. 29:

50 48 Florida Entomologist 79(1) March, 1996 SAS INSTITUTE SAS/STAT guide for personal computers, version 6 edition. SAS Institute, Cary, North Carolina. SORIA, R. G Flora de Morelos. Descripción de especies vegetales de la selva baja caducifolia del Cañon de Lobos, Mpio. de Yautepec; Serie Ciencias Naturales y de la Salud. Programa Florístico-Faunístico. Universidad Autónoma del Estado de Morelos. Cuernavaca, Morelos, Mexico. 165 p. WOLCOTT, G. N An economic entomology of the West Indies. Clay and Sons, Bungay, Suffolk.

51 48 Florida Entomologist 79(1) March, 1996 DISTINGUISHING FALL ARMYWORM (LEPIDOPTERA: NOCTUIDAE) STRAINS USING A DIAGNOSTIC MITOCHONDRIAL DNA MARKER YANGJIANG LU AND MICHAEL J. ADANG Department of Entomology, University of Georgia Athens, GA ABSTRACT The fall armyworm, Spodoptera frugiperda (J. E. Smith), includes morphologically indistinguishable corn and rice strains. The two strains were surveyed for diagnostic restriction patterns in mitochondrial DNA (mtdna) using 25 restriction endonucleases. Polymorphic mtdna restriction patterns were identified for BstNI, HinfI and MspI. The MspI pattern was the most distinctive since the molecular size of each DNA fragment differed between the two strains. Analyses of laboratory and field-collected insects showed the MspI mtdna pattern to be a diagnostic marker for corn and rice strain insects. Strain identification by the MspI mtdna profile correlated exactly with nuclear DNA markers. Since no HaeIII sites are present in fall armyworm mtdna, a double-digest of total fall armyworm DNA using HaeIII and MspI allowed the direct detection of mtdna restriction fragments from total DNA on a stained agarose gel. In contrast to conventional techniques utilizing mtdna markers, this rapid and simple procedure does not require the isolation of mtdna, or avoids the use of DNA blots and labeled mtdna. Key Words: Spodoptera, fall armyworm, mitochondrial DNA, strain identification. RESUMEN El gusano trozador, Spodoptera frugiperda (J. E. Smith), posee cepas de maíz y arroz morfológicamente indistiguibles. Esas dos cepas fueron muestreadas para el diagnóstico mediante patrones de restricción en DNA mitocondrial (mtdna), utilizando 25 endonucleasas de restricción. Los patrones polimórficos del mtdna de restricción fueron identificados para BstNI, HinfI y MspI. El patrón MspI fue el más distintivo debido a que el tamaño molecular de cada fragmento de DNA difirió entre las dos cepas. Los análisis de insectos de laboratorio y colectados en el campo mostraron que el patrón del MspI mtdna es un marcador diagnóstico para los insectos de las cepas de maíz y arroz. La identificación del perfil del MspI mtdna se correlacionó exactamente con los marcadores nucleares de DNA. Debido a la ausencia de sitios HaeIII en el mtdna del gusano trozador, una doble digestión del total del DNA usando HaeIII y MspI permitió la detección de los fragmentos de restricción del

52 Lu & Adang: Fall Armyworm Strain Identification 49 mtdna a partir del DNA total en un gel de agarosa teñido. En contraste con las técnicas convencionales que utilizan marcadores de mtdna, este proceso rápido y simple no requiere del aislamiento del mtdna, o evita el uso de blots de DNA y de mtdna marcado. The fall armyworm, Spodoptera frugiperda (J. E. Smith), is a major pest on corn (Zea mays L.), sorghum (Sorghum vulgare pers.) and bermudagrass (Cynodon dactylon pers.) in the southeastern United States. The insect is an occasional pest on many other crops, including cotton (Gossypium hirsutum L.), peanut (Arachis hypogea L.), millet (Pennisetum glaucum Pers.), alfalfa (Medicago sativa L.), rye (Secale cereale L.), rice (Oryza sativa L.) and soybean (Glycine max Merr.) (Sparks 1979). Difficulties in the control of fall armyworms have been attributed to its wide range of host plants, vast geographical distribution, and rapid and long distance movement (Knipling 1980). Pashley (1986) discovered by allozyme analyses that fall armyworm populations consist of two host strains. The corn-strain prefers corn, sorghum and cotton, while the rice-strain prefers rice and bermudagrass. Subsequent investigations of the nuclear and mitochondrial genomes using restriction fragment length polymorphism (RFLP) techniques revealed significant differences between the two strains (Pashley 1989, Lu et al. 1992). Further evidence for the genetic separation of the strains was the discovery of a unique repeated DNA sequence in the genome of rice strain insects (Lu et al. 1994). Reproductive isolation mechanisms exist between the two fall armyworm strains (Pashley et al. 1987; Pashley et al. 1992). Genetic separation and barriers to interbreeding support the species status for the fall armyworm strains (Pashley et al. 1992). The practical impact of the sympatric strains is not clear, but several studies suggest the strain diversity complicates pest management. Quisenberry & Whitford (1988) demonstrated that bermudagrass bred for resistance to the corn strain insects was still susceptible to rice strain. As efforts are directed towards developing fall armyworm-resistant corn (Wiseman & Davis 1979, Williams et al. 1989, Wiseman & Isenhour 1988), the use of characterized fall armyworm strains may be crucial. Other biological differences between the strains, including dispersal pattern (Pashley et al. 1992) and response to pesticides (Pashley et al. 1988), may influence population monitoring studies (Barfield et al. 1980, Pair et al. 1986) and pest control strategies. The major objective of this study was to identify restriction enzyme patterns for mtdna that distinguish the corn and rice strain insects. Pashley (1989) reported differences in the number of restriction fragments for several restriction enzymes. We tested additional enzymes and report that MspI digested mtdna reveals a diagnostic pattern for the strains. This MspI pattern is directly detected on agarose gels when a second endonuclease HaeIII is included to digest nuclear DNA. Due to its simplicity this MspI/HaeIII co-digestion method will be useful in the identification of fall armyworm strains. MATERIALS AND METHODS Insect Sources Both laboratory-reared and field-collected fall armyworms were used. Sources described in Lu et al. (1992) include the corn strain (designated as C) and the rice strain

53 50 Florida Entomologist 79(1) March, 1996 (R) colonies from Louisiana State University (a gift from S. Quisenberry); a colony (M) from the USDA-ARS laboratory at Mississippi State University (a gift from F. Davis); and colony (I) formerly maintained by D. Isenhour at the University of Georgia, Coastal Plains Experimental Station. Population P was collected from corn plants in Tifton, Georgia, in Other insect sources included a laboratory colony established with bermudagrass-collected fall armyworms (provided by R. Mcpherson, USDA Tifton, Georgia) and fall armyworms collected from 2 sorghum and 5 corn fields near Athens, Georgia in Total DNA Isolation Total DNA was isolated from individual fifth instar larvae as described in Lu et al. (1992). Isolated DNA samples in TE buffer (10 mm Tris, 1 mm EDTA, ph 8.0) were stored at -20 C until needed. Mitochondrial DNA Isolation and 32 P-labeling Fifteen larvae (4.2 g) were homogenized with about 20 strokes in a Dounce tissue homogenizer with 20 ml of homogenization buffer [200 mm mannitol, 70 mm sucrose, 500 mm Tris-HCl ph 7.5, 100 mm EDTA, and 0.2 mg/ml proteinase K (Sigma, St. Louis, MO) (Zehnder et al. 1992)]. The homogenate was centrifuged at 1,600 g for 10 min to pellet the cellular debris. The supernatant was centrifuged again at 17,000 g for 30 min to pellet mitochondria. The mitochondrial pellet was washed once by resuspending the pellet with 20 ml of homogenization buffer, centrifuged at 1,600 g for 10 min, and the supernatant decanted and centrifuged at 40,000 g for 15 min. The final mitochondrial pellet was resuspended in 3 ml of buffer containing 100 mm NaCl, 500 mm Tris-HCl ph 8.0, 10 mm EDTA ph 8.0, and then mixed with ml of 10% sodium dodecyl sulfate (SDS). To the above mixture, 6.8 g CsCl and 0.25 ml of ethidium bromide solution (10 mg/ml) were added, and the total volume was adjusted to 7.15 ml with TE buffer. The mixture was centrifuged at 161,000 g for 20 h. The DNA bands were visualized under UV light and the lower mtdna band was removed. Isolated mtdna was extracted with 1-butanol to remove ethidium bromide according to procedures described (Sambrook et al. 1989). Purified mtdna was labeled with 32 P- dctp using the random primer procedure (Feinberg & Vogelstein 1984), and labeled mtdna was used as a probe to hybridize with restriction digested total DNA. Restriction Digestion and Electrophoresis Complete digestion of total DNA was carried out by using 2 units of enzyme (Boehringer-Mannheim Laboratories, Indianapolis, Ind.) per µg of DNA in the supplied buffer at 37 C for 16 h. In the case of double digestion, the same enzyme/dna ratio was used for each enzyme. Digested DNA was electrophoretically fractionated in a 1% agarose gel (Bio-Rad, Richmond, Calif) in TAE buffer (40 mm Tris-acetate, 1 mm EDTA, ph 7.5). After electrophoresis, the agarose gel was stained with ethidium bromide and DNA visualized under UV. Southern Blot Procedures Following electrophoresis, DNA was transferred to nylon membranes (GeneScreen Plus, DuPont) by the method of Southern (1975). DNA was immobilized on membranes by baking at 80 C for 2 h.

54 Lu & Adang: Fall Armyworm Strain Identification 51 Hybridizations were carried out as follows: membranes were prehybridized with sheared salmon sperm DNA (100 g/ml) in hybridization buffer consisting of 3X SSC (1X SSC=0.15 M NaCl, M Na citrate) containing 0.1% SDS at 65 C for 6 h. Labeled mtdna was denatured by boiling for 20 min and then added to the prehybridized filter in hybridization buffer. Hybridization was for at least 6 h at 65 C. Filters were washed once in 2X SSC, 0.1% SDS for 20 min and twice in 1X SSC, 0.1% SDS for 20 min at 65 C. Filters were air-dried and exposed to X-ray film (Kodak) at -80 C with intensifying screens. RESULTS We screened for polymorphic mtdna restriction patterns between the corn and rice strains of fall armyworm by digesting total DNA from both strains with 25 restriction enzymes followed by probing with 32 P-labeled mtdna. The preliminary screening data is not shown. Among the restriction enzymes tested, a single cleavage site was observed for EcoRV, HincII, PvuII, SalI and ScaI. The size of fall armyworm mtdna was estimated as approximately 14.8 kb. This value is slightly smaller than the 16.3 kb size estimated by Ke & Pashley (1992). BamHI, HaeIII, HpaI and SmaI did not digest fall armyworm mtdna. Restriction enzymes having 2 or more cleavage sites included AluI, BstNI, DraI, EcoRI, FokI, HhaII, HinfI, HindIII, HpaII, MspI, NdeI, NruI, PstI, RsaI, TaqI and XbaI. Of the 16 restriction enzymes that generated more than two mtdna fragments, only three enzymes (BstNI, HinfI and MspI) produced restriction profiles that differed between the corn and rice strains (Fig. 1). The MspI and HpaII patterns are the same because the two enzymes share the same recognition site (designated the MspI pattern). The BstNI and HinfI mtdna patterns were previously described by Pashley (1989), while the MspI pattern is first described in this study. The MspI pattern is very distinctive between the two strains, with the corn strain pattern consisting of 4 restriction fragments of 5.4, 4.3, 3.8 and 1.3 kb, and the rice strain pattern consisting of 2 fragments of 10.4 and 4.4 kb (Fig. 1). Further, we explored the MspI mtdna restriction pattern as a diagnostic marker for corn and rice strain insects. We determined the MspI mtdna patterns for fall armyworms from five populations identified previously as rice (R) or corn (C, I, M, and P) strain populations (Lu et al. 1992). Fifteen to 20 individual larvae were selected from each population and total DNA extracted. Fig. 2 shows a Southern blot of representative samples of MspI-digested DNA probed with 32 P-labeled mtdna. Each insect belonged to one of two mtdna haplotypes; one represented by the corn strain and the other by the rice strain (Fig. 2). We also examined the mtdna from insects of unknown strain status. Larvae were collected from corn and sorghum fields, and from a laboratory colony initiated from bermudagrass-collected insects. The analyses showed that 32 of 36 fall armyworms (over 85%) collected from the corn and sorghum fields had the corn strain MspI restriction pattern (data not shown). All of the 38 insects from the bermudagrass-originated colony had the rice strain pattern (data not shown). These results are consistent with the host preference of strains as reported by Pashley et al. (1988). Corn strain fall armyworms prefer corn and sorghum, while the rice strain insects prefer rice and bermudagrass. However, the above result also showed that a small fraction of rice strain insects can be found on corn and sorghum, indicating some overlap in host usage occurs (Pashley 1989). Total DNA preparations from the field-collected and bermudagrass colony insects were analyzed using a repeated DNA marker found only in rice strain insects (Lu et al. 1994). All the total DNA samples that reacted with the rice strain-specific marker showed the rice strain mtdna pattern, while the remainder displayed the corn strain

55 52 Florida Entomologist 79(1) March, 1996 Figure 1. Polymorphic mtdna restriction patterns between the fall armyworm corn (C) and rice (R) strain generated by BstNI, HinfI and MspI, and detecting by probing total DNA blot with 32 P-labeled mtdna. Molecular sizes (bp) are indicated at the left. mtdna pattern (data not shown). These results correlate the MspI mtdna pattern with known nuclear DNA markers and establish the MspI mtdna pattern as a diagnostic marker for fall armyworm host strains. We felt that the MspI strain marker would be more useful if the time consuming and expensive Southern blotting steps could be eliminated. This was accomplished using the following rationale and approach. We observed that fall armyworm mtdna is uncut by HaeIII, while genomic DNA is digested by HaeIII to fragments smaller than 4 kb in molecular size (data not shown). Also, after MspI digestion of total fall armyworm DNA, several mtdna restriction fragments were partially discernible in an ethidium bromide stained agarose gel (data not shown). These observations led us to postulate that the diagnostic MspI mtdna pattern might be visible on a stained gel if the molecular size of the background genomic DNA fragments was reduced. Figure 3 depicts a stained agarose gel of fall armyworm DNA samples after treatment with MspI and HaeIII. As expected the genomic DNA fragments migrated further down the agarose gel and uncovered the mtdna fragments. The two mtdna fragments (10.4 and 4.4 kb) of the rice strain pattern, and 3 of the 4 fragments (5.4, 4.3 and 3.8 kb) of the corn strain pattern were easily detected. The 1.4 kb fragment in the corn strain pattern was hidden by the bulk genomic DNA fragments in the lower part of the gel. We have termed this the 3-band pattern for the corn strain and the 2-band pattern for the rice strain.

56 Lu & Adang: Fall Armyworm Strain Identification 53 Figure 2. A Southern blot showing the MspI mtdna pattern of fall armyworms from different populations as indicated. Population R is the rice strain. Populations C, M, I and P were previously identified as corn strain populations (Lu et al. 1992). DISCUSSION The results of a previous study (Pashley 1989) and this study show that the corn and rice strains of fall armyworm have distinct mtdna RFLPs for BstNI, HinfI and MspI. Since the diagnostic MspI mtdna pattern can be detected by a simple double- Figure 3. A stained agarose gel of MspI and HaeIII digested total fall armyworm DNA, showing the diagnostic mtdna restriction pattern on the upper part of the gel. Fall armyworm populations and DNA samples are the same as in Fig. 2.

57 54 Florida Entomologist 79(1) March, 1996 digest of total DNA using MspI and HaeIII, the method does not require the isolation of mtdna, or the use of Southern blot analysis. Thus, this very straightforward MspI/ HaeIII digestion technique facilitates the rapid detection of fall armyworm strains. Since mtdna is maternally inherited, a concern is that mtdna markers may not correctly identify intraspecific pest populations, particularly when these populations are sympatrically distributed and interbred. In these situations, insect populations with certain mtdna genotypes may simply represent maternal lineages, but not nuclear genotypes. To address this concern, we analyzed fall armyworms with both mtdna and nuclear DNA markers (this study, Lu et al. 1992, Lu et al. 1994). Our results demonstrated that the two types of markers agreed completely, indicating that fall armyworm mtdna genotypes correspond to their nuclear genotypes. These results support Pashley s conclusion that gene flow between the sympatric fall armyworm strains is very limited (Pashley 1989). Fall armyworm strains were shown to be selective in host usage (Pashley 1986), however, it should not be assumed that all the insects collected from a given host will belong to the host strain. Pashley (1989) reported finding a small number of corn strain insects on bermudagrass and rice strain insects on corn and sorghum plants. By using the mtdna and nuclear DNA markers, we also identified a small number of fall armyworms taken from corn and sorghum plants to be rice strain insects. Further studies are needed to gain insight into the mechanism of strain differentiation and its impact on pest management. ACKNOWLEDGMENT The authors are grateful to Drs. D. Isenhour, G. Kochert, and M. Bass for their support. This research was funded by a HATCH project through the University of Georgia College of Agriculture and Environmental Sciences. REFERENCES CITED BARFIELD, C. S., J. L. STIMAC, AND M. A. KELLER State-of-art for predicting damaging infestations of fall armyworm. Florida Entomol. 63: FEINBERG, A. P., AND B. VOGELSTEIN A technique for labeling DNA restriction fragments to a high specific activity. Anal. Biochem. 132: KE, L. D., AND D. P. PASHLEY Characterization of fall armyworm mitochondrial DNA (Lepidoptera: Noctuidae). Arch. Insect. Bio. Phisil. 21: KNIPLING, E. F Regional management of the fall armyworm A realistic approach? Florida Entomol. 63: LU, Y. J., M. J. ADANG, D. J. ISENHOUR, AND G. D. KOCHERT RFLP analysis of genetic variation in North American populations of the fall armyworm moth Spodoptera frugiperda (Lepidoptera: Noctuidae). Molecular Ecology 1: LU, Y. J., G. D. KOCHERT, D. J. ISENHOUR, AND M. J. ADANG Molecular characterization of a strain specific repeated DNA sequence in fall armyworm Spodoptera frugiperda (Lepidoptera: Noctuidae). Insect Molecular Biology PAIR, S. D., J. R. RAULSTON, A. N. SPARKS, J. K. WESTBROOK, AND G. K. DOUNCE Fall armyworm distribution and population dynamics in the southeastern states. Florida Entomol. 69: PASHLEY, D. P Host-associated genetic differentiation in fall armyworm (Lepidoptera: Noctuidae): A sibling species complex? Ann. Entomol. Soc. America 79: PASHLEY, D. P Host-associated differentiation in armyworms (Lepidoptera: Noctuidae): An allozymic and mitochondrial DNA perspective, pp in

58 Lu & Adang: Fall Armyworm Strain Identification 55 H. D.Loxdale, and J. den Hollander, [ed.], Electrophoretic Studies on Agricultural Pests. Clarendon Press, Oxford. PASHLEY, D. P., A. M. HAMMOND, AND T. N. HARDY Reproductive isolating mechanisms in fall armyworm host strains (Lepidoptera: Noctuidae). Ann. Entomol. Soc. America. 85: PASHLEY, D. P., AND J. A. MARTIN Reproductive incompatibility between host strains of the fall armyworm (Lepidoptera: Noctuidae). Ann. Entomol. Soc. America 80: PASHLEY, D. P., T. C. SPARKS, S. S. QUISENBERRY, T. JAMJANYA, AND P. F. DOWD Two fall armyworm strains feed on corn, rice and bermudagrass. Louisiana Agriculture 30: 8-9. QUISENBERRY S. S., AND F. WHITFORD Evaluation of bermudagrass resistance to fall armyworm (Lepidoptera: Noctuidae): influence of host strain and dietary conditioning. J. Econ. Entomol. 81: SAMBROOK, J., E. F. FRITSCH, AND T. O. MANIATIS Molecular cloning: A Laboratory Manual. 2nd ed. Cold Spring Harbor Press, New York. SOUTHERN, E. M Detection of specific sequences among DNA fragments separated by gel electrophoresis. J. Mol. Biol. 98: SPARKS, A. N A review of the biology of the fall armyworm. Florida Entomol. 62: WILLIAMS, W. P., P. M. BUCKLEY, AND F. M. DAVIS Combining ability for resistance in corn to fall Armyworm and southwestern corn borer. Crop Sci. 29: WISEMAN, B. R., AND F. M. DAVIS Plant resistance to the fall armyworm. Florida Entomol. 62: WISEMAN, B. R., AND D. J. ISENHOUR Feeding responses of fall armyworm larvae on excised green and yellow whorl tissue of resistant and susceptible corn. Florida Entomol. 71: ZEHNDER, G. W., L. SANDALL, A. M. TISLER, AND T. O. POWERS Mitochondrial DNA diversity among 17 geographic populations of Leptinotarsa decemlineata (Coleoptera: Chrysomelidae). Ann. Entomol. Soc. America. 85:

59 56 Florida Entomologist 79(1) March, 1996 DESCRIPTION OF THE MATURE LARVAE OF CHRYSIS GRACILLIMA AND OMALUS BIACCINCTUS AND NEW DATA ON THE BIOLOGY OF TRICHRYSIS CYANEA (HYMENOPTERA: CHRYSIDIDAE) J. TORMOS 1, J. D. ASIS 1, S. F. GAYUBO 1 AND E. MINGO 2 1 Departamento de Zoología. Facultad de Biología. Universidad de Salamanca Salamanca (Spain) 2 Museo Nacional de Ciencias Naturales. C/J. Gutiérrez Abascal, Madrid (Spain) ABSTRACT The mature larvae of Chrysis gracillima Förster, 1853 and Omalus biaccinctus (Buysson, 1893) are described and compared with others known in the tribes Chrysidini and Elampini, respectively. Additionally, new data are reported on the biology of Trichrysis cyanea (L., 1758). Key Words: Cleptoparasites, biology, preimaginal states. RESUMEN Se describen las larvas maduras de Chrysis gracillima Förster, 1853 y Omalus biaccinctus (Buysson, 1893) y se comparan con otras conocidas de las tribus Chrysidini y Elampini, respectivamente. Adicionalmente, son reportados nuevos datos sobre la biología de Trichrysis cyanea (L., 1758). The family Chrysididae includes some 3,000 species (Kimsey & Bohart 1990), most of which are ectoparasitoids or cleptoparasites of Hymenoptera Aculeata. Their biology is not very well known, most contributions referring to isolated data on their possible hosts. However, particularly outstanding are the workers in the case of the Chrysididae that attack stem-nesting Sphecidae and Eumenidae: Trautmann (1927), Enslin (1929), Clausen (1940), Van Lith (1953, 1955, 1956, 1958), Grandi (1959, 1961), Krombein (1967), Danks (1970), Spradbery (1973), Iwata (1976), and Medvedev (1978). The mature larvae of 15 Chrysidinae have been described. These species belong to eight genera, five of which are Chrysidini and three are Elampini. Studies on the mature larvae of Palearctic Chrysis ignita (L., 1758); Chrysura dichroa (Dahlbom, 1854); Hedychrum rutilans Dahlbom, 1854; Praestochrysis shanghaiensis (Smith, 1874); Trichrysis cyanea (L., 1758); or Holarctic Omalus aeneus (F., 1758) and Pseudomalus auratus (L., 1758)), have been carried out by: Enslin (1929); Maréchal (1923); Soika (1934); Maneval (1936); Parker (1936); Grandi (1959, 1961); Yamane (1976); Evans (1987) and Asís et al. (1994). Janvier (1933) described the mature larva of a Neotropical species (Chrysis grandis (Brullé, 1846)), and Evans (1987) dealt with the major morphological features of all the above mentioned species (with the exception of Chrysis ignita, Hedychrum rutilans and T. cyanea), the Nearctic species Chrysis cembricola Krombein, 1958; Chry-

60 Tormos et al.: Larvae of Chrysididae 57 sis cessata Buysson, 1891; Chrysis nitidula F., 1775; Chrysis inflata Aaron, 1885; Chrysis inaequidens Dahlbom, 1854, and the Nearctic-Neotropical species Caenochrysis doriae (Gribodo, 1874) and Chrysis smaragdula F., In the course of a study on the fauna of stem-nesting Hymenoptera on the northern subplateau of the Iberian Peninsula, data were obtained on the biology of T. cyanea. Mature larvae of Chrysis gracillima Förster, 1853 and Omalus biaccinctus (Buysson, 1893) were also collected. For Omalus Panzer, Pemphredoninae (Sphecidae) are accepted as hosts (Kimsey & Bohart 1990). For hosts of T. cyanea, these authors cite Larrinae and Pemphredoninae (Sphecidae). The mature larvae of C. gracillima and O. biaccinctus have been previously undescribed. MATERIALS AND METHODS A total of 2,043 trap nests was placed in different sites of the provinces of Burgos, Cuenca, Soria, Teruel and Valencia (Spain). The trap nests, comprising stems of Ailanthus altissima (Miller) Swingle and of Phragmites australis (Cav.) Trin ex Steudel (length=20-30 cm, diam=2-5 mm), were placed in the field in mid spring of 1992 and 1993 and were collected at the end of autumn of the same year. After opening them in the laboratory, the contents of each cell were transferred to glass vials that were kept at 6-8 C over the winter. During the following spring the vials were transferred to a culture chamber at 28 C to elicit emergence of the imagos; it was thus possible to know the identity of the occupants of the nests and their parasitoids. Some of the mature larvae were fixed and preserved in 70% alcohol for later study and description. The methodology used in their preparation was similar to that employed by Asís et al. (1994). The descriptions employ the terminology and organization used by Evans (1987). The cells were numbered from the exterior to the interior (cell 1 outermost) even though chronologically the innermost cell was the one first completed. RESULTS In all of the nests examined by us that contained two or more parasitized cells, the parasites were always of the same species. Chrysis gracillima Förster Three larvae were collected from a nest with three cells, all of them parasitized. Of these, two were reared to obtain imagos while the other one was placed in 70% alcohol for the description that follows. The absolute measurements refer to the specimen from Vilviestre (Soria) (reference: ). Mature larva. General aspect (Fig. 1). Body robust (length = 4.3 µm, width = 1.9 µm), with the abdominal segments divided into two annulets by a transverse crease. Anus small, terminal, as a transverse slit. Pleural lobes developed. Integument with eighteen minute setae (l = 12 µm) on each segment, distributed in a line between spiracles (Fig. 2). Spiracles (Fig. 3). With a well-developed peritreme; atrium simple, naked. First pair (mean diam = 42 µm, n=10) slightly larger than the rest (mean diam = 35.7 µm, n=18). Head (Fig. 4). Width = 838 mm, height = 600 µm, with sparse and short setae (l = 19 µm) very numerous near insertion of mandibles. Coronal suture and parietal

61 58 Florida Entomologist 79(1) March, 1996 Figures Mature larva of C. gracillima Förster: 1, general aspect; 2, segment (dorsum) with the distribution of setae; 3, spiracle (atrium and subatrium); 4, head in frontal view; 5, labrum; 6, labium and maxilla (oral face). Mature larva of Omalus biaccinctus (Buysson). bands absent. Antennal orbits elliptical (57 µm 47 µm); antennal papilla short (l=9.5 µm), with three sensillae at center. Mouthparts. Labrum (Fig. 5) (w = 209 mm) emarginate, with fourteen marginal setae (l = 19 µm). Epipharynx naked. Mandibles (l = 266 µm, w = 142 µm) tridentate. Maxillae (l = 260 µm, w = 104 µm), with three setae on external part (l = 28.5 µm), mesal margin papillose. Maxillary palpi much wider than long (l = 9.5 µm, w = 28.5 µm);

62 Tormos et al.: Larvae of Chrysididae 59 galeae present (l = 9.5 µm, w = 9.5 µm). Labium (Fig. 6) (l = 95.2 µm, w=228 µm) with short palpi (l=19 µm, w=38 µm); spinneret a transverse slit (l = 57.1 µm). Maxillae and labium with pigmented bands. Omalus biaccinctus (Buysson) Imagos were obtained from three nests: one of Passaloecus gracilis (Curtis, 1834) which had six cells (cell 1 being parasitized, from which a male emerged). Another nest of Passaloecus sp. contained three cells, all of them parasitized; the mature larva of cell 2 was transferred to 70% alcohol for later study; a male emerged from cell 1 and in cell 3 the development of the chrysidid was not completed although its larva did construct a cocoon. The third nest contained nine parasitized cells; the mature larvae of cells 1, 2 and 3 were transferred to 70% alcohol for later study; two females emerged from cells 4 and 9, and in cells 5, 6, 7, and 8 the development of the chrysidid was not completed, although the larvae did construct a cocoon. The absolute measurements of the description of the larva refer to the specimen from Cofrentes (Valencia) (reference: 3I94004A1). Mature larva. General aspect (Fig. 7). Body robust (l =4.3 µm, w = 1.8 µm); abdominal segments not divided into annulets. Fourth abdominal segment humped dorsolaterally. Anus terminal, as a transverse slit. Pleural lobes scarcely developed. Integument microspinulose (l = 4 µm), with sparse minute setae (l = 14 µm) (Fig. 8a). Spiracles (Fig. 8b) with a well-developed peritreme; atrium simple, lined with weak ridges. First pair (mean diam = 75.3 µm, n = 10) slightly larger than the rest (mean diam = 66.6 mm, n = 18). Head (Fig. 9). W = 914 µm, h = 885 µm with sparse, short setae (l = 19 µm). Coronal suture and parietal bands present. Antennal orbits circular (d = 47.5 µm) located below middle of head; antennal papilla (l = 19 µm, w = 18 µm) rather long, with three small sensillae at center (Fig. 10). Clypeolabral suture slightly emarginate. Mouthparts. Labrum (Fig. 11) W = 324 µm, emarginate, with ten short setae (l = 10 µm) and six marginal sensillae (w = 9 µm). Epipharynx naked. Mandibles (l = 257 µm, w = 125 µm) tridentate. Maxillae (l = 105 µm, w = 63 µm) with a few setae on external part (l = 9 µm); margin mesally papillose. Maxillary palpi slightly wider than long (l = 29 µm, w = 30.4 µm); galeae present (l = 19 µm, w = 9 µm). Labium (Fig. 12) (l = 142 µm, w = 171 µm) with four setae at lower face (l = 10 µm); palpi short (l = 19 µm, w = 29 µm); spinneret a transverse slit (l = 42 µm). Hypopharynx with four sensory pores (d = 7 µm). Maxillae and labium with strong, pigmented bands. Trichrysis cyanea (Linnaeus) Data on the biology of this species (hosts, mechanisms of parasitism, sex-ratio) have been provided by Enslin (1929) and Danks (1970). Various hosts have been cited by Trautmann (1927), Micheli (1929), Hamm & Richards (1930), Grandi (1931), and Medvedev (1978). Individuals were obtained from nests of Trypoxylon attenuatum Smith, 1851; T. beaumonti Antropov, 1991; T. figulus (L., 1758); Trypoxylon sp.; Pemphredon lethifera (Schuckard, 1837) and Psenulus pallipes (Panzer, 1978). Table 1 shows the incidence of the Chrysididae in each of the hosts. This was the most abundant cleptoparasite, occurring in 42 nests with 177 cells, of which 62 were parasitized (35%). The number of cells affected per nest and the parasitism index as a function of the position of cell in the nest are shown in Tables 2 and 3, respectively. With respect to the mortality observed in the 62 parasitized cells, in 7

63 60 Florida Entomologist 79(1) March, 1996 Figures Mature larvae of O. biaccinctus: 7, general aspect; 8a, setae of integument (detail); 8b, spiracle (atrium and subatrium); 9, head in frontal view; 10, antennal papilla; 11, labrum; 12, labium and maxilla (oral face).

64 Tormos et al.: Larvae of Chrysididae 61 TABLE 1. NUMBER OF NESTS AND CELLS OF THE DIFFERENT HOSTS AFFECTED BY TRICHRYSIS CYANEA. Host Number of Nests Affected Total Number of Cells Number of Parasitized Cells Trypoxylon attenuatum (27%) Trypoxylon beaumonti (42%) Trypoxylon figulus (22%) Trypoxylon sp (65%) Pemphredon lethifera 1 2 1(50%) Psenulus pallipes 1 8 1(13%) (11%) the development of the Chrysididae was not completed, although in all cases the larva constructed a cocoon. Of the 51 imagos obtained (four larvae were preserved in 70% alcohol), 21 were male and 30 female. Although the sex-ratio obtained is 0.7 m: 1.0 f, this is not significantly different from a 1:1 ratio as observed in many haplo-diploid species (X 2 = 1.607; d.f.= 1; 0.3 > p > 0.2). Furthermore, no significant differences were seen in the sex-ratio as a function of the position of the parasitized cell (X 2 = 9.82; d.f.= 6; 0.2 > p > 0.1). Comparing these data with those obtained by Krombein (1967) for other species of Chrysididae, it may be concluded that T. cyanea has a low degree of host specificity and that it is quite effective as a parasitoid. The proportion of parasitized nests with respect to the total number of nests of each species obtained was as follows: T. attenuatum (16%), T. beaumonti (10%), T. figulus (38%), Trypoxylon sp. (38%), P. lethifera (2%), and P. pallipes (17%); only 8 of the 38 affected nests with two or more cells had 50% or more of the cells parasitized and the mean number of cells affected per parasitized nest was 1.48 (35%). The sex-ratio was similar to that reported by Danks (1970). DISCUSSION Like the rest of the described species of Elampini (Hedychrum rutilans, O. aeneus and Pseudomalus auratus), O. biaccinctus displays antennal orbits located near or below middle of head; well-developed antennal papillae; and entire (not divided into two rings) body segments that are slightly humped in the dorsolateral direction. In this species, the back of the abdominal segments is very humped. The scanty and protuberant marginal sensilla of the labrum and the presence of galeae are morphological characters shared with O. aeneus and Pseudomalus auratus. TABLE 2. NUMBER OF CELLS AFFECTED PER PARASITIZED NEST BY TRICHRYSIS CYANEA. Nests with 1 affected cell 28 (67%) Nests with 2 affected cells 10 (24%) Nests with 3 affected cells 3 (7%) Nests with 4 affected cells 0 Nests with 5 affected cells 1 (2%)

65 62 Florida Entomologist 79(1) March, 1996 TABLE 3. INDEX OF PARASITISM AND NUMBER OF MALES AND FEMALES OBTAINED FROM CELLS ACCORDING TO THEIR POSITION IN THE NEST (C1 IS THE EXTERNAL ONE) IN TRICHRYSIS CYANEA. Cell Position Total Number of Cells Number of Parasitized Cells male Sex female C (55%) 9 12 C (37%) 2 9 C (41%) 3 7 C (25%) 5 C (19%) 1 C (23%) 1 1 C7 5 1 (20%) 1 Total (36%) Although their presence is quite probable in Hedychrum rutilans, these characters are not noted in the work of Maneval (1936). The number of mandibular teeth is variable: 4-5 in Pseudomalus auratus, 3 in Hedychrum rutilans and O. biaccinctus, and 2 in O. aeneus. Chrysis gracillima has the segments of the body divided into two annulets, antennal orbits situated in normal position and marginal sensillae of the labrum numerous and minute. These characteristics are shared by the described species of Chrysidini: Caenochrysis doriae, Chrysis cembricola, C. cessata, C. grandis, C. ignita, C. inaequidens, C. inflata, C. nitidula, C. smaragdula, Chrysura dichroa, Praestochrysis shangaiensis, and T. cyanea. All species have tridentate mandibles with the exception of T. cyanea, in which the mandibles are quadridentate. Other generalized characters are the presence of galeae and short antennal papillae, although both structures are absent in Chrysura dichroa (Grandi, 1961). Praestochrysis shangaiensis does not have antennal papillae, and Parker (1936) does not represent the galeae. Our scanty knowledge of the mature larvae of the Chrysidoidea does not permit the establishment of primitive vs. specialized states of the various characters. In this respect, it is highly significant that in his cladistic analysis of the Chrysidoidea, Carpenter (1986) did not use any preimaginal morphological character. However, from the characters studied it may be deduced that in some Chrysidinae (sensu Kimsey & Bohart 1990), the galeae are apparently lacking, thus a specialized feature. In the Elampini the low position of the antennal orbits and long antennal papillae may also represent states of specialized character. ACKNOWLEDGMENTS We are much indebted to H. E. Evans (Colorado State University, U.S.A.), L. S. Kimsey (University of California, Davis, U.S.A.) and Karl V. Krombein (National Museum of Natural History, Washington, U.S.A.), for their comments on the manuscript. Grants from the DGICYT (PB C02) supported the study.

66 Tormos et al.: Larvae of Chrysididae 63 REFERENCES CITED ASÍS, J. D., J. TORMOS, AND S. F. GAYUBO Biological observations on Trypoxylon attenuatum and description of its mature larva and its natural enemy Trichrysis cyanea (Hymenoptera: Sphecidae: Chrysididae). J. Kansas Entomol. Soc. 67: CARPENTER, J. M Cladistics of the Chrysidoidea (Hymenoptera). J. New York Entomol. Soc. 94: CLAUSEN, C. P Entomophagous Insects. Hafner Pub. Co. New York. DANKS, H. V Biology of some stem-nesting aculeate Hymenoptera. Trans. R. Entomol. Soc. London 122: ENSLIN, E Beiträge zur Metamorphose der Goldwespen. Zeitschr. wissenschaftl. Insektenbiologie XXIV: EVANS, H. E Order Hymenoptera, pp in F. W. Stehr [ed.] Immature Insects. Kendall/Hunt Publishing Company. Dubuque. Iowa. GRANDI, G Contributi alla conoscenza biologica e morfologica degli imenotteri melliferi e predatori XII. Bol. Lab. Entomol. R. Ist. Sup. Agr. Bologna 4: GRANDI, G Contributi alla conoscenza degli Imenotteri Aculeati. Bol. Inst. Entomol. Univ. Bologna 28: GRANDI, G Studi di un entomologo sugli imenotteri superiori. Boll. Inst. Entomol. Univ. Bologna 25. HAMM, A. H., AND O. W. RICHARDS The biology of the British fossorial wasps of the wasps of the families Mellinidae, Gorytidae, Philanthidae, Oxybelidae and Trypoxylonidae. Trans. Entomol. Soc. London 78: IWATA, K Evolution of Instinct. Comparative Ethology of Hymenoptera. Amerind Publishing Co. Pt. Ltd. New York. JANVIER, H Etude biologique de quelques Hyménoptères du Chili. Ann. Sci. Nat. Zool. 16: KARLIN, S., AND S. LESSARD Theoretical studies on Sex Ratio Evolution. Princeton University Press. New Jersey. KIMSEY, L. S., AND R. M. BOHART The Chrysidid Wasps of the World. Oxford University Press. Oxford. KROMBEIN, K. V Trap-nesting Wasps and Bees: Life Histories, Nests and Associates. Smithsonian Press. Washington. MARÉCHAL, P Note sur l état larvaire et l état nymphal de Chrysis ignita L. Bull. Soc. Entomol. Belg. 5: MANEVAL, H Nouvelles notes sur divers Hyménoptères et leurs larves. Rev. Fran. Entomol. 3: MEDVEDEV, G. S Keys for the identification of the insects of the European part of the USSR. Acad. Sciencies.USSR. Zoological Institute. (in Russian). MICHELI, L Note biologiche e morfologiche sugli imenotteri (I). Bol. Soc. Entomol. Ital. 7: PARKER, D. E Chrysis shangaiensis Smith, a parasite of the oriental moth. J. Agric. Res. 52: SOIKA, A. G Etudes sur les larves des hyménopteres. Ann. Soc. Entomol. France. 103: SPRADBERY, J. P Wasps: An account of the biology and natural history of social and solitary wasps. Sidwick & Jackson Biology Series. London. TRAUTMANN, W Die Goldwespen Europas. G. Uschmann Weimar. VAN LITH, J. P De Nederlandse metselwespen. Lev. Nat. 12: VAN LITH, J. P De Nederlandse Spilomena-soorten (Hym.: Sphecidae). Entomol. Ber. 15: VAN LITH, J. P Hoplomerus (Hoplomerus) spinipes (L.) en H. (Spinicoxa) reniformis (Gmel.). Entomol. Ber. 16: VAN LITH, J. P Opmerkingen over Chrysididae (3). Entomol. Ber. 58: YAMANE, S Mature larva of Chrysis ignita (Hymenoptera: Chrysididae). New Entomol. 25:

67 64 Florida Entomologist 79(1) March, 1996 PODISUS MACULIVENTRIS (HEMIPTERA: PENTATOMIDAE) PREDATION ON LADYBIRD BEETLES (COLEOPTERA: COCCINELLIDAE) J. HOUGH-GOLDSTEIN, J. COX AND A. ARMSTRONG Delaware Agricultural Experiment Station Department of Entomology and Applied Ecology College of Agricultural Sciences, University of Delaware Newark, DE ABSTRACT Adult Podisus maculiventris (Say) consumed no adult ladybird beetles (Harmonia axyridis Pallas) and 30-35% of larval ladybird beetles in no-choice laboratory trials. Nymphal P. maculiventris appeared to be more agile than adults in petri dishes and attacked 65% of both larval and adult H. axyrides; however, they only killed and consumed 50% of larvae and 20% of adult ladybird beetles. Ladybird beetle larvae were aggressive and often escaped, whereas adult beetles were usually rejected by the nymphal predators, suggesting unpalatability. Overall, P. maculiventris took more than four times longer to capture ladybeetle larvae than to capture fall armyworms (Spodoptera frugiperda [J. E. Smith]). In nature, while some ladybird beetles are undoubtedly consumed by P. maculiventris, most probably escape predation, either by evasion or through lack of palatability. Key Words: Harmonia axyrides, prey selection, defense, palatability, refugia, nochoice tests. RESUMEN Los adultos de Podisus maculiventris (Say) no consumieron adultos del coccinélido Harmonia axyridis Pallas y sólo depredaron el 30-35% de sus larvas en ensayos de alimentación obligada. Las ninfas de P. maculiventris parecieron ser más ágiles que los adultos en las placas de petri y atacaron el 65% de las larvas y adultos de H. axyridis; sin embargo, estas sólo mataron y consumieron el 50% de las larvas y el 20% de los adultos del coccinélido. Las larvas del coccinélido fueron agresivas y a menudo escaparon, mientras que los adultos usualmente eran rechazados por las ninfas predatoras, lo cual sugiere su impalatabilidad. En general, P. maculiventris necesitó 4 veces más tiempo para capturar un coccinélido que para capturar un gusano trozador (Spodoptera frugiperda (J. E. Smith)). En la naturaleza, mientras algunos coccinélidos son sin duda consumidos por P. maculiventris, la mayoría probablemente escape a la depredación, ya sea por evasión o por falta de palatabilidad. The predaceous hemipteran, Podisus maculiventris (Say), has a broad host range, consisting primarily of soft-bodied, slow-moving larvae of Coleoptera and Lepidoptera (Esselbaugh 1948, Mukerji & LeRoux 1965). Among those listed as prey species, however, are at least four species of predaceous Coccinellidae (McPherson 1980). These would seem to be unlikely prey for P. maculiventris, because of their typically agile, rapid movements and the variety of chemical defenses generally present in coccinellids (Dettner 1987). Podisus maculiventris is a candidate for augmentative release (e.g., Hough-Goldstein 1996), but before large numbers are released, its potential impact on beneficial

68 Hough-Goldstein et al.: Podisus Predation on Coccinellids 65 arthropods should be evaluated. In this study, we examined the potential for P. maculiventris adults and nymphs to feed on larval and adult ladybird beetles. MATERIALS AND METHODS Podisus maculiventris were provided initially by the Maryland Department of Agriculture (Plant Protection Section), Annapolis, MD; these were reared in our laboratory for one or two generations on fall armyworm, Spodoptera frugiperda (J. E. Smith), before use in trials. Fall armyworms were obtained from the DuPont Stine- Haskell Laboratory, Newark, DE, where they were reared on artificial diet. Colorado potato beetles, Leptinotarsa decemlineata (Say), were field-collected as adults, and larvae were reared on greenhouse-grown potatoes ( Superior ). Ladybird beetle (Harmonia axyridis Pallas) adults and larvae were collected from mixed, weedy vegetation on the University of Delaware Experiment Station Farm immediately before use. Four no-choice trials were conducted between 11 and 25 July, No-choice experiments were used exclusively, because these are more likely to result in consumption and thus are a more stringent test of prey suitability than choice tests. Predators were fed in the colony 24 h prior to the start of each experiment. For each trial, P. maculiventris were placed individually in plastic petri dishes approximately 2 h before the trial began. Between 8 and 20 (usually 10) dishes were set up for each prey type, and one potential prey was placed in each dish. The dishes were observed continuously for 1 h after prey were introduced and then checked for prey consumption every half hour for the next 6 h. Dishes were left at room temperature (approximately 24 C) overnight and checked again for consumption 20 h after set-up. Two trials were run using P. maculiventris adults and two using fourth instars. For each P. maculiventris life stage, half the trials used arenas consisting of 9 by 1.3 cm plastic petri dishes, and half used 11 by 2.5 cm plastic dishes containing a piece of 9- cm filter paper, folded in half with the edges folded up as a refugium. The first trial compared fall armyworm larvae, Colorado potato beetle larvae, ladybird beetle larvae, and ladybird beetle adults as potential prey. Subsequent trials eliminated Colorado potato beetle, since both it and fall armyworm were palatable. Results of encounters (consumption or no consumption) were evaluated using Fisher s Exact Test (SAS Institute 1990). Times to capture (in min) of different types of prey for those P. maculiventris that consumed prey during each 20-h trial were compared using Analysis of Variance (ANOVA; PROC GLM, SAS Institute 1990). Time of capture was assigned as soon as the predator inserted its proboscis firmly and securely into the prey without either the prey escaping or the predator rejecting the prey. Attacks that occurred between 1 h and 6 h were assigned times as soon as they were observed (e.g., a capture noted at 1.5 h was assigned 90 min, even though it may have occurred somewhat sooner). Attacks that occurred overnight were (conservatively) assigned a time of 7 h (420 min), even though they may actually have occurred somewhat earlier or considerably later. Time of capture over all trials for all predators that consumed fall armyworms or ladybird beetle larvae was also analyzed by predator stage (adult or fourth instar), type of arena (9-cm or 11-cm with refugium) and prey type (fall armyworm or ladybird beetle larva), using ANOVA (SAS Institute 1990). RESULTS In small dishes without refugia, P. maculiventris adults ate 100% of the fall armyworms and Colorado potato beetle larvae, 35% of the ladybird beetle larvae, and none of the ladybird beetle adults (Table 1). Consumption was similar in the larger dishes

69 66 Florida Entomologist 79(1) March, 1996 with refugia where adult P. maculiventris ate 100% of the fall armyworms, 30% of the ladybird beetle larvae, and none of the ladybird beetle adults. In both trials, consumption rate varied significantly by prey (P<0.0001, Fisher s Exact Test). The time to capture was significantly shorter for fall armyworms than for ladybird beetle larvae in the 11-cm arenas (F = 17.27, df = 1, P = ; ANOVA, Table 1). A similar trend was observed in the 9-cm dishes (F = 3.27, df = 2, P = , Table 1). In 9-cm and 11-cm dishes, respectively, P. maculiventris fourth instars killed and consumed 60 and 80% of the fall armyworms, 30 and 70% of the ladybird beetle larvae, and 10 and 30% of the ladybird beetle adults during the 20-h observation period (Table 1). Differences in nymphal consumption by prey type were not significant for the 9-cm dish test (P = 0.080) or the 11-cm dish test (P = 0.111); however, the difference was significant for the two nymph tests considered together (P = 0.007, Fisher s Exact Test). The time to capture (for those that consumed prey) did not differ significantly by prey type for nymphs in 9-cm dishes (ANOVA, F = 0.32, P = , df = 1), but approached significance in 11-cm dishes (F = 3.36, P = , df = 2) with fall armyworms on average captured more quickly than ladybird beetle larvae or adults (Table 1). In addition to ladybird beetles that were killed and consumed by P. maculiventris nymphs, several other ladybird beetle larvae (2 in the 9-cm arena and 1 in the 11-cm arena) and nearly half of the ladybird beetle adults (4 in the 9-cm arena and 5 in the 11-cm arena) were attacked and apparently sucked by the P. maculiventris nymphs for one or more brief periods, and then abandoned. Thus, the nymphs actually caught and sampled 65% of the ladybird beetle larvae and 65% of the ladybird beetle adults (over both dish sizes), but only consumed and killed 50% of the larvae and 20% of the adult ladybird beetles. Ladybird beetle larvae often attempted to bite the predator s TABLE 1. CONSUMPTION AND TIME TO CAPTURE OF DIFFERENT PREY BY PODISUS MACU- LIVENTRIS. Predator Stage Arena Prey 1 (N) No. Consumed Time (min.) to Capture (mean ± SEM) Adult 9-cm FAW 8 (8) 48 ± 23 CPB 9 (9) LBBL 7 (20) 253 ± 79 LBBA 0 (10) Adult 11-cm, refugia FAW 10 (10) 84 ± 42 LBBL 6 (20) 367 ± 53 LBBA 0 (10) Fourth instar 9-cm FAW 6 (10) 23 ± 19 LBBL 3 (10) 7 ± 4 LBBA 1 (10) Fourth instar 11-cm, refugia FAW 8 (10) 77 ± 33 LBBL 7 (10) 289 ± 70 LBBA 3 (10) 146 ± FAW, fall armyworm; CPB, Colorado potato beetle; LBBL, ladybird beetle larva; LBBA, ladybird beetle adult.

70 Hough-Goldstein et al.: Podisus Predation on Coccinellids 67 proboscis and frequently escaped, whereas adult ladybird beetles were usually rejected by the predator. Analysis of variance over all experiments on time of capture by predators that consumed fall armyworms or ladybird beetle larvae indicated a trend toward more rapid capture by P. maculiventris nymphs than by adults (Table 2). Prey were captured more rapidly in 9-cm dishes without refugia than in 11-cm dishes with refugia (Table 2); fall armyworms were captured much more quickly than ladybird beetle larvae (Table 2). DISCUSSION Adult P. maculiventris appeared to be unable to capture ladybird beetle larvae or adults successfully because of their lack of agility (in small petri dishes) compared with the quick-moving prey. Those adult predators that did capture ladybird beetle larvae, however, did not subsequently reject them. Podisus maculiventris nymphs appeared to be considerably more agile (in petri dishes) than P. maculiventris adults or, perhaps, they were simply more highly motivated to feed than the adult predators. Both ladybird beetle larvae and adults were caught by the predator nymphs. Once caught, however, nearly a quarter of the ladybird beetle larvae and 30% of the ladybird beetle adults either escaped or were rejected by the nymphs after one or more feeding bouts. Ladybird beetle larvae responded aggressively to predation attempts in a manner similar to that described by Marston et al. (1978) for corn earworm, Heliothis [Helicoverpa] zea (Boddie). Adult ladybird beetles did not respond aggressively to attack, and thus were apparently unpalatable to the predator. Some of the sampled ladybird beetle larvae or adults may have died eventually due to injection of salivary digestive enzymes (Cohen 1990, 1995), although we did not observe any obvious ill effects during several days in the laboratory. TABLE 2. PODISUS MACULIVENTRIS TIME TO CAPTURE BY PREDATOR LIFE STAGE (ADULT OR NYMPH), TYPE OF ARENA (9-CM WITH NO REFUGIUM OR 11-CM WITH REF- UGIUM), AND TYPE OF PREY (FALL ARMYWORM OR LADYBIRD BEETLE LARVA). Variable N Time (min) to Capture (mean ± SE) F 1 P 1 Predator Stage Adult ± Fourth instar ± 33 Arena 9-cm ± cm, refugia ± 33 Prey 2 FAW ± LBBL ± 40 1 Overall F value = (df=3, 51), P = (PROC GLM, SAS Institute 1990). 2 FAW, fall armyworm; LBBL, ladybird beetle larva.

71 68 Florida Entomologist 79(1) March, 1996 Over all trials, P. maculiventris took more than four times longer to capture ladybird beetle larvae than to capture fall armyworms, probably due to the agility and speed of the ladybird beetle larvae in evading the predators. Fall armyworms, in contrast, were slow-moving, with little defense evident against predation. Because the predators were reared on fall armyworms prior to the trials, however, it is also possible that some adaptation to this prey species may have occurred. Arena size and presence of refugia primarily affected time to capture rather than total prey captured, with prey overall captured nearly twice as quickly in the smaller dishes that lacked refugia than in the larger dishes with refugia. The larger dishes with refugia undoubtedly reflect natural conditions somewhat better than the small empty dishes. However, even the larger dishes did not allow flight, an obvious means of escape for adult ladybird beetles in the field. In summary, this species of ladybird beetle (Harmonia axyrides), which was by far the most abundant coccinellid species in our study area in July of 1995, was clearly not a preferred prey for P. maculiventris. Undoubtedly some ladybird beetles are eaten by the predator in nature; indeed, during specimen collection for these experiments, one P. maculiventris nymph was observed feeding on an H. axyrides larva. Under some circumstances, local losses of coccinellids due to P. maculinventris predation could be high. In general, however, most larvae and especially adults of this coccinellid species are likely to escape predation by most P. maculiventris. We predict that other predaceous coccinellids will show similar behavioral and chemical predation avoidance characteristics. ACKNOWLEDGMENTS We thank C. Holko and P. Tipping, Maryland Department of Agriculture for providing Podisus maculiventris to initiate our colony, and D. Tallamy for valuable comments on a draft of the manuscript. Published as Paper No in the Journal Series of the Delaware Agricultural Experiment Station, Contribution No. 681 of the Department of Entomology and Applied Ecology, University of Delaware, Newark. REFERENCES CITED COHEN, A. C Feeding adaptations of some predaceous Hemiptera. Ann. Entomol. Soc. America 83: COHEN, A. C Extra-oral digestion in predaceous terrestrial Arthropoda. Annu. Rev. Entomol. 40: DETTNER, K Chemosystematics and evolution of beetle chemical defenses. Annu. Rev. Entomol. 32: ESSELBAUGH, C. O Notes on the bionomics of some midwestern Pentatomidae. Entomol. America 28: HOUGH-GOLDSTEIN, J. A Use of predaceous pentatomids in integrated management of the Colorado potato beetle (Coleoptera: Chrysomelidae), in M. Coll and J. Ruberson [eds.], Predatory Heteroptera in agroecosystems: their ecology and use in biological control. Thomas Say Publ., Entomol. Soc. America, Lanham, MD. MARSTON, N. L., G. T. SCHMIDT, K. D. BIEVER, AND W. A. DICKERSON Reaction of five species of soybean caterpillars to attack by the predator, Podisus maculiventris. Environ. Entomol. 7: MCPHERSON, J. E A list of the prey species of Podisus maculiventris (Hemiptera: Pentatomidae). Great Lakes Entomol. 13: MUKERJI, M. K., AND E. J. LEROUX Laboratory rearing of a Quebec strain of the pentatomid predator, Podisus maculiventris (Say) (Hemiptera: Pentatomidae). Phytoprotection 46: SAS INSTITUTE SAS user s guide, version 6, 4th ed. SAS Institute, Cary, NC.

72 Scientific Notes 69 THE CICADA DICEROPROCTA DELICATA (HOMOPTERA: CICADIDAE) AS PREY FOR THE DRAGONFLY ERYTHEMIS SIMPLICICOLLIS (ANISOPTERA: LIBELLULIDAE) ALLEN F. SANBORN Barry University, School of Natural and Health Sciences, N.E. Second Avenue, Miami Shores, FL 33161, USA While working on the coastal dunes at Holly Beach in Cameron Parish, Louisiana during the summer of 1995, I had the opportunity to observe predation by the dragonfly Erythemis simplicicollis (Say) on the cicada Diceroprocta delicata (Osborn). An individual D. delicata that had just flown from its perch was captured by an E. simplicicollis and was being consumed in the surrounding tall vegetation. The dragonfly appeared to have been drawn to the movement of the cicada as it flew from its perch. Under similar conditions I witnessed another cicada being attacked by two dragonflies: both dragonflies rose from their perches and collided with the cicada as it approached the edge of a dune. However, this attack was unsuccessful and the cicada escaped. The specific identity of these dragonflies was not determined. A colleague of mine counted at least 17 species of Odonata at Holly Beach that day, and we could not make a positive identification of the individuals that attacked the cicada. I have been unable to locate other references to dragonflies using a non-periodical cicada species as prey in North America. Fitch (1855), Riley (1885), Marlatt (1907), Felt (1912), and McAtee (1921) have reported dragonflies feeding on periodical cicadas (Magicicada spp.). However, most carnivorous animal species (see list in Marlatt 1907) use the superabundant food source that periodical cicadas represent during an emergence where local population densities are often greater than three million cicadas per acre (Dybas & Davis 1962). Diceroprocta delicata is apparently the first nonperiodical North American cicada species reported to be prey for dragonflies. Cicadas have also been reported as prey of dragonflies in New Zealand (Myers 1929), Afghanistan (Hay 1840), and South Africa (Distant 1897). The relatively low population numbers of non-periodical cicadas combined with their large body size (D. delicata body length is about 20.5 mm and length to wingtip is 29.8 mm; the captured E. simplicicollis was 50.8 mm long) may make cicadas difficult targets for capture and perhaps, therefore, influence the paucity of reports of dragonflies using cicadas as prey. Sidney W. Dunkle identified the E. simplicicollis specimen, and Sr. John Karen Frei of Barry University provided financial support. SUMMARY This paper reports the dragonfly Erythemis simplicicollis capturing the cicada Diceroprocta delicata as prey. It is unusual in that it represents a non-periodical cicada being used as prey by a dragonfly. REFERENCES CITED DYBAS, H. S., AND D. D. DAVIS A population census of seventeen-year periodical cicadas (Homoptera: Cicadidae: Magicicada). Ecology 43: DISTANT, W. L Zoological rambles in and around the Transvaal. Zoologist (4)1: FELT, E. P Twenty-seventh report of the state Entomologist New York St. Mus. Bull. 155:

73 70 Florida Entomologist 79(1) March, 1996 FITCH, A The seventeen-year locust, Cicada septendecim, Linnaeus. Trans. New York St. Agric. Soc. 14: HAY, R. G Notes on the wild sheep of the Hindoo Koosh, and a species of cicada. Jour. Asiatic Soc. Bengal 53: MARLATT, C. L The periodical cicada. USDA Bur. Entomol. Bull. 71: MCATEE, W. L The periodical cicada, 1919; brief notes for the District of Columbia region. Proc. Entomol. Soc. Washington 23: MYERS, J. G Insect singers: A natural history of the cicadas. George Routledge & Sons, London. RILEY, C. V The periodical cicada. USDA Div. Entomol. Bull. 8: 1-46.

74 70 Florida Entomologist 79(1) March, 1996 AN EXPLORATORY INSECT SURVEY OF TROPICAL SODA APPLE IN BRAZIL AND PARAGUAY J. C. MEDAL 1, R. CHARUDATTAN 2, J. J. MULLAHEY 3 AND R. A. PITELLI 4 1 Entomology and Nematology Department, University of Florida Gainesville, FL Plant Pathology Department, University of Florida, Gainesville, FL Wildlife and Range Science Department, South West Florida Research and Education Center, Immokalee, FL Universidade Estadual Paulista do Jaboticabal, Sao Paulo, Brazil Tropical soda apple, Solanum viarum Dunal (Solanaceae), a herbaceous annual weed (Aranha et al. 1982) native to South America (Nee 1991), is considered a serious weed threat and has been included in the list of noxious weeds by the Florida Department of Agriculture and Consumer Services (Florida Dept. Agric. and Consumer Services 1994). This weed has spread into other geographical regions including Central America, the Caribbean, India, China, Africa (Coile 1993, Chandra & Srivastava 1978), and south Florida (Mullahey et al. 1993). It was recently found (July 1994) in southern Mississippi by J. Byrd (Extension weed specialist, Mississippi State Univ., pers. comm.). Solanum viarum is spreading rapidly, invading pasture lands in south Florida. In 1992, infestations of pastures by S. viarum were estimated to be near 150,000 acres (Mullahey et al. 1993). A census of Florida ranchers in 1993 (J. J. M. unpublished) found that nearly 400,000 acres of pastures and natural systems were infested by S. viarum causing an annual production loss to Florida ranchers of $11 million. The rapid spread in south Florida can be partially attributed to the plant s ability to grow in sandy loam soils (Chandra & Srivastava 1978), great reproductive potential (Mullahey et al. 1993), and effective seed dispersal by cattle and wildlife (Mullahey & Colving 1993). How this South American plant entered Florida is unknown. Earliest records indicate that it was initially observed in Hendry county (southwest Florida) in the 1980s (Mullahey et al. 1993). No insects have been reported feeding on this plant in Florida. An exploratory survey was conducted by the authors during 8-14 June, 1994 to record insects feeding on S. viarum plants in Sao Paulo and Parana states, Brazil and

75 Scientific Notes 71 southeast Paraguay. Collections were made on natural stands of S.viarum on roadsides and grasslands. Insects found on the plant were trapped by hand. Plant parts (fruits, leaves, stems) were examined for insects. Fruits and stems were dissected to collect insects feeding on these internal parts. Identification to species level was determined mainly with those insects that were actually feeding on S. viarum. The insects collected on S. viarum at the different sites are listed in Table 1. The major insects collected included Neoleucinodes elegantalis Guenee (Pyralidae), a fruit borer that feeds inside S. viarum fruits destroying a great proportion of the seeds. The hostfeeding range of this insect is not known (J. Vasconcelos, Universidade Estadual Paulista do Campinas, pers. comm.). Gallo et al. (1988) listed N. elegantalis as a tomato pest in Brazil. Based on host data provided by Buzzi (1994), S. viarum represents a new host record for the cassidines Metriona elatior Klug and Gratiana boliviana Spaeth (Chrysomelidae), the larvae and adults of which feed on S. viarum leaves. Sweetpotato, Ipomoea batatas L. (Convolvulaceae), is also listed as host of M. elatior (Buzzi 1994). The genus Metriona includes six species (M. argentina Spaeth, M. tenella Klug, M. bifossulata Boheman, M. erratica Boheman, M. vilis Boheman, M. bilimeki Spaeth) that occur in Central America and/or South America (Buzzi 1988, Maes & Staines 1991) and for most of them the host ranges are unknown or incompletely known. Likewise, the host-feeding range of G. boliviana is unknown. A complete description of this cassidine is provided by Buzzi (1995). Few attempts have been made to use cassidines for biological control of Solanum weeds. Gratiana lutescens Boheman and G. pallidula Boheman were studied in South Africa (Siebert 1975) as potential biological control agents of Solanum elaeagnifolium Cav. but research efforts were not continued because the insects host range included eggplant, S. melongena L. In June 1994, the lace bugs (Hemiptera: Tingidae), Corythaica sp. near cyanthicollis (Costa) formed small aggregations on the underside of S. viarum leaves. This insect was most commonly found on S. viarum leaves. The genus Corythaica includes thirteen species that occur in the neotropics (Drave & Ruhoff 1965) and for most of them the plant host ranges are unknown or incompletely known. Amblyophallus macualatus Funkhonser (Homoptera: Membracidae) was the second most abundant species found on S. viarum stems and leaves. The butterfly Mechanitis lysimnia Fabricius (Nymphalidae: Ithominae), probably polyphagous, was collected by Pitelli, a junior author (R. A. P.), around Jaboticabal, Sao Paulo. The larvae seem to be associated with extensive S. viarum leaf-chewing damage. The known geographical distribution of M. lysimnia extends from Mexico to northern Argentina (D Abrera 1984, DeVries 1987), including Brazil. The other minor insects collected on S. viarum were not causing significant damage to the plants and most are probably polyphagous (Coccidae, Coreidae, Fulgoridae, Pentatomidae, Rhopalidae). Diabrotica speciosa Germar is a key pest in Brazil. Larvae and adults feed on roots and foliage of common beans, Phaseolus vulgaris L., field corn, Zea mays L., and potatoes, Solanum tuberosum L. This survey in Brazil and Paraguay indicated that a diverse group of insects (phytophagous and others) is associated with S. viarum and at least 5 species, M. elatior, N. elegantalis, and M. lysimnia, the tingid Corythaica sp. and the membracid A. maculatus, probably cause significant damage to S. viarum plants in South America. Paraguayan technicians (Centro Tecnologico Agropecuario) indicated that S. viarum is not considered an important weed in the region. In Brazil, S. viarum is an occasional invader of field crops and its geographical distribution includes Goias, Minas Gerais, and the southern states (Costa et al. 1985). Further surveys in the weed s native region, basic biological studies, and preliminary host range tests need to be implemented to start a biological control project for S. viarum in Florida.

76 72 Florida Entomologist 79(1) March, 1996 TABLE 1. INSECTS ASSOCIATED WITH SOLANUM VIARUM IN BRAZIL AND PARAGUAY. Family Species Plant Part Collected Insect Stage 1 Collection Site 2 Pyralidae Neoleucinodes elegantalis fruit L Taiuva, SP Divinolandia, SP Ciudad del Este Nymphalidae Mechanitis lysimnia leaf L,P Jaboticabal, SP Noctuidae leaf L Divinolandia, SP Chrysomelidae Metriona elatior leaf P,A Divinolandia, SP Ciudad del Este Chrysomelidae Gratiana boliviana leaf A Divinolandia, SP Chrysomelidae Diabrotica speciosa leaf A Taiuva, SP Divinolandia, SP Ciudad del Este Cerambycidae leaf A Ciudad del Este Tingidae Corythaica sp. leaf A Taiuva, SP Divinolandia, SP Ciudad del Este Coreidae leaf A Divinolandia, SP Pentatomidae leaf A Taiuva, SP Rhopalidae leaf A Ciudad del Este Coccidae stem N,A Taiuva, SP Fulgoridae leaf A Taiuva, SP Membracidae Amblyophallus maculatus leaf-stem N,A Taiuva, SP Divinolandia, SP Ciudad del Este Acrididae leaf A All sites Formicidae leaf-stem A All sites 1 L=larva, N=nymph, P=pupa, A=adult. 2 SP=Sao Paulo state, Brazil; Ciudad del Este in south-east Paraguay. We thank G. R. Buckingham (USDA-ARS), J. H. Frank, and D. H. Habeck (Department of Entomology and Nematology, University of Florida) for reviewing the manuscript. We also thank J. Vasconcelos (Universidade Estadual Paulista do Campinas, Sao Paulo) for identifying M. elatior, N. elegantalis, and M. lysimnia; Z. J. Buzzi and A. Sakakibara (Universidade Federal do Parana in Curitiba, PR) for identifying G. boliviana and A. maculatus, respectively; S. E. Halbert (State of Florida Department of Agriculture & Consumer Services-Division of Plant Industry, Gainesville) for identifying the Corythaica sp. Specimens of several of the species identified were retained by the taxonomists responsible for identification. Others were deposited in the insect museum of the Universidade Federal do Parana in Curitiba, Brazil, and the University of Florida Entomology & Nematology Department insect collection in Gainesville, Florida. This manuscript is published as Florida Agricultural Experiment Station Journal No. R

77 Scientific Notes 73 SUMMARY An exploratory survey was conducted in Brazil and Paraguay to record insects feeding on Solanum viarum Dunal (Solanaceae). A list of insects collected is included. The survey indicated that a diverse group of phytophagous insects is associated with S. viarum, and some of them may have potential as biocontrol agents of S. viarum in Florida. REFERENCES CITED ARANHA, C., O. BACCHI, AND H. F. LEITAO Plantas invasoras de culturas. Vol. II. Instituto Campineiro de Ensino Agricola. Campinas, Brasil. 597 p. BUZZI, Z. J Biology of neotropical Cassidinae, p , in P. Jolivet, E. Petitpierre, T. H. Hsiao [eds.]. Biology of Chrysomelidae. Kluwer Academic Publishers, Netherlands. BUZZI, Z. J Host plants of neotropical Cassidinae, p in P. H. Jolivet, N. L. Cox, and E. Petitpierre [eds.]. Novel aspects of the biology of Chrysomelidae. Kluwer Academic Publishers, Netherlands. BUZZI, Z. J Redescricao de Gratiana boliviana Spaeth, (Coleoptera: Chrysomelidae: Cassidinae). Revista Brasileira de Zoologia (In press). CHANDRA, V., AND S. N. SRIVASTAVA Solanum viarum Dunal syn. Solanum khasianum Clarke, a crop for production of solasadine. Indian Drugs 16: COILE, N. C Tropical soda apple, Solanum viarum Dunal: the plant from hell. Botany Circular No. 27. Florida Dept. Agric. & Consumer Services, Division of Plant Industry. COSTA, J., E. SANTOS, E. FROMM, N. L. COSTA, AND M. C. CUNHA Ervas daninhas do Brasil. Solanaceae I. EMBRAPA. Brasilia, Brasil. 58 p. D ABRERA, B Butterflies of South America. Hill House. Victoria, Australia. 256 p. DEVRIES, P. J The butterflies of Costa Rica and their natural history. Princeton University Press, New Jersey. 328 p. DRAVE, C. J., AND F. A. RUHOFF Lacebugs of the world: A catalog (Hemiptera: Tingidae). Smithsonian Institution. Bull Washington D.C. 634 p. FLORIDA DEPARTMENT OF AGRICULTURE AND CONSUMER SERVICES Biological control agents. Chapter 5B-57. Introduction or release of plant pests, noxious weeds, arthropods, and biological control agents. Vol. 2, p GALLO, D., O. NAKANO, S. SILVEIRA, R. PEREIRA, G. CASADEI, E. BERTI, J. R. POSTALI, R. A. ZUCCHI, S. BATISTA, AND J. DJAIR Manual de Entomologia Agricola. Editora Agronomica Ceres Ltda. Sao Paulo. 649 p. MAES, J. M., AND C. L. STAINES Catalogo de los Chrysomelidae (Coleoptera) de Nicaragua. Revista Nicaraguense de Entomologia 18: MULLAHEY, J. J., AND D. L. COLVING Tropical soda apple: A new noxious weed in Florida. University of Florida, Florida Cooperative Extension Service, Fact Sheet WRS-7. MULLAHEY, J. J., M. NEE, R. P. WUNDERLIN, AND K. R. DELANEY Tropical soda apple (Solanum viarum): A new weed threat in subtropical regions. Weed Technology 7: NEE, M Synopsis of Solanum section Acanthophora: A group of interest for glyco-alkaloides, pp in J. G. Hawkes, R. N. Lester, M. Nee, N. Estrada [eds.] Solanaceae III: Taxonomy, chemistry, evolution. Royal Botanic Gardens Kew, Richmond, Surrey, UK. SIEBERT, M. W Candidates for the biological control of Solanum elaeagnifolium Cav. (Solanaceae) in South Africa. 1. Laboratory studies on the biology of Gratiana lutescens (Boh.) and Gratiana pallidula (Boh.) (Coleoptera, Cassididae). J. Entomol. Soc. South Africa. 38:

78 74 Florida Entomologist 79(1) March, 1996 AN ERIOPHYID TEGOLOPHUS PERSEAFLORAE (ACARI:ERIOPHYIDAE) NEW TO FLORIDA AND THE USA J. E. PEÑA 1 AND H. A. DENMARK 2 1 University of Florida, Tropical Research and Education Center, SW 280th Street, Homestead, FL Florida Department of Agriculture & Consumer Services, Division Plant Industry, Gainesville, FL Keifer (1969) described an eriophyid mite Tegolophus perseaflorae sent to him by C. W. Fletchman collected from Persea gratissima from Recife, Pernambuco, Brazil. Dr. Fletchman reported this mite caused flower damage and decreased fruit production. In 1977, Dr. R. Baranowski (UF/TREC, Homestead, FL) collected this mite in the bud of avocado, Persea americana Miller, at Irupana, Bolivia. In May 1991, excessive flower drop and fruit deformation was observed in avocado trees in the vicinity of Homestead, Dade Co., Florida. In a preliminary survey, an avocado orchard was sampled in May of Sampling consisted of collections of ten floral clusters, fruitlets and buds from each of 10 trees. Tegolophus perseaflorae were collected from buds and fruits. The mites were observed feeding on buds, causing necrotic spots on apical leaves, and subcircular, irregular openings on mature leaves. Mites were also found in petioles, the underside of leaves and fruitlets (Fig. 1) The mite is also reported to feed on the peduncle, calyx and stylar area (Medina et al. 1978, Jeppson et al. 1975). Feeding by this mite on fruitlets may cause fruit deformation and discoloration. A preliminary survey was initiated in June 1991 through May 1992 to determine the relative frequency of T. perseaflorae on fruits, leaves and flowers. Initially, ten fruits, buds, and flower clusters were collected twice a month from each orchard. They were placed in an ice chest (about 10 C) and transported to the laboratory where the mites were counted. Voucher specimens were identified by the junior author. A total of 508 T. perseaflorae were collected from leaf buds and fruitlets. A significant difference in frequency of mites was observed between buds and fruits [Chi-square 0.05 (1) = 3.84]. More mites were observed on the buds (x ± SE = 5.85 ± 1.06) than in fruits (2.29 ± 0.72). Population peaks (18-35 mites per bud) were observed from March to May These warm dry months correspond with blooming and fruit formation on avocado in the area. No mites were observed on flowers, fruits or leaves from June through February. The above accounts would imply that warm weather is most favorable for this eriophyid and that the presence of developing avocado plant organs might influence its development. Description. (See Figs. 2-7.) Adult females are µ long, about 37 µ thick, abdominal thanosome with about rings; rostrum 20 µ long, curved down, shield design not clear; forelegs 24 µ long; tibia 5 µ long, tarsus 4.5 µ, and claw 5.5 µ long; featherclaw 5-rayed; lateral sets 18 µ long on about 5-7 behind the shield; first ventral seta 38 µ long on ring 20, second ventral seta 31 µ long on ring 36. Telesome with 5-6 rings, the granules fine. Accessory seta 3 µ long. Female genitalia 19 µ wide, 12 µ long, covering flap with about 18 close-set longitudinal ribs; genital seta 13 µ long. The presence of this mite in Florida represents another piece in the puzzle in the continuous appearance of neotropical pests in this state. Since, avocado grafting material is often transported by commerce, it may be that T. perseaflorae is an immigrant species that was introduced in Florida, unintentionally, by human transport. Florida Experimental Station Journal Series No. R

79 Scientific Notes 75 Figure 1. Tegolophus perseaflorae damage to buds (A) and fruit (B).

80 76 Florida Entomologist 79(1) March, 1996 Figures 2-7. Adult female Tegolophus perseaflorae (2), side of anterior section(3), left foreleg (4), featherclaw (5), lateral rings and microtubercles on thanosome (6), female genital structures and coxae (7) (after Keiffer). SUMMARY The status, damage and description of Tegolophus perseaflorae Keifer, a newly introduced mite into southern Florida, are discussed. REFERENCES CITED JEPPSON, L. R., H. H. KEIFER, AND E. E. BAKER Mites injurious to economic plants. Berkeley, California, University of California Press. 614 pp. KEIFER, H. H Eriophyid studies, California Dept. of Agric. C-1:1-20. MEDINA, C., E. BLEINROTH, J. TANGO, AND W. DO CANTO Abacate. Inst. Tecnol. Alimentos, Secr. Agric., Sao Paulo 212 p.

81 Scientific Notes 77 TEN YEAR PERSISTENCE OF A NON-AUGMENTED POPULATION OF THE BROWNBANDED COCKROACH (ORTHOPTERA: BLATTELLIDAE) PARASITOID, COMPERIA MERCETI (HYMENOPTERA: ENCYRTIDAE) A. HECHMER AND R. G. VAN DRIESCHE Department of Entomology, University of Massachusetts, Amherst, MA, The oothecal parasitoid Comperia merceti (Compere) has been used as a biological control agent against the brownbanded cockroach, Supella longipalpa (F.) (Orthoptera: Blattellidae), in environments where pesticide use is undesirable (Slater et al. 1980). In an experimental study of brownbanded cockroach populations, Coler et al. (1984) released C. merceti in 1978 in two insect rearing rooms at the University of Massachusetts at Amherst. In one room, cockroach populations were augmented during the study by provision of food. Parasitism at that site increased as the brownbanded cockroach population density increased, followed by the collapse of the cockroach population (Coler et al. 1984). After the end of the Coler et al. study in 1983, no further parasitoid releases were made in the building in which these rearing rooms were located. This note reports the status of the parasitoid populations in 1993 at the two sites used by Coler et al. (1984), fifteen years after the initial release and ten years after the last parasitoid release. Insect rearing has been conducted continuously in these same rearing rooms with little change since the initial parasitoid releases. In this note we demonstrate the ability of this parasitoid to persist under such conditions without augmentation for long periods and to cause high levels of mortality to host oothecae. On 8 and 22 November, 1993, S. longipalpa oothecae were collected from each of the two insect rearing rooms used in the Coler et al. study. These rooms were of moderate size (39.8 m 3 and 35.7 m 3 ) with loose construction providing cockroach harborage. Rooms were searched thoroughly for oothecae; search areas included walls, shelving, molding, along electrical conduits, inside light timing boxes, and the underside of tables. Oothecae inside electrical conduits, in deep wall cracks, or around insect rearing cages could not be retrieved. All oothecae (both currently live oothecae and older, emerged or dead ones) encountered in accessible areas were collected and examined for parasitism. Most oothecae contained neither live cockroaches nor parasitoid stages, but rather were oothecae which had either died, or from which cockroaches or parasitoids had previously emerged. These oothecae were probably no more than two to three years old, as the rearing rooms are periodically repainted. In the laboratory, all oothecae were examined under a stereoscope for either wasp emergence holes or the oviposition stalks present on the surface of parasitized oothecae. Oothecae from which cockroach nymphs had emerged were easily identified by opening the oothecae and observing the remnants of hatched eggs. Oothecae from which neither cockroaches nor wasps emerged were held for days and those producing either cockroach nymphs or parasitoids recorded. Oothecae from which no emergence occurred were dissected. Of the 556 oothecae collect, only three were classified as having died, with nothing emerged. The remaining oothecae all resulted in emergence of hosts or parasitoids, either in the field prior to collection, or during rearing. As a separate measure of parasitism, oothecae from a laboratory colony were exposed at one study site as trap hosts for a 14 day period (22 November-6 December, 1993). Trap host oothecae were evenly distributed on walls and shelving. These oothecae were then recovered and reared to determine the rate of parasitoid attack during

82 78 Florida Entomologist 79(1) March, 1996 TABLE 1. LEVELS OF PARASITISM OF OOTHECAE OF THE BROWNBANDED COCKROACH, SU- PELLA LONGIPALPA, BY COMPERIA MERCETI IN TWO INSECT REARING ROOMS IN FERNALD HALL, UNIVERSITY OF MASSACHUSETTS, AMHERST, MASSACHU- SETTS, USA, IN 1993, FIFTEEN YEARS AFTER INITIAL RELEASE OF THE PARASI- TOID. % Parasitism Type of Host Room No. 1 Room No. 2 Old oothecae 35.5 (6.4) 1 (214) (2.8) (312) Live oothecae 46.2 (27.1) (13) 88.2 (15.4) (17) Trap host oothecae 45.8 (19.9) (24) % confidence interval for parasitism. 2 Number of oothecae examined. 3 Number of oothecae deployed as trap hosts, 22 Nov.-6 Dec., the exposure interval, which was approximately half of the total period for oothecal development. When comparing parasitism of the total live oothecae versus total oothecae showing previous emergence, no significant difference was observed (Chi 2 =0.004, d.f.=1, p- value=0.95) (Table 1). Percent parasitism of oothecae deployed as trap hosts was lower than that of live oothecae collected in the same room. However, trap host oothecae were exposed for 14 days, approximately half of the normal exposure period of Figure 1. Comperia merceti preparing to oviposit in ootheca of Supella longipapla (Photograph courtesy of R. Coler).