Differential Physiological Responses of Dalmatian Toadflax, Linaria Dalmatica L

Differential Physiological Responses of Dalmatian Toadflax, Linaria dalmatica L. Miller, to Injury from Two Insect Biological Control Agents: Implications for Decision-Making in Biological Control Author(s) :Robert K. D. Peterson, Sharlene E. Sing, and David K. Weaver Source: Environmental Entomology, 34(4):899-905. 2005. Published By: Entomological Society of America DOI: URL: http://www.bioone.org/doi/full/10.1603/0046-225X-34.4.899

BioOne (www.bioone.org) is a a nonprofit, online aggregation of core research in the biological, ecological, and environmental sciences. BioOne provides a sustainable online platform for over 170 journals and books published by nonprofit societies, associations, museums, institutions, and presses. Your use of this PDF, the BioOne Web site, and all posted and associated content indicates your acceptance of BioOne’s Terms of Use, available at www.bioone.org/page/ terms_of_use. Usage of BioOne content is strictly limited to personal, educational, and non-commercial use. Commercial inquiries or rights and permissions requests should be directed to the individual publisher as copyright holder.

BioOne sees sustainable scholarly publishing as an inherently collaborative enterprise connecting authors, nonprofit publishers, academic institutions, research libraries, and research funders in the common goal of maximizing access to critical research. PLANTÐINSECT INTERACTIONS Differential Physiological Responses of Dalmatian Toadﬂax, Linaria dalmatica L. Miller, to Injury from Two Insect Biological Control Agents: Implications for Decision-Making in Biological Control

1, 2 3 1 ROBERT K.D. PETERSON, SHARLENE E. SING, AND DAVID K. WEAVER

Environ. Entomol. 34(4): 899Ð905 (2005) ABSTRACT Successful biological control of invasive weeds with specialist herbivorous insects is predicated on the assumption that the injury stresses the weeds sufÞciently to cause reductions in individual Þtness. Because plant gas exchange directly impacts growth and Þtness, characterizing how injury affects these primary processes may provide a key indicator of physiological impairmentÑ which then may lead to reductions in Þtness. The objective of this study was to use physiological methods to evaluate how the invasive weed, Linaria dalmatica L. Miller (Dalmatian toadßax), is affected by two introduced biological control agents within different injury guilds: the stem-boring weevil, Mecinus janthinus Germar, and the defoliating moth, Calophasia lunula Hufnagel. All studies with M. janthinus were conducted under Þeld conditions at two sites in Montana in 2003 and 2004. For C. lunula evaluations, a total of Þve greenhouse studies in 2003 and 2004 were used. One Þeld study

in 2003 and two studies in 2004 also were conducted. Variables measured included net CO2 exchange rate, stomatal conductance, and transpiration rate. Results from both Þeld sites revealed that the primary physiology of Dalmatian toadßax was deleteriously affected by M. janthinus larval injury. There were no signiÞcant differences among treatments for any of the gas exchange variables measured in all eight experiments with C. lunula. Our results indicate that insect herbivores in two distinct injury guilds differentially affect Dalmatian toadßax physiology. Based on the primary physiological parameters evaluated in this study, M. janthinus had more impact on Dalmatian toadßax than C. lunula. With such information, improved risk-beneÞt decisions can be made about whether to release exotic biological control agents.

KEY WORDS Calophasia lunula, Mecinus janthinus, herbivory, photosynthesis, plant gas exchange

BIOLOGICAL CONTROL OF WEEDS through the intentional An improved set of measurable indicators of bio- introduction of nonindigenous herbivorous insects logical control impact on weed densities would aid has reached a crossroads, both in terms of research and decision-makers in objective evaluation of tangible application. The research community has long ac- beneÞts versus potential risks when deciding whether knowledged the potential for classical biological con- to release nonindigenous organisms. Therefore, a trol of weeds to result in emerging or increased en- quantitative evaluation of beneÞts should be an im- vironmental risks (Harris and Zwo¨lfer 1968, Wapshere portant part of the overall risk assessment for agents. 1974, Howarth 1991, Louda et al. 2003, Sheppard et al. The lack of demonstrable beneÞts from the release of 2003). However, current regulatory attitudes and Þs- biological control agents can have substantial conse- cal shortfalls (Briese 2004) are reßected in the narrow quences (Thomas and Willis 1998). SpeciÞcally, if a focus of agent prerelease evaluations on host speci- nonindigenous organism does not deleteriously affect Þcity, at the expense of a more holistic screening and the targeted weed population, the economic costs or assessment process (but see Louda 1998). The lack of environmental risks associated with its release and evaluation of agent efÞcacy, as well as potential eco- establishment may always be greater than its beneÞts. logical risks, emphasize the need for formal, well- Therefore, it is crucial that agents approved for release quantiÞed risk-beneÞt evaluations of insect agents in- will actually reduce target weed populations and not troduced to manage invasive weeds. simply proliferate on them. Although several strategies exist for determining 1 Department of Entomology, Montana State University, Bozeman, the potential efÞcacy of weed biological control, using MT 59717Ð3020. methods that characterize changes in weed growth 2 Corresponding author: Department of Entomology, 333 Leon and Þtness can be very costly and time-consuming. We Johnson Hall, Bozeman, MT 59717Ð3020 (e-mail: bpeterson@ believe that the characterization of plant physiological montana.edu). 3 U.S. Forest Service, Rocky Mountain Research Station, Bozeman, response to herbivory provides a tenable alternative MT 59717. approach as a valuable indicator of the ability of bi-

0046-225X/05/0899Ð0905$04.00/0 ᭧ 2005 Entomological Society of America 900 ENVIRONMENTAL ENTOMOLOGY Vol. 34, no. 4 ological control agents to reduce target weed popu- oviposited in the target weed. Our physiological stud- lations. The delineation of physiological mechanisms ies were conducted 3Ð19 July 2003 and 7Ð14 July 2004. underlying plant responses to insect injury has been Two study groups were evaluated: injured and un- crucial to the explanation of yield loss and to the injured plants. In 2003, there were 18 group replicates development of general models of insect-induced at the Melstone site and 24 replicates at the Boulder plant stress response in crop plants (Boote 1981; Peter- site. In 2004, there were 18 replicates per group at each son 2001; Peterson and Higley 2001). Plant gas ex- site. Individual plants served as sample units at all sites. change processes such as photosynthesis, water vapor Injured plants were chosen based on the presence of transfer, and respiration represent a subset of a plantÕs ovipositional scars and swelling of stems, which indi- primary physiological processes. Understanding how cated the presence of larvae in the stems. Injured insect injury inßuences these parameters is important plants that were chosen had similar amounts of injury because these are the primary processes determining within each location, but plants at the Boulder site plant growth, development, and, ultimately, Þtness were not as visually injured as those at the Melstone (Peterson and Higley 1993). Although individual leaf site. Uninjured plants, those absent of oviposition scars photosynthetic rates typically are not accurate pre- and stem swelling, were chosen near infested plants. dictors of plant yield and Þtness (e.g., Irvine 1975, All plants at both sites were at approximately the same Elmore 1980, Baker and Ort 1992, Higley 1992), they developmental stage: early to mid ßower. can be used to objectively quantify physiological im- The primary limitation in our Þeld research with M. pairmentÑwhich may lead to reductions in Þtness. janthinus was that it did not involve experimental Boote (1981), Pedigo et al. (1986), and Higley et al. manipulation to create treatments. We chose injured (1993) emphasized the use of categorizing plant biotic plants based on the presence of oviposition scars and stressors, such as weed biological control agents, based stem swelling, rather than caging a group of plants and on injury type and physiological response, rather than assigning treatments for two main reasons: (1) Dal- on the taxonomic classiÞcation of the stressors or phys- matian toadßax plants are deleteriously affected (al- ical appearance of injury (as conventionally had been terations in stem and leaf thickness and whole plant done). Furthermore, Peterson and Higley (2001) ar- architecture) by caging whole plants (N. J. Irish, per- gued that similarities of plant response to speciÞc sonal communication), and (2) the Þeld sites were injury types, also known as injury guilds, are effective inaccessible when the adults Þrst emerged and fe- foci for addressing many basic and applied research males began ovipositing. questions. Calophasia lunula. We conducted a total of Þve The objective of this study therefore was to use greenhouse experiments in 2003 and 2004 (Table 1). physiological methods to evaluate how the invasive The experimental design for all greenhouse experi- weed, Dalmatian toadßax, Linaria dalmatica L. Miller, ments was a randomized complete block, with the area is affected by two introduced biological control agents under individual metal halide lamps serving as the in different injury guilds: the stem-boring weevil, blocking factor. Four of the experiments incorporated Mecinus janthinus Germar, and the defoliating moth, repeated measures into the experimental design. Calophasia lunula Hufnagel. Characterizing Dalma- Treatments consisted of C. lunula larval injury, artiÞ- tian toadßax primary metabolic impairment to her- cial defoliation injury, and uninjured control. We in- bivory would provide an initial step toward determin- corporated an artiÞcial defoliation treatment to de- ing the impact of biological control agents on the termine if manual defoliation could be used as a weedÕs Þtness and population dynamics. surrogate for C. lunula feeding. Treatments were replicated Þve or more times, and an individual pot containing one plant grown from seed served as the ex- Materials and Methods perimental unit. In two greenhouse experiments, we Mecinus janthinus. All studies with M. janthinus were not able to include the actual insect injury treat- were conducted under Þeld conditions at two sites in ment because of a lack of larvae. In one greenhouse both 2003 and 2004. The two sites were located near study, we did not include the artiÞcial defoliation Boulder (elevation ϭ 1798 m) and Melstone (eleva- injury treatment because of a lack of sufÞcient plants. tion ϭ 939 m), MT, in rangeland that had been ex- We also conducted three Þeld experiments: one in tensively and intensively burned by wildÞre in 2000. 2003 (Bozeman, MT) and two in 2004 (Missoula, MT). The Boulder site has a lower average temperature and The experimental design for all studies was completely receives slightly more precipitation per year than the random (Table 1). Treatments for two experiments Melstone site. Dalmatian toadßax was the dominant consisted of C. lunula larval injury, artiÞcial defolia- plant species at both sites. Observers with pre- and tion injury, and uninjured control. In a third experi- postÞre knowledge of the vegetation community at ment, only two treatments, C. lunula injury and con- these sites suggest that toadßax had increased signif- trol, were used (Table 1). Treatments were replicated icantly in both density and biomass postburn, sup- six times in 2003. In 2004, treatments were replicated porting the inference that Dalmatian toadßax is a Þre- six times in one experiment and eight times in the enhanced invasive weed (Jacobs and Sheley 2003aÐc, other. Individual plants served as experimental units. Phillips and Crisp 2000). Insects were released at each For seven of the eight experiments, a leaf at the bottom site in 2002, and follow-up observations made later in one-third of each stem also was measured for physi- the same year conÞrmed that the insects had fed and ological parameters to determine if there were inter- August 2005 PETERSON ET AL.: PHYSIOLOGICAL RESPONSES TO BIOCONTROL AGENTS 901

Table 1. Experimental details and statistical summaries for each C. lunula herbivory experiment

Treatment Treatment ϫ Mean percentage Treatment Treatment Year Type Treatments Experimental design ϫ leaf leaf position defoliation by C. Fd P Ͼ F position F P Ͼ F lunula Ϯ SEM 2003 Field 3a CRD, factorial, repeated 2.23 0.13 0.32 0.73 38 Ϯ 6.16 measures 2004 Field 3a CRD, factorial 1.97 0.15 0.66 0.52 45 Ϯ 5.98 2004 Field 2b CRD, factorial 1.94 0.19 0.001 0.96 75 Ϯ 15.44 2003 Greenhouse 3a RCBD, factorial, repeated 0.24 0.79 3.31 0.07 52 Ϯ 6.37 measures 2003 Greenhouse 2c RCBD, factorial, repeated 0.57 0.47 1.91 0.2 NA measures 2003 Greenhouse 2c RCBD, factorial, repeated 0.47 0.51 0.91 0.37 NA measures 2003 Greenhouse 3a RCBD, factorial, repeated 2.73 0.11 0.25 0.78 43 Ϯ 7.4 measures 2004 Greenhouse 2b RCBD 1.67 0.25 NA NA 13 Ϯ 2.8

a Treatments ϭ C. lunula larval injury, artiÞcial defoliation injury, and control. b Treatments ϭ C. lunula larval injury and control. c Treatments ϭ artiÞcial defoliation injury and control. d Treatment effect. CRD, completely random design; RCBD, randomized complete block design; NA, not applicable.

actions between defoliated (or undefoliated control) tance rate, and transpiration rate. The leaves were upper leaves and undefoliated lower leaves. illuminated with a light intensity of 1400 ␮mol pho- For most greenhouse and Þeld experiments, pre- tons/m2/s from a light source inside the 2-cm2 leaf

starved, C. lunula fourth instars were securely con- chamber. The leaf chamber reference CO2 concen- ␮ tained within leaf cages made from stiff, Þne-mesh tration was 400 mol CO2/mol, generated from a 12-g nylon netting (tuile) and allowed to feed for at least CO2 cylinder connected to the LI-6400. 12 h. The leaf cages were Ϸ11 by 11 cm and covered Statistical Analysis. Because gas exchange variables the top portion of the stem, usually enclosing from usually were observed on the same leaf over time, four to six large leaves. The open end of each cage was most analyses were conducted using a repeated mea- closed tightly around the stem with a drawstring. Leaf sures multivariate analysis of variance (MANOVA; cages could not enclose single leaves because the ␣ ϭ 0.05; SAS 9.0, Cary, NC). Dalmatian toadßax leaves are small and lack a petiole. The leaf-cage fabric intercepted Ͻ5% of photosyn- Results thetically active radiation (Peterson et al. 1998). The artiÞcial defoliation injury treatment was imposed by Mecinus janthinus. Although establishment of M. clipping leaf tissue with scissors from the plant in a janthinus was relatively recent at both sites, injury was spatial and temporal pattern consistent with C. lunula more widespread at Melstone than at Boulder. Evi- feeding; however, because of variability in C. lunula dence of M. janthinus injury to plants was compara- feeding (Table 1), we standardized artiÞcial defolia- tively intermittent at Boulder. Results from both Þeld tion injury at 50% of leaf area within each cage. Leaves sites over the 2-yr study period revealed that the of both the control and artiÞcially defoliated plants primary physiology of Dalmatian toadßax was signif- also were enclosed in leaf cages, ensuring appropri- icantly affected by M. janthinus larval injury (Table 2). ately comparable treatments. Larvae were removed, In particular, over both years and locations, photo- and gas exchange variables were recorded on injured synthetic rates for plants with injury from larvae were leaves at 1 and 48 h after the termination of injury. signiÞcantly lower than rates for uninjured plants (Ta- Plant Primary Metabolism. Repeated gas exchange ble 2; Fig. 1). At the Melstone site, the mean reduction measurements were made from the same leaf per plant in photosynthetic rates associated with larval injury at approximately the apical one-third of the stem. was 29.5% in 2003 and 27.3% in 2004. At the Boulder When it was not possible to take readings from the site, the mean photosynthetic rate reduction associ- same leaf per plant because of breakage, the nearest ated with larval injury was 38.6% in 2003 and 17% in leaf was measured. For all C. lunula experiments, mea- 2004. There was a signiÞcant time effect between surements were recorded from injured (partially de- measurement dates in both years at the Boulder site, foliated) leaves. If defoliated leaves had too little leaf but only a signiÞcant time effect in 2003 at the Mel- area remaining to measure, the nearest undefoliated stone site. leaf was used. At the Melstone site, transpiration rates and stoma- All gas exchange measurements were made within tal conductances were signiÞcantly lower for injured 3 h of solar noon using a portable photosynthesis plants in 2004, but not in 2003 (Table 2). These trends system (model LI-6400; LI-COR, Lincoln, NE). Vari- were reversed at the Boulder site, where we found that

ables measured and analyzed included: net CO2 ex- transpiration rates and stomatal conductances were change rate (photosynthetic rate), stomatal conduc- signiÞcantly lower for injured plants in 2003, but not 902 ENVIRONMENTAL ENTOMOLOGY Vol. 34, no. 4

Table 2. Mean gas exchange responses ؎ SEM of Dalmatian toadﬂax to M. janthinus larval injury

2003 2004 Location treatment 3 Jul 7 Jul 7 Jul 12 Jul Melstone Site 2 Stomatal conductance (mol H2O/m /s) Uninjured 0.51 Ϯ 0.025 0.23 Ϯ 0.014 0.19 Ϯ 0.032 0.16 Ϯ 0.017 Injured 0.48 Ϯ 0.035 0.17 Ϯ 0.031 0.12 Ϯ 0.019 0.11 Ϯ 0.014 F 1.29 6.34 P Ͼ F 0.27 0.017 df 1,22 1,34 2 Transpiration rate (mmol H2O/m /s) Uninjured 17.3 Ϯ 0.68 4.5 Ϯ 0.66 6.5 Ϯ 0.76 5.3 Ϯ 0.41 Injured 16.2 Ϯ 0.74 4.1 Ϯ 0.69 4.4 Ϯ 0.48 3.6 Ϯ 0.36 F 1.12 12.97 P Ͼ F 0.3 0.001 df 1,28 1,31 14-Jul 19-Jul 8-Jul 14-Jul Boulder Site 2 Stomatal conductance (mol H2O/m /s) Uninjured 0.17 Ϯ 0.01 0.17 Ϯ 0.01 0.25 Ϯ 0.01 0.22 Ϯ 0.02 Injured 0.12 Ϯ 0.02 0.13 Ϯ 0.01 0.22 Ϯ 0.02 0.22 Ϯ 0.02 F 9.9 0.2 P Ͼ F 0.003 0.65 df 1,36 2 Transpiration rate (mmol H2O/m /s) Uninjured 3.5 Ϯ 0.26 6.4 Ϯ 0.42 4.5 Ϯ 0.18 6.5 Ϯ 0.62 Injured 3.3 Ϯ 0.52 4.8 Ϯ 0.34 3.8 Ϯ 0.23 6.6 Ϯ 0.39 F 9.28 0.41 P Ͼ F 0.004 0.53 df 1,37 1,34

2004. The inconsistency in these responses suggests may be affected if stem injury by M. janthinus larvae that stem-boring injury by M. janthinus larvae may not prevents translocation of carbohydrates to the roots. be severely impairing the ability of the leaves to up- Our results support the contention that M. janthinus take CO2 and transpire H2O (stomatal limitations). injury affects photosynthesis of Dalmatian toadßax. Disruption of xylem tissues because of larval feeding More detailed physiological measurements will be within the stem apparently was insufÞcient to consis- required to determine the mechanisms underlying tently close stomata and reduce transpiration. There- reductions in photosynthetic rates of Dalmatian toad- fore, larval injury did not seem to physiologically ßax. Measurements of light-response curves, assimila- mimic drought stress. Reductions in stomatal conduc- tion-intercellular CO2 curves, and chlorophyll ßuo- tances and transpiration, when they occurred, seemed rescence can be used to determine precisely where to follow, rather than cause, photosynthetic reduc- biochemical limitations to photosynthesis are occur- tions. This type of primary metabolic response also ring and therefore can provide mechanistic explana- was observed with injury from Epilachna varivestis tions of physiological impairment. Stem-boring injury Mulsant on Phaseolus vulgaris L. and Glycine max L. is so poorly understood that, to date, the injury guild Merrill (Peterson et al. 1998). Jeanneret and Schroe- is characterized by the physical appearance of injury der (1991) suggested that photosynthetic mechanisms rather than on physiological effects of the injury (Wel-

Fig. 1. Photosynthetic responses of L. dalmatica to injury by M. janthinus larvae (A, Melstone, MT; B, Boulder, MT). Note statistically signiÞcant differences in all tests between injured/uninjured plants. August 2005 PETERSON ET AL.: PHYSIOLOGICAL RESPONSES TO BIOCONTROL AGENTS 903 ter 1989; Peterson and Higley 1993). Additional stud- ation of whole plants is known to alter the pattern of ies within this system therefore will assist in charac- normal progressive leaf senescence of some plant spe- terizing plant physiological responses to this injury cies (Higley 1992, Peterson et al. 1992). Consequently, guild. Dalmatian toadßax may respond differently to defo- Because we did not experimentally impose the liation injury at different levels of biological organi- treatments, we cannot discount the possibility that the zation, as has been observed for other plant-insect plants containing M. janthinus larvae were previously systems (Peterson and Higley 1993, Peterson and stressed. Adult females may have selected speciÞc Higley 2001). plants for oviposition because those plants were al- Because there were no statistically signiÞcant phys- ready stressed. Therefore, the gas exchange responses iological differences between leaves from C. lunula observed in this study may not have been caused by injured and artiÞcial defoliation injured plants (Table larval injury, but by other factors. However, that ex- 1), it is possible to simulate C. lunula defoliation. This planation is unlikely for the following reasons: (1) the is important because simulating insect defoliation of- uninjured and injured plants were often Ͻ20 cm apart, ten allows for better experimental control and a more indicating that abiotic or other biotic stress factors rapid quantiÞcation of injury than using actual insects likely were not different between treatments, and (2) (Pedigo et al. 1986, Peterson et al. 2004). However, if the injured plants were chosen based only on the the objective of a study is to characterize whole plant presence of ovipositional scars and stem swelling; no and Þtness responses of Dalmatian toadßax to C. lunula other visual differences between plants were evident. injury, the artiÞcial defoliation techniques must ade- Future experimental research will need to determine quately reproduce the spatial and temporal pattern of conclusively if the responses we observed are solely injury on the plant (Pedigo et al. 1986). the result of larval injury. Calophasia lunula. Defoliation by C. lunula larvae Discussion varied considerably among plants within and among experiments (Table 1). There were signiÞcant time Injury Guilds. The injury guild concept has evolved effects in all Þve of the experiments in which repeated from grouping herbivores by taxonomic category to measures statistical analysis was used. In all eight ex- grouping them by categories of plant physiological periments, there were no signiÞcant differences impact (Peterson 2001). Our results indicate that in- among treatments for any of the gas exchange vari- sect herbivores in two distinct injury guilds differen- ables measured (Table 1). Therefore, our results sug- tially impact Dalmatian toadßax physiology. Based on gest that defoliation injury did not affect Dalmatian the primary physiological parameters evaluated in this toadßax primary metabolism (i.e., photosynthesis, sto- study, the stem borer, M. janthinus, seemed to affect matal conductance, transpiration) over the observa- Dalmatian toadßax more than the defoliator, C. lunula. tion period. Larval defoliation reduced leaf area, but The differences we observed in host-plant physio- did not affect the photosynthetic apparatus of remain- logical response to the two herbivore species, repre- ing tissue. There was no statistically signiÞcant inter- senting two different injury guilds, may help explain action between injury of upper leaves and photosyn- their relative effectiveness in controlling the target thesis of lower, uninjured leaves on the same plant weed. Weed ecological studies indicate that crude (Table 1). This indicates that there was no photosyn- tissue removal or consumption, such as is produced by thetic response by lower leaves linked to defoliation C. lunula larvae, seldom signiÞcantly affects trenchant injury on upper leaves. weed infestations (Cousens and Mortimer 1995; My- Many studies have reported that there are no ers and Bazely 2003); this also holds true for Dalmatian changes in photosynthetic rate in the remaining leaf toadßax (Robocker et al. 1972; Vujnovic and Wein tissue of injured leaves in response to insects in the 1997). Furthermore, weed species with both vegeta- defoliator injury guild (e.g., Davidson and Milthorpe tive and seed reproduction, such as Dalmatian toad- 1966, Poston et al. 1976, Syvertsen and McCoy 1985, ßax, are particularly resilient to simple biomass reduc- Welter 1989, 1991, Higley 1992, Peterson et al. 1992, tion (Burdon et al. 1980, Burdon and Marshall 1981, Peterson et al. 1996, Peterson and Higley 1996, Burk- Lajeunesse et al. 1993). ness et al. 1999, Peterson et al. 2004). The similarity in Jeanneret and Schroeder (1992) reported that lar- response across these studies suggests that the pho- val mining in high density or outbreak populations of tosynthetic processes of many plant species is not M. janthinus caused premature stem wilting and sup- affected directly by defoliation. The principal effect of pression of ßower formation, and consequently, re- this injury type seems to be linked more to the reduction in seed production (Jeanneret and Schroeder duction in photosynthesizing leaf area than to a re- 1991). Larval feeding also was correlated with reduced duction or enhancement of photosynthetic capacity of stem biomass and with an increased incidence of veg- remaining tissue of injured leaves (Peterson and etative stem mortality during prerelease tests (Saner Higley 1996, Peterson 2001, Peterson et al. 2004). et al. 1994). According to Robocker (1970), the veg- Measurement of the potential impact from the de- etative stems of Dalmatian toadßax contribute signif- foliation injury in these experiments was limited to icantly to both the root carbohydrate reserves and relatively few leaves on the distal end of one stem of vigor of each individual toadßax plant. a plant. Therefore, the results may not translate to Conclusions. Successful biological control of inva- whole plant performance. Indeed, substantial defoli- sive weeds with herbivorous insects is predicated on 904 ENVIRONMENTAL ENTOMOLOGY Vol. 34, no. 4 the assumption that insect-induced injury will stress less costly than measurements of biomass, yield, and the weed sufÞciently to cause reductions in individual Þtness reduction, potentially improve the process of Þtness and, eventually, weed population density. agent selection from among numerous available can- These reductions in Þtness can be caused through didate species. direct and/or indirect consequences of herbivory (Crawley 1989, Sheppard 1996). Herbivores may cause physiological impairment to the weed host, Acknowledgments which may lead directly to reductions in Þtness through feeding activity. Alternatively, herbivores This study was funded, in part, by the National Fire Plan, may impair the weed hostÕs ability to compete effec- Project 01.RMS.B.3, the USDA-USFS Rocky Mountain Re- tively with other plants. Regardless of the mechanism search Station RM-4151, Project 03-JV-11222022-249, and the Montana Agricultural Experiment Station, Montana State underlying Þtness loss, because plant gas exchange University. D.K.W. would like to acknowledge funding in directly affects growth, development, and Þtness, we support of biological control of Dalmatian toadßax from argue that characterizing how agent injury impacts three United States Department of the Interior coopera- these primary physiological processes may be a key tors, speciÞcally Cooperative Agreement 98-FC-60-10740 indicator of the agentÕs potential efÞcacy. However, with the Bureau of Reclamation, Cooperative Agreement plant physiological impacts by the agent do not nec- ESA04DD31 with the Bureau of Land Management, and essarily predict agent efÞcacy. Indeed, the type of Memorandum of Agreement AG4C5000584 with the Bureau injury does not solely determine yield or Þtness loss. of Indian Affairs. We thank D. FitzGerald and J. Dinkins for In addition to injury type (impact per unit injury), the technical support. magnitude and duration of injury (intensity of injury) also is an important determinant (Pedigo et al. 1986, Peterson and Higley 2001). The intensity of injury References Cited depends heavily on the density of herbivores on the Baker, N. R., and D. R. Ort. 1992. Light and crop photo- host. For example, although defoliation injury may not synthetic performance, pp. 289Ð312. In N. R. Baker and alter photosynthetic rates of remaining leaves, high H. Thomas (eds.), Crop photosynthesis and temporal densities of herbivores, and therefore high rates of determinants. Elsevier, London, UK. defoliation, may lead to reduced Þtness. Therefore, to Boote, K. J. 1981. Concepts for modeling crop response to predict efÞcacy, it also is necessary to understand the pest damage. American Society of Agricultural Engineers, potential intensity of injury as well as the physiological St. Joseph, MI. Briese, D. T. 2004. Weed biological control: applying sci- impact per unit of injury. ence to solve seemingly intractable problems. Austral. J. With such information, improved risk-beneÞt and Entomol. 43: 304Ð317. cost-effective decisions can be made about whether to Burdon, J. J., and D. R. Marshall. 1981. Biological control release exotic biological control agents. For example, and the reproductive mode of weeds. J. Appl. Ecol. 18: if both biological control agent species were being 649Ð658. evaluated for release, gas exchange measures would Burdon, J. J., D. R. Marshall, and R. H. Groves. 1980. As- indicate that M. janthinus would have a higher prob- pects of weed biology important to biological control. In ability of being efÞcacious. Consequently, further E. S. Delfosse (ed.), Proc. Fifth Intern. Symp. Biol. Cont. studies on C. lunula would be needed to evaluate Weeds, Brisbane, Australia, 22Ð29 July 1980. Burkness, E. C., W. D. Hutchison, and L. G. Higley. 1999. whether the intensity of injury (population density Photosynthetic response of ÔCarolinaÕ cucumber to sim- potential) would be sufÞcient to reduce invasive Dal- ulated and actual striped cucumber beetle (Coleoptera: matian toadßax populations. Chrysomelidae) defoliation. Entomol. Sinica. 6: 29Ð38. To our knowledge, this study is the Þrst to show Cousens, R., and M. Mortimer. 1995. Dynamics of weed primary metabolic impacts by insects introduced as populations. Cambridge University Press, London, UK. biological control agents against an introduced weed Crawley, M. J. 1989. Insect herbivores and plant population species. Furthermore, these results contribute to and dynamics. Annu. Rev. Entomol. 34: 531Ð564. support current knowledge of plant gas exchange re- Davidson, J. L., and F. L. Milthorpe. 1966. The effect of sponses to insect injury types (Boote 1981, Welter defoliation on the carbon balance in Dactylis glomerata. Ann. Bot. 30: 185Ð198. 1989, Peterson and Higley 1993, Peterson 2001). Ad- Elmore, C. D. 1980. The paradox of no correlation between ditional research within the system we studied will leaf photosynthetic rates and crop yield, pp 49Ð68. In J. D. need to address directly the links between host-plant Hesketh and J. W. Jones (eds.), Predicting photosynthesis physiological impairment and changes in the hostÕs for ecosystem models, vol. II. CRC, Boca Raton, FL. population. Harris, P., and H. Zwo¨lfer. 1968. Screening of phytopha- Understanding plant physiological responses to her- gous insects for biological control of weeds. Can. Ento- bivory and tracking the impact of those responses on mologist. 100: 295Ð303. the weedÕs population and community dynamics can Higley, L. G. 1992. New understandings of soybean defoli- provide a quantitative and systematic method for eval- ation and their implications for pest management, pp. 56Ð65. In L. G. Copping, M. B. Green, and R. T. Rees uating agent efÞcacy. If physiological impairment (eds.), Pest management in soybean. Elsevier, London, from herbivores can be related to Þtness effects, the UK. metabolic measurements could be used as a diagnostic Higley, L. G., J. A. Browde, and P. M. Higley. 1993. Moving tool during the exploration phase of classical biological towards new understandings of biotic stress and stress control. These physiological methods, which often are interactions, pp. 749Ð754. In D. R. Buxton (ed.), Inter- August 2005 PETERSON ET AL.: PHYSIOLOGICAL RESPONSES TO BIOCONTROL AGENTS 905

national crop science I. Crop Science Society of America, Peterson, R.K.D., L. G. Higley, F. J. Haile, and J.A.F. Barri- Madison, WI. gossi. 1998. Mexican bean beetle (Coleoptera: Coccinel- Howarth, F. G. 1991. Environmental impacts of classical lidae) injury affects photosynthesis of Glycine max and biological control. Annu. Rev. Entomol. 36: 485Ð509. Phaseolus vulgaris. Environ. Entomol. 27: 373Ð381. Irvine, J. E. 1975. Relations of photosynthetic rates and leaf Peterson, R.K.D., L. G. Higley, and S. M. Spomer. 1996. canopy characters to sugarcane yield. Crop. Sci. 15: 671Ð Injury by Hyalophora cecropia (Lepidoptera: Saturni- 676. idae) and photosynthetic responses of apple and Jacobs, J. S., and R. L. Sheley. 2003a. Combination of burn- crabapple. Environ. Entomol. 25: 416Ð422. ing and herbicides may favor establishment of weedy Peterson, R.K.D., S. D. Danielson, and L. G. Higley. 1992. species in rangeland restoration. Ecol. Restor. 21: 329Ð Photosynthetic responses of alfalfa to actual and simu- 330. lated alfalfa weevil (Coleoptera: Curculionidae) injury. Jacobs, J. S., and R. L. Sheley. 2003b. Prescribed Þre effects Environ. Entomol. 21: 501Ð507. on Dalmatian toadßax. J. Range Manag. 56: 193Ð197. Phillips, B. G., and D. Crisp. 2000. Dalmatian toadßax, an Jacobs, J. S., and R. L. Sheley. 2003c. Testing the effects of invasive exotic noxious weed, threatens Flagstaff penny- herbicides and prescribed burning on Dalmatian toadßax. royal community following prescribed Þre. In J. Maschin- Ecol. Restor. 21: 138Ð139. ski and H. Holter (tech. coords.), Southwestern rare and Jeanneret, P., and D. Schroeder. 1991. Final Report: Meci- endangered plants: proceedings of the third conference. nus janthinus Germar (Col.: Curculionidae): a candidate U.S. Department of Agriculture, Forest Service, Rocky for biological control of Dalmatian and yellow toadßax in Mountain Research Station, Ft. Collins, CO. North America. International Institute of Biological Con- Poston, F. L., L. P. Pedigo. R. B. Pearce, and R. B. Hammond. trol, European Station, Delemont, Switzerland. 1976. Effects of artiÞcial and insect defoliation on soy- Jeanneret, P., and D. Schroeder. 1992. Biology and host bean net photosynthesis. J. Econ. Entomol. 69: 109Ð112. speciÞcity of Mecinus janthinus Germar (Col.: Curculion- Robocker, W. C. 1970. Seed characteristics and seedling idae), a candidate for the biological control of yellow and emergence of Dalmatian toadßax. Weed Sci. 18: 720Ð725. Dalmatian toadßax, Linaria vulgaris (L.) Mill. and Linaria Robocker, W. C., R. Schirman, and B. A. Zamora. 1972. dalmatica (L.) Mill. (Scrophulariaceae) in North Amer- Carbohydrate reserves in roots of Dalmatian toadßax. ica. Biocontrol Sci. Technol. 2: 25Ð34. Weed Sci. 20: 212Ð214. Lajeunesse, S. E., P. K. Fay, D. Cooksey, J. R. Lacey, R. M. Saner, M. A., P. Jeanneret, and H. Muller-Scharer. 1994. Interaction among two biological control agents and the Nowierski, and D. Zamora. 1993. Dalmatian and yellow developmental stage of their target weed, Dalmatian toadßax: weeds of pasture and rangeland. Montana State toadßax, Linaria dalmatica (L.) Mill. (Scrophulariaceae). University Extension Service, Bozeman, MT. Biocontrol Sci. Technol. 4: 215Ð222. Louda, S. M. 1998. Ecology of interactions needed in bio- Sheppard, A. W. 1996. The interaction between natural en- logical control practice and policy. Bull. Brit. Ecol. Soc. emies and interspeciÞc plant competition in the control 29: 8Ð11. of invasive pasture weeds, pp. 47Ð53. In V. C. Moran and Louda, S. M., R. W. Pemberton, M. T. Johnson, and P. A. J. H. Hoffmann (eds.), Proc. IX Intern. Symp. Biological Follett. 2003. Nontarget effectsÑthe AchillesÕ heel of Cont. Weeds. University of Cape Town, Stellenbosch, biological control? Retrospective analyses to reduce risk South Africa. associated with biocontrol introductions. Annu. Rev. En- Sheppard, A. W., R. Hill, R. A. DeClerck-Floate, A. McClay, tomol. 48: 365Ð396. T. Olckers, P. C. Quimby Jr., and H. G. Zimmerman. Myers, J. H., and D. R. Bazely. 2003. Ecology and control of 2003. A global review of risk-beneÞt-cost analysis for the introduced plants. Cambridge University Press, London, introduction of classical biological control agents against UK. weeds: a crisis in the making? Biocontrol. 24: 77Ð94. Pedigo, L. P., S. H. Hutchins, and L. G. Higley. 1986. Eco- Syvertsen, J. P., and C. W. McCoy. 1985. Leaf feeding injury nomic injury levels in theory and practice. Annu. Rev. to citrus by root weevil adults: leaf area, photosynthesis, Entomol. 31: 341Ð368. and water use efÞciency. Fla. Entomol. 63: 386Ð393. Peterson, R.K.D. 2001. Photosynthesis, yield loss, and injury Thomas, M. B., and A. J. Willis. 1998. BiocontrolÑrisky but guilds, pp. 83Ð97. In R.K.D. Peterson and L. G. Higley necessary? Trends Ecol. Evol. 13: 325Ð329. (eds.), Biotic stress and yield loss. CRC, Boca Raton, FL. Vujnovic, K., and R. W. Wein. 1997. The biology of Cana- Peterson, R.K.D., and L. G. Higley. 1993. Arthropod injury dian weeds. 106. Linaria dalmatica (L.) Mill. Can. J. Plant and plant gas exchange: current understandings and ap- Sci. 77: 483Ð491. proaches for synthesis. Trends Agric. Sci. Entomol. 1: Wapshere, A. J. 1974. A strategy for evaluating the safety of 93Ð100. organisms for biological weed control. Ann. Appl. Biol. 77: Peterson, R.K.D., and L. G. Higley. 1996. Temporal changes 201Ð211. in soybean gas exchange following simulated insect de- Welter, S. C. 1989. Arthropod impact on plant gas exchange. foliation. Agron. J. 88: 550Ð554. In E. A. Bernays (ed.), Insect-plant interactions, vol. 1. Peterson, R.K.D., and L. G. Higley. 2001. Illuminating the CRC, Boca Raton, FL black box: the relationship between injury and yield, pp. Welter, S. C. 1991. Responses of tomato to simulated and 1Ð12. In R.K.D. Peterson and L. G. Higley (eds.), Biotic real herbivory by tobacco hornworm. Environ. Entomol. stress and yield loss. CRC, Boca Raton, FL. 20: 1537Ð1541. Peterson, R.K.D., C. L. Shannon, and A. W. Lenssen. 2004. Photosynthetic responses of legume species to leaf-mass Received for publication 22 December 2004; accepted 12 consumption injury. Environ. Entomol. 33: 450Ð456. April 2005.