Mechanical and chemical control of the invasive cordgrass Spartina densiflora and native plant community responses in an estuarine salt marsh

106 J. Aquat. Plant Manage. 50: 2012. J. Aquat. Plant Manage. 50: 106-111 Mechanical and chemical control of the invasive cordgrass Spartina densiflora and native plant community responses in an estuarine salt marsh EnriquE MatEos-naranjo, jEsús CaMbrollé, juan GarCía dE loMas, raquEl Parra, and susana rEdondo-GóMEz* AbstrAct the south american cordgrass, Spartina densiflora, has invaded a wide range of saltmarsh areas in southern spain. a field experiment to examine physical and chemical control of S. densiflora, including mowing, herbicide (glyphosate), mowing plus herbicide combination, and the breaking of rhizomes, was conducted in low-gradient marsh invaded by S. densiflora to find a means of controlling this invasive species. the growth parameters of density, as well as species richness and diversity, were used to assess the efficacy of different treatments in december 2007 and 2008. all treatments reduced live tiller density of S. densiflora after 1 and 2 years of treatment. Compared to the control, the reductions in tiller density with rhizome breaking, mowing plus herbicide, mowing, and herbicide application were 85, 65, 56, and 38% and 66, 70, 52, and 52% after 1 and 2 years of treatment respectively. despite a reduction in S. densiflora abundance, none of the treatments eradicated this species completely. However, rhizome breaking and mowing plus herbicide treatments proved to have the highest control efficiency, and plots treated with these treatments contained the highest values of native species richness and diversity. Key words: breaking rhizomes, herbicide, invasive species, marshes, mowing IntroductIon the south american cordgrass, Spartina densiflora brongn. (Poaceae), is invading salt marshes from southern Europe (Figueroa and Castellanos 1988), north africa (Fennane and Mathez 1988), and north america (Kittelson and boyd 1997). in its native south america, S. densiflora is a salt-marsh dominant over a wide latitudinal range and exhibits considerable morphological variation (bortolus 2006). it has been postulated that this species was accidentally introduced into southwest spain by means of lumber trade from south america. in southwest spain, S. densiflora has emerged as a vigorous invader and ecosystem engineer that spreads by prolific seed production and consolidates its stands by clonal growth. it can be a formidable competitor; it produces dense tussocks with tall canopies and persistently high aboveand belowground biomass (Figueroa and Castellanos 1988, Castillo et al. 2008, Mateos-naranjo et al. 2008). invasion by S. densiflora is one of the most important conservation problems affecting the Gulf of Cadiz in the southwestern iberian Peninsula because this species alters the composition of plant and animal communities, reducing their biological diversity (Kittelson and boyd 1997). it has become the dominant plant species on recent tidal marsh restorations in the doñana national Park (Gallego-Fernández and Garcíanovo 2007), threatening to spread to other marsh systems in southern Europe. learning how to effectively manage populations of S. densiflora is vital; thus, research must be conducted to find methods that will either control or eradicate this species, as has been already suggested for other species of Spartina (an et al. 2007). Chemical and mechanical methods may constitute an important tool for the control of S. densiflora, as they have for other species of Spartina (Patten 2004, roberts and Pullin 2008). Hedge et al. (2003) explained the establishment and considerable limitations of a wide range of control techniques, including physical removal, mowing, and herbicide, for the management of nonnative Spartina plants (S. alterniflora, S. anglica, and S. patens) in Washington state. they indicated that the combination of mowing and herbicide application provided the greatest control efficacy. For S. densiflora only long-term flooding and glyphosate application under controlled environmental conditions have been studied as control techniques (Mateos-naranjo et al. 2007, 2009). therefore, the aims of this study were to (1) evaluate the effects of different methods on the control of S. densiflora under field conditions and (2) determine whether these control methods lead to enhanced plant diversity in the long term for restoration of invaded areas. MAterIAl And Methods study location this experiment was performed at tinto Marshes, situated on the joint estuary of the odiel and tinto rivers near the *First, second, fourth, and fifth authors: departamento de biología Vegetal y Ecología, Facultad de biología, universidad de sevilla, apartado 1095, 41080 sevilla, spain; second author: departamento de biología. Facultad de Ciencias del Mar y ambientales, universidad de Cádiz. rio san Pedro s/n, 11510 Cádiz, spain. Corresponding author email: [email protected]. received for publication july 7, 2011 and in revised form February 7, 2012. J. Aquat. Plant Manage. 50: 2012. 107 town of Huelva, on the atlantic coast of southwestern spain (37°15’n, 6°58’W). the salt marsh is subject to a Mediterranean climate with oceanic influences. Winter is wet and mild (mean temperature about 11 °C in january) and summer is long and dry (mean temperature about 25 °C). Mean annual rainfall is 510 mm, with an interannual variation coefficient of 31%. the semidiurnal tides have a mean range of 2.10 m and a mean spring tidal range of 2.97 m, representing 0.40 to 3.37 m above spanish Hydrographic zero (sHz). Mean sea level is +1.85 m relative to sHz (Mateos-naranjo et al. 2008). to test the efficacy of the different methods in the control of Spartina and the damage to native plants, we restricted our experiment to low-gradient marsh invaded by S. densiflora with an area of 0.5 ha with a height difference of 40 cm between its lower (+2.8 m sHz) and upper (+3.2 m sHz) limit of study area. in this location, S. densiflora shared habitat with the native species Halimione portulacoides, Arthrocnemum macrostachyum, Sarcocornia perennis, and Salicornia ramosissima. [a detailed examination of the physicochemical properties of the study site is given in table 1.] environmental measurements Environmental soil characteristics were analysed. Measurements of sediment conductivity, pH (n = 15), and redox potential (n = 10) were made in low tide on the upper 10 cm at the sediment in december 2006. Conductivity of sediment was determined in the laboratory with a conductivity meter (Crison-522, spain) after mixing the sediment with distilled water (1:1; redondo-Gómez et al. 2007). redox potential and pH of the sediment were obtained with a portable meter and electrode system calibrated in the field (Crison pH/ mV p-506, spain). soil water content was determined from samples taken from the upper 10 cm of sediment (n = 15). samples were weighed before and after drying at 80 C for 48 h. experimental design and treatments in december 2006, nine replicate plots of 2.5 by 2.5 m were positioned 2 m apart in the study area, and four treatments (mowing, herbicide, mowing plus herbicide combination, and the breaking of rhizomes) and one control were randomly assigned to each plot. Mechanical treatments (mowing and breaking) of plots were carried out by personnel from the university of seville using various hand-held brush cutters and shovels. all mowing plots were mown to within 10 cm of the substrate. For the breaking of rhizomes treatment, the above-ground portions of the plants in the plots were removed and then the belowground sediments, rhizomes, and roots were dissected into 20 cm long sediments with shovels. Herbicide was applied homogeneously to the plot surface using a backpack agricultural sprayer (Matabi 5-l, Goizper s.C., spain) with a hand-held wand and an adjustable brass nozzle at a speed of 1 m s-1 and at a pressure of 200 kPa (250 ml of spray volume). the herbicide used was glyphosate at 7200 g a.i. ha-1 (Glialka® 36; 360 g a.i. l-1, Presmar sl, spain). in the mowing plus herbicide treatment, glyphosate was applied immediately after 24 h of mowing at the lowest tidal level to provide the necessary time for herbicide uptake and translocation. the chemical treatment was likewise performed at low tides, allowing 3 to 6 h of drying time before inundation of 50% of the plant. Weather conditions were optimal, with air temperatures ranging between 12 and 15 C and wind speeds from 0 to 5 Km h-1. also, in december the application period coincided with a period of neap tides and with plants dropping seeds and entering senescence (nieva et al. 2005). data collection to estimate the control efficacy, four 0.2 by 0.2 m quadrats were randomly selected in each plot in december 2007 and 2008, and the number of live tillers of S. densiflora was recorded. native species richness and plant species diversity were measured for each treatment. species diversity was calculated using shannon’s index (H’; shannon and Weaver 1949) formula H’= −Σpi · lnpi, where pi = relative abundance of each species divided by the total number of species observed in each plot. statistical analysis statistical analyses were carried out using statistica v. 6.0 (statsoft inc.). data were analyzed using a one-way analysis of variance (F-test). data were first tested for normality with the Kolmogorov-smirnov test and for homogeneity of variance with the brown-Forsythe test. We used a normal error distribution, so alive tiller density and native species richness were ln (x) transformed, respectively, for one-way analysis of variance. significant test results were followed by tukey tests for identification of important contrasts (P < 0.05). results control of invasive S. densiflora all mechanical and chemical treatments had significant effects on the control efficacy of S. densiflora after 1 and 2 years of treatment, with all treated plots showing lower live tiller density than control plots (table 2; Figure 1). breaking rhizome treatment recorded the lowest tiller density of S. densiflora after 1 year of treatment, followed by mowing plus herbicide, mowing, and herbicide treatment (Figure 1). Compared to the control, the reductions in tiller density with these treatments were 85, 65, 56, and 38% respectively. Table 1. Physicochemical ProPerTies of The Three conTrasTing siTes from TinTo marshes. Values are mean ± se of 15 rePlicaTes. excePT for redox PoTenTial (Values are mean ± se of 5 rePlicaTes in The uPPer and lower limiT of sTudy area). Parameter values Conductivity (ms cm-1) 15.6 ± 0.8 pH 6.7 ± 0.1 redox potential (mV) 87 ± 6.3 -146 ± 19.9 soil water content (%) 27 ± 0.4 108 J. Aquat. Plant Manage. 50: 2012. Table 2. resulTs of a one-way analysis of Variance (anoVa) for differenT conTrol TreaTmenTs on s. densiflora liVe Tiller densiTy and sPecies richness and naTiVe PlanT diVersiTy of TreaTed PloTs (shannon’s index, h´) one and Two years afTer TreaTmenT. TreaTmenTs: c, conTrol PloT; h, herbicide; b, breaking rhizome; m, mowing; and m+h, mowing Plus herbicide combinaTion. asTerisk indicaTe significanT difference; Tukey TesT, P < 0.05). 1st year one way anova d.f. F P 2nd year one way anova d.f. F P variable tiller density (tiller m-2) 4 67.60 0.000 variable tiller density (tiller m-2) 4 37.32 0.000 treatment C H b M treatment C H b M H 0.000* — — — H 0.000* — — — b 0.000* 0.000* — — b 0.000* 0.025* — — M 0.000* 0.006* 0.000* — M 0.000* 1.000 0.018* — M+H 0.000* 0.000* 0.000* 0.616 M+H 0.000* 0.000* 0.619 0.000* 1st year one way anova d.f. F P 2nd year one way anova d.f. F P variable species richness 4 9.59 0.000 variable species richness 4 8.40 0.000 treatment C H b M treatment C H b M H 1.000 — — — H 0.998 — — — b 0.002* 0.002* — — b 0.002* 0.004* — — M 0.006* 0.006* 0.986 — M 0.133 0.234 0.341 — M+H 0.007* 0.007* 0.976 1.000 M+H 0.003* 0.007* 0.999 0.473 1st year one way anova d.f. F P 2nd year one way anova d.f. F P variable native plant diversity 4 13.04 0.000 variable native plant diversity 4 6.55 0.001 treatment C H b M treatment C H b M H 0.954 — — — H 0.880 — — — b 0.000* 0.000* — — b 0.005* 0.041* — — M 0.011* 0.049* 0.341 — M 0.101 0.472 0.657 — M+H 0.001* 0.004* 0.941 0.803 M+H 0.004* 0.033* 1.000 0.597 J. Aquat. Plant Manage. 50: 2012. 109 However, rhizome breaking treatment showed a tiller density similar to mowing plus herbicide treatment after 2 years of treatment (Figure 1). after 2 years, breaking and mowing plus herbicide treatments recorded the lowest tiller densities (Figure 1), with percentage reductions compared to the control of 66 and 70%, respectively. Herbicide and mowing treatments each showed a tiller reduction of 52% and were not statistically different (table 2). recolonization by native vegetation breaking, mowing, and mowing plus herbicide treatments showed the highest natives species richness after 1 year of treatment (table 2; Figure 2); however, diversity values were not similar to those of the mowing treatment due to the presence of dominant species (table 2; Figure 3). after 2 years, rhizome breaking and mowing plus herbicide treatments showed the highest natives species richness (Figure 2). these treatments also recorded the highest diversity values (table 2; Figure 3). Finally, species richness and diversity values of plots treated with herbicide were similar to the control plot as a consequence of the wider coverage by S. densiflora. dIscussIon our results show that breaking rhizomes and mowing plus herbicide combination treatments had the highest control efficacy on live tiller density of S. densiflora compared to the control plot. Glyphosate application was the least efficacious treatment (in terms of autochthonous plant species colonization), especially in some plots where the coverage by S. densiflora was 100% (as happened in the control plot). although glyphosate did reduce live tiller density, the shading effect of the large number of dead and erect tillers of S. densiflora might account for the lower presence of other species in herbicide treatment plots. this species exhibits a phalanx type of growth characterized by the production of dense tussocks, which can reduce light at soil surface and thereby inhibit colonization by native species (Castellanos et al. 1998). Mateos-naranjo et al. (2009) demonstrated that glyphosate at doses as high as 7200 g a.i. ha-1 has a negative effect on the photosynthetic apparatus and growth of S. densiflora under controlled environmental conditions, so this may reduce Figure 1. live tiller density in treated plots at tinto marshes 1 and 2 years after treatment. treatments: C, control plot; H, herbicide; b, breaking rhizome; M, mowing; and M+H, mowing plus herbicide combination. Values represent mean ± sE, n = 36. different letters indicate means that are significantly different from each other (capital letters for first year and lowercase for the second year of treatment; tukey test, P < 0.05). Figure 2. Mean species richness in treated plots at tinto marshes 1 and 2 years after treatment. treatments: C, control plot; H, herbicide; b, breaking rhizome; M, mowing; and M+H, mowing plus herbicide combination. Values represent mean ±sE, n = 18. different letters indicate means that are significantly different from each other (one-way anoVa, species × treatment, p < 0.05). Figure 3. native plant diversity (shannon’s index, H´) in treated plots at tinto marshes 1 and 2 years after treatment. treatments: C, control plot; H, herbicide; b, breaking rhizome; M, mowing; and M+H, mowing plus herbicide combination. Values represent mean ± sE, n = 36. different letters indicate means that are significantly different from each other (capital letters for the first year and lowercase for the second year of treatment; tukey test, P < 0.05). 110 J. Aquat. Plant Manage. 50: 2012. its competitive ability. However, Patten (2002) found that the time between application and tidal inundation over the canopy affected the efficacy of glyphosate in the control of S. alterniflora, and zanatta et al. (2007) observed that soil water content influenced the glyphosate efficacy in the control of Euphorbia heterophylla. thus, the effect of environmental factors such as tidal influence and soil water content could alter the efficacy of glyphosate application and partly explain our results. in addition, to increase the efficacy of glyphosate application, repetition of the treatment in subsequent seasons is necessary, as previously suggested for different Spartina species (roberts and Pullin 2008) as well as other invasive species of wetland ecosystems (ailstock et al. 2001). Many studies have demonstrated variable responses of Spartina to mowing, depending on the species (li and zhang 2008, roberts and Pullin 2008). accordingly, S. alterniflora showed a significant reduction in the density with an overall mean percentage decline of 68%, whereas S. anglica and S. townsendii showed an over-compensation effect with a mean increase of 42.8 and 14.7%, respectively (roberts and Pullin 2008). For S. densiflora, we recorded mean percentage reductions of about 56 and 52% after 1 and 2 years of treatment, respectively. the removal of above-ground parts by mowing might have greatly reduced the energy allocation to its belowground structures, leading to a reduced regrowth potential (Haferkamp and Karl 1999). Mowing was performed only once at the beginning of the experiment, although repeated mowing has been shown to effectively reduce growth in other Spartina species (li and zhang 2008). Major et al. (2003) showed that mowing might be the least efficacious control treatment for Spartina alterniflora, the most labour intensive, and the most destructive to the surrounding mudflat. Moreover, its efficacy increases with the frequency of mowing and the use of larger machinery, two aspects that add up to greater damage to the associated mudflat. in addition, mechanical control methods are costly and require highly specialized equipment (li and zhang 2008). therefore, further testing of repeated mowing is needed to better assess the control of S. densiflora with this method and ascertain whether it constitutes a feasible option. Finally, rhizome breaking and mowing plus herbicide treatments reduced the density of S. densiflora between 85 and 65% after 1 year of treatment, and between 66 and 70% after 2 years. li and zhang (2008) observed similar reductions for S. alterniflora treated with rhizome breaking in the first season, but growth had almost recovered to the control level by the end of the second growing season. other techniques such as the use of roto-tilling, has produced >90% efficacy for the control of S. alterniflora during winter trials but was <70% effective during spring trials (Hedge et al. 2003). in comparison, roberts and Pullin (2008) found that the use of mowing followed by glyphosate application decreased the density of S. alterniflora by 91%. in contrast, the same intervention, when used to control S. anglica, increased densities by 19% per plot. Major et al. (2003) found that one-time mowing seemed to yield a more consistently uniform application of the herbicide and was likely to provide an initial reduction in the plant’s energy reserves before chemical treatment. Furthermore, in rhizome breaking and mowing plus herbicide plots, S. densiflora was replaced by a higher number of native species than in herbicide, mowing, and control plots, resulting in the highest diversity values. conclusIons our data indicated that S. densiflora has a strong capacity to resist mechanical and chemical control interventions; a single application of the various control techniques tested in this study seemed incapable of eradicating S. densiflora and returning invaded marshes to a pre-invasion state. However, this study provides valuable information for the management of this exotic species in the invaded habitats. rhizome breaking and mowing plus herbicide application could be useful for the control of the invasion of this species; both treatments decreased S. densiflora biomass and favored an increased diversity and native species richness. AcknowledgMents We are grateful to antonio j. ruiz rico for revision of the English in the manuscript. We also thank the spanish Environmental and science and technology Ministries for their support (project 042/2007 organismo autónomo Parques nacionales and project CtM2008-04453). lIterAture cIted ailstock Ms, norman CM, bushmann Pj. 2001. Common reed Phragmites australis control and effects upon biodiversity in freshwater nontidal wetlands. restor. Ecol. 9:49-59. an sq, Gu bH, zhou CF, Wang zs, deng zF, zhi Yb, li Hl, Chen l, Yu dH, liu YH. 2007. Spartina invasion in China: implications for invasive species management and future research. Weed res. 47:183-191. bortolus a. 2006. the austral cordgrass Spartina densiflora brong.: its taxonomy, biogeography and natural history. j. biogeogr. 33:158-168. Castellanos EM, Heredia C, Figueroa ME, davy aj. 1998. tiller dynamics of Spartina maritima in successional and non-successional Mediterranean salt marsh. Plant Ecol. 137:213-225. Castillo jM, Mateos-naranjo E, nieva Fj, Figueroa E. 2008. Plant zonation at salt marshes of the endangered cordgrass Spartina maritima invaded by Spartina densiflora. Hydrobiologia. 614:363-371. Fennane M, Mathez j. 1988. nouveaux matériaux pour la flore de Maroc. naturalia Montspeliensia. 52:135-141. Figueroa ME, Castellanos EM. 1988. Vertical structure of Spartina maritima and Spartina densiflora in Mediterranean marshes, pp. 105-108. in : M. j. a. Werger, P. j. M. Van der aart, H. j. during, and j. t. a. Verhoeven (eds.). Plant form and vegetation structure. sPb academic Publishing, the Hague. Gallego-Fernández jb, García-novo F. 2007. High-intensity versus low-intensity restoration alternatives of a tidal marsh in Guadalquivir estuary, sW spain. Ecol. Eng. 30:112-121. Haferkamp Mr, Karl MG. 1999. Clipping effects on growth dynamics of japanese brome. j. range Manage. 52:339-345. Hedge P, Kriwoken lK, Patten K. 2003. a review of spartina management in Washington state, us. j. aquat. Plant Manage. 41:82-90. Kittelson PM, boyd Mj. 1997. Mechanisms of expansion for an introduced species of cordgrass, Spartina densiflora, in Humboldt bay, California. Estuaries. 20:770-778. li H, zhang l. 2008. an experimental study on physical controls of an exotic plant Spartina alterniflora in shangai, China. Ecol. Eng. 32:11-21. Major WW, Grue CE, Grassley jM, Conquest l. 2003. Mechanical and chemical control of smooth Cordgrass in Willapa bay, Washington. j. aquat. Plant Manage. 41:6-12. Mateos-naranjo E, redondo-Gómez s, silva j, santos r, Figueroa ME. 2007. Effect of prolonged flooding on the invader Spartina densiflora brong. j. aquat. Plant Manage. 45:121-123. Mateos-naranjo E, redondo-Gómez s, luque Cj, Castellanos EM, davy aj, Figueroa ME. 2008. Environmental limitations on recruitment from seed J. Aquat. Plant Manage. 50: 2012. 111 in invasive Spartina densiflora on a southern European salt marsh. Estuar. Coast. shelf s. 79:727-732. Mateos-naranjo E, redondo-Gómez s, Cox l, Cornejo j, Figueroa ME. 2009. Effectiveness of glyphosate and imazamox on the control of the invasive cordgrass Spartina densiflora. Ecotox. Environ. safe. 72:1694-1700. nieva Fjj, Castellanos EM, Castillo jM, Figueroa ME. 2005. Clonal growth and tiller demography of the invader cordgrass Spartina densiflora brongn. at two contrasting habitats in sw European salt marshes. Wetlands. 25:122129. Patten K. 2002. smooth cordgrass (Spartina alterniflora) control with imazapyr. Weed technol. 16:826-832. Patten K. 2004. Comparison of chemical and mechanical control efforts for invasive spartina in Willapa bay, Wa, report. the Washington state department of agriculture. redondo-Gómez s, Mateos-naranjo E, davy aj, Fernández-Muñoz F, Castellanos EM, luque t, Figueroa ME. 2007. Growth and photosynthetic responses to salinity of the salt-marsh shrub Atriplex portulacoides. ann. bot-london 100:555-563. roberts Pd, Pullin as. 2008. the effectiveness of management interventions for the control of Spatina species: a systematic review and meta-analysis. aquat. Conserv. 18:592-618. shannon CE, Weaver W. 1949. the mathematical theory of communication. university of illinois Press, urbana, il. zanatta jF, Procopio so, Manica r, Pauletto Ea, Carnelutti a, Vargas l, sganzerla dC, rosenthal Mda, Pinto jjo. 2007. soil water contents and glyphosate efficacy in controlling Euphorbia heterophylla. Planta daninha. 25:799-811. J. Aquat. Plant Manage. 50: 111-116 Response of target and nontarget floating and emergent aquatic plants to flumioxazin CHristoPHEr r. MudGE and WilliaM t. HallEr* AbstrAct the effects of subsurface and foliar flumioxazin {2-[7-fluoro-3,4-dihydro-3-oxo-4-(2-propynyl)-2H-1,4-benzoxazin6-yl]-4,5,6,7-tetrahydro-1H-isoindole-1,3(2H)-dione} treatments were evaluated on the floating weeds waterhyacinth (Eichhornia crassipes (Mart.) solms), water lettuce (Pistia stratiotes l.), and landoltia (Landoltia punctata [G. Mey] d.H. les and d.j. Crawford) as well as the nontarget emergent species eleocharis (Eleocharis interstincta (Vahl) roem & j.a. schult), maidencane (Panicum hemitomon schult.), pickerelweed (Pontederia cordata l.), and sagittaria (Sagittaria lancifolia l.). all subsurface treatments (≥100 µg a.i. l−1) and foliar application rates >143 g a.i. ha−1 provided complete water lettuce control. Conversely, both flumioxazin application techniques provided <30% control of waterhyacinth. no injury symptoms were exhibited by landoltia treated with foliar flumioxazin applications, and in water, concentrations ≥200 µg a.i. l−1 were required to provide more than 50% control. sagittaria was the most sensitive nontarget emergent species to subsurface flumioxazin applications, followed by maidencane, eleocharis, and pickerelweed. sagittaria dry weight was reduced 100% at herbicide concentrations ≥800 µg a.i. l−1 compared to a 73 to 83% dry weight reduction in eleocharis, maidencane, and pickerelweed. Conversely, all emergent species were highly tolerant to foliar flumioxazin treatments, yielding calculated EC50 values ≥1320 g a.i. ha−1 for dry weight. results of this study indicate differential efficacy and selectivity among floating and emergent target and nontarget aquatic plant species when treated with flumioxazin. Key words: chemical control, dose response, EC50: Effective Concentration 50, Eichhornia crassipes, Eleocharis interstincta, Landoltia punctata, Panicum hemitomon, Pistia stratiotes, Pontederia cordata, protoporphyrinogen oxidase inhibitor, Sagittaria lancifolia, selectivity IntroductIon invasive floating aquatic plants, including waterhyacinth (Eichhornia crassipes (Mart.) solms) and water lettuce (Pistia stratiotes l.), spread by vegetative reproduction, forming extensive free floating mats that often interfere with navigation, hydroelectric generation, irrigation, and fishing as well as lowering the dissolved oxygen and pH of the water (Weldon and blackburn 1966, Harley et al. 1984, owens and Madsen 1995). they may also harbor mosquitoes, which are vectors for diseases like dengue fever, malaria, and encephalitis (Holm et al. 1977). Conversely, native emergent aquatic plants may provide a diverse and valuable food source and habitat for animals, can improve water clarity and quality, reinforce shorelines, and protect soil against erosion from wind and wave action (savino and stein 1982, Heitmeyer and Vohs 1984, smart 1995, dibble et al. 1996). damage to emergent nontarget and native plants species is a major consideration in herbicide selection; favorable aquatic herbicides are able to selectively remove unwanted plants while minimizing damage to nontarget native plants. one of the primary goals of aquatic weed management is to control invasive plants while maintaining a diverse native plant community. native plant density and diversity have been shown to increase when canopy-forming exotic plants *Former Graduate research assistant and Professor, Center for aquatic and invasive Plants, institute of Food and agricultural sciences, university of Florida, Po box 110610, Gainesville, Fl 32611. Current address of first author: us army Engineer research and development Center, Vicksburg, Ms 39180. Corresponding author’s E-mail: Christopher.r.Mudge@usace. army.mil. received for publication February 10, 2012 and in revised form june 28, 2012. 112 J. Aquat. Plant Manage. 50: 2012. are removed (Getsinger et al. 1997), and continued presence of native vegetation allows diversity of invertebrate and fish habitats to be maintained (dibble et al. 1996). therefore, it is beneficial to selectively remove floating invasive species to maintain native vegetation and, in turn, wildlife habitat. in 2010, flumioxazin received FiFra-section 3 registration for control of submersed, emergent, and floating aquatic weeds in the united states (Valent usa Corporation 2011). Flumioxazin is a very fast-acting contact herbicide that inhibits protoporphyrinogen oxidase (PPo; protoporphyrin iX:oxygen oxidoreductase, EC 1.3.3.4). it inhibits chlorophyll biosynthesis by preventing transformation of protoporphyrinogen iX into protoporphyrin iX, a precursor to heme and chlorophyll production (Matringe et al. 1989, Cobb 1992, aizawa and brown 1999). it has been evaluated in greenhouse and field studies for control of hydrilla (Hydrilla verticillata [l.f.] royle) and other invasive aquatic species (Mudge 2007, richardson et al. 2008, Mudge et al. 2010). the high costs associated with registering an herbicide for a new market (i.e., aquatics) may be overcome by maximizing the market potential; therefore, one objective of this research was to determine if flumioxazin has utility as a foliar or a subsurface treatment to control floating aquatic weeds. Emergent nontarget aquatic plants could be impacted by flumioxazin applications, so the second objective was to quantify the effects of foliar and subsurface flumioxazin treatments on common nontarget emergent aquatic plants. MAterIAls And Methods Floating aquatic plants. Waterhyacinth and water lettuce. Plants were collected from rodman reservoir near interlachen, Florida, and established in 95 l high-density polyethylene (HdPE) tanks filled with 80 l of tap water (pH 8.0) in april 2006 at the university of Florida (Center for aquatic and invasive Plants) in Gainesville, Florida. tap water was supplemented with 1 ml of Chelated iron Plus1 (12-0-0) and 150 mg l−1 Miracle-Gro®2 (24-8-16) fertilizer prior to herbicide treatment. nutrients were added again at 1 and 3 weeks after treatment. Waterhyacinth (5 plants per tank) and water lettuce (20 plants per tank) were allowed to acclimate for 3 weeks before treatment. this study was repeated in august 2006 on the main campus of the university of Florida with water from biven’s arm lake (pH 7.5). both studies were completely randomized designs with 4 replications (tanks) for each treatment. all studies were conducted under full sunlight. For the foliar treatments, flumioxazin3 was applied to the foliage with a forced air Co2-powered sprayer calibrated to deliver a spray volume of 935 l ha−1 through a single teejet®4 80-0067 nozzle at 0, 36, 72, 143, 286, 572, and 1144 g a.i. ha−1 plus a non-ionic surfactant5 (0.25% v/v). subsurface flumioxazin treatments were conducted concurrently at each location. Flumioxazin was applied at 0, 100, 200, 400, 800, and 1600 µg a.i. l−1 as static treatments. all live waterhyacinth and water lettuce tissue was harvested 34 d after treatment (dat), placed in a drying oven at 90 C for about 1 week, and weighed. Plant dry weight data were analyzed using nonlinear regression (exponential decay, y = b0e−bx) with the ProC nlin procedure (sas institute 2002), and regression models were used to determine the effective concentration 50 (EC50), which is the concentration of flumioxazin required to cause a 50% reduction in dry weight compared to control plants. because there were no differences between the slopes of the regression lines at the 95% confidence level, data were pooled for each repeated study. Landoltia. a population of landoltia (Landoltia punctata [G. Mey] d.H. les and d.j. Crawford) was collected from a pond with no history of herbicide treatments in alachua County, Florida, and cultured at the Center for aquatic and invasive Plants in 266 l fiberglass tanks in a greenhouse (light intensity of 1200 µmol m−2 s−1). Plants were cultured in tap water (pH 8.2) amended with Miracle-Gro (24-8-16, 150 mg l−1) and allowed to acclimate for 2 weeks before treatment. a 10 g aliquot (fresh weight; 1.3 ± 0.07 g dry weight) of landoltia was placed in 3 l HdPE (17.1 cm diameter by 13.3 cm deep) pots filled with tap water (pH 8.0) and Miracle-Gro. Plants were allowed to acclimate in the pots for an additional 2 d prior to herbicide treatment. all pots were amended with MiracleGro at 2 and 14 dat. the subsurface experiment was conducted in april and May 2007. landoltia was treated with flumioxazin at 0, 10, 25, 50, 100, 200, 400, 800, and 1600 µg a.i. l−1 as static treatments. as a comparison treatment, diquat6 was applied as a foliar treatment at 1.1 kg a.i. ha−1 using the described methods for foliar flumioxazin treatments. this experiment was a randomized design with five replicates. Foliar flumioxazin trials were conducted in october 2005, april 2007, and May 2007. Flumioxazin was applied to landoltia at 0, 36, 72, 143, 286, 572, and 1144 g a.i. ha−1 plus a non-ionic surfactant (sunstream; 0.25% v/v) using the same methods as described in the water lettuce and waterhyacinth studies. due to the difficulty of removing large quantities of necrotic or chlorotic and dead landoltia plants, visual estimates of control (% control) were determined on a scale of 0 to 100%, where 0 = no chlorosis or necrosis and 100 = plant death. Percent control ratings were based on nontreated control plants. there were no differences in control between the two experiments (Fisher’s Protected lsd, p ≤ 0.05); therefore, the data from the two experiments were pooled for analysis and means were separated using 95% confidence intervals. Emergent aquatic plants. the sensitivity of the nontarget emergent aquatic plants eleocharis, maidencane, pickerelweed, and sagittaria were similarly evaluated against subsurface and foliar flumioxazin application techniques. all plants were purchased from a local plant nursery in august 2006 and april 2007 for the subsurface and foliar studies, respectively. two healthy stems (30 to 38 cm) of each species were planted in a mixture of 2:1 potting media7 to masonry sand in 3 l HdPE pots amended with osmocote®8 (15-9-12) fertilizer at a rate of 1g kg−1 soil. the subsurface flumioxazin experiment was a randomized design with five replicates (tanks). Each species was cultured outdoors under shade cloth (70% sunlight) for 4 weeks in 95 l HdPE tanks. Water level in the tanks was maintained at 25 cm, and pH remained at or near 7.5 throughout the study. Plants were cultured for 1 month when flumioxazin was applied at 0, 50, 100, 200, 400, 800, and 1600 µg a.i. l−1 as static J. Aquat. Plant Manage. 50: 2012. 113 treatments. in the foliar flumioxazin trial, all five emergent replicates (pots) were placed in one 266 l fiberglass tank (72 by 82 by 45 cm) prior to treatment. Flumioxazin was applied as a foliar treatment at 0, 36, 72, 143, 286, 572, and 1144 g a.i. ha−1 with a non-ionic surfactant (sunstream; 0.25% v/v) using the same methods as the floating plants. at 40 dat, all live tissue from the foliar and subsurface trials was harvested at the soil line, placed in a drying oven at 90 C for about 1 week, and weighed. Plant dry weight data was analyzed using no-linear regression (ProC nlin, sas institute 2002), and EC50 values were determined for dry weight. data were pooled across experimental runs when no statistical differences between the slopes of regression lines were observed. results And dIscussIon Floating aquatic plants. Waterhyacinth and water lettuce. Water lettuce was much more sensitive to foliar applications of flumioxazin than waterhyacinth, with EC50 values of 69 and 1435 g a.i. ha−1, respectively (Figure 1). all foliar rates ≥286 g a.i. ha−1 resulted in complete control of water lettuce. treated water lettuce plants exhibited chlorosis and necrosis on the leaves 3 to 5 dat and defoliation 12 to 15 dat. Foliar flumioxazin rates ≥286 g a.i. ha−1 resulted in complete plant decay 21 dat, whereas sublethal rates (36 to 143 g a.i. ha−1) resulted in regrowth of young plants of water lettuce (ramets) from the central meristematic region. Previous research (richardson et al. 2008) demonstrated flumioxazin provided 97% control of water lettuce plants (9 cm diameter) with 34 g a.i. ha−1 and 100% with higher rates. our research evaluated flumioxazin on larger and more mature plants (15+ cm diameter), which likely explains the lower level of flumioxazin sensitivity in our research compared to that of richardson et al. (2008). Waterhyacinth biomass was reduced by only 41% of the nontreated control at the highest foliar flumioxazin rate evaluated (1144 g a.i. ha−1) 34 dat. the projected EC50 was 1435 g a.i. ha−1 (Figure 1), about three times the maximum label rate. treated waterhyacinth plants exhibited blackening on younger leaves only, which is similar to injury symptoms noted on water lettuce and waterhyacinth treated with the PPo inhibitor carfentrazone-ethyl (Koschnick et al. 2004). subsurface flumioxazin applications provided 100% water lettuce control at concentrations ≥100 µg a.i. l−1 (data not shown). in contrast, flumioxazin applied to the water column failed to reduce waterhyacinth biomass by more than 30% of the nontreated control plants at any rate evaluated (data not shown) and confirms that waterhyacinth is more tolerant of flumioxazin than water lettuce. these results are similar to those reported for waterhyacinth and water lettuce treated with the PPo inhibitor carfentrazone-ethyl (Koschnick et al. 2004). Landoltia. the effects of a subsurface application of flumioxazin to landoltia (Figure 2) show that most flumioxazin treatments caused foliar bleaching within 7 to 10 dat, but none of the treatments resulted in complete control of landoltia. Each flumioxazin treatment was different as indicated by the 95% confidence intervals. landoltia colonies treated at concentrations >25 µg a.i. l−1 began to separate, and roots became detached from individual fronds. Koschnick (2005) reported landoltia treated in the dark with diquat underwent root detachment without chlorosis. the primary function of roots of plants in the lemnaceae family is stabilization of fronds (landolt 1986). diquat applied as a comparison treatment resulted in 100% control less than 5 dat when the herbicide was foliar applied at 1.1 kg a.i. ha−1. duckweed is extremely sensitive to diquat and has a typical EC50 of 4 µg a.i. l−1 (Peterson et al. 1997), the current industry standard for duckweed control. the foliar applied flumioxazin landoltia study was conducted three times; treated plants were visually similar to control plants at all rates up to 1144 g a.i. ha−1 and showed no dose response, and therefore were not harvested (data not shown). the foliar treatments to landoltia and waterhyaFigure 1. the effect of a foliar flumioxazin application on the dry weight of waterhyacinth and water lettuce 34 d after treatment. data are shown as dry weight means ± standard error (n = 8). EC50 = effective concentration 50, concentration of flumioxazin (g a.i. ha−1) required to reduce waterhyacinth and water lettuce biomass by 50%. Figure 2. Percent control (visual) of landoltia 21 d after a foliar diquat (kg a.i. ha−1) and subsurface flumioxazin application (µg a.i. l−1 a.i.). Percent control ±95% confidence interval (n = 10). overlapping confidence interval bars indicate no significant difference. 114 J. Aquat. Plant Manage. 50: 2012. cinth were unsuccessful, and foliar rates ≥286 g a.i. ha-1 were needed to control water lettuce. these results suggest that flumioxazin uptake is limited by the leaf cuticle or occurs primarily through root uptake or absorption by the underside of the plant. Further research is needed to determine if flumioxazin applied as a subsurface treatment is as efficacious to water lettuce in higher pH water (9.0). Mudge et al. (2010) demonstrated flumioxazin in neutral pH water is more injurious to submersed aquatic plants than when applied to high pH (9.0) water. Emergent aquatic plants. Emergent aquatic plants had highly variable sensitivity to flumioxazin aqueous concentrations up to 1600 µg a.i. l−1 (Figure 3). sagittaria dry weight was reduced 100% at concentrations ≥800 µg a.i. l−1 compared to a 73 to 83% dry weight reduction in eleocharis, maidencane, and pickerelweed. sagittaria was the most sensitive species to a subsurface flumioxazin treatment followed by maidencane, eleocharis, and pickerelweed based on calculated EC50 values for dry weight (table 1). these data indicate eleocharis, pickerelweed, and maidencane would be injured by the maximum flumioxazin concentration of 400 µg a.i. l−1; however, eleocharis and pickerelweed would likely recover from the treatment. Visual injury symptoms observed 2 weeks after the subsurface flumioxazin application to emergent plants included interveinal chlorosis (sagittaria and pickerelweed), reddening on leaf margins (maidencane), and minor chlorosis (eleocharis). Flumioxazin and other PPo-inhibiting herbicides are absorbed primarily by plant roots with some absorbance in the shoots when applied to the soil, but translocation is limited once herbicides are absorbed into foliar tissue because of the rapid desiccation (Fadayomi and Warren 1977, ritter and Coble 1981, unland et al. 1999, senseman 2007). Pots without holes were used in these studies, and few roots were visible above the soil line. therefore, flumioxazin uptake occurred either through the underwater stem, submersed leaves, or roots. Previous research (Fadayomi and Warren 1977, ritter and Coble 1981) found little translocation of PPo-inhibiting herbicides in plants, but the subsurface treatment of emergent aquatic plants in this study suggested movement of flumioxazin from the soil or lower stem into the leaves. if translocation of flumioxazin was limited, this herbicide should have girdled the plant at the soil line and produced injury symptoms such as necrosis of the stem and leaves without veinal chlorosis first appearing in the leaves. Ferrell et al. (2007) showed flumioxazin in combination with MsMa (monosodium salt of Maa) resulted in a 94% yield reduction when applied as a high post-direct treatment to 20 cm tall cotton. symptomology of flumioxazin-treated cotton included necrotic lesions on leaves, reddening stems, stem girdling, and eventual lodging. Previous research also found that as cotton matures, plants become more tolerant to flumioxazin because of greater bark development and metabolic capacity (Ferrell and Vencill 2003). Foliar flumioxazin treatments were much less injurious to emergent aquatic plants than subsurface treatments (Figure 4). Maidencane and sagittaria would require foliar application rates >1884 and 1320 g a.i. ha−1, respectively, to reduce dry weight by 50% based on the calculated EC50 values (table 1). an EC50 value could not be calculated for dry weight for both eleocharis and pickerelweed because increased flumioxazin concentrations resulted in an increase in dry weight (positive regression slope). Postemergent applications of flumioxazin are generally recommended for actively growing weeds <5 cm in height (Valent usa Corporation 2009), so the minimal foliar injury and substantial lack of reduction in biomass observed in this foliar study were probably due to the advanced maturity of these plants. injury symptoms (including chlorotic and necrotic lesions on the leaves) were similar to those described for other PPo-inhibiting herbicides (Peterson et al. 2001). tolerant species have reduced or no symptoms, whereas the leaves of susceptible species rapidly desiccate and die (Peterson et al. 2001). the limited injury symptoms exhibited by plants in the foliar experiment were similar to plants exposed to subsurface treatments. selective control of invasive weed species is the goal of aquatic weed managers. Herbicide applicators target invasive plants through the use of specifically formulated herbicides, seasonally timed herbicide applications, and/or preemptive spot treatments before weeds become a problem (university of Florida 2011). the native emergent plants tested in this study were tolerant to foliar flumioxazin applications, evidenced by no observed mortality. in contrast, subsurface applications resulted in more injury to nontarget plants, especially when treatments were made in low pH water. Flumioxazin is rapidly degraded by hydrolysis, with an average half-life of 4.1 d, 16.1 h, and 17.5 min at pH 5.0, 7.0, and 9.0, respectively (Katagi 2003, senseman 2007). Mudge et al. (2010) demonstrated that water pH does not directly influence flumioxazin activity, but through an impact on aqueous flumioxazin degradation rates, pH of the treated water can have a profound impact on efficacy. Mesocosm and field trials have demonstrated differences in flumioxazin efficacy and selectivity when applied to water with a pH >8.0 (Mudge 2007, Mudge and Haller 2010). Figure 3. the effect of flumioxazin concentration on the dry weight of the nontarget emergent aquatic plants eleocharis (), maidencane (▫), pickerelweed (), and sagittaria () 40 d after treatment. data are shown as actual dry weight means ± standard error (n = 10). J. Aquat. Plant Manage. 50: 2012. 121 the study; however, the low dose followed by the high dose treatment resulted in an additional 50% decrease in biomass compared to the aforementioned treatment (Figure 2). these results indicate additional penoxsulam may be applied at higher concentrations a few weeks after initial treatment to increase control. the threshold for initial and secondary treatments will need to be further researched to determine the effectiveness of high dose follow up treatments. both fluridone treatments were applied at 20 µg l−1 followed by 5 µg l−1 at 4 or 8 Wat (table 4). although initial and follow up treatments were the same concentration, the additional 4 wk exposure at the higher concentration resulted in an additional 87% control. these data indicate the extra 4 weeks of fluridone exposure at the higher concentration are necessary to control this weed, compared to the temporary growth regulation observed with the low dose follow up treatment at 4 Wat. Penoxsulam concentrations were 4 and 21.5 µg l−1 for the penoxsulam 5 and 20 µg l−1 static treatments, respectively, 1 dat (table 5). Herbicide concentrations remained relatively stable throughout the course of the study for the 5 µg l−1 static treatment, whereas the 20 µg l−1 static concentrations declined at a much faster rate. the shorter half-life of the 20 µg l−1 treatment could be attributed to greater efficacy of the higher dose as plants were controlled and desiccated at a much faster rate, increasing uV light penetration into the water column and aiding herbicide degradation. Fluridone concentrations decreased to <1 µg l−1 by 8 Wat for the F20Ex4Fb5 treatment; therefore, experimental units in this treatment received an additional 3 µg l−1 of fluridone 9 Wat to supplement the loss of herbicide (data not shown). the fluridone re-treatment concentration was chosen based on the half-life of fluridone in the CEt experiment. significant rainfall occurred in Vicksburg throughout the month of july, which may have contributed to the rapid dilution of the herbicide. However, during this time the P5Ex16 treatment concentration (2.8 ± 0.27) was less than the F20Ex4Fb5 concentration 4 Wat, and by 8 Wat, penoxsulam remained stable while fluridone decreased to <1 µg l−1. the giant salvinia mat remained intact for the F20Ex4Fb5 treatment from 4 to 8 Wat; thus, increased photolytic degradation was unlikely to be the cause of rapid fluridone loss. results from this experiment indicate giant salvinia control can be achieved by implementing multiple applications or maintaining penoxsulam or fluridone concentrations for an extended period of time (>12 wk). low dose repeat applications are commonly used to manage hydrilla and Eurasian watermilfoil (Myriophyllum spicatum l.) with fluridone or penoxsulam (Getsinger et al. 2001, Koschnick et al. 2003). low dose static penoxsulam treatments (P5Ex16) can result in 76% control and completely suppress growth during the exposure period. increased control can be attained by exposing plants to higher penoxsulam concentrations (10 or 20 µg l−1) for longer periods of time (16 wk). low concentrations of slow-acting herbicides such as penoxsulam and fluridone may temporarily growth regulate or stunt giant salvinia for several weeks, but plants will ultimately recover once concentrations fall below this threshold. Herbicides and plant growth regulators have been proposed and investigated to achieve a balance of controlling invasive aquatic plants while preventing negative ecological effects from unchecked growth of these species (lembi and Chand 1992, netherland and lembi 1992, nelson 1996, 1997, 2012 forthcoming). Growth regulating concentrations of slow-acting herbicides may also be beneficial for aquatic plant management, including giant salvinia. this form of management may aid in preventing development of dense infestations and their negative effects, or by slowing the recovery of target invasive plants from other required forms of management. this could include foliar herbicide applications or stress from biocontrol agents, such as the giant salvinia weevil (Cyrtobagous salviniae), in an integrated pest management program (Mudge and Harms 2012). recent work has shown positive response to integrating penoxsulam treatments and multiple biocontrol agents on water hyacinth (Moran 2012). Subsurface vs. foliar experiment. at the conclusion of the study (11 Wat), all subsurface and foliar herbicide treatments reduced giant salvinia dry weight 27 to 67% of the nontreated control (Figure 3). in particular, penoxsulam foTable 4. herbicide concenTraTions measured from gianT salVinia TreaTed wiTh subsurface aPPlicaTions of Penoxsulam and fluridone aT 20 µ g l−1 in The ceT exPerimenT. Herbicide sampling Period Concentration (µg a.i. l−1 ± s.E.) Penoxsulam 1 data26.0 ± 0.08 1 Wat 24.6 ± 0.11 2 Wat 20.0 ± 3.37 4 Wat 20.3 ± 0.74 8 Wat 13.7 ± 0.46 Fluridone 1 dat 10.8 ± 0.23 1 Wat 17.0 ± 0.27 2 Wat 15.6 ± 0.20 4 Wat 12.4 ± 0.68 Table 5. herbicide concenTraTions measured in TreaTed waTer following subsurface aPPlicaTions of Penoxsulam To gianT salVinia in The single sTaTic Vs. mulTiPle aPPlicaTion exPerimenT. Herbicide treatmentasampling Period Concentration (µg a.i. l−1 ± s.E.) Penoxsulam 5 µg a.i. l−1 1 datb4.0 ± 0.17 1 Wat 4.4 ± 0.16 2 Wat 3.3 ± 0.03 4 Wat 2.8 ± 0.27 8 Wat 3.7 ± 0.11 10 Wat 3.4 ± 0.25 12 Wat 2.7 ± 0.25 16 Wat 1.8 ± 0.11 Penoxsulam 20 µg a.i. l−1 1 dat 21.5 ± 1.64 1 Wat 21.2 ± 1.69 2 Wat 18.3 ± 0.90 4 Wat 17.4 ± 0.73 8 Wat 11.8 ± 0.29 10 Wat 9.1 ± 0.79 12 Wat 7.3 ± 0.71 16 Wat 1.9 ± 1.85 aPenoxsulam applied as a onetime treatment and plants exposed for 16 wk; n = 4. babbreviations: dat, days after treatment; Wat, weeks after treatment. 122 J. Aquat. Plant Manage. 50: 2012. liar (24 h) plus fluridone subsurface, penoxsulam foliar (24 h), and penoxsulam foliar (static) reduced mean plant dry weight to below pretreatment level. the penoxsulam foliar (24 h) plus fluridone subsurface treatment provided better control than all penoxsulam or fluridone stand alone or combination subsurface treatments. although the glyphosate plus diquat mix was initially highly efficacious, plants began to recover within 3 Wat and displayed no injury symptoms by 11 Wat. Previous research demonstrated diquat at a much higher foliar rate (1.12 kg ha−1) plus a methylated seed oil and organosilicone surfactant blend provided 100% control to a single layer of giant salvinia 6 Wat (nelson et al. 2001). the amount of plant material and density of the mat in the subsurface versus foliar experiment was much greater (i.e., thicker) at the inception of the experiment compared to the single layer of giant salvinia treated in previous research by nelson et al. (2001). the additional layers of plant material in this study likely prevented some of the herbicide spray solution from reaching the plant material below the water surface. Fluridone concentrations failed to reach the target dose of 20 µg l−1 and never exceeded 6.1 µg l−1 throughout the course of the experiment (table 6). the reason for the low concentration, despite the “bump” at 5 Wat is unknown. Conversely, penoxsulam concentrations remained above 30 µg l−1 for the first week of the study; however, concentrations decreased to 22.8 µg l−1 at 2 Wat and remained relatively stable throughout the remainder of the study (table 6). the penoxsulam 59.57 g ha−1 foliar rate (24 h and static) was equivalent in amount of active ingredient to a 20 µg l−1 subsurface penoxsulam treatment. Partial migration of herbicide into underlying water was a planned effect of the static foliar treatment. although the theoretical in-water concentration of 20 µg l−1 was never achieved by this static foliar treatment, some of the herbicide solution was absorbed by the foliage, and the remainder reached the water column and was available for uptake by the submersed foliage. the penoxsulam static foliar treatment was 12.0 µg l−1 at 1 Wat and decreased to 7.8 µg l−1 by 11 Wat (table 4). Giant salvinia treated with subsurface or foliar static penoxsulam treatments began to exhibit similar injury symptoms as early as 2 Wat. Plants treated with penoxsulam at 20 µg l−1 and 59.57 g ha−1 (static) exhibited growth regulation, and older tissue became necrotic through 6 Wat. the penoxsulam foliar (24 h) application resulted in witches broom symptoms on all new tissue in addition to necrosis of older tissue 2 Wat. Fluridone-treated plants exhibited chlorosis by 1 Wat; however, minimal bleaching of the foliage remained by 4 Wat, and plants were symptom free by 6 Wat. the rapid decrease in injury symptoms was probably due to the low fluridone concentrations in the water. although the fluridone treatment was targeted at 20 µg l−1, the concentration was 5.9 ± 0.62 to 3.2 ± 0.20 between 1 dat and 4 Wat, respectively (table 6). the decrease in injury symptoms and concentrations prompted a bump treatment 5 Wat to increase the dose to 20 µg l−1 in all experimental units containing fluriFigure 3. Effect of subsurface and foliar penoxsulam (P), fluridone (F), glyphosate (G), and diquat (d) applications on mature giant salvinia mean dry weight (± s.E.) 11 weeks after treatment (Wat). Plants exposed to static treatments remained for the duration of the experiment, while 24 h indicates plants were removed from treatment and placed in fresh water. Horizontal line represents pretreatment biomass for giant salvinia. Means with the same letter are not significant according to Fisher’s protected lsd test at p = 0.05; n = 4. Table 6. herbicide concenTraTions measured following sTaTic subsurface or foliar aPPlicaTions of Penoxsulam and fluridone To maTure gianT salVinia in The subsurface Vs. foliar exPerimenT. Herbicide treatment sampling Period Concentration (µg a.i. l−1 ± s.E.) Fluridonea (20 µg a.i. l−1) 1 datb5.9 ± 0.62 1 Wat 5.2 ± 0.28 2 Wat 4.6 ± 0.24 4 Wat 3.2 ± 0.20 8 Wat 6.1 ± 0.53 11 Wat 4.6 ± 0.50 Penoxsulam (20 a.i. l−1) 1 dat 30.3 ± 1.81 1 Wat 30.9 ± 3.55 2 Wat 22.8 ± 1.76 4 Wat 19.4 ± 4.79 8 Wat 16.2 ± 1.39 11 Wat 14.1 ± 1.32 Penoxsulam (59.57 g a.i. ha−1) 1 dat 7.3 ± 0.36 1 Wat 12.0 ± 1.14 2 Wat 13.2 ± 0.78 4 Wat 14.8 ± 1.63 8 Wat 10.6 ± 0.86 11 Wat 7.8 ± 1.26 aFluridone applied on day of treatment and reapplied 5 Wat to increase concentration to 20 µg a.i. l−1. babbreviations: dat, days after treatment; Wat, weeks after treatment; n = 4. J. Aquat. Plant Manage. 50: 2012. 123 done. the fluridone bump resulted in an increase in bleaching symptoms through the remainder of the study. the subsurface penoxsulam plus subsurface fluridone treatment resulted in a variety of injury symptoms throughout the course of the experiment. Plant injury symptoms included necrosis and growth regulation (1 to 2 Wat) along with minimal chlorosis and witches broom. the combination of glyphosate plus diquat resulted in faster and more intense injury symptoms than any other herbicide treatment in this trial. Plants treated with this herbicide combination exhibited necrosis <1 Wat. the rapid activity of this combination was not surprising because diquat injures giant salvinia as early as 1 dat (nelson et al. 2001). although this combination seemed to be highly efficacious at quickly desiccating older tissue, new plant growth was observed as early as 3 Wat. the single static versus multiple application study demonstrated penoxsulam at 20 µg l−1 reduced giant salvinia biomass 100% when plants were exposed for 16 wk, whereas plants in this study were continuing to die and lose buoyancy at 11 Wat. an additional 3 to 5 wks of exposure should have resulted in near complete to complete control based on the response of giant salvinia to the penoxsulam in the single static versus multiple applications study. this notion is supported by previous research, which indicated that alsand Pds-inhibiting herbicides penoxsulam and fluridone, respectively, are relatively slow acting and require long exposures (60+ d) to effectively control target species (netherland and Getsinger 1995, Koschnick et al. 2007b). the intent was to conclude the experiment 16 Wat, but it was shortened because control plants began to decline in health after temperatures were unusually cooler than normal in september 2010. although fluridone-treated plants were minimally controlled in this study (Fig. 3), the CEt and single static versus multiple application studies demonstrated the effectiveness of this product (Fig. 1 and 2). Fluridone was highly efficacious (99% control) when plants were exposed to 20 µg l−1 for at least 8 wk (Fig. 2). the focus of the third year of research was to extend the exposure time beyond 8 wk to achieve 100% control, but fluridone concentrations failed to reach or be maintained at the target concentration (table 6). the tank mix of glyphosate plus diquat plus two surfactants (nonionic and buffering agent + nonionic organo-silicone) is currently one of the recommended foliar treatments for giant salvinia in louisiana (d. E. sanders and a. j. Perret, 2012, pers. comm.). one or two plant layers of giant salvinia are controlled with this mixture; however, multiple levels of plant material are difficult to penetrate with a single application of any foliar applied herbicide or herbicide combination; therefore, multiple applications are often necessary to effectively control or eradicate dense giant salvinia infestation (nelson et al. 2007). both foliar penoxsulam treatments (static and 24 h exposure) provided similar control to the glyphosate plus diquat tank mix evaluated in this study. the penoxsulam 59.57 g ha−1 foliar rate was equivalent to a 20 µg l−1 subsurface treatment if all the herbicide spray reached the water column and failed to come in contact with the plant canopy. in comparison, the 24 h foliar penoxsulam treatment was designed to limit any potential herbicide uptake from the water column. although the dry weight data reflected no differences, many new healthy leaves were developing from plants exposed for 24 h, whereas only a few healthy fronds were witnessed with the static foliar penoxsulam treatment. because penoxsulam was still present 11Wat (7.8 ± 1.26 µg l−1), an additional few weeks of herbicide exposure may have separated these treatments, allowing older plant tissue in the static penoxsulam foliar treatment to desiccate and allow more new plant growth in the 24 h penoxsulam foliar and glyphosate plus diquat treatments. in addition, penoxsulam is recommended at 35.04 to 98.12 g ha−1 as a foliar application, which is equivalent to 2.0 to 5.6 oz product a-1. Future research should be conducted to determine if a higher foliar rate can provide greater efficacy as well as faster activity. Previous research has shown that penoxsulam is an effective herbicide when applied subsurface to control hydrilla and variable-leaf watermilfoil (Myriophyllum heterophyllum Michx.; Koschnick et al. 2007a, Glomski and netherland 2008). our data indicate penoxsulam as a foliar or subsurface application can be a viable alternative to the standard tank mix of glyphosate plus diquat plus two surfactants for controlling various sized infestations of giant salvinia. Penoxsulam may have the potential to provide improved, longer-term control over previous standard foliar treatments under certain use scenarios, particularly for large, dense infestations with high recovery potential. subsurface applications should be maintained for a minimum of 8 wks to provide acceptable control, but 12+ wk of exposure generally provided excellent control. Penoxsulam applied at 5 to 20 µg l−1 under extended exposures, can provide growth regulation or control of giant salvinia. along with lethal control outcomes, the ability to use low-dose penoxsulam for growth regulation is an additional use characteristic that may complement other control techniques such as biological control or foliar applications where otherwise re-growth potential would preclude effective management. depending on potential for dilution or other forms of dissipation, multiple applications or bump treatments may be necessary to maintain effective concentrations of penoxsulam in the water column. a foliar or subsurface penoxsulam treatment may be a beneficial treatment depending on the locale of plants (open water vs. backwater), presence of nontarget plant species, or the number of layers/thickness of the giant salvinia mat. the penoxsulam and fluridone data generated in these three experiments need to be further investigated on an operational level in field sites infested with giant salvinia. sources of MAterIAls 1Miracle-Gro® 36-6-6, the scott’s Company, Marysville, ohio 2aquashade®, applied biochemists, Germantown, Wisconsin 3Galleon sC®, sePro Corporation, Carmel, indiana 4sonar as®, sePro Corporation, Carmel, indiana 5aquaPro®, sePro Corporation, Carmel, indiana 6reward® landscape and aquatic Herbicide, syngenta Professional Products, Greensboro, nC 7aqua-King Plus®, Winfield solutions, llC, st. Paul, Minnesota 8thoroughbred®, Winfield solutions, llC, st. Paul, Minnesota 9teejet technologies, Wheaton, il AcknowledgMents this research was supported by sePro Corporation/ aquatic Ecosystem restoration Foundation and the us army aquatic Plant Control research Program. appreciation is 124 J. Aquat. Plant Manage. 50: 2012. extended to l. nelson, C. Grodowitz, j. namanny, M. robertson, j. smith, K. derossette, M. sternberg, and a. Poovey for technical assistance throughout the research. d. sanders (lsu agCenter) and a. Perret (louisiana department of Wildlife and Fisheries) kindly provided recommendations on giant salvinia control in louisiana. M. netherland and l. Glomski kindly provided reviews of this manuscript. Permission was granted by the Chief of Engineers to publish this information. Citation of trade names does not constitute endorsement or approval of the use of such commercial products. lIterAture cIted anderson WP. 1996. Weed science: Principles and applications, 3rd ed. West Publishing Company, st. Paul, Mn. Cary Pr, Weerts PGj. 1983. Growth of Salvinia molesta as affected by water temperature and nutrition. iii. nitrogen-phosphorus interactions and effect of pH. aquat. bot. 19:171-182. Forno iW, Harley Kls. 1979. the occurrence of Salvinia molesta in brazil. aquat. bot. 6:185-187. Getsinger Kd, Madsen jd, Koschnick tj, netherland Md, stewart rM, Honnell dr, staddon aG, owens Cs. 2001. Whole-lake applications of sonar tM for selective control of Eurasian watermilfoil. technical report ErdC/El tr-01-7, us army Engineer research and development Center, Vicksburg, Ms. Glomski lM, netherland Md. 2008. Efficacy of fluridone, penoxsulam, and bispyribac-sodium on variable-leaf milfoil. j. aquat. Plant Manage. 46:193-196. jacono C. 1999. Salvinia molesta (d. s. Mitchell) invades the united states. aquatics. 21(1):4-9. jacono C, Pitman b. 2001. Salvinia molesta: around the world in 70 years. ans digest 4(2):13-16. johnson s, sanders d, Eisenberg l, Whitehead K. 2010. Fighting the blob: efforts to control giant salvinia. louisiana agriculture. 5(1)3:6-9. Koschnick tj, Haller Wt, Vandiver VV, santra u. 2003. Efficacy and residue comparisons between two slow-release formulations of fluridone. j. aquat. Plant Manage. 41:25-27. Koschnick t, Heilman Ma, Miller s. 2007a. Penoxsulam: a new aquatic herbicide for large-scale aquatic plant management. Proc. aquatic Plant Management soc. 47:41. Koschnick tj, netherland Md, Haller Wt. 2007b. Effects of three als-inhibitors on five emergent native plant species in Florida. j. aquat Plant Manage. 45:47-51. lembi Ca, Chand t. 1992. response of hydrilla and Eurasian watermilfoil to flurprimidol concentrations and exposure times. j. aquat. Plant Manage. 30:6-9. Madsen jd, Wersal rM. 2009. Giant salvinia. invasive Plant atlas of the Midsouth. http://www.gri.msstate.edu/ipams/species.php?sname= salvinia+ molesta&Cname=. accessed 9 aug 2010. McFarland dG, nelson ls, Grodowitz Mj, smart rM, owens Cs. 2004. Salvinia molesta d. s. Mitchell (giant salvinia) in the united states: a review of species ecology and approaches to management. sr04-2. u.s. army Engineer research and development Center, Environmental laboratory, Vicksburg, Ms. 40 pp. Moran Pj. 2012. influence of biological control damage on efficacy of penoxsulam and two other herbicides on waterhyacinth. j. aquat. Plant Manage. 50:32-38. Mudge Cr, Harms nE. 2012. development of an integrated pest management approach for controlling giant salvinia using herbicides and insects. us army Engineer research and development Center, Vicksburg, Ms. aPCrP bulletin. ErdC aPCrP-a-12-1. Murai n, skoog F, doyle ME, Hanson rs. 1980. relationships between cytokinin production, presence of plasmids, and fasciation caused by strains of Corynebacterium fascians. Proc. natl. acad. sci. usa. 77:619-623. nelson ls. 1996. Growth regulation of Eurasian watermilfoil with flurprimidol. j. Plant Growth regul. 15:33-38. nelson ls. 1997. response of hydrilla and american pondweed to flurprimidol. j. aquat. Plant Manage. 35:50-54. nelson ls, skogerboe jG, Getsinger Kd. 2001. Herbicide evaluation against giant salvinia. j. aquat. Plant Manage. 39:48-53. nelson ls, Glomski lM, Gladwin dn. 2007. Effect of glyphosate rate and spray volume on control of giant salvinia. j. aquat. Plant Manage. 45:5861. netherland Md. 2011. Comparative susceptibility of fluridone resistant and susceptible hydrilla to four als inhibiting herbicides under laboratory and greenhouse conditions. j. aquat. Plant Manage. 49:100-106. netherland Md, Getsinger Kd. 1995a. laboratory evaluation of threshold fluridone concentrations under static conditions for controlling hydrilla and Eurasian watermilfoil. j. aquat. Plant Manage. 33:33-36. netherland Md, lembi Ca. 1992. Gibberellin synthesis inhibitor effects on submersed aquatic weed species. Weed science. 40:29-36. owens Cs, smart rM. 2010. Effects of nutrients, salinity, and pH on Salvinia molesta (Mitchell) growth. Vicksburg, Ms: us army Engineer research and development Center. aPCrP technical notes Collection. ErdC/ tn aPCrP-Ea-23. owens Cs, smart rM, stewart rM. 2004. low temperature limits of giant salvinia. j. aquat. Plant Manage. 42:91-94. richardson rj, Gardner aP. 2007. Evaluation of penoxsulam for water hyacinth [Eichhornia crassipes (Mart.) solms] and giant salvinia (Salvinia molesta Mitchell) control. Weed sci soc am abstr. 47:58. http://wssa.net/ Meetings/Wssaabstracts/abstractsearch.php. accessed 20 Mar 2012. senseman sa. 2007. Herbicide Handbook. 9th ed. lawrence, Ks: Weed science society of america. 458 p. theel Hj, nelson ls, Mudge Cr. 2012. Growth regulating hydrilla and subsequent effects on habitat complexity. j. aquat. Plant Manage. In press. thomas P, room PM. 1986. taxonomy and control of Salvinia molesta. nature. 320(17):581-584. tranel Pj, Wright tr. 2002. resistance of weeds to als-inhibiting herbicides: what have we learned? Weed science. 50:700-712. Wersal rM, Madsen jd. 2007. Comparison of imazapyr and imazamox for control of parrotfeather (Myriophyllum aquaticum (Vell.) Verdc.). j. aquat. Plant Manage. 45:132-136. Wersal rM, Madsen jd. 2010. Combinations of penoxsulam and diquat as foliar applications for control of waterhyacinth and common salvinia: evidence of herbicide antagonism. j. aquat. Plant Manage. 48:21-25.