BIOLOGICAL CONTROL 8, 52–57 (1997) ARTICLE NO. BC960484 Growth and Development of the Biological Control Agent Bagous hydrillae as Influenced by Hydrilla (Hydrilla verticillata) Stem Quality G. S. WHEELER AND T. D. CENTER USDA/ARS Aquatic Weed Research Unit, 3205 College Avenue, Ft. Lauderdale, Florida 33314 Received March 11, 1996; accepted August 28, 1996 that reduces the impact of insects imported for weed Plant quality of dioecious hydrilla was studied as a biological control. factor that may influence larval survival, growth, and The Australian weevil Bagous hydrillae O’Brien (Bal- development of the biological control agent Bagous ciunas and Purcell, 1991) was introduced into the hydrillae. Nitrogen content and stem toughness of United States for biological control of hydrilla. Release hydrilla varied among the five sites studied and be- of this species began in 1991, and to date, at least two tween summer and fall collections. The nitrogen con- field populations have established, one in Florida and tent of hydrilla collected during summer ranged from another in Texas (Center et al., unpublished data). 1.2 to 3.6% (dry weight) and during fall from 1.6 to 2.9%. Considerable difficulty has been experienced in estab- Stem toughness ranged from 487 to 940 g/mm2 during lishing this species despite release of several thousand the summer and from 418 to 1442 g/mm2 during the fall. individuals throughout the area. Among the factors The larvae of this weevil species required more time to that could influence weevil performance and establish- complete development when fed hydrilla containing ment, the quality of hydrilla, which varies greatly at lower levels of nitrogen and tougher stemmed plants. different sites (Zimba et al., 1993), may be of critical Moreover, pupal and adult fresh weights were also importance. The goal of this study was to determine the reduced when the larvae were fed the poorer quality effects of plant nitrogen and stem toughness on B. plants. Relative growth rates were reduced in larvae hydrillae larval survival, growth, and developmental fed hydrilla of lower nitrogen level and tougher plants. rates. Hydrilla plant quality constitutes an important factor that may limit the establishment and impact of this potential biological control agent. r 1997 Academic Press MATERIALS AND METHODS KEY WORDS: biological control of weeds; dioecious hydrilla; Curculionidae; Bagous hydrillae; leaf tough- Sites. Dioecious hydrilla samples were collected at ness; nitrogen. five sites in southeastern Florida that comprised a range of plant nitrogen content and plant toughness (Wheeler and Center, 1996). These sites included Lake Helen Blazes (HNB), located in Brevard County, and INTRODUCTION Sky Lake (SL), Miami Canal (MC), Hacienda Village (HV), and L-Lake (LL), all located in Broward County. It is widely recognized that plant quality influences These water bodies were flowing rivers or canals (HNB herbivore survival, growth, and development (Rosen- and MC) and artificial lakes (SL, HV, and LL), all of thal and Berenbaum, 1991, 1992; Stamp and Casey, which have been heavily infested with hydrilla consis- 1993). In addition to plant defenses, plant nutrient tently each year. Four hydrilla samples (about 500 g), content (Slansky, 1993; White, 1993) and physical separated by about 100 m, were collected from each characteristics (Grubb, 1986) may also influence herbi- site, consisting of a contiguous section of the upper 20 vore performance. For example, larvae of the ephydrid cm of the bed. From each sample, five apical shoot fly Hydrellia pakistanae Deonier fed relatively poor cuttings ca 20 cm in length, were measured for stem quality (low nitrogen levels and high leaf toughness) toughness; the remainder of the sample was used for plants of the aquatic weed hydrilla Hydrilla verticillata nitrogen determinations. The hydrilla at each site was (L.f.) Royle had increased mortality, decreased growth sampled during the summer (July) and fall (November– rates, and reduced adult weight compared with insects December). fed higher quality plants (Wheeler and Center, 1996). Plant quality. Hydrilla samples were analyzed for We predict that low plant quality is a common factor stem toughness with a gram gauge (Halda, Stockholm, 1049-9644/97 $25.00 52 Copyright r 1997 by Academic Press All rights of reproduction in any form reserved. B. hydrillae PERFORMANCE IS AFFECTED BY HYDRILLA PLANT QUALITY 53 Sweden) modified with a 0.53-mm-diameter blunt probe, (F 5 46.9; df 5 1,6; P 5 0.0005) during summer and which measured the pressure required to completely greater at HV (F 5 8.1; df 5 1,4; P 5 0.05) and LL penetrate each stem (Sands and Brancatini, 1991; (F 5 14.4; df 5 1,6; P 5 0.009; Fig. 1) during fall. Wheeler and Center, 1996). Stems were analyzed while Among-site analyses indicated that hydrilla from HNB on a Plexiglas stage equipped with a lid that held the had the greatest percentage of nitrogen during both stems in place. Holes in the lid and base acted as guides seasons. The hydrilla from HV and LL had the lowest for the insertion of the probe. Hydrilla stem toughness percentage of nitrogen during the summer, as did the was analyzed sequentially, starting with the point 5 cm hydrilla from MC and LL during fall (Fig. 1). from the apex of the shoot and cutting toward the base Stem toughness differed significantly by season at 5-cm intervals (i.e., 5, 10, 15, and 20 cm). Each (F 5 27.7; df 5 1,39; P , 0.0001), site (F 5 258.7; sample was submersed in a water-filled shoe box for 0 df 5 4,39; P , 0.0001), and distance from the tip to 3 days until ready for use. Percentage nitrogen of the (F 5 39.7; df 5 3,39; P , 0.0001; Fig. 2). Linear regres- hydrilla dry weight was determined with a modified sion indicated that stem toughness increased from the Kjeldahl method (Allen et al., 1974). tip toward the base during both seasons at all sites Weevil mortality, growth, and development. Bagous except HNB during summer and at SL during both hydrillae weevils were collected from our laboratory summer and fall (Fig. 2). colony (approximately 3 years in culture without infu- Weevil mortality, growth, and development. Percent- sion of wild individuals). Eggs were inserted individu- age mortality (mean 6 SE) of weevil larvae (summer, ally by hand into hydrilla stems within 5 cm of the tip. 26.0 6 4.8%; fall, 28.5 6 3.6%) was not influenced sig- Larvae (15 per site) were reared to pupation in petri nificantly by season or site. Larval development time dishes (15 3 3 cm) lined with moist filter paper and differed significantly by season (F 5 38.3; df 5 1,136; sealed with parafilm. The larvae were checked daily, P , 0.0001) and site (F 5 20.5; df 5 4,136; P , 0.0001). mortality was recorded, and fresh hydrilla stems were Moreover, the interaction of these two effects was added as needed. The length of time for the larvae to significant (F 5 7.8; df 5 4,136; P , 0.0001). Within- reach the prepupal, pupal, and adult stage was re- site comparisons indicated that larval development corded, as were pupal and adult weights using an times were significantly longer when fed hydrilla col- analytical balance (610 µg). Relative growth rates lected during the summer at all sites except HNB (Fig. (5RGR) of larvae were calculated according to the 3). Among-site comparisons indicated that during the following formula: fresh weight gained (mg)/average summer the larvae required the longest development fresh weight of the larva during the experiment (mg) · time when fed hydrilla from LL and the shortest time development time (Kogan, 1986). Adult gender was when fed hydrilla from HNB. Larvae fed the fall- determined by dissection and examination of the geni- collected hydrilla had the longest development times talia. when fed material from HV and MC, whereas the Data analysis. All analyses were conducted with shortest development times occurred on HNB hydrilla. SAS/PC, PROC GLM unless otherwise noted (SAS Pupal fresh weight differed significantly by season Institute, Inc., 1988). Nitrogen content and stem tough- (F 5 6.9; df 5 1,123; P 5 0.01) and site (F 5 7.3; ness were analyzed with a two-way analysis of variance df 5 4,123; P , 0.0001). Moreover, the interaction of (ANOVA), where site and season were the main effects. The influence of plant quality on larval developmental rates and adult weight was analyzed as a three-way factorial design, where site, season, and weevil sex were the main effects. Means were compared by the Ryan Q test (P 5 0.05). Linear regression (PROC REG; SAS Institute, 1988) was used to assess the relation- ship between RGR and leaf quality. RESULTS Plant quality. Tissue nitrogen concentrations var- ied among the five sites (F 5 130.1; df 5 4,27; P , 0.0001). As the interaction of the two effects FIG. 1. Mean (6SE) percentage of nitrogen of hydrilla collected (site 3 season) was significant (F 5 14.9; df 5 4,27; during two seasons from five sites in south Florida 1994. Within-site P , 0.0001), each effect was analyzed at fixed levels of differences in percentage of nitrogen between summer- and fall- collected hydrilla are indicated with an asterisk. Among-site compari- the other (Montgomery, 1984). Analysis of the effect of sons were not significantly different if the same uppercase letter season within each site indicated that nitrogen content appears above open bars (summer) or if the same lowercase letter was significantly greater in hydrilla from HNB appears above solid bars (fall).