PLANTÐINSECT INTERACTIONS Effect of Host Plants on Successful Parasitism by Haeckeliania sperata (: Trichogrammatidae) on Diaprepes abbreviatus (Coleoptera: Curculionidae) Eggs

1,2 1 3 DANIEL CARRILLO, JORGE E. PEN˜ A, AND JOHN L. CAPINERA

Environ. Entomol. 37(6): 1565Ð1572 (2008) ABSTRACT Diaprepes abbreviatus L. is a highly polyphagous root weevil that causes damage to several fruit crops, ornamental plants, and other naturally occurring plants in Florida. Haeckeliania sperata Pinto is a gregarious endoparasitoid that attacks D. abbreviatus eggs. We hypothesized that the reproductive success of H. sperata is affected by the host plant of D. abbreviatus. Six host plants with varying degrees of pubescence were used to determine the effect of leaf trichomes on the searching behavior and success of parasitism by H. sperata. No-choice tests showed that H. sperata was able to parasitize Diaprepes eggs laid on the six host plants. However, the plants with a high trichome density on their leaves had a lower percent of parasitism than the plants with smoother leaves. Removing trichomes from a host plant showed that the presence of some leaf trichomes had a negative effect on the overall searching efÞciency of H. sperata. The presence of trichomes decreased the total distance traveled and increased the total search time of H. sperata females, resulting in a lower searching speed. Multiple regression analysis was used to establish the relationship between percent parasitism and leaf trichome density, leaf thickness, and the searching speed of H. sperata. Searching speed was the best predictor of percent parasitism; this parameter could be used to predict the suitability of host plants for the establishment of H. sperata. Our Þndings suggest that the reproductive success of H. sperata is affected by the host plant of D. abbreviatus.

KEY WORDS Diaprepes root weevil, egg , host plants, trichomes, searching speed

Diaprepes abbreviatus L. is a highly polyphagous root The pest status of D. abbreviatus is partly attribut- weevil that is native to the Lesser Antilles and was able to its polyphagy. The list of host plants of D. unintentionally introduced to Florida in the mid-1960s abbreviatus includes Ͼ270 plant species from 60 plant (Woodruff 1985). Since its Þrst detection in Apopka families (Simpson et al. 1996, Mannion et al. 2003). (Orange County) in 1964, this weevil has spread Many Diaprepes host plants are important agricultural throughout the central and southern part of the Flor- crops such as citrus (all varieties), papaya, sugarcane, ida peninsula. Diaprepes is now considered established peanut, sorghum, sweet potato, cassava, loquat, guava, in 23 counties, infesting Ͼ100,000 acres of citrus mango, avocado, banana, and corn. Some of the other groves and many other agricultural, ornamental, and reported hosts are widely used by the ornamental wild plants (Nguyen et al. 2003, Weissling et al. 2004). plant industry, and they include native plants and Estimates show that this pest has increased production invasive species, all of great ecological signiÞcance costs Ͼ70 million dollars annually for the citrus in- (Simpson et al. 1996). The broad host range of D. dustry in Florida (Stanley 1996, Muraro 2000). More- abbreviatus represents a signiÞcant challenge for de- over, the infestation of this weevil has spread to Texas veloping efÞcient and sustainable management tac- and California (Grafton-Cardwell et al. 2004, CDFA tics. 2006), resulting in drastic measures to restrict the Classical biological control of Diaprepes abbreviatus introduction of ornamental plants from Florida (TDA is viewed as one of the components of integrated pest 2001). management (IPM) programs with potential to effec- tively reduce weevil numbers (Pen˜ a et al. 2001, 2004). Several efforts have been made to identify and intro- duce classical biological control agents of D. abbre- 1 Department of Entomology and Nematology, University of Flor- ida, Tropical Research and Education Center, 18905 SW 280 St., viatus in an attempt to reduce weevil populations to Homestead, FL 33031. tolerable levels. Three species of egg have 2 Corresponding author: University of Florida, IFAS, Department of been introduced to Florida. As a result of these intro- Entomology and Nematology, Bldg. 970, PO Box 110620, Gainesville, ductions, two eulophid parasitoids [Quadrastichus FL 32611 (e-mail: dancar@uß.edu). 3 University of Florida, Department of Entomology and Nematol- haitiensis (Gahan) and Aprostocetus vaquitarum (Wol- ogy, Bldg. 970 Natural Area Dr., Gainesville, FL 32611. cott)] have successfully established in the southern

0046-225X/08/1565Ð1572$04.00/0 ᭧ 2008 Entomological Society of America 1566 ENVIRONMENTAL ENTOMOLOGY Vol. 37, no. 6 portion of the Florida peninsula where they have unpublished data). Thus, physical and chemical leaf become an important mortality factor of Diaprepes surface characters can alter parasitoid searching and eggs (Pen˜ a et al. 2004, 2006). Another promising bio- host encounter rates. control agent against D. abbreviatus is Haeckeliania Another factor that can affect the ability of H. spe- sperata (Hymenoptera: Trichogrammatidae), a gre- rata to parasitize D. abbreviatus eggs is the leaf thick- garious, egg-endoparasitoid that was collected in Do- ness of the host plant. H. sperata inserts its ovipositor minica in 2003 (Pen˜ a et al. 2006). This species was through the leaf to reach the weevil egg mass ce- found as a primary parasitoid parasitizing Ͼ50% of egg mented between two leaves (Pen˜ a et al. 2004). This masses found in citrus groves and citrus nurseries in aspect of H. sperata’s behavior suggests that the rela- Dominica (Pen˜ a et al. 2006). Jacas et al. (2008) studied tionship between the length of the ovipositor and leaf the thermal requirements of H. sperata and found that thickness could be critical to the successful parasitism this parasitoid has a greater thermal plasticity than the of D. abbreviatus in certain host plants. Given the wide parasitoids already established in south Florida, which range of host plants of D. abbreviatus, knowledge of could increase their chances of establishment in the the potential antagonism between the host plants and central portion of the state. Furthermore, H. sperata H. sperata is particularly important. was collected in areas of high altitude (509 m), which The study presented here hypothesizes that the suggests that it could have more cold hardiness than reproductive success of H. sperata is affected by the other imported parasitoids. In addition to the adapt- host plant of D. abbreviatus. The objectives of this ability of this parasitoid to FloridaÕs climate, other study were to (1) determine the successful parasitism important factors could inßuence its Þtness in its new of D. abbreviatus eggs by H. sperata on different host environment. plants, (2) determine the effect of leaf trichomes on An array of factors of the host plant can affect the the searching behavior of H. sperata, and (3) deter- colonization, phenology, and effectiveness of natural mine the relationship between searching speed, enemies used in biological control (Bottrell and Bar- trichome density and leaf thickness to parasitism of H. bosa 1998). Plants use many physical and chemical sperata on Diaprepes eggs. defenses against herbivores that are not neces- sarily compatible with natural enemies. Physical de- fenses may include structural traits such as surface Materials and Methods waxes, leaf trichomes, and/or spines, whereas chem- Stock Colonies ical defenses are often secondary metabolites that act as repellents, toxins, and/or digestibility reducers Haeckeliania sperata used in each of the experi- (Jeffree 1986, Southwood 1986). Plant defense ments described below were collected in Dominica in mechanisms may directly or indirectly inhibit nat- 2003 and reared for several generations at the Tropical ural enemies. Research and Education Center (TREC) insectary Direct plant effects on parasitoids may involve sim- (L:D: 12:12 h, 26.5 Ϯ 1ЊC, and 75% RH). Parasitoids ple mechanisms such as reduced parasitoid searching were reared on D. abbreviatus eggs from adult weevils efÞciency caused by trichomes (Bottrell and Barbosa that were collected from a pesticide-free commercial 1998).An interesting case is the negative effect of nursery in Homestead, FL. The weevils were placed glandular trichomes of some alfalfa cultivars on the randomly in groups of 300 in Plexiglas cages (30 by 30 parasitism of Empoasca fabae (Harris) (Hemiptera: by 30 cm) and fed on green buttonwood, Conocarpus Cicadellidae) by Anagrus nigriventris Girault (Hyme- erectus L. (Combretaceae), 2 wk before the beginning noptera: Mymaridae) (Lovinger et al. 2000). In an- of the experiments. Several collections of weevils were other study, Mulatu et al. (2006) found that the pres- made during the course of all the experiments. The ence of trichomes on tomato leaves had a direct parasitoids used in the experiments were presumably adverse effect on the parasitoid Diadegma pulchripes mated, fed, naõ¨ve with respect to hosts, and were 1 d (Kokujev) (Hymenoptera: ). In con- old at the time of testing. trast, Styrsky et al. (2006) found that the presence of trichomes in soybean plants did not inhibit Þre Plant Material [Solenopsis invicta Buren (Hymenoptera: Formici- dae)] from foraging on plants and that predation of Six host plants with different trichome densities and herbivores was greater in pubescent plants when com- leaf thickness were chosen (Table 1). The fruit crops pared with glabrous plants. selected for this study were lime, Citrus aurantifolia, a The effects of host plants have proven to be espe- many-branched, spiny shrub with elliptic ovate gla- cially important for relatively small parasitoids (Rabb brous leaves (Bayley and Bayley 1976); and loquat, and Bradley1968, Jeffree 1986, Lukianchuk and Smith Eriobotrya japonica, a small tree with alternate, short- 1997, Su¨ tterlin and Van Lenteren 1997, 1998, 2000, petioled leaves with a high density of long, unicellular, Romeis et al. 1998, 1999, 2005, Styrsky et al. 2006). nonglandular trichomes (Bayley and Bayley 1976). Because of the small size of H. sperata, trichomes and Ornamental host plants selected were pygmy palm, trichome exudates could inhibit the parasitoidÕs move- Phoenix roebelinii, a sometimes clustered slender palm ments. As with many other trichogrammatids with pinnate narrow pinnae with sparse unicellular, (Schmidt 1994, Romeis et al. 2005), H. sperata searches nonglandular trichomes; button mangrove, C. erectus, for weevil eggs by walking on the leaf surface (D.C., a native erect shrub or tree, with elliptic to ovate December 2008 CARRILLO ET AL.: EFFECT OF D. abbreviatus HOST PLANTS ON PARASITISM BY H. sperata 1567

Table 1. Trichome density and leaf thickness of six host plants Statistical Analysis. Analysis of covariance (ANCOVA; means ؎ SEM) SAS Institute 1999) was used to detect the effects of ,10 ؍ of D. abbreviates (N host plants on the level of parasitism by H. sperata in Leaf No. Plant thickness the no-choice tests. Egg mass size (number of Dia- trichomes/cm2 (mm) prepes eggs per egg mass) was used as a covariate to Citrus aurantifolia (Christm.) 0 0.14 Ϯ 0013 control variation associated with host abundance. (Rutaceae) Means were compared using TukeyÕs honest signiÞ- Conocarpus erectus L. 22 Ϯ 2.5 0.25 Ϯ 0.021 cance test (P Ͻ 0.05). (Combretaceae) Phoenix roebelinii OÕBrien 29 Ϯ 0.7 0.21 Ϯ 0.011 (Arecaceae) Effect of Leaf Trichomes on the Searching Pennisetum purpureum 92.4 Ϯ 2.7 0.07 Ϯ 0.006 Behavior of H. sperata Schumach (Poaceae) Eriobotrya japonica Lindl. 429.6 Ϯ 17.5 0.22 Ϯ 0.011 This experiment was designed to determine the (Rosaceae) C. erectus var. sericeus Fors 39,030 Ϯ 3,487 0.17 Ϯ 0.013 effect of leaf trichomes in two host plants on the ex. DC (Combretaceae) walking speed of H. sperata. Only the two biotypes of C. erectus were used in this experiment. The searching speed of H. sperata females was measured on leaves nearly glabrous leaves; and silver buttonwood, C. erec- kept intact or leaves where the trichomes were re- tus variety. sericeus Fors ex. DC with silky-hairy leaves moved by two methods. A set of leaves of each C. (Bayley and Bayley 1976). The sixth host plant se- erectus biotype was washed with a 0.1% Triton X-100 lected for this study was elephant grass, Pennisetum solution, a nonphytotoxic detergent that helps remov- purpureum, an invasive grass with often branched ing trichomes and leaf-surface linked chemicals. stems and elongate blades with a high density of long, Trichomes were removed from another set of leaves of unicellular trichomes. C. erectus variety sericeus by shaving them with a razor blade under a microscope. We included wax paper (Cut-Rite; Reynolds, Richmond, VA) as a nonleaf sur- Successful Parasitism by H. sperata on D. abbreviatus face without the potential confounding effects from Eggs on Six Host Plants chemicals and epicuticular waxes. The mean search- No-Choice Tests. Leaves free of any damage were ing speed on each substrate was measured for 10 removed from each host plant, leaving a 10-cm petiole, mated, 1-d-old and honey-fed H. sperata females. Leaf stem, or other plant structure to hold the cutting Þrmly disks (6.3 cm2) of each surface (abaxial and adaxial) upright when formed into bouquets with exposed were excised and tightly placed in the bottom of a petri leaves (Ϸ1,500 cm2 of leaf surface). The bouquets dish of the same size. Female were individual- were formed by inserting the stems into wet ßorist ized in similar petri dishes 30 min before each obser- foam covered with aluminum foil, leaving the leaves vation. Females were placed on the middle of each exposed. The bouquets were placed separately into surface, and the petri dish was left uncovered, which Plexiglas oviposition cages (30 by 30 by 30 cm) along allowed ventilation and the females to leave the sur- with 250 adult weevils (50:50 (:Ǩ) for 24 h. After that face freely. The walking patterns were monitored us- period, the egg-infested bouquets were removed from ing a video camera (Video Flex 7000 series; Ken-a- the weevil oviposition cages, and the leaf area and vision, Inc., Kansas City, MO), equipped with a macro number of egg masses were standardized (250 cm2 of lens mounted 5 cm above the plane of the surface. leaf surface and Þve egg masses per bouquet). The Walking patterns were traced onto acetate sheets at- standardized bouquets were placed individually in tached to a television screen. The path and location smaller cages (17 by 10 by 10 cm), which were pre- were recorded every 10 s until the left the sur- viously provisioned with a small drop of honey as food face or after 2 min had passed. The acetate sheets were source for the parasitoids. The control treatment con- scanned to convert the walking traces into digital sisted of opened egg masses laid on wax paper that images. The traces were processed using the image were obtained by hanging doubled strips of wax paper processing and analyzing software Scion Image. This on the sides of a weevil cage during the oviposition procedure allowed us to calculate the total search time period. Twenty 1-d-old adults of H. sperata (1:1 Ǩ:() (s), the time spent motionless (s), the total distance were introduced into each cage, and parasitism was traveled (mm), the walking speed (mm/s), and the allowed for 24 h. The bouquets were removed from searching speed of each female on each surface (mm/ the cages and checked under a microscope to make s). Searching speed is the average velocity of H. spe- sure that no parasitoids were left in the bouquets. rata during the whole observational period (consid- Seven days later, the leaves were torn open, and par- ering the time spent walking and immobile). Walking asitized eggs were identiÞed by the presence of four speed refers to the average velocity only while walk- to six compartments within a single Diaprepes egg, ing. Searching speed is usually the variable of interest each containing one H. sperata individual (Ulmer et al. because it is a quantitative measure of how far a para- 2006). Percent parasitism was determined by dividing sitoid searches in a given time (Van Hezewijk et al. the number of parasitized eggs by the total number of 2000). eggs on each egg mass. This experiment was replicated Statistical Analysis. This experiment was replicated seven times, each time with new parasitoids. 13 times, each time with new parasitoids and leaves. 1568 ENVIRONMENTAL ENTOMOLOGY Vol. 37, no. 6

Fig. 1. Effect of host plant on successful parasitism of D. abbreviatus eggs by H. sperata. No-choice tests on six host plants and control treatment (wax paper). Different letters represent signiÞcant differences (P Ͻ 0.05; ANOVA, TukeyÕs studentized range test); error bars represent SE.

ANOVAs (SAS institute 1999) were used to detect the Trichome density and leaf thickness were mea- effect of trichomes on the total search time, the time sured on 10 newly expanded leaves of the same plant spent motionless, the total distance traveled, the walk- material used in the previous experiments. Two ing speed, and the searching speed of H. sperata fe- squares (1 cm2 each) from the middle part of the males on C. erectus leaves kept intact or leaves where leaves were excised; one square was used to deter- the trichomes were removed. Means were compared mine the leaf thickness, and the other one was used using TukeyÕs honest signiÞcance test (P Ͻ 0.05). for the trichome density for each leaf surface (ab- axial and adaxial). Leaf thickness was determined by placing the excised square standing on a piece of Relationship of Searching Speed, Trichome Density, clay and measuring its thickness under a microscope and Leaf Thickness to Parasitism (40ϫ) equipped with a reticule. Trichome density This experiment was designed to measure the was established by counting the total number of degree to which variation in searching speed, trichomes on each surface (abaxial and adaxial) and trichome density, and leaf thickness explained the making an average of the two values. Because of the variation in parasitism. The six host plants used in high density of trichomes on C. erectus variety seri- the Þrst experiment were included. All observations ceus leaves, the 1-cm2 squares were divided in four were made between 1300 and 1700 hours. Leaves equal 0.25-cm2 sections and trichomes counted in from each plant were collected and divided in three one of the four sections. groups. The Þrst group was used to test the propor- Statistical Analysis. Multiple regression (PROG tion of hosts parasitized on each plant, the second REG) analysis was used to establish the relationship group was used to determine the searching speed on between three explanatory variables (searching each leaf surface, and the third group was used to speed, trichome density, and leaf thickness) and the measure the trichome density and leaf thickness of response variable level of parasitism (SAS Institute each plant biotype. 1999). The proportion of parasitized hosts was determined in an experiment similar to the no-choice tests. Bou- quets of each host plant bearing Þve egg masses were Results introduced in 2-liter square jars (17 by 10 by 10 cm) provisioned with a drop of honey as food source for Successful Parasitism by H. sperata on D. abbreviatus the parasitoids. Ten 1-d-old females of H. sperata were Eggs on Six Host Plants introduced into each cage. After the parasitism period, the bouquets were removed from the cages and The no-choice test showed that H. sperata is able to checked under a microscope to make sure that no parasitize D. abbreviatus eggs laid on all six host plants parasitoids were on the bouquets. Seven days later, but failed to parasitize egg masses on wax paper (Fig. parasitized eggs were counted and percent parasitism 1). The size of egg masses did not vary signiÞcantly was determined by dividing the number of parasitized among the plants (F ϭ 1.08; df ϭ 6,47; P ϭ 0.29). eggs by the total number of eggs on each egg mass of However, eggs on leaves with higher densities of each host plant. Searching speed of H. sperata on the trichomes had signiÞcantly lower parasitism com- different host plant leaves was measured in the same pared with those on glabrous leaves (C. aurantifolia) way as in the previous experiment. This experiment or a low density of trichomes (P. roebelinii and C. was replicated 10 times, each time with new parasi- erectus; F ϭ 17.22; df ϭ 6,47; P Ͻ 0.01; Fig. 1; Table 1). toids and plants. Furthermore, percent parasitism was Ͼ10 times December 2008 CARRILLO ET AL.: EFFECT OF D. abbreviatus HOST PLANTS ON PARASITISM BY H. sperata 1569

(Table 2. Effect of leaf trichomes on the searching behavior of H. sperata (means ؎ SEM

Total distance Total time on Walking speed Searching speed Leaf surface traveled (mm) surface (s) (mm/s) (mm/s) Wax paper 79.0 Ϯ 7.8 a 59.2 Ϯ 5.5 c 7.8 Ϯ 3.9 a 1.9 Ϯ 0.3 a C. erectus 59.0 Ϯ 4.3 b 72.3 Ϯ 4.7 bc 5.7 Ϯ 0.6 b 1.0 Ϯ 0.2 b C. erectus var. sericeus 17.9 Ϯ 1.2 c 94.2 Ϯ 6.3 a 1.9 Ϯ 0.2 c 0.2 Ϯ 0.04 c C. erectus var. sericeus Shaved 54.9 Ϯ 3.8 b 61.2 Ϯ 6.1 c 5.1 Ϯ 0.5 b 1.5 Ϯ 0.4 ab C. erectus ϩ Triton X-100 56.0 Ϯ 3.8 b 76.9 Ϯ 5.1 ab 5.5 Ϯ 0.5 b 0.9 Ϯ 0.2 bc C. erectus var. sericeus ϩ 50.8 Ϯ 3.2 b 82.7 Ϯ 6.2 ab 5.2 Ϯ 0.5 b 0.8 Ϯ 0.1 bc Triton X-100

Means within a column followed by the same letter are not signiÞcantly different (P Ͻ 0.05; ANOVA, TukeyÕs studentized range test). greater in the glabrous morph of C. erectus compared associated with them are more important than the with the pubescent biotype. number per unit of area.

Effect of Leaf Trichomes on the Searching Discussion Behavior of H. sperata Haeckeliania sperata was able to parasitize eggs laid The overall results of this experiment show that the on the leaves of all the host plants. However, the levels presence of trichomes has a signiÞcant negative effect of parasitism were affected by variations in the leaf in the searching efÞciency of H. sperata. There was no surface structures of the host plants. We tested six host effect of the leaf surface orientation (abaxial and adax- plants with various degrees of pubescence and found ial) on any of the variables measured (F ϭ 0.03; df ϭ that the reproductive success of H. sperata is much 1,141; P ϭ 0.86). The total distance traveled by H. higher in plants with smooth surfaces, like C. erectus or sperata females was three-fold greater on any of the C. aurantifolia, than on those with pubescence (C. surfaces that had no trichomes than on the pubescent erectus variety sericeus and E. japonica). Considering surface (Table 2). When the trichomes were removed that H. sperata was originally collected from citrus from the pubescent morph by shaving them or treating groves and nurseries where it acted as a primary para- the leaf with Triton X-100, the traveled distance in- sitoid, we could expect better chances of control and creased to levels similar to the glabrous green morph establishment in citrus plants which are one of the (Table 2; F ϭ 23.07; df ϭ 5,141; P Ͻ 0.001). In contrast, major hosts of D. abbreviatus in central and south H. sperata females spent more time on the pubescent Florida. Conocarpus erectus is used for rearing H. spe- surface and less on the smooth leaf surfaces. When the rata; our results reafÞrm that this plant is the best trichomes were shaved from the pubescent morph, known option for rearing this parasitoid. Besides being the total time spent on the surface decreased (Table available throughout the year in south Florida, a high 2; F ϭ 4.55; P Ͻ 0.01). The walking speed was signif- parasitism rate was found on this plant in all the ex- icantly higher on all the surfaces with no trichomes periments. In contrast, parasitism was lowest on the compared with the pubescent surfaces. When the pubescent morph of C. erectus. These results are sim- trichomes were removed by any method, the walking ilar to those obtained by Amalin et al. (2005) working speed increased to levels similar to the smooth sur- with another trichogrammatid parasitoid of D. abbre- faces (F ϭ 22.9; P Ͻ 0.001). A similar pattern was viatus, Ceratogramma etiennei Delvare. In a no-choice observed in the searching speed analysis; however, the experiment with four of the same host plants that we searching speed on surfaces treated with Triton was used, C. etiennei showed a high parasitism of D. ab- signiÞcantly lower, which could suggest a confound- breviatus eggs on C. aurantifolia, P. roebelinii, and C. ing effect of this detergent that made the parasitoids stay longer on the surface (F ϭ 4.93; P Ͻ 0.01). Table 3. Relationship of searching speed, trichome density, and leaf thickness to parasitism Relationship of Searching Speed, Trichome Density, Pc Model Slope SE Pa R2b and Leaf Thickness to Parasitism (overall) Multiple regression analyses showed an association Parasitism between the response variable (percent of parasitism) Searching speed 2.8 0.16 Ͻ0.01 0.89 Ͻ0.01 Ϫ6 Ϫ5 and the full set of explanatory variables (Mean Square Trichome density 1.2 ϫ 10 2.1 ϫ 10 0.53 Leaf thickness 12.02 7.16 0.016 Error ϭ 708.41; P Ͻ 0.001). However, when testing for association between percent of parasitism and the Multiple regression analysis of the effect of searching speed, individual explanatory variables, we found that trichome density, and leaf thickness on mean percent of parasitism of searching speed was useful in predicting parasitism, H. sperata on D. abbreviatus eggs laid on six host plants. a but neither trichome density nor leaf thickness ex- P values associated with each parameter and signiÞcance of the variable in the predicted response. plained the variation on parasitism (Table 3). It is b R2, proportion of the variance explained by the model. likely that the type of trichomes and the exudates c P(overall) is for the test of signiÞcance of the model using GLM. 1570 ENVIRONMENTAL ENTOMOLOGY Vol. 37, no. 6 erectus and a low parasitism on the pubescent C. erec- aniaÕs Þtness. Further studies are needed to evaluate tus variety sericeus. the importance of these factors. Schoener (1987) suggested that the presence of Our study underlines the importance of habitat se- trichomes on C. erectus represents a putative defense lection on the success of inoculative and augmentative against defoliation by herbivores. He also suggested biological control programs. To increase the chances that trichomes are costly, occurring differentially of establishment of H. sperata, releases should be done where the threat of defoliation is greater. Interest- where suitable host plant species exist. As seen in ingly, Mannion et al. (2003) found that D. abbreviatus, other trichogrammatids (Flanders 1937), it is possible when given a choice, prefers to oviposit on C. erectus that H. sperata will become more prevalent in certain variety sericeus, where it has better larval survivor- habitats or on speciÞc plants. Also, parasitism levels of ship. This suggests that the presence of trichomes in D. abbreviatus can vary widely, depending on the C. erectus variety sericeus has no adverse effects on plants on which the eggs are found. In addition, par- the Þtness of D. abbreviatus. Moreover, results pre- asitism by H. sperata may vary with the plant structure sented here and those reported by Amalin et al. on which the host eggs are located. For instance, most (2005) suggest that the presence of trichomes in C. of the Florida Þelds that are known to be infested with erectus could be beneÞcial for D. abbreviatus in Diaprepes are either citrus monocultures or open nurs- terms of avoiding parasitism by H. sperata and C. eries. Citrus monocultures seem to be suitable sce- etiennei, respectively. narios for releasing H. sperata. However, citrus groves The presence of trichomes on C. erectus variety also have weeds that could serve as reservoirs of Dia- sericeus has a negative effect on the searching efÞ- prepes eggs were a low parasitism by H. sperata could ciency and therefore on the Þtness of H. sperata. be expected, such as P. purpureum. The situation in Trichomes can alter the walking patterns of many open nurseries could be more complex. South Flori- parasitoids by changing the distribution of turning daÕs open nurseries are often composed of a mosaic of angles, reducing their searching speed, and ultimately plants that are randomly placed in the Þelds. Under affecting their host Þnding ability (Price et al. 1980, these conditions, we could expect a highly variable parasitism by H. sperata depending on the plant com- Su¨ tterlin and Van Lenteren 1997, Mulatu et al. 2006, Olson and Andow 2006). In theory, faster-walking position of the nursery. Our results suggest that dif- ferences in the establishment and levels of control by females have a higher chance of Þnding hosts because H. sperata in different habitats with varying plant com- they can search a larger surface area per unit of time position can be expected. (Olson and Andow 2003). In fact, searching speed has been adopted by the IOBC working group on “quality control of mass reared organisms” as an important criterion for preintroductory selection of suitable Acknowledgments strains of spp. parasitoids (Limburg and This study is partial fulÞllment of requirements for the Pak 1991). Searching speed has been used to select MSc, degree, University of Florida, by D.C. We thank J. Jacas strains and/or releasing sites of several trichogram- (U. Jaume, I, Spain), H. Frank (University of Florida, Gaines- matid species (i.e., Bezdenko, ville, FL), and B. Ulmer (Syngenta) for helpful suggestions Trichogramma nubilale Ertle and Davis, Trichogramma to improve this manuscript and D. Long, R. Duncan, Z. exiguum Pinto and Platner, and Trichogramma chilonis Alegria, and J. Alegria for assistance. This work was supported Ishii) (Keller 1987). However, Van Hezewijk et al. in part by a CSREES grant to J.E.P. (2000) reported that the level of parasitism in Tri- chogramma minutum Riley was positively affected by host acceptance but was not related to the parasiteÕs References Cited searching speed. Thus, searching speed cannot always Amalin, D. M., J. E. Pen˜ a, and R. E. Duncan. 2005. Effects be used to predict parasitoid quality or to select plants of host age, female parasitoid age, and host plant on or release sites. However, our results suggest that parasitism of Ceratogramma etiennei (Hymenoptera: Tri- searching speed is a parameter tightly linked to par- chogrammatidae). Fla. Entomol. 88: 77Ð82. asitism and that it is feasible to use this parameter to Bayley, L. H., and E. Z. Bayley. 1976. Hortus third, a concise predict the suitability of host plants for releasing H. dictionary of plants cultivated in the United States and Canada. Macmillan, New York. sperata. Bottrell, D. G., and P. Barbosa. 1998. Manipulating natural Diaprepes has many host plants in Florida that rep- enemies by plant variety selection and modiÞcation: a resent an array of leaf structures and types of realistic strategy? Annu. Rev. Entomol. 43: 347Ð367. trichomes that could hinder the reproductive success Burbutis, P. P., G. D. Curl, and C. P. Davis. 1977. Host of H. sperata. The degree of control of Diaprepes pop- searching behavior by Trichogramma nubilale on corn. ulations by H. sperata may depend on, among other Environ. Entomol. 6: 400Ð402. factors, the interaction of the parasitoids with the [CDFA] California Department of Food and Agriculture. microenvironment created by the plants on the leaves 2006. Pest advisory number 04Ð06. CDFA, Sacra- mento, CA. surfaces. Other factors such as the total leaf area, host Flanders, S. E. 1937. Habitat selection by Trichogramma. plant foliage structural complexity (Gingras and Ann. Entomol. Soc. Am. 30: 208Ð210. Boivin 2002, Gingras et al. 2003), and surface-associ- Gingras, D., and G. Boivin. 2002. Effect of plant structure, ated chemicals could also have an effect on Haeckeli- host density and foraging duration on host Þnding by December 2008 CARRILLO ET AL.: EFFECT OF D. abbreviatus HOST PLANTS ON PARASITISM BY H. sperata 1571

Trichogramma evanescens (Hymenoptera: Trichogram- Pen˜ a, J. E., D. G. Hall, R. Nguyen, C. W. McCoy, D. matidae). Environ. Entomol. 31: 1153Ð1157. Amalin, P. Stansly, R. Adair, S. Lapointe, R. Duncan, Gingras, D., P. Dutilleul, and G. Boivin. 2003. Effect of and A. Hoyte. 2004. Recovery of parasitoids (Hyme- plant structure on host Þnding capacity of lepidopterous noptera: and Trichogrammatidae) released pests of crucifers by two Trichogramma parasitoids. Biol. for biological control of Diaprepes abbreviatus (Co- Control 27: 25Ð31. leoptera: Curculionidae) in Florida. Proc. Int. Citrus. Grafton-Cardwell, E. E., K. E. Godfrey, J. E. Pen˜ a, C. W. Congress 3: 879Ð884. McCoy, and R. F. Luck. 2004. Diaprepes root weevil. Pen˜ a, J. E., B. J. Ulmer, J. A. Jacas, R. E. Duncan, and C. W. University of California, Oakland, CA. McCoy. 2006. Biological control of neotropical citrus Jacas, J. A., J. E. Pen˜ a, R. E. Duncan, and B. J. Ulmer. 2008. root weevils. Memorias del IV Congreso Internacional de Thermal requirements of Fidiobia dominica (Hymenop- Control Biolo´gico, 31 May to 2 June 2006, Palmira, Co- tera: ) and Haeckeliania sperata (Hyme- lombia. noptera: Trichogrammatidae), two exotic egg parasitoids Price, P. W., C. E. Bouton, P. Gross, B. A. McPheron, J. N. of Diaprepes abbreviatus (Coleoptera: Curculionidae). Thompson, and A. E. Weis. 1980. Interactions among BioControl 53: 451Ð460. three trophic levels: inßuence of plants on interactions Jeffree, C. E. 1986. The cuticle, epicuticular waxes and between insect herbivores and natural enemies. Annu. trichomes, with reference to their structure, functions Rev. Ecol. Syst. 11: 41Ð65. and evolution, pp. 23Ð64. In B. E. Juniper and T.R.E. Rabb, R. L., and J. R. Bradley. 1968. The inßuence of host Southwood (eds.), and the plant surface. Edward plants on parasitism of eggs of the tobacco hornworm. J. Arnold Publishers, London, United Kingdom. Keller, M. A. 1987. Inßuence of leaf surfaces on movements Econ. Entomol. 61: 1249Ð1252. by the hymenopterous parasitoid Trichogramma exiguum. Romeis, J., T. G. Shanower, and C.P.W. Zebitz. 1998. Phys- Entomol. Exp. Appl. 43: 55Ð59. ical and chemical characters inhibiting the searching be- Limburg, H., and G. A. Pak. 1991. Genetic variation in the havior of Trichogramma chilonis. Entomol Exp. Appl. 87: walking behaviour of the egg parasitoid Trichogramma, 275Ð284. pp. 47Ð55. In F. Bigler (ed.), Proceedings, 5th workshop Romeis, J., T. G. Shanower, and C.P.W. Zebitz. 1999. of the IOBC global working group on quality control of Why Trichogramma (Hymenoptera: Trichogrammati- mass reared . Swiss Federal Research Station dae) egg parasitoids of Helicoverpa armigera (Lepi- for Agronomy, Zu¨ rich, Switzerland. doptera: Noctuidae) fail on chickpea. Bull. Entomol. Lovinger, A., D. Liewehr, and W. O. Lamp. 2000. Glandular Res. 89: 89Ð95. trichomes on alfalfa impede searching behavior of the Romeis, J., D. Babendreier, F. L. Wa¨ckers, and T. G. potato leafhopper parasitoid. Biol. Control 18: 187Ð192. Shanower. 2005. Habitat and plant speciÞcity of Tri- Lukianchuk, J. L., and S. M. Smith. 1997. Inßuence of plant chogramma egg parasitoids-underlying mechanisms and structural complexity on the foraging success of Tri- implications. Basic Appl. Ecol. 6: 215Ð236. chogramma minutum: a comparison of search on arti- SAS Institute. 1999. SAS version 9.1. SAS Institute, Cary, Þcial and foliage models. Entomol. Exp. Appl. 84: 221Ð NC. 228. Schmidt, J. M. 1994. Host recognition and acceptance by Mannion, C. M., A. Hunsberger, J. E. Pen˜ a, and L. S. Osborne. Trichogramma, pp. 165Ð200. In E. Wajnberg and S. S. 2003. Oviposition and larval survival of Diaprepes abbre- Hassan (eds.), Biological control with egg parasitoids. viatus (Coleoptera: Curculionidae) on select host plants. CAB International, Wallingford, UK. Fla. Entomol. 86: 165Ð173. Schoener, T. W. 1987. Leaf pubescence in buttonwood: Mulatu, B., S. W. Applebaum, and M. Coll. 2006. Effect of community variation in a putative defense against defo- leaf traits on the potato tuber moth and its predominant liation. Proc. Natl. Acad. Sci. U.S.A. 84: 7992Ð7995. larval parasitoid: a mechanism for enemy-free space. Biol. Simpson, S. E., H. N. Nigg, N. C. Coile, and R. A. Adair. 1996. Control 37: 231Ð236. Diaprepes abbreviatus (Coleoptera: Curculionidae): host Muraro, R. P. 2000. Cost beneÞt analysis for controlling plant associations. Environ. Entomol. 25: 333Ð349. Diaprepes, pp. 105Ð110. In S. H. Futch (ed.). Proceedings, Southwood, T.R.E. 1986. Plant surfaces and insects-an over- Diaprepes short course. University of Florida, Institute of view, pp. 1Ð22. In B. E. Juniper and T.R.E. Southwood Food and Agricultural Sciences, Gainesville, FL. (eds.), Insects and the plant surface. Edward Arnold Nguyen, R., D. G. Hall, J. E. Pena, D. Amalin, C. W. ˜ Publishers, London, United Kingdom. McCoy, S. L. Lapointe, B. Adair, and P. Stansly. 2003. Stanley, D. 1996. Suppressing a serious citrus pest. Agric. Rearing methods for Quadrastichus haitiensis (Gahan) Res. 44: 22. (Hymenoptera: Eulophidae) for biological control of Diaprepes abbreviatus (L.) (Coleoptera: Curculion- Styrsky, J. D., I. Kaplan, and M. D. Eubanks. 2006. Plant idae). (http://www.doacs.state.fl.us/pi/methods/ trichomes indirectly enhance tritrophic interactions in- diaprepesposter.html). volving a generalist predator, the red imported Þre . Olson, D. M., and D. A. Andow. 2003. Inheritance of host Biol. Control 36: 375Ð384. Þnding ability on structurally complex surfaces. Oecolo- Su¨ tterlin, S., and J. C. Van Lenteren. 1997. Inßuence of gia (Berl.) 136: 324Ð328. hairiness of Gerbera jamesonii leaves on the searching Olson, D. M., and D. A. Andow. 2006. Walking pattern of efÞciency of the Parasitoid Encarsia formosa. Biol. Con- Trichogramma nubilale Ertle & Davis (Hymenoptera: Tri- trol 9: 157Ð165. chogrammatidae) on various surfaces. Biol. Control 39: Su¨tterlin, S., and J. C. Van Lenteren. 1998. Foraging behavior 329Ð335. of the parasitoid Encarsia formosa on Gerbera jamesonii Pen˜ a, J. E., D. Hall, R. Nguyen, R. E. Duncan, D. Amalin, P. leaves. J. Insect. Behav. 12: 105Ð120. Stansly, C. W. McCoy, A. Adair, S. E. Lapointe, H. Brown- Su¨ tterlin, S., and J. C. Van Lenteren. 2000. Pre- and post- ing, and J. Knapp. 2001. Efforts toward establishment of landing response of the parasitoid Encarsia formosa to biological control agents of Diaprepes root weevil. whiteßy hosts on Gerbera jamesonii. Entomol. Exp. Appl. (http://edis.ifas.uß.edu/BODY IN122). 96: 299Ð307. 1572 ENVIRONMENTAL ENTOMOLOGY Vol. 37, no. 6

[TDA] Texas Department of Agriculture. 2001. Texas ad- (Insecta: Coleoptera: Curculionidae). EENY-24. Fea- ministrative code, title 4, chapter 19, subchapter P, rule tured creatures. (http://creatures.ifas.uß.edu/citrus/ ¤19.161. TDA, Austin, TX. sugarcane_rootstock_borer_weevil.htm). Ulmer, B. J., J. E. Pen˜ a, S. L. Lapointe, and G. Mathurin. Woodruff, R. E. 1985. Citrus weevils in Florida and the West 2006. The occurrence of parasitoids attacking citrus wee- Indies: preliminary report on systematics, biology, and vil eggs on Saint Lucia. Fla. Entomol. 89: 407Ð409. distribution (Coleoptera: Curculionidae). Fla. Entomol. Van Hezewijk, B. H., R. S. Bourchier, and S. M. Smith. 2000. 68: 370Ð379. Searching speed of Trichogramma minutum and its potential as a measure of parasitoid quality. Biol. Control 17: 139Ð146. Weissling, T. J., J. E. Pen˜ a, R. M. Giblin-Davis, and J. Knapp. 2004. Diaprepes abbreviatus (Linnaeus) Received 25 March 2008; accepted 2 September 2008.