RESEARCH REPORT

Prospects for Disrupting Rhizome Apical Dominance Prior to Chemical Treatment of Phalaris arundinacea Craig A. Annen

ABSTRACT Reed canarygrass (Phalaris arundinacea) is a widely distributed invasive species that dominates many natural areas and restoration sites. Cost-effective suppression and restoration strategies need to be developed for plant communities affected by this species. Pretreatments designed to disrupt rhizome apical dominance may augment herbicide performance by making reed canarygrass rhizomes more susceptible to herbicide applications. I tested whether coupling pretreatment disking or kinetin application to herbicide application would enhance chemical control relative to only solitary herbicide application. I also evaluated the relative performance of two grass-selective herbicides, sethoxydim and fluazifop. All treatments suppressed reed canarygrass and indirectly led to improvements in existing native species abundance com- pared to the untreated control. In terms of reed canarygrass suppression, non–reed canarygrass aboveground biomass, and species diversity (Shannon’s diversity), fluazifop performed as well as sethoxydim. Reed canarygrass biomass was consistently lower in plots where either disking or kinetin pretreatments were coupled with herbicide application than in plots receiving only herbicide treatment, though the degree of additional suppression varied with choice of herbicide. When sethoxydim was used for follow-up herbicide applications, disking reduced reed canarygrass biomass more than the kinetin pretreatment, but when fluazifop was used, kinetin pretreatments and disking were similar in their suppres- sive effect. Results of this study suggest that coupling these pretreatments with herbicide application can improve grass- selective herbicide performance on reed canarygrass. Keywords: apical dominance, fluazifop, kinetin,UW Press reed canarygrass / Ecological (Phalaris Restorationarundinacea), sethoxydim

eed canarygrass (Phalaris arun- 11 different plant community types bank and/or rhizome bud bank (Reyes Rdinacea) is a widely distributed (Curtis 1959). Reed canarygrass inva- 2004, Annen 2008). For this reason, perennial grass that invades and sions displace native plant species and I propose adopting the term “sup- dominates natural areas and restora- alter restoration trajectories and vege- pression” rather than “control” when tion sites that have been disturbed tation succession patterns (Apfelbaum discussing herbicide effects on reed by sedimentation (Werner and Zedler and Sams 1987, Maurer et al. 2003, canarygrass to take into account the 2002), nutrient enrichment (Wetzel Annen et al. 2008). Long-term, sus- fact that these effects often do not and van der Valk 1998, Green and tained control of this species continues persist beyond one or two growing Galatowitsch 2001), hydrological to be a difficult management objec- seasons (and also to prevent confu- instability or modification (Galatow- tive to achieve, particularly at sites sion when the term “control” is used itsch et al. 2000, Bonilla-Warford and where the disturbances that contribute in experimental design terminology). Zedler 2002, Miller and Zedler 2003), to reed canarygrass invasions cannot Post-application regrowth in reed or any combination of these factors be remedied prior to implementing canarygrass has been documented in (Kercher and Zedler 2004). !is inva- control programs. several herbicide studies (Preuninger sive species covers more than 202,000 Herbicide application, alone or in and Umbanhowar 1994, Kilbride and ha of Wisconsin, USA (Bernthal and combination with other treatments, Paveglio 1999, Rachich and Reader Hatch 2008), where it has fidelity for is the method most commonly 1999, Reinhardt and Galatowitsch employed for reed canarygrass control. 2004, Lesica and Martin 2004, Wilcox Ecological Restoration Vol. 28, No. 3, 2010 Herbicide applications are effective 2004, Annen et al. 2005, Hovick and ISSN 1522-4740 E-ISSN 1543-4079 in the short term, but repeated treat- Reinartz 2007, Wilcox et al. 2007, ©2010 by the Board of Regents of the ments are usually required because Annen 2008), and a mechanism to University of Wisconsin System. this species can resprout from its seed explain resprouting from rhizomes was

September 2010 ECOLOGICAL RESTORATION 28:3 ! 291

4-ER28.3 Annen (291-99).indd 291 6/18/10 5:35 AM Figure 1. The effects of apical dominance and pretreatments on the distribution of herbicide within a perennial grass rhizome. When apical dominance is active, lateral buds are unaffected by herbicide and will develop into new tillers that will perpetuate the clone unless retreated with additional herbicide applications. When apical dominance is disrupted prior to herbicide application, herbicides are more uniformly distributed throughout the rhizome/tiller system, resulting in enhanced suppression and reduced regrowth.

Herbicide application only (apical dominance in place)

Quiescent lateral buds Active apical bud

Unaffected lateral buds Carbohydrate and herbicide distributions are polarized within the rhizome Regrowth of new tillers Disking (apical dominance disrupted)

Disking slices rhizomes into isolated fragments

Ɣ Herbicide molecule Bud-bearing rhizome fragments resprout into new tillers

and follow up herbicide application affects more of the clone Direction of herbicide translocation

Kinetin pretreatment (apical dominance disrupted) Reduced regrowth

Carbohydrate and herbicide distributions are uniform within the rhizome

Figure 1. The effects of apical dominance and pretreatments on the distribution of herbicide within a perennial grass rhizome. When apical dominance is active, lateral buds are unaffected by herbicide and will develop into new tillers that will perpetuate the clone unless retreated with additional herbicide applications. When apical dominance is disrupted prior to herbicide application, herbicides are more uniformly distributed throughout the rhizome/tiller system, resulting in enhanced suppression and reduced regrowth.

postulated by Annen (2008). Herbi- UWreed Presscanarygrass / Ecological rhizomes Restorationmore sus- reed canarygrass clones. Apical domi- cide performance depends on several ceptible to the effects of herbicides nance gives rise to rhizome buds that factors, including choice of herbicide, (Hillman 1985, Harker and Vanden are physiologically, anatomically, and proper use of additives, and appli- Born 1997, Annen 2004). Coupling morphologically heterogeneous (Hill- cation timing, and further depends accessory treatments (such as tillage man 1985, Sachs 2002). As a result on how an herbicide is used in the or growth regulator pretreatment) of apical dominance, perennial grass context of the life history traits of the with herbicide application has been rhizomes possess two types of buds: target species. Life history traits such shown to improve reed canarygrass actively growing apical buds that give as perenniality and clonal expansion, suppression and reduce the magnitude rise to new tillers, and metabolically both of which are conferred by rhizome of regrowth in several recent studies inactive (dormant) lateral buds that systems, complicate control efforts. In (Kilbride and Paveglio 1999, Paveglio initiate renewed growth whenever a order to completely kill a perennial and Kilbride 2000, Hovick and Rein- disturbance affects top growth or the species such as reed canarygrass, foliar- artz 2007, Annen 2008). rhizome apex. !ese dormant lateral applied herbicides need to be translo- rhizome buds may contribute substan- cated to rhizomes in toxic quantities Rhizome Apical tially to reed canarygrass resurgence and must also be distributed through- Dominance and Reed capacity: Reyes (2004) measured an out the entire rhizome system. Regret- average of 820–900 dormant rhizome Canarygrass Control tably, systemic herbicides tend to con- buds per square meter (47%–76% of centrate only in the distal, actively Apical dominance is the inhibitory the total buds) in a well-established growing portions of perennial grass influence of the rhizome apex on lat- reed canarygrass stand. rhizomes because of apical dominance eral rhizome bud outgrowth and rhi- Apical dominance causes carbo- (Harker and Dekker 1988). Acces- zome branching. Holt (1954), Mar- hydrates and inorganic nutrients to sory treatments that disrupt rhizome quis and others (1979), and Reyes concentrate at the rhizome apex rather apical dominance and stimulate lateral (2004) presented evidence that an being uniformly distributed through- rhizome bud outgrowth may make apical dominance system operates in out the entire rhizome (Figure 1).

292 ! September 2010 ECOLOGICAL RESTORATION 28:3

4-ER28.3 Annen (291-99).indd 292 6/18/10 5:35 AM Herbicides applied to aboveground sites with remnant sod and no prior Evaluating Fluazifop as an leaves and culms likewise accumulate history of tillage), no-till methods that Alternative to Sethoxydim in rhizome apices because they are can elicit similar effects on rhizome transported along with the carbohy- suppression (such as plant growth Sethoxydim and fluazifop are two drate assimilate stream. Harker and regulator applications) are potential postemergence systemic graminicides Dekker (1988) measured the distal/ pretreatment options, but further with the same mode of action (non- basal accumulation ratio (a measure research is required before protocols competitive inhibition of lipid synthe- of the degree of uniformity of her- can be established for their use. Harker sis) but from different chemical fami- bicide distribution) for a series of and Taylor (1994) tested a mixture lies. Annen and others (2005), Wilcox postemergence herbicides in quack- of the growth regulators chlormequat and others (2007), Annen (2008), and grass (Elytrigia repens) rhizomes. !eir chloride and ethephon for enhancing Healy and Zedler (2010) evaluated results indicated that postemergence quackgrass control with sethoxydim, the efficacy of sethoxydim on reed herbicides translocated toward the and Annen (2008) evaluated this canarygrass, but to date no empirical tips of rhizomes and that lateral rhi- same mixture for reed canarygrass studies have documented the effects of zome buds were largely unaffected by control. Although Harker and Taylor fluazifop on reed canarygrass or com- herbicide applications to top growth. (1994) reported a 60% increase in pared its performance to sethoxydim Similarly, Marquis and others (1979) suppression of aboveground biomass in the same experiment. Herbicide reported that glyphosate applied to when this growth regulator system activity follows a dose-response model reed canarygrass leaves accumulated in was applied as a pretreatment, Annen that can be quantified by an IC₅₀ distal meristematic tissues. As a conse- (2008) found that when this mixture value—the concentration of herbicide quence of apical dominance, systemic was applied at the same rate it lessened molecules required to inhibit 50% of herbicides kill only a portion of the reed canarygrass regrowth by only target enzyme activity; a more effective rhizome. Unaffected lateral buds can 26% and was cost prohibitive as a herbicide will have a lower IC₅₀ value. resprout into new tillers when the treatment option. Sachs and !imann Devine (1997) assayed the IC₅₀ of herbicide degrades. From a manage- (1967) and McIntyre (1971) tested the several graminicides applied to yard- Eleusine indica ment perspective, apical dominance synthetic cytokinin growth regulator grass ( ) and reported a + necessitates multiple-year herbicide kinetin (6N-furfurylaminopurine) for value of 1.4 mol/L for fluazifop and + applications to deplete the pool of releasing quackgrass lateral buds from 1.3 mol/L for sethoxydim. He con- dormant lateral rhizome buds within apical dominance. A 20 ppm kine- cluded that sethoxydim and fluazifop a reed canarygrass stand. tin solution applied directly to dor- had a comparable level of activity on grasses. However, individual species Two accessory treatments that dis- UWmant Press lateral / rhizomeEcological buds Restorationresulted in rupt rhizome apical dominance are till- activation of growth and elongation, can vary in their responses to herbicide age and plant growth regulator appli- though the effects diminished within treatments (cf. Marquis et al. 1979). cation (Figure 1). Tillage enhances two weeks. Cline (1991) reviewed Harker and Dekker (1988) reported herbicide performance by decapitat- the literature and reported that inac- that sethoxydim had a higher distal/ ing rhizome apices and slicing rhi- tive lateral buds began to synthesize basal accumulation ratio than fluazi- zomes into isolated multinodal frag- proteins similar to those produced by fop when applied to quackgrass, and ments, which initiates active growth actively elongating apical buds after based on their results we could predict in dormant lateral buds (Leakey et al. treatment with kinetin. Since herbi- that sethoxydim performance will be 1975). Resprouting makes the rhi- cide transport follows carbohydrate more affected by apical dominance zome more susceptible to complete source-sink patterns, herbicide trans- than fluazifop, the latter being a more herbicide translocation. Pretreatment location to dormant lateral buds might effective chemical treatment. Since it is with a growth regulator biochemically be enhanced by pretreatment with not clear which graminicide will per- promotes lateral bud outgrowth and the growth regulator kinetin. X-Cyte form better against reed canarygrass, increases herbicide performance. growth regulator (Stoller Enterprises, I wanted to test if fluazifop was more In recent field experiments (Kilbride Houston TX) is a water-soluble pos- effective than sethoxydim at sup- and Paveglio 1999, Paveglio and temergence formulation of kinetin pressing reed canarygrass. Fluazifop Kilbride 2000, Hovick and Reinartz (a synthetic cytokinin) that is used offers two additional advantages over 2007, Annen 2008), the combined in agriculture to promote tillering, sethoxydim: it is less expensive on a use of tillage and herbicide applica- enhance translocation of substances per hectare basis ($12 vs. $20, 2008 tion suppressed reed canarygrass from leaves, increase leaf surface area, prices, not including additives) and abundance more effectively than and increase root initiation in cereal more resistant to ultraviolet degrada- only solitary herbicide application. grasses. !is growth regulator may tion (see Matysiak and Nalewaja 1999 For sites where tillage is troublesome have potential for enhancing herbi- or Annen 2006) owing to double bond or impractical (such as wet sites or cide performance on reed canarygrass. resonance in its molecular structure.

September 2010 ECOLOGICAL RESTORATION 28:3 ! 293

4-ER28.3 Annen (291-99).indd 293 6/18/10 5:35 AM !e objectives of this study were to with graminicides were tested in a agent and Dyne-Amic MSO/NIS were determine if disking or kinetin pre- randomized complete block split-plot added to fluazifop spray mixtures at treatment followed by graminicide experiment in 2006 and 2007. Four the same rate as for sethoxydim to application enhances chemical control main effect treatments were tested standardize treatments and ensure that of reed canarygrass relative to solitary in 100 m² whole-plot experimental any treatment differences were not due graminicide application and to com- units: 1) tillage (light disking to a to additives, which can substantially pare relative reed canarygrass suppres- depth of 10–15 cm) followed by her- affect herbicide performance on reed sion with the graminicides sethoxydim bicide application (at the 3- to 4-leaf canarygrass (Annen 2006). To prevent and fluazifop. growth stage, approx. 21 d treatment cross-contamination of herbicides, interval); 2) two sets of kinetin appli- separate spray tanks were used to apply Methods cations to reed canarygrass during the sethoxydim and fluazifop, and tanks 2- to 3-leaf growth stage (approx. 12 d were flushed out thoroughly between Study Site and Design interval) followed by herbicide appli- each subplot application to prevent cation at the 3- to 4-leaf growth stage buildup of herbicide concentrations. !is study was conducted within a 23 (5 d after the second kinetin applica- X-Cyte (0.04% a.i. kinetin) was ha restored wet prairie at the Swamplo- tion); 3) herbicide application only applied at a rate of 1.0 pint/acre (1.2 vers Foundation Nature Preserve, a (at the 3- to 4-leaf growth stage); and L/ha). Kinetin applications were made 186 ha land trust located in the Drift- 4) untreated control. In addition, two after 4:00 p.m. because this growth less Area of southwestern Wisconsin, herbicide split-plot effects were tested regulator formulation is light sensi- USA (89°40'N, 43°7½'W). Hydro- in 50 m² experimental subunits: 1) tive and requires an uptake period of logical input for the wet prairie is pri- sethoxydim (Sethoxydim E Pro) and several hours (Leo Brostowitz, profes- marily surface flow from the adjacent 2) fluazifop butyl ester (Fusilade DX). sional crop consultant, pers. comm.). landscape, and the site has a 227 ha Treatments were randomly assigned to A nonionic surfactant and sticking drainage basin. During the course of subplots and split plots and replicated agent (Induce pH, Helena Chemical, this investigation, surface water was three times. Memphis TN) was added to kinetin present at the site for only one week spray mixtures at a rate of 4 mL/L following early spring rainfall in 2006, Application Protocol (0.375% v/v) to encourage penetra- and for three weeks following autumn Sethoxydim E Pro (13% a.i. seth- tion of the growth regulator into reed rainfall in 2007. !e research site was oxydim) was applied at a rate of 3.75 canarygrass leaves, prevent premature used for row-crop agriculture for sev- pints/acre (4.45 L/ha). A water con- loss of applied growth regulator from eral decades prior to being restored to UWditioning Press agent/ Ecological (ReQuest, Restoration Helena leaf washing and rewetting, prevent wet prairie in 1987. Although 79 spe- Chemical, Memphis TN) was added evaporation of spray solutions from cies were originally planted at the site, to sethoxydim spray mixtures at leaf surfaces, and stabilize tank mix- at the beginning of the experiment a rate of 2.5 mL/L (0.25% v/v) to ture pH. McIntyre (1971) reported reed canarygrass dominated the plant- sequester hardwater cations that can that the effects of kinetin application ing at ca. 85% cover. !e next most accelerate physical decomposition on lateral bud outgrowth wore off abundant species were, in decreasing of sethoxydim (Shoaf and Carlson after 12 days. For that reason, kine- order of abundance, Canada golden- 1992, Annen 2006). A methylated tin was applied twice with a 12-day rod (Solidago canadensis), tall white seed oil/nonionic surfactant blend interval between applications. !e beard-tongue (Penstemon digitalis), (MSO/NIS, Dyne-Amic, Helena 21-day interval between disking and New England aster (Aster novae- Chemical, Memphis TN) was added subsequent herbicide application rep- angliae), common ragweed (Ambrosia to sethoxydim spray mixtures at a rate resented the time required for reed artemisiifolia), and Virginia mountain- of 4 mL/L (0.375% v/v) to enhance canarygrass regrowth to reach the 3- to mint (Pycnanthemum virginianum) foliar penetration and protect against 4-leaf growth stage after disking. (nomenclature follows Gleason and UV degradation of sethoxydim, which In 2006, kinetin was applied on 21 Cronquist 1991). Twenty-five addi- can inactivate 50% of applied herbi- May and 2 June, plots were disked on tional species (22 planted native and cide within ten minutes (Matysiak 2 June, herbicide was applied to kine- 3 weedy non-natives) were also pres- and Nalewaja 1999). Fusilade DX tin and herbicide-only plots on 7 June, ent in low density at the beginning of (24.5% a.i. fluazifop butyl ester) was and herbicide was applied to disked the experiment. As part of ongoing applied at a rate of 24 fluid ounces/ plots on 22 June. In 2007, kinetin was management at the preserve, the entire acre (1.75 L/ha). Both herbicides were applied on 21 May and 1 June, plots experimental area was burned on 22 applied as a 30 psi broadcast spray were disked on 31 May, herbicide was April 2006 and 8 May 2007. from a small capacity tank with a cone applied to kinetin and herbicide-only !e effects of tillage and kinetin pre- nozzle adjusted to provide a wide spray plots on 8 June, and herbicide was treatments on reed canarygrass control pattern. ReQuest water conditioning applied to disked plots on 28 June.

294 ! September 2010 ECOLOGICAL RESTORATION 28:3

4-ER28.3 Annen (291-99).indd 294 6/18/10 5:35 AM Table 1. Summary of mean (± SEM) treatment effects for an experiment to suppress reed canarygrass (RCG) bio- mass (g/0.125 m2) and enhance biomass (g/0.125 m2) and Shannon’s Diversity (H’) of native plant in a wet prairie in the Driftless Area of Wisconsin. Means with different letters were separate at ƴ = 0.05.

Response Treatment RCG biomass (g) non-RCG biomass (g) Diversity (eH ’) Sethoxydim Control 29.78 (2.0) a 6.93 (1.5) c 2.08 (0.05) a Herbicide only 5.16 (1.6) b 46.43 (2.4) a 4.94 (0.08) a Kinetin + herbicide 1.70 (0.6) c 41.87 (1.6) a 4.73 (0.03) a Disk + herbicide 0.23 (0.3) d 33.47 (2.0) b 4.92 (0.11) a

Fluazifop Control 29.65 (1.6) a 15.37 (2.4) c 2.12 (0.05) a Herbicide only 4.88 (1.5) b 53.60 (3.3) a 3.96 (0.08) a Kinetin + herbicide 1.20 (0.8) c 39.38 (1.5) b 4.36 (0.06) a Disk + herbicide 1.17 (0.7) c 42.08 (1.8) b 2.75 (0.12) a

Disked plots were treated with her- species per 0.125 m²) and Shannon’s Results bicide later than herbicide-only and Diversity (H ') were calculated for each Compared to the untreated control, kinetin-herbicide plots so that reed main effect treatment and split plot reed canarygrass aboveground biomass canarygrass phenology was standard- treatment. Shannon’s Diversity was was lower and non–reed canarygrass ized at the 3- to 4-leaf growth stage calculated as H ' = Y p (ln p ), where i i abundance was higher in treated during herbicide application. p corresponds to the proportional i plots (Table 1). !ere was no differ- abundance of the ith species. For ence in reed canarygrass suppression clarity, H' estimates were converted Response Variables (F ₁₂ = 0.004, p = 0.957), non–reed into the same scale as species richness ( ) and Data Analysis canarygrass biomass (F ₁₂ = 1.334, with MacArthur’s N₁ (where N₁ = eH ') ( ) p = 0.367), or herbaceous species Treatment responses were measured (MacArthur 1965). diversity (F ₁₂ = 2.115, p = 0.283) due on 4–6 September 2007. Aboveg- Data were tested for normality (r² ( ) to choice of herbicide. Species rich- round biomass (hereafter, biomass) goodness-of-fit test) and homoscedas- ness was 55% greater in disked plots of herbaceous species was sampled ticity (Bartlett’s Test) (TOXSTAT vers. UW Press / Ecological Restorationtreated with sethoxydim than disked in four (0.5 × 0.25 m) rectangular 3.0, University of Wyoming, Laramie). plots treated with fluazifop (9.0 vs. quadrats/subplot. Quadrat shape and Main effect (disking + herbicide, kine- 5.8 species per sampling frame; F ₁₂ size were appropriate for this type of tin + herbicide, herbicide only, and ( ) = 25.00, p = 0.038) (see Appendix vegetation (Brummer et al. 1994). All control) and split effect (sethoxydim available at uwpress.wisc.edu/journals/ herbaceous species present within each and fluazifop) treatment means were journals/er_suppl.html). quadrat were sampled. Biomass was compared with a parametric analysis Compared to the untreated control, determined on a dry-mass basis. Plants of variance (ANOVA) for a random- reed canarygrass biomass was nearly were clipped at the plant-soil inter- ized block split-plot design (SPSS vers. five times lower in plots treated with face, separated by species, trimmed 14.0, SPSS Inc., Chicago IL). Means sethoxydim, equal to a suppressive to ca. 10-cm pieces, and then dried were separated by the degree of over- effect of 197 g/m² (Table 1). Reed to constant mass. Plant biomass was lap in the 95% confidence interval canarygrass biomass was even lower measured to the nearest 0.1 g with an for each response (Day and Quinn when either disking or kinetin pre- Acculab EC-211 analytical balance 1989). When constructing confidence treatment was coupled to sethoxydim (Acculab Sartorius Group, Edge- intervals for Shannon’s Entropy, vari- application. !e combined suppres- wood NY). Since biomass harvest is ance estimates were calculated follow- sive effect on reed canarygrass of cou- a destructive sampling technique that ing methods outlined by Magurran pling treatments was equivalent to 236 can alter treatment responses (Krebs (1988). Statistical significance was g/m² when disking was the pretreat- 1989), and since Annen (2008) found tested at _ = 0.05. Treatment effects ment and 225 g/m² when kinetin was that a two-year lag time existed to sizes were calculated as [(treatment the pretreatment (disk + sethoxydim > detect effects of similar treatments, mean − control)/treatment mean] × kinetin + sethoxydim > sethoxydim plots were sampled only at the conclu- 100. sion of the experimental treatments. only > control). Non–reed canarygrass Species richness (the number of biomass was higher in treated plots

September 2010 ECOLOGICAL RESTORATION 28:3 ! 295

4-ER28.3 Annen (291-99).indd 295 6/18/10 5:35 AM relative to the untreated control (Table herbicide plots, and highest in tilled lowering seed production capability 1). Although diversity was statistically plots. !ree species had a mean abun- and fecundity. similar among all treatments, it was dance of at least 1.0 g/quadrat in the Although species richness was great- twice as high in treated plots as in the untreated control plots, while treated est in disked plots, floristic quality was untreated control and was ecologically plots had between six and nine species lower and several common agricul- significant (Table 1). Reed canarygrass with mean abundance of at least 1.0 tural weeds were present, for example, comprised 81.2%, 10.0%, 3.9%, g/quadrat. velvet-leaf (Abutilon theophrasti), black and 0.7% of the total aboveground mustard (Brassica nigra), bull thistle biomass in the control, sethoxydim Discussion (Cirsium vulgare), Canada thistle only, kinetin + sethoxydim, and disk (C. arvense), yellow foxtail (Setaria + sethoxydim treatments, respectively. Reed canarygrass abundance was glauca), green foxtail (S. viridis), and Mean species richness per sampling consistently lower and non–reed clover (Trifolium spp.), which was frame was highest in disked plots, canarygrass abundance, species rich- likely a result of the agricultural land- similar in kinetin and sethoxydim- ness, and diversity consistently higher use history of the site and the fact only plots, and lowest in control plots in all treated plots compared to the that the wet prairie was a ten-year- (online appendix). untreated controls (Table 1). In terms old restored community. In contrast, Compared to the untreated con- of reed canarygrass suppression and Kilbride and Paveglio (1999), Paveg- trol, reed canarygrass biomass was improvements in native species abun- lio and Kilbride (2000), and Annen five times lower in plots treated with dance and diversity, fluazifop per- (2008) reported increases in floris- fluazifop, a suppressive effect of 198 formed as well as sethoxydim. Both tic quality after remnant sites were g/m² (Table 1). As with sethoxydim, herbicide formulations reduced reed tilled. !e size and composition of reed canarygrass biomass was even canarygrass aboveground biomass the native species propagule bank may lower when pretreatments were cou- by a factor of five (Table 1). Reed affect treatment outcomes. Tillage can pled to fluazifop application. !e canarygrass abundance was even lower either bury (Combroux et al. 2002) combined suppressive effect on reed in plots where either disking or kine- or expose (!ompson and Luthin canarygrass of coupling treatments tin application was coupled to herbi- 2004) seed banks, and these agricul- was equivalent to 228 g/m² when cide application, suggesting that these tural weeds were nearly absent in other either disking or kinetin pretreat- combinations of treatments were more treatment plots. !erefore, another ment was used in conjunction with conducive to desired endpoints than disadvantage of tillage is the potential fluazifop (disk + fluazifop = kinetin + solitary herbicide use, at least in the for secondary weed outbreaks, particu- fluazifop > fluazifop only > control). UWshort Press term. / Ecological Restorationlarly in restored sites. Disking adds $8 Non–reed canarygrass biomass was Use of Disking Prior to (2008 USD) per hectare to control higher in all treated plots compared costs (although the cost of using this Herbicide Application to untreated controls (Table 1). As technique varies). with sethoxydim, diversity was statisti- As reported in previous studies cally similar among all treatments, but (Kilbride and Paveglio 1999, Paveg- Use of Kinetin Pretreatments was 30% to 106% higher in treated lio and Kilbride 2000, Hovick and Coupling kinetin pretreatments to plots compared to the untreated Reinartz 2007, Annen 2008), reed graminicide application also resulted control (Table 1). Reed canarygrass canarygrass was less abundant after in lower reed canarygrass biomass comprised 65.9%, 8.3%, 2.9%, and sequencing pretreatment disking with compared to solitary herbicide appli- 2.7% of the total aboveground bio- herbicide application to augment cation. Kinetin pretreatments doubled mass in the control, sethoxydim only, herbicide performance. Non–reed reed canarygrass suppression when kinetin + sethoxydim, and disk + canarygrass biomass in disked plots sethoxydim was used as the follow- sethoxydim treatments, respectively. was intermediate between the control up treatment and tripled it when Species richness per unit area was simi- and the other treatment plots (Table 1), fluazifop was used. Coupling kinetin lar among treated plots and lowest probably a result of the restart associ- pretreatments to graminicide appli- in the untreated control (see online ated with disking. !is restart affected cation was as effective as preapplica- appendix). both reed canarygrass and established tion disking when fluazifop was used !e online appendix summarizes prairie species. !e nontarget planted but not sethoxydim (Table 1). Annen the composition of post-treatment prairie species, however, were allowed (2008) reported a 26% decrease in non-reed canarygrass vegetation. to recover from tillage by use of a selec- reed canarygrass resurgence capac- !e total number of species sampled tive graminicide on reed canarygrass ity and Harker and Taylor (1994) among all replications was lowest in regrowth. Nevertheless, repeated till- described a 60% additional decrease in the untreated control plots, interme- age could have negative consequences quackgrass biomass when 2:1 mixtures diate in herbicide only and kinetin + on native perennial species, such as of the growth regulators chlormequat

296 ! September 2010 ECOLOGICAL RESTORATION 28:3

4-ER28.3 Annen (291-99).indd 296 6/18/10 5:35 AM chloride and ethephon were used as and Zedler (2010), posttreatment veg- Conclusions and an herbicide pretreatment. In this etation composition was weedy or of Management Implications study, use of kinetin as a growth-reg- low ecological quality when gramini- ulator pretreatment suppressed reed cides were used in restored sites with 1. Reed canarygrass abundance was canarygrass up to two times more agricultural use histories. For Annen lower and non–reed canarygrass than growth regulators previously and colleagues (2005), it consisted of abundance and diversity were higher tested by Annen (2008) and up to midsuccessional native species in low in treated plots than in the untreated four times more than those tested by abundance when graminicides were control. Harker and Taylor (1994). Not only used in degraded natural areas. For 2. Fluazifop was an effective substitute were kinetin applications more effec- Annen (2008), posttreatment veg- for sethoxydim in this experiment tive, they were considerably less expen- etation consisted of a diverse mix of and may have some economic advan- sive: paired kinetin treatments added semiconservative native species when tages. However, these herbicides may only $14 per hectare to suppression graminicides were used in high-quality not be suitable for all abatements, and costs (as compared to $175 per ha for remnant sites with low initial reed it is therefore important to read and CCC/ethephon pretreatments) (2008 canarygrass abundance. Posttreat- understand the labeling of both her- prices in USD, not including labor or ment vegetation abundance depends bicide products before deciding on a additives). on site factors such as land-use history formulation to use. It is possible that additional growth- and density and composition of rem- 3. Coupling pretreatment disking and enhancing effects of kinetin may have nant vegetation stands and propagule kinetin applications with herbicide been responsible for the results mea- banks, and also on active revegetation use augmented reed canarygrass sup- sured in this field experiment. Cyto- efforts conducted in conjunction with pression with sethoxydim and fluazi- kinins are known to promote cell divi- herbicide applications. High-quality fop relative to solitary herbicide use. sion and elongation, and increasing remnant sites typically respond more 4. Secondary weed outbreaks and weed leaf surface area for herbicide contact positively to graminicide treatments shifts can occur when tillage is used to could have resulted in absorption of than restored sites with a long history augment reed canarygrass suppression greater concentrations of herbicide by of chronic disturbance and off-site in restored settings, and no-till meth- reed canarygrass. Follow-up studies impacts (pers. obs.). ods of disrupting apical dominance should consider incorporating a kine- In this experiment, increases in may be more advisable in restored tin-only treatment with leaf charac- non–reed canarygrass abundance communities where reed canarygrass teristics as response variables to deter- could have been the result of an indi- is problematic or in remnant sites mine if kinetin application increases UWrect Press competitive / Ecological release mechanism, Restoration with no prior history of tillage. leaf surface area of reed canarygrass although this experiment was not 5. Kinetin pretreatments were an effec- and associated species. designed to specifically address this tive substitute for pretreatment dis- hypothesis. In this framework, the king when fluazifop was used as the Posttreatment Species abundance of planted species initially follow-up herbicide. Composition present in the study area was sup- 6. Kinetin growth regulators were a less Several species that subdominated the pressed by resource competition with expensive, more efficacious alterna- planting prior to initiating the experi- reed canarygrass. Reed canarygrass tive to chlormequat chloride/ethe- ment were among the most abundant thatch has a mulching effect on native phon growth regulators reported by species sampled at its conclusion (see species, which further contributed to Harker and Taylor (1994) and Annen Appendix available at uwpress.wisc. reed canarygrass dominance. Com- (2008). edu/journals/journals/er_suppl.html). binations of prescribed burning (for 7. Kinetin has potential for enhancing With the exception of the weedy litter removal) followed by selective reed canarygrass suppression with ruderal species that appeared in the reed canarygrass suppression treat- graminicides and warrants further tilled plots after disking, no species ments (and treatment combinations) investigation. were sampled at the conclusion of may have opened up niche space in the the experiment that were not present vegetation matrix and enabled existing when the experiment was initiated. species to expand in response to new Acknowledgments Previous studies (Annen et al. 2005, resource opportunities. !e Swamplovers Foundation and Inte- Annen 2008, Healy and Zedler 2010) grated Restorations provided funding for documented higher post-treatment this study. Helena Chemical and Wiscon- non–reed canarygrass abundance con- sin River Agronomy provided herbicide additives and X-Cyte growth regulator. !e cordant with graminicide applications author thanks H. Lee Swanson and Gerald in mixed vegetation stands. For Healy R. Goth for support, Mike McClyman and

September 2010 ECOLOGICAL RESTORATION 28:3 ! 297

4-ER28.3 Annen (291-99).indd 297 6/18/10 5:35 AM Leo Brostowitz for technical advice, Todd Combroux, I.C.S., G. Bornetta and L. Iowa State College Journal of Science Shumate for assistance with vegetation C. Amoros. 2002. Plant regenera- 28:603–621. sampling, Paul Michler for assistance with tion strategies after a major distur- Hovick, S.M. and J.A. Reinartz. 2007. species identification, and Daniela Rödel bance: !e case of a riverine wetland Restoring forest in wetlands domi- for assistance with figures. Dr. !omas D. restoration. Wetlands 22:234–246. nated by reed canarygrass: !e effects Brock and Art Kitchen offered comments on Curtis, J.T. 1959. !e Vegetation of of pre-planting treatments on early earlier drafts of this manuscript, and I also Wisconsin. Madison: University of survival of planted stock. Wetlands thank three anonymous reviewers for their Wisconsin Press. 27:1–26. input, which greatly improved the clarity of Day, R.W. and G.P. Quinn. 1989. Com- Kercher, S.M. and J.B. Zedler. 2004. Mul- this manuscript. parisons of treatments after an anal- tiple disturbances accelerate invasion ysis of variance in ecology. Ecological of reed canarygrass (Phalaris arundina- Monographs 59:433–463. cea L.) in a mesocosm study. Oecologia References Devine, M.D. 1997. Mechanisms of resis- 138:455–464. Annen, C.A. 2004. Will tillage and tance to acetyl-coenzyme A carbox- Kilbride, K.M. and F.L. Paveglio. 1999. plant growth regulator pretreat- ylase inhibitors: A review. Pesticide Integrated pest management to con- ments enhance herbicide effects on Science 51:259–264. trol reed canarygrass in seasonal wet- reed canarygrass? Proceedings of the Galatowitsch, S.M., D.C. Whited, lands of southwestern Washington. North American Prairie Conference R. Lehtinen, J. Husveth and K. Schik. Wildlife Society Bulletin 27:292–297. 19:171–177. 2000. !e vegetation of wet mead- Krebs, C.J. 1989. Ecological Methodology. ___. 2006. Using sethoxydim to control ows in relation to their land-use. Envi- New York: Harper-Collins. reed canarygrass: Observed effects of ronmental Monitoring and Assessment Leakey, R.B., R.J. Chancellor and additives and field conditions (Wis- 60:121–144. D. Vince-Prue. 1975. Parental factors consin). Ecological Restoration 24:53. Gleason, H.A. and A. Cronquist. 1991. in dominance of lateral buds on rhi- ___. 2008. Effects of tillage and growth Manual of Vascular Plants of Northeast- zomes of Agropyron repens (L.) Beauv. regulator pretreatments on reed ern United States and Adjacent Canada. Planta 123:267–274. canarygrass (Phalaris arundinacea L.) New York: New York Botanical Lesica, P. and B. Martin. 2004. Study control with sethoxydim. Natural Garden. evaluates methods for control- Areas Journal 28:6–13. Green, E.K. and S.M. Galatowitsch. 2001. ling reed canarygrass in sedge mead- Annen, C.A., E.M. Kirsch and R.W. Tyser. Differences in wetland plant com- ows (Montana). Ecological Restoration 2008. Reed canarygrass invasions alter munity establishment with addi- 22:227–228. succession patterns and may reduce tions of nitrate-N and invasive spe- MacArthur, R.H. 1965. Patterns of spe- habitat quality in wet meadows. cies (Phalaris arundinacea and Typha cies diversity. Biological Review Ecological Restoration 26:190–193. × glauca). Canadian Journal of Botany 40:510–533. Annen, C.A., R.W. Tyser and E.M. 79:170–178. Magurran, A.E. 1988. Ecological Diver- Kirsch. 2005. Effects of a selec- Harker, K.N. and J. Dekker. 1988. Effects sity and Its Measurement. Princeton: tive herbicide, sethoxydim, on reed UWof Press phenology / Ecological on translocation Restoration pat- Princeton University Press. canarygrass. Ecological Restoration terns of several herbicides in quack- Marquis, L.Y., R.D. Comes and C.P. 23:99–102. grass, Agropyron repens. Weed Science Yang. 1979. Selectivity of glypho- Apfelbaum, S.I. and C.E. Sams. 36:463–472. sate in creeping red fescue and 1987. Ecology and control of reed Harker, K.N. and J.S. Taylor. 1994. reed canarygrass. Weed Research canarygrass (Phalaris arundinacea L.). Chlormequat chloride (CCC) pre- 19:335–342. Natural Areas Journal 7:9–17. treatments may enhance quack- Matysiak, R. and J.D. Nalewaja. 1999. Bernthal, T.W. and B.K. Hatch. 2008. grass (Elytrigia repens) control Temperature, adjuvants, and UV light Mapping Wisconsin wetlands with sethoxydim. Weed Technology affect sethoxydim phytotoxicity. Weed dominated by reed canary grass 8:499–507. Technology 13:94–99. (Phalaris arundinacea L.): A land- Harker, K.N. and W.H. Vanden Born. Maurer, D.A, R. Lindig-Cisneros, K.J. scape-level assessment. Wisconsin 1997. Glyphosate or sethoxydim for Werner, S. Kercher, R. Miller and J.B. Department of Natural Resources quackgrass (Elytrigia repens) control Zedler. 2003. !e replacement of wet- PUB-WT-900–2008. in two tillage regimes. Weed Science land vegetation by reed canarygrass Bonilla-Warford, C.M. and J.B. Zedler. 45:812–823. (Phalaris arundinacea). Ecological 2002. Potential for using native Healy, M.T. and J.B. Zedler. 2010. Set- Restoration 21:116–119. plant species in stormwater wet- backs in replacing Phalaris arundina- McIntyre, G.I. 1971. Apical dominance in lands. Environmental Management cea monotypes with sedge meadow the rhizome of Agropyron repens: Some 29:385–394. vegetation. Restoration Ecology factors affecting the degree of domi- Brummer, J.E., J.T. Nichols, R.K. Engel 18:155–164. nance in isolated rhizomes. Canadian and K.M. Eskridge. 1994. Efficiency Hillman, J.R. 1985. Apical dominance. Journal of Botany 49:99–109. of different quadrat sizes and shapes Pages 127–148 in M.B. Wilkins (ed), Miller, R.B. and J.B. Zedler. 2003. for sampling standing crop. Journal of Advanced Plant Physiology. London: Responses of native and invasive wet- Range Management 47:84–89. Pitman Publishing. land plants to hydroperiod and water Cline, M.G. 1991. Apical dominance. Holt, I.V. 1954. Initiation and develop- depth. Plant Ecology 137:57–69. Botanical Review 57:318–358. ment of the inflorescences of Phalaris Paveglio, F.L. and K.M. Kilbride. 2000. arundinacea L. and Dactylis glomerata Response of vegetation to control of

298 ! September 2010 ECOLOGICAL RESTORATION 28:3

4-ER28.3 Annen (291-99).indd 298 6/18/10 5:35 AM reed canarygrass in seasonally man- Sachs, T. 2002. Developmental processes Wilcox, J.C. 2004. Challenges of replacing aged wetlands of southwestern Wash- and the evolution of plant clonality. reed canary grass (Phalaris arundina- ington. Wildlife Society Bulletin Evolutionary Ecology 15:485–500. cea L.) with native species. MS thesis, 28:730–740. Sachs, T. and K.V. !imann. 1967. !e University of Wisconsin–Madison. Preuninger, J.S. and C.E. Umbanhowar Jr. role of auxins and cytokinins in the Wilcox, J.C., M.T. Healy and J.B. Zedler. 1994. Effects of burning, cutting, and release of buds from apical domi- 2007. Restoring native vegetation to spraying on reed canary grass studied nance. American Journal of Botany an urban wet meadow dominated by (Minnesota). Restoration & Manage- 45:136–144. reed canarygrass (Phalaris arundina- ment Notes 12:207. Shoaf, A.R. and W.C. Carlson. 1992. Sta- cea L.) in Wisconsin. Natural Areas Rachich, J. and R.J. Reader. 1999. Inter- bility of sethoxydim and its degrada- Journal 27:354–365. active effects of herbivory and com- tion products in solution, in soil, and petition on blue vervain (Verbena on surfaces. Weed Science 40:384–389. hastata L.: Verbenaceae). Wetlands !ompson, A. and C. Luthin. 2004. Wet- 19:156–161. land Restoration Handbook for Wis- Reinhardt, C. and S.M. Galatowitsch. consin Landowners, 2nd ed. Madi- Craig A. Annen is Operations Manager 2004. Best management practices for son: Wisconsin Department of Nat- and Director of Research for the firm Inte- the invasive Phalaris arundinacea L. ural Resources, Bureau of Integrated grated Restorations LLC and is a member (reed canary grass) in wetland resto- Science Services. of the USDA-NRCS Reed Canarygrass rations. Minnesota Department of Werner, K.J. and J.B. Zedler. 2002. How Working Group Committee. His research Transportation Report No. 2004–36. sedge meadow soils, microtopog- interests include integrated management Reyes, C.M. 2004. !e feasibility of raphy, and vegetation respond to strategies for invasive species, conservation using prescribed burning to con- sedimentation. Wetlands 22:451–466. ecology of threatened and endangered spe- trol reed canary grass (Phalaris arun- Wetzel, P.R. and A.G. van der Valk. 1998. cies, gradient analysis, and mathematical dinacea L.) populations in Wiscon- Effects of nutrient and soil moisture ecology. He can be reached at Integrated sin wetlands. MS thesis, University of on competition between Carex stricta, Restorations LLC, 228 South Park St, Wisconsin–Madison. Phalaris arundinacea, and Typha Belleville, WI 53508, 608/424-6997, latifolia. Plant Ecology 138:179–190. [email protected].

UW Press / Ecological Restoration

September 2010 ECOLOGICAL RESTORATION 28:3 ! 299

4-ER28.3 Annen (291-99).indd 299 6/18/10 5:35 AM