Plant reproduction is altered by simulated herbicide drift to constructed plant communities.

Herbicide drift may have unintended impacts on native vegetation, adversely affecting individual species and plant communities. To determine the potential ecological effects of herbicide drift, small plant community plots were constructed using 9 perennial species found in different Willamette Valley (OR, USA) grassland habitats. Studies were conducted at 2 Oregon State University (Corvallis, OR, USA) farms in 2 separate years, with single and combined treatments of 0.01 to 0.2× field application rates (f.a.r.) of 1119 g ha-1 for glyphosate (active ingredient [a.i.] of 830 g ha-1 acid glyphosate) and 560 g ha-1 a.i. for dicamba. Plant responses were percentage of cover, number of reproductive structures, mature and immature seed production, and vegetative biomass. Herbicide effects differed with species, year, and, to a lesser extent, farm. Generally, 0.1 to 0.2× f.a.r. of the herbicides were required to affect reproduction in Camassia leichtlinii, Elymus glaucus, Eriophyllum lanatum, Festuca idahoensis, Iris tenax, and Prunella vulgaris. Eriophyllum lanatum also had a significant increase in percentage of immature seed dry weight with 0.01× f.a.r. of dicamba or the combination of glyphosate plus dicamba. Other species showed similar trends, but fewer significant responses. These studies indicated potential effects of low levels of herbicides on reproduction of native plants, and demonstrated a protocol whereby species growing in a constructed plant community can be evaluated for ecological responses. Environ Toxicol Chem 2017;36:2799-2813. Published 2017 SETAC. This article is a US government work and, as such, is in the public domain in the United States of America.

Glyphosate and dicamba herbicide tank mixture effects on native plant and non-genetically engineered soybean seedlings.

Crops engineered to contain genes for tolerance to multiple herbicides may be treated with several herbicides to manage weeds resistant to each herbicide. Thus, nearby non-target plants may be subjected to increased exposure to several herbicides used in combination. Of particular concern are native plants, as well as adjacent crops which have not been genetically engineered for tolerance to herbicides. We evaluated responses of seven species of native plants grown in a greenhouse and treated less than field application rates of glyphosate and/or dicamba: Andropogon gerardii, Asclepias syriaca, Eutrochium purpureum, Oenothera biennis, Polyganum lapathifolium, Solidago canadensis and Tridens flavus, and non-herbicide resistant soybean (Glycine max, Oregon line M4). Herbicide concentrations were 0.03 or 0.1 × field application rates of 1122 g ha(-1) active ingredient (a.i) (831 g ha(-1) acid glyphosate) for glyphosate and 562 g ha(-1) a.i. for dicamba. In general, plant growth responses to combinations of glyphosate and dicamba were less than the sum of growth responses to the individual herbicides (i.e., antagonistic effect), primarily when one or both herbicides alone caused a large reduction in growth. E. purpureum, P. lapathifolium and S. canadensis were the most sensitive species to both herbicides, while A. gerardii was the most tolerant, with no response to either herbicide. The combinations of herbicides resulted in responses most similar to that from dicamba alone for G. max and from glyphosate alone for T. flavus. The results of this study indicated the need for more data such as effects on native plants in the field to assess risks to non-target plants from combinations of herbicides.

Effects of single and multiple applications of glyphosate or aminopyralid on simple constructed plant communities.

To determine effects of multiple applications of herbicides on small constructed plant communities, Prunella vulgaris L.var. lanceolata Fern, Festuca roemeri (Pavlick) Alexeev, Clarkia amoena (Lehm.) Nels., and Cynosurus echinatus L. were grown together in small field plots. Plants were treated with glyphosate at target concentrations of 0 × , 0.01 × , 0.1 × , and 0.2× a field application rate (FAR) of 1122 g ha(-1) active ingredient (a.i.) for 3 yr in 1 location, and for 2 yr in a second location. Plants also were treated with aminopyralid at 0 × , 0.037 × , 0.136 × , and 0.5× FAR of 123 g ha(-1) a.i. for 2 yr in 2 locations. Plants received 1, 2, or 3 applications of each herbicide each year. Species and community responses depended on herbicide concentration and number of applications. With glyphosate, plant volume (modified formula for a cone) tended to decrease for all species (especially C. echinatus), and the decreases generally became larger with more applications. Plant communities exposed to the 2 greatest concentrations initially differed from controls but then appeared to recover. With aminopyralid, C. amoena was essentially eliminated from the communities, especially at the 2 greatest FARs, whereas the other 3 species tended to have significant increases in volume, especially at the 2 smallest FARs. With aminopyralid, increasing numbers of applications produced variable results, and the plant community volume never tended to recover.

Effects of low levels of herbicides on prairie species of the Willamette Valley, Oregon.

The relative sensitivity of 17 noncrop plant species from Oregon's Willamette Valley was determined in response to glyphosate, tribenuron methyl (tribenuron), and fluazifop-p-butyl (fluazifop) herbicides. For glyphosate, Elymus trachycaulus, Festuca arundinacea, Madia elegans, Potentilla gracilis, and Ranunculus occidentalis were the most sensitive species, based on a concentration calculated to reduce shoot dry weight by 25% (IC25 values) of 0.02 to 0.04 × a field application rate of 1112 g active ingredient (a.i.) per hectare. Clarkia amoena and Lupinus albicaulis were the most tolerant to glyphosate, with IC25 values near the field application rate. Clarkia amoena, Prunella vulgaris, and R. occidentalis were the most sensitive to tribenuron, with IC25 values of 0.001 to 0.004 × a field application rate of 8.7 g a.i. ha(-1) for shoot dry weight. Five grass species were tolerant to tribenuron with no significant IC25 values. For fluazifop, 2 native grasses, E. trachycaulus and Danthonia californica, were the most sensitive species, with IC25 values of 0.007 and 0.010 × a field application rate of 210 g a.i. ha(-1) , respectively, for shoot dry weight, while a native grass, Festuca roemeri, and nearly all forbs showed little or no response. These results also indicated that the 3 introduced species used in the present study may be controlled with 1 of the tested herbicides: glyphosate (F. arundinacea), tribenuron (Leucanthemum vulgare), and fluazifop (Cynosurus echinatus).

The effects of glyphosate and aminopyralid on a multi-species plant field trial.

In the United States, the US EPA has the responsibility for the registration of pesticides. For the protection of nontarget terrestrial plants this requires two simple greenhouse tests (seedling emergence and vegetative vigor), each done with ten species grown individually. Indications of unacceptable effects levels equivalent to environmental exposure can lead to field testing which is not well-defined. Our objective was to develop a regional field test that is simple, economical, geographically flexible and with endpoints of ecological significance and compare the results with the standard greenhouse tests. Three native Oregon plant species were grown together with an introduced species. The experiment was replicated at two locations and repeated for 3 years with glyphosate applied at 0, 0.01 (8.3 g/ha), 0.1 (83.2 g/ha), and 0.2 (166.4 g/ha) × FAR (Field Application Rate of 832 gm/ha acid equivalent) and 2 years with aminopyralid applied at 0, 0.037 (4.6 g/ha), 0.136 (16.7 g/ha), and 0.5 (61.5 g/ha) × FAR (123 g/ha acid equivalent). With glyphosate, plant height and volume decreased with increasing herbicide concentration for all species, and for nearly all farm × year combinations. With aminopyralid, one species died at nearly all concentrations, sites and years, while the effects on the other three species were less pronounced and variable. The relative rank in glyphosate sensitivity among species in the field studies differed from the ranking from greenhouse studies, with Cynososurs echinatus the most sensitive in the field but Prunella vulgaris the most sensitive in the greenhouse. With aminopyralid, sensitivity generally was similar for all species in the greenhouse as in the field. The results suggest that a simple field test can be successfully designed to investigate the ecological effects of herbicides on plant communities and supplement information gained from greenhouse tests performed in controlled environments.

Comparing effects of low levels of herbicides on greenhouse- and field-grown potatoes (Solanum tuberosum L.), soybeans (Glycine max L.), and peas (Pisum sativum L.).

Although laboratory toxicology tests are generally easy to perform, cost effective, and readily interpreted, they have been questioned for their environmental relevance. In contrast, field tests are considered realistic while producing results that are difficult to interpret and expensive to obtain. Toxicology tests were conducted on potatoes, peas, and soybeans grown in a native soil in pots in the greenhouse and were compared to plants grown outside under natural environmental conditions to determine toxicological differences between environments, whether different plant developmental stages were more sensitive to herbicides, and whether these species were good candidates for plant reproductive tests. The reproductive and vegetative endpoints of the greenhouse plants and field-grown plants were also compared. The herbicides bromoxynil, glyphosate, MCPA ([4-chloro-2-methylphenoxy] acetic acid), and sulfometuron-methyl were applied at below field application rates to potato plants at two developmental stages. Peas and soybeans were exposed to sulfometuron-methyl at similar rates at three developmental stages. The effective herbicide concentrations producing a 25% reduction in a given measure differed between experimental conditions but were generally within a single order of magnitude within a species, even though there were differences in plant morphology. This study demonstrated that potatoes, peas, and soybeans grown in pots in a greenhouse produce phytotoxicity results similar to those grown outside in pots; that reproductive endpoints in many cases were more sensitive than vegetative ones; and that potato and pea plants are reasonable candidates for asexual and sexual reproductive phytotoxicity tests, respectively. Plants grown in pots in a greenhouse and outside varied little in toxicity. However, extrapolating those toxicity results to native plant communities in the field is basically unknown and in need of research.

Potato (Solanum tuberosum) greenhouse tuber production as an assay for asexual reproduction effects from herbicides.

The present study determined whether young potato plants can be used as an assay to indicate potential effects of pesticides on asexual reproduction. Solanum tuberosum (Russet Burbank) plants were grown from seed pieces in a mineral soil in pots under greenhouse conditions. Plants were treated with herbicides (cloransulam, dicamba, glyphosate, imazapyr, primsulfuron, sulfometuron, or tribenuron) at simulated drift levels [

Phytotoxicity assay for seed production using Brassica rapa L.

Although pesticide drift can affect crop yield adversely, current plant testing protocols emphasize only the potential impacts on vegetative plant growth. The present study was conducted to determine whether a plant species with a short life cycle, such as Brassica rapa L. Wisconsin Fast Plants®, can be used to indicate potential effects on seed production of herbicides applied at relatively low levels (e.g., low field application rates [FAR]). The effects of ≤0.1 × FAR of aminopyralid, cloransulam, glyphosate, primisulfuron, or sulfometuron applied 14 d after emergence (DAE), were evaluated for B. rapa grown in mineral soil in pots under greenhouse conditions. Effects were expressed as the effective concentration of the herbicide producing a 25% reduction in a response (EC25) based on nonlinear regression. Brassica rapa seed dry weight was reduced by sulfometuron at an EC25 of 0.00014 × a field application rate (FAR) of 53 g active ingredient (a.i.) ha(-1), primisulfuron at 0.008 (experiment 1) or 0.0050 (experiment 2) × FAR of 40 g a.i. ha(-1), cloransulam at 0.022 × FAR of 18 g a.i. ha(-1), glyphosate at 0.0399 × FAR of 834 g a.i. ha(-1), and by aminopyralid at 0.005 × FAR of 123 g a.i. ha(-1), but only for 1 of 2 experiments. Reduced seed production occurred at less than the FAR that reduced shoot dry weight with sulfometuron and primisulfuron, whereas neither aminopyralid, cloransulam, nor glyphosate affected shoot dry weight. A short life cycle form of B. rapa could be used to indicate reduced seed production with plants grown only 1 week longer (∼35 DAE) than as the current vegetative vigor test for nontarget herbicide effects on plants.

A composite transcriptional signature differentiates responses towards closely related herbicides in Arabidopsis thaliana and Brassica napus.

In this study, genome-wide expression profiling based on Affymetrix ATH1 arrays was used to identify discriminating responses of Arabidopsis thaliana to five herbicides, which contain active ingredients targeting two different branches of amino acid biosynthesis. One herbicide contained glyphosate, which targets 5-enolpyruvylshikimate-3-phosphate synthase (EPSPS), while the other four herbicides contain different acetolactate synthase (ALS) inhibiting compounds. In contrast to the herbicide containing glyphosate, which affected only a few transcripts, many effects of the ALS inhibiting herbicides were revealed based on transcriptional changes related to ribosome biogenesis and translation, secondary metabolism, cell wall modification and growth. The expression pattern of a set of 101 genes provided a specific, composite signature that was distinct from other major stress responses and differentiated among herbicides targeting the same enzyme (ALS) or containing the same chemical class of active ingredient (sulfonylurea). A set of homologous genes could be identified in Brassica napus that exhibited a similar expression pattern and correctly distinguished exposure to the five herbicides. Our results show the ability of a limited number of genes to classify and differentiate responses to closely related herbicides in A. thaliana and B. napus and the transferability of a complex transcriptional signature across species.

Pea (Pisum sativum) seed production as an assay for reproductive effects due to herbicides.

Even though herbicide drift can affect plant reproduction, current plant testing protocols emphasize effects on vegetative growth. In this study, we determined whether a short-growing season plant can indicate potential effects of herbicides on seed production. Pea (Pisum sativum cv. Dakota) plants were grown in mineral soil in pots under greenhouse conditions. Plants were treated with a variety of herbicides (dicamba, clopyralid, glufosinate, glyphosate, 2-methyl-4-chlorophenoxyacetic acid, primisulfuron, or sulfometuron) at below standard field application rates applied at a vegetative stage of growth (approximately 14 d after emergence) or at flowering (approximately 20 d after emergence). Pea seed production was greatly reduced by sulfometuron at the minimum concentration used (0.001 x field application rate), with an effective concentration producing a 25% reduction in seed dry weight of 0.00007 x field application rate. Primisulfuron and glyphosate had a 25% reduction in seed dry weight for seed dry weight of 0.0035 and 0.0096 x field application rate, respectively. Clopyralid and dicamba reduced pea seed dry weight at a 25% reduction in seed dry weight of approximately 0.07 x field application rate. Glufosinate only reduced pea seed weight in one experiment, with a 25% reduction in seed dry weight of 0.07 and 0.008 x field application rate at vegetative growth and flowering stages, respectively. Pea seed dry weight was not affected by 2-methyl-4-chlorophenoxyacetic acid. Plant developmental stage had no consistent effect on herbicide responses. Reduced seed production occurred with some herbicides (especially acetolactate synthase inhibitors), which caused little or no reduction in plant height or shoot biomass and little visible injury. Thus, pea may be a model species to indicate seed reproductive responses to herbicides, with seed production obtained by extending plant growth for usually only 7 d longer than the period usually used in the vegetative vigor test.

Effects of low concentrations of herbicides on full-season, field-grown potatoes.

Current phytotoxicity plant test protocols for US pesticide registration require testing for effects on seedling emergence and early growth without regard to other important factors, such as plant reproduction. Yield and quality reduction can have significant economic and ecological effects. Therefore, field trials were conducted to determine if potato (Solanum tubersum L.) vegetative growth and tuber yield and quality were affected by herbicides at below recommended field rates. Potatoes were grown in fields at the Oregon State University Horticulture Farm with herbicides applied at below recommended field application rates 14 d after emergence (DAE) or at 28 DAE. Plant height was measured before and 14 d after application. Visual foliar injury was rated 14 d after application, and tuber yield and quality parameters were measured at harvest (120 DAE). Some tubers were grown in the greenhouse the following year to determine if there were carry-over effects. Potato vegetation and tuber yield quality were generally more affected by herbicides applied at 14 DAE than at 28 DAE. Tuber yield and quality parameters were more affected by lower herbicide rates than were plant height or injury. There were significant yield losses caused by low rates of sulfometuron methyl and imazapyr and, to a lesser extent, with glyphosate and cloransulam-methyl. Bromoxynil and MCPA ((4-chloro-2-methylphenoxy)acetic) acid had little effect on the plants. Vegetative responses did not accurately predict yield and quality responses of tubers; therefore, reproductive responses should be considered in phytotoxicity test protocols for pesticide registration in the USA.

Can artificial soil be used in the vegetative vigor test for U.S. pesticide registration?

Current testing guidelines for pesticide registration for the protection of nontarget plants calls for the use of sterilized, standardized soil consisting of primarily sandy loam, loamy sand, loamy clay, or clay loam that contains up to 3% organic matter. Low organic matter soils can be difficult to manage in a greenhouse setting because when soils dry, they contract, causing impeded water infiltration, or when overwatered, poor drainage increases the chances of anaerobic conditions. The purpose of this study was to determine if the results for the vegetative vigor test differed when using either natural or artificial soils. The herbicide sulfometuron methyl was applied 14 d after emergence at 0.1 and 0.0032 of the suggested field application rate. Six plant species were tested, 4 of the common test species, Zea mays (corn), Glycine max (soybean), Avena sativa (oat), and Lactuca sativa (lettuce), and 2 native plants of the Willamette Valley, Oregon prairie, Bromus carinatus (California brome) and Ranunculus occidentalis (western buttercup). Herbicide application rate was the most significant factor in the experiment regardless of soil type. The different soils generally produced different results, even though the 2 native soils, one from Oregon and the other from Maryland, are both acceptable soils for the pesticide registration tests. The plants grown on artificial soil produced results generally between the Oregon and Maryland soil results. This study indicates that artificial soils may produce results similar to or more sensitive than soils currently used in the vegetative vigor test.

Selecting and evaluating native plants for region-specific phytotoxicity testing.

In this study, we evaluated methodology to determine risks to terrestrial native plant species from potential herbecide drift, focusing on 1) selection of native species for testing, 2) growth of these species, and 3) variability in herbicide response among native species and compared with crop plants. Native plant species were selected for initial testing on the basis of spatial analysis, which indicated that species from Illinois, USA, were at potential risk for off-target effects of herbicide drift. On the basis of preliminary seed germination tests, 5 native plant species (Andropogon gerardi, Polygonum lapathifolium, Solidago canadensis, Symphyotrichum lateriflorum, and Tridens flavus) were selected for comparison with crops grown in Illinois, normally used in the US Environmental Protection Agency's (USEPA's) Vegetative Vigor Test (Avena sativa, Daucus carota, Glycine max, Solanum lycopersicon, and Zea mays), or both. When treated with low concentrations of a test herbicide, sulfometuron methyl, 2 native species, P. lapathifolium and S. canadensis, were as sensitive as the 5 crop species. The effective herbicide concentrations producing a 25% reduction in shoot dry weight (EC25) for these species, ranged from 0.00015 to 0.0014 times a field application concentration of 52 g/ha active ingredient of sulfometuron methyl. S. lateriflorum and T. flavus were less sensitive than the other native species, whereas A. gerardi was tolerant to sulfometuron methyl with no growth reduction at any herbicide concentration tested. This study indicated that native species can be successfully selected and grown, used in the suite of species used in the USEPA's phytotoxicity test to assess risks of chemical herbicides to nontarget plants. It also showed (with a limited number of species) that native species varied more in sensitivity to simulated herbicide drift than crop species often used in phytotoxicity testing and that a Weibull function was useful to calculate EC25 values when low concentrations of herbicides was used.

Using a geographic information system to identify areas with potential for off-target pesticide exposure.

In many countries, numerous tests are required as part of the risk assessment process before chemical registration to protect human health and the environment from unintended effects of chemical releases. Most of these tests are not based on ecological or environmental relevance but, rather, on consistent performance in the laboratory. A conceptual approach based on Geographic Information System (GIS) technology has been developed to identify areas that are vulnerable to nontarget chemical exposure. This GIS-based approach uses wind speed, frequency of those winds, pesticide application rates, and spatial location of agricultural crops to identify areas with the highest potential for pesticide exposure. A test scenario based on an incident in Idaho (USA) was used to identify the relative magnitude of risk from off-target movement of herbicides to plants in the conterminous United States. This analysis indicated that the western portion of the Corn Belt, the central California valley, southeastern Washington, the Willamette Valley of Oregon, and agricultural areas bordering the Great Lakes are among those areas in the United States that appear to have the greatest potential for off-target movement of herbicides via drift. Agricultural areas, such as the Mississippi River Valley and the southeastern United States, appears to have less potential, possibly due to lower average wind speeds. Ecological risk assessments developed for pesticide registration would be improved by using response data from species common to high-risk areas instead of extrapolating test data from species unrelated to those areas with the highest potential for exposure.