Application skip followlng ground application of clopyralid in Aerial treatment of clopyralid on right and untreated plot on left. yellow starthistle infested field. New growth regulator herbicide provides excellent control of yellow starthistle

Joseph M. DiTomaso P Guy B. Kyser o Steve B. Orloff P Stephen F. Enloe o Glenn A. Nader

Yellow starthistle is the most through April, but treatments in tebuthiuron, oxyfluorfen and pro- widespread invasive weed in Cali- February maximized desirable metone. All these compounds are reg- fornia. An exotic that invaded forage production, particularly istered for use on rights-of-way or in- California in the Gold Rush days, grasses. Late-season applications dustrial sites, but few can be used in starthistle was once a minor an- of glyphosate and clopyralid were rangeland, pastures or wildlands. In noyance but is now out of control. effective for control of starthistle rangeland, no preemergence herbicide One reason is that few herbicides plants in late rosette and bolting provides selective control of yellow are registered for use on Califor- stages. starthistle without injuring desirable grasses. Chlorsulfuron is the only se- nia rangelands, pastures and wild- Yellow starthistle has become com- lective preemergence herbicide cur- , lands. Of the registered herbi- mon in open areas on roadsides, rently available that effectively con- cides, the majority are effective rangeland, wildlands, hay fields, pas- trols yellow starthistle in noncrop only when applied to the foliage of tures and waste areas. It can form areas, but it does not have post- target plants; season-long control dense infestations that rapidly deplete emergence activity on yellow star- of yellow starthistle is not pro- soil moisture, thus preventing the es- thistle and must be used in combina- vided because seedlings that tablishment of other species. Effective tion with 2,4-D, dicamba or triclopyr emerge after application of the long-term management of yellow to provide some level of control. herbicide escape injury. In a com- starthistle will likely be achieved with A major problem with postemer- parison of herbicides, the newly the use of an integrated approach. This gence herbicides to control yellow registered herbicide clopyralid could include the use of biological starthistle (Centaureu solstitidis L.) has provided excellent control at low control agents, mowing, prescribed been the weed's ability to germinate use rates and worked equally well burning, properly timed grazing, re- continuously throughout winter, whether applied to leaves or to seeding programs and the judicious spring and into summer whenever soil to control germinating seed- use of herbicides. moisture is available. As a result, re- lings. In soil, it showed residual Yellow starthistle can be controlled peated applications of broadleaf selec- activity throughout the season. with a number of nonselective pre- tive, postemergence herbicides such as Complete yellow starthistle con- emergence herbicides, including si- 2,4-D, dicamba or triclopyr are re- trol was achieved with applica- mazine, diuron, atrazine, sulfometuron- quired. One late application at the end tions made from December methyl, chlorsulfuron,bromacil, of the rainy season is not sufficient be- cause many plants are too large and evaluations were made at least 2 production in yellow starthistle infes- are not completely controlled. months following treatment. tations, this level of control is not con- Clopyralid (Transline) is a growth- Chlorsulfuron typically provides sidered acceptable for long-term regulator herbicide that arrests devel- excellent postemergence and preemer- starthistle management. Application opment of the growing points of the gence control of many weed species. of 2,4-D in March followed by a sec- plants. It was recently registered for However, in our first 1996 trial it gave ond treatment in May controlled 96% use in noncrop areas of California, in- little postemergence control of yellow of the yellow starthistle. Although not cluding pastures, rangeland and wild- starthistle, even at the highest rate, 2 statistically significant, 2,4-D was con- lands. It is not expected to impact the ounces active ingredient per acre (oz sistently more effective than triclopyr. biological control agents and has been ai/A) with or without a surfactant Clopyralid at the lowest rate tested in demonstrated to be very safe on (table 1).When chlorsulfuron was this experiment (1 oz acid equivalent/ grasses but effective for the control of combined with 2,4-D or triclopyr, yel- acre, 1/6 pint product/acre) was not many members of the sunflower fam- low starthistle control improved to as statistically different from repeated ily (Asteraceae), as well as legumes much as 91%. Because of the high seed applications of triclopyr or 2,4-D, but (Fabaceae), nightshades (Solanaceae), and some species in the knotweed (Polygonaceae) and carrot (Apiaceae) families. In contrast, many other broadleaf species, including those in the mustard family (Brassicaceae)and filarees (Erodium spp.), are tolerant to the herbicide. Unlike other postemer- gence compounds, clopyralid also has excellent preemergence activity and is effective at very low use rates. In these experiments, we compared rates and timing of several noncrop herbicides for yellow starthistle control, from be- fore emergence through reproductive maturity. Starthistle seedling control We established field trials in Davis in March 1996 to compare several postemergence herbicides and the preemergence herbicide chlorsulfuron on yellow starthistle seedlings. Each plot was 5 feet by 10 feet and all treatments were applied using a COz-pressurized backpack sprayer delivering 40 gallons per acre of spray mixture at a pressure of 20 pounds per square inch (psi). A similar experiment was also estab- lished in Davis in November 1996 to determine the most effective applica- tion timing (November to April) for control of yellow starthistle seedlings and rosettes using clopyralid, 2,4-D, triclopyr, and combinations of chlor- sulfuron and 2,4-D or triclopyr. In this experiment, plots were 8 feet by 15 feet and treatments were applied us- ing a COz-pressurized backpack sprayer delivering 20 gallons per acre at 30 psi. All treatments in both experi- ments were replicated four times. Vi- sual evaluations were used to assess percent control. Because clopyralid acts slowly on starthistle seedlings, County, however, we also considered May. Both control and forage quantity the possible effect of the previous were determined in October 1997, and year’s starthistle skeletal remains on the following spring 1998, by collect- the efficacy of clopyralid. To test this, ing all aboveground plant biomass in we established 10-feet-by-20-feetplots three 2.7 ft2 (0.25 m’) quadrat samples along a roadside heavily infested with per plot (four replicates per treatment) yellow starthistle. The skeletons were and separating uncontrolled living mowed and removed in half of each starthistle, dead starthistle (1997 only) plot, and the entire plot was treated and other forage species, primarily an- with the appropriate rate of clopyralid nual grasses. Data are reported as dry as described in the previous experi- weight. ment. Each trial was established as a Poor control was obtained with 0.5 randomized, complete block experi- oz ae/acre clopyralid at all treatment Comparison of March (/eft) and May (right) clopyralid treatments on grass ment with four replicates. times. In contrast, complete control of forage production in Siskiyou County. In Calaveras County, the highest yellow starthistle was achieved in Feb- Both treatments resulted in complete rate tested, 2 oz ae/acre, was required ruary with clopyralid rates of 1 oz ae/ control of yellow starthistle. for acceptable control (table 3). We acre and higher (fig. la). Yellow demonstrated similar results in Butte starthistle control was insufficient at consistently provided complete con- County. The control of yellow 1 oz ae/acre at the late application trol of yellow starthistle for the entire starthistle was better in the presence of dates, as the rosettes were larger at the season. skeletons compared to the areas where time of application. Both 2 and 4 oz In a subsequent experiment, we skeletons were removed. With skeletal ae/acre clopyralid provided near com- tested clopyralid and other herbicides remains present, both 0.5 and 1 oz ae/ plete control at all treatment timings, for yellow starthistle control every acre clopyralid provided better than including the initial bolting stage in month from November 1996 through 90% starthistle control. The difference May. April 1997 (table 2). In this experi- between the mowed and unmowed Forage quantity progressively de- ment, we used clopyralid at rates as plots was attributed to the reduced creased with later treatment times (fig. low as 0.25 oz ae/acre (1/24 pt prod- number of seedlings present in the lb). Desirable forage biomass was uct/acre), which is well below the area shaded by the previous year’s maximized with a February treatment lowest registered rate of 1.5 oz ae/acre skeletons. Consequently, fewer seed- at 4 oz ae/acre clopyralid, which rep- (1/4 pt product/acre). Clopyralid lings were available to escape herbi- resents the highest labeled Transline rates greater or equal to 1 oz ae/acre cide injury in the unmowed plots. rate allowed in California (2/3 pt of were necessary for acceptable yellow product/acre). At this timing, yellow starthistle control with applications Timing for maximum forage starthistle was in the early rosette made from November through April. We also established a trial in the stage. Reduced forage at later treat- Although not always statistically sig- intermountain region, near Yreka ment times is likely due to the com- nificant, control of yellow starthistle (Siskiyou County) to determine the ef- petitive effects of yellow starthistle on with 1 to 6 oz ae/acre clopyralid was fect of clopyralid rate and treatment grass development during the early consistently superior to control with timing on yellow starthistle control spring months. At higher rates, yellow 2,4-D, triclopyr, and the combination and forage production during fall and starthistle injury occurred more rap- of these herbicides with chlorsulfuron, the following spring. Plots were 10 idly and control was more complete, in January, February and March. feet by 30 feet and spray equipment increasing light penetration and water These results indicate that clopy- was similar to that previously de- availability to developing grasses and ralid provides excellent control of yel- scribed. Treatments included 0.5, 1,2 other desirable forage. low starthistle seedlings and rosettes and 4 oz ae/acre clopyralid applied in The following spring, yellow at between 1 and 6 oz ae/acre. Fur- every month from February through starthistle biomass tended to be lower thermore, there is a wide window for application timing, extending from December through April. To test the effectiveness of clopyralid on yellow starthistle in other Northern California regions, we conducted trials in April 1997 in Calaveras County, 5 miles north of San Andreas, and in Butte County, 7 miles north of Oroville. The Calaveras County plots were 10 feet by 10 feet and were treated identically to those in the Davis experiment. In Butte with earlier application timing (fig. 2a). Al- though not statistically significant, higher rates led to consistently less yellow starthistle com- pared with lower rates. However, this response did not appear to be due entirely to residual her- bicide activity, as later application timings were not significantly differ- ent from early-season treatments. Early-season treatments also tended to increase desirable for- age production com- pared with untreated plots and late-season treatments (fig. 2b). Fur- Fig. 1. Effect of clopyralid rate and timing on sum- Fig. 2. Effect of clopyralid rate and timing on mer (A) biomass of yellow starthistle and (B) desir- spring (A) biomass of yellow starthistle and (B) thermore’ in Some able forage in Siskiyou County, 1997. Desirable for- desirable forage in Siskiyou County, 1998. Desir- higher rates also in- age consisted of annual and perennial grasses, and able forage consisted of annual and perennial creased desirable forage forbs other than yellow starthistle. grasses, and forbs other than yellow starthistle. production compared- with lower rates. The early-season treatment at higher la- pot, postemergence applications control of yellow starthistle. Several beled rates not only provides excellent were made to plants after two ”true rates of glyphosate, triclopyr, control in the year of treatment, but leaves” developed. Each treatment clopyralid, 2,4-D and dicamba were may also give both increased yellow was replicated four times. Visual in- applied to starthistle in the following starthistle control and desirable jury evaluations were used for rela- late rosette stages: just prior to bolting spring forage biomass 1 year after tive comparisons between pre- and (1997 only), bolting, spiny (bud) and application. postemergence applications. early flowering (2% to 5% flower). Although clopyralid can effectively Our results indicate that the V150 Plants were treated as previously de- control yellow starthistle from Decem- (rate which caused a level of visible scribed with a minimum of 10 repli- ber through May, the optimum timing injury of 50°/0 relative to untreated cates (individual plants) per treatment for treatment appears to be between seedlings) for pre- and postemer- in 1995 and four replicates (plots con- January and March in most locations gence (V150= 0.140 oz ae/acre and sisting of six plants each) per treat- in California. This is particularly true 0.113 oz ae/acre, respectively, with- ment in 1997. All treatments, except if the primary object of a starthistle out surfactant) applications were glyphosate, were applied with the ad- management program is to enhance similar (fig. 3). This is supported by dition of 0.1% Sylgard as a surfactant. forage quality and quantity in range- the wide timing window for control Because late-season control typically land and grassland environments. of starthistle seedlings under field conditions. Furthermore, the addi- Foliar and soil sensitivity tion of an organosilicone surfactant 100 4 / We further compared the dose re- (0.1% Sylgard) seemed to provide sponse of yellow starthistle seedlings somewhat lower levels of yellow under greenhouse conditions to sev- starthistle seedling control compared VIw = 0.140 oz aelA eral pre- and postemergence rates of with treatments without surfactant. f 40 0 Postemergence VIm = 0.113 or aelA clopyralid. Postemergence treatments Two to three times more herbicide A Postemergence + suriactant were conducted with and without the (V150= 0.342 oz ae/acre) was re- L< Vlm = 0.342 oz aelA addition of a surfactant. Ten to 20 quired for the same level of injury 0 1, I,,, I seeds were planted in 4-inch pots. when an organosilicone surfactant 0.1 1 Rate of clopyralld, oz adA Preemergence treatments were made was added. (applied in series dilutions from 1116 to 2 02) immediately after sowing using a compressed-air, table-spray chamber Late-season starthistle control Fig. 3. Effect of application technique on delivering 17.7 gallons per acre at 30 We conducted experiments in clopyralid control of greenhouse-grown yellow starthistle seedlings. VI5o = rate psi through an 8002E nozzle. After 1995 and 1997 to determine the most which caused 50% visible injury to seed- thinning to three to five seedlings per effective herbicide for late-season lings in comparison to untreated plants. occurs during the warmer spring or with glyphosate is after annual grasses desirable forage production in the year early summer months, amine formula- or broadleaf species have completed of treatment, as well as the following tions of both 2,4-D and triclopyr were their life cycle, but prior to yellow spring. Clopyralid, glyphosate and used. starthistle seed production (<5%flow- dicamba also give effective late-season In 1995, we compared two rates of ering). Control is less effective when control of yellow starthistle, provided glyphosate and triclopyr at three mature plants show physical signs of applications are not made when plants stages of yellow starthistle develop- drought-stress. When clopyralid was are drought-stressed. ment. Glyphosate at both rates pro- previously applied in late winter or It is important to recognize the pos- vided excellent control at all stages early spring, glyphosate can be used in sibility that other undesirable species, (table 4). In contrast, triclopyr pro- a broadcast or spot treatment follow- particularly annual grasses such as vided only marginal control in the up program to kill uncontrolled plants medusahead (Taeniatkerum caput- bolting stage and poor control at later before they produce seed. It can also medusae [L.] Nevski) or barbed goat- growth stages. In 1997, control of late- be used to prevent the proliferation of grass (Aegilops triuncialis L.), may re- rosette plants was excellent with all potential clopyralid-resistant plants. place yellow starthistle in clopyralid- herbicides tested except for the lowest Broadcast treatment with glyphosate treated areas. Under these conditions, rate of glyphosate or triclopyr (table is not recommended when desirable other management strategies may be 5). Glyphosate, dicamba, and higher perennial grasses or broadleaf species necessary. In addition, continuous rates of clopyralid and triclopyr con- are present. clopyralid use over many years may tinued to be effective when plants have a long-term detrimental effect on were in the bolting stage. At the spiny Advantages and warnings legume populations, and potentially and early flowering stages, yellow There are several advantages to us- soil nitrogen levels. Consequently, starthistle plants were noticeably ing clopyralid for yellow starthistle other control options should be ro- drought-stressed. Control was poor control over other postemergence her- tated in the overall yellow starthistle for all herbicides tested, regardless of bicides registered for use in rangeland, management program. Also, the po- the application rate. Differences be- pastures and wildland areas. It has tential exists for the evolution of resis- tween the results for glyphosate in both pre- and postemergence activity tance to clopyralid if the herbicide is 1995 and 1997 could be attributed to at very low use rates, a low toxicology used year after year without employ- the lack of spring rainfall in 1997, com- profile and no grazing restrictions. ing other methods. The potential for pared with 1995. Treatments in December or later gave resistance can be minimized by using If late-season yellow starthistle con- season-long control of starthistle, with other strategies or through late-season trol is necessary, the best time to treat earlier applications resulting in greater applications of glyphosate to control escapes.

J.M.DiTomaso is Non-Crop Extension Weed Ecologist, G.B. Kyser is Staff Re- search Associate, and S.F. Enloe is Gradu- ate Student, Department of Vegetable Crops, UC Davis; S.B. Orloff is Farm Ad- visor, Siskiyou County; and G.A. Nader is Farm Advisor, Sutter-Yuba-Butte Counties. This project was partiallyfunded by Dow AgroScience.

Further reading Carrithers VF. 1998. Transline herbicide in rangeland and wildlife habitat restoration Pro- ceedings, Calif. Weed Sci SOC.50:138-41. DiTomaso JM, Lanini WT, Thomsen CD, Prather TS, et al. 1999. Pest Notes: Yellow starthistle. UC Division of Ag and Nat Re- sources, No. 7402. 4 p. DiTomaso JM, Orloff SB, Kyser GB and Nader GA. 1998. Influence of timing and chemical control on yellow starthistle. Pro- ceedings, Calif. Weed Sci SOC.50:190-3. Lass LW and Callihan RH. 1994. Herbi- cide evaluation for yellow starthistle control. Res Prog Rept. Westem Soc Weed Sci. p 142. Northam FE and Callihan RH. 1991. Ef- fects of herbicides on yellow starthistle den- sity and vegetative biomass. Res Prog Rept. Western SOCWeed Sci. p 43-6.