1262 Environmental Toxicology and Chemistry, Vol. 22, No. 6, pp. 1262–1268, 2003 Printed in the USA 0730-7268/03 $12.00 + .00 Environmental Toxicology METHYL PARATHION TOXICITY IN VEGETATED AND NONVEGETATED WETLAND MESOCOSMS RALF SCHULZ,*² M ATT T. MOORE,² E RIN R. BENNETT,‡ JERRY L. FARRIS,§ SAMMIE SMITH,JR.,² and CHARLES M. COOPER² ²U.S. Department of Agriculture–Agricultural Research Service, National Sedimentation Laboratory, P.O. Box 1157, Oxford, Mississippi 38655, USA ‡GLIER, University of Windsor, Windsor, Ontario N9B 3P4, Canada §Environmental Science Program, Arkansas State University, P.O. Box 847, State University, Arkansas 72467, USA ( Received 18 June 2002; Accepted 1 September 2002) Abstract—Methyl parathion (MeP) was introduced into constructed wetlands for the purpose of assessing the influence of emergent vegetation on transport and toxicity of the pesticide. Two vegetated (90% cover, mainly Juncus effusus) and two nonvegetated wetland cells (each with a water body of 50  5.5  0.2 m) were each dosed with 6.5 m 3 of water containing active ingredient of MeP at 6.6 mg/L associated with suspended soil at 400 mg/L to simulate a storm runoff event. Acute toxicity was assessed by sampling benthic macroinvertebrates at 5, 10, 20, and 40 m from the inlet before and 96 h after contamination and by in situ exposure of Chironomus tentans (Diptera) up to 24 h after contamination. Methyl parathion was detected throughout the nonvegetated wetland cells (70 g/L at 20 m, 8 g/L at 40 m), whereas the pesticide was not transported through the vegetated wetland cells (20 g/L at 20 m, 0.1 g/L at 40 m). A three-way analysis of variance using contamination (repeated measure variable), location, and vegetation indicated significant negative effects of contamination on various insect taxa, such as mayfly nymphs and caddisfly larvae. Seven out of the total of 15 species revealed a significant contamination  vegetation effect, with individuals in the vegetated wetlands being less affected. Four species showed a significant contamination  location effect, confirming a higher toxicity in the inlet area of the wetlands. A significant three-way interaction of contamination  vegetation  location was detected in Chironomus sp., which was most strongly affected at the inlet area of the nonvegetated wetland cells. The in situ bioassay employing C. tentans confirmed the positive effect of wetland vegetation on MeP toxicity. These results demonstrate the importance of vegetation for pesticide mitigation in constructed wetlands. Keywords—Insecticides Risk mitigation Non–point-source pollution Vegetation Wetland communities INTRODUCTION Constructed wetlands recently have been shown to have the ability to retain non–point-source insecticide pollution and pre- venting it from entering receiving aquatic habitats [1–3]. The implementation of retention ponds in agricultural watersheds was mentioned by Scott et al. [4] as one strategy to reduce the amount and toxicity of runoff-related insecticide pollution discharging into estuaries. The usefulness of aquatic plants for removal of insecticides from water has been shown [5] and the effects of the organophosphate phorate have been assessed by using littoral mesocosms in South Dakota (USA) wetlands [6]. However, few other studies in the open literature deal with the fate or effects of agricultural insecticide input in con- structed wetlands. Processes important for removal of non–point-source pes- ticide runoff in wetlands may include adsorption, decomposi- tion, and microbial metabolism [7]. The macrophytes present in the wetland may play an important role in providing an increased surface area for sorption as well as for microbial activity [8,9]. Furthermore, they may contribute directly to chemical metabolism [10]. Emergent vegetation was demon- strated to reduce the resuspension of sediments in wetlands [11]. Spray drift and runoff are important routes for non–point- source pesticide pollution of aquatic habitats, and runoff has * To whom correspondence may be addressed (ralf.schulz@syngenta.com). The current address of R. Schulz is Eco- logical Sciences, Syngenta Crop Protection AG, Jealott’s Hill Inter- national Research Centre, Bracknell, Berkshire RG42 6EY, United Kingdom. been shown to possibly contribute to greater concentrations and loads of insecticides than spray drift [12]. Runoff is the major source of aquatic insecticide input in the intensively cultivated Mississippi River (USA) delta region [13]. Con- structed wetlands could serve as a suitable risk mitigation strategy for agricultural runoff, given that enough information on their effectiveness with specific reference to the importance of the wetland vegetation is available. Biological effects of pesticides in wetlands have been studied under experimental conditions with mesocosms [14], in littoral enclosures [6], or in the field by employing organisms in situ [1,2,4]. The need now exists to link wet- land characteristics such as the presence of emergent mac- rophyte vegetation with the transport of non–point-source pesticide contamination and the resulting biological effects. The following study was undertaken for this purpose. Meth- yl parathion, an organophosphate insecticide primarily ap- plied to cotton, was chosen as the test substance for a sim- ulated runoff event. The use of MeP in the lower Mississippi River delta averages approximately 400,000 kg of active ingredient per year [15] and MeP has been detected at high levels in agricultural runoff [16]. Methyl parathion has an organic carbon partition coefficient (K OC ) of 5,100 and a water solubility of 55 mg/L. MATERIALS AND METHODS Description of the wetland mesocosms Constructed wetlands (water body: 50  11  0.2 m) at the University of Mississippi Field Station (Abbeville, MS,