Field and Greenhouse Experiments
Field studies were conducted in 2012, 2013, and 2015 to evaluate the sequestration potential of five agricultural hose types and different cleanout procedures while using dicamba. In 2012, a preliminary study was conducted to determine whether the five hose types led to any differences with respect to injury in soybean. After preliminary results indicated differences among hose types, the experiment was replicated and data from the preliminary trial were omitted. In 2013 and 2015, the experiment was conducted at the Black Belt Branch Experiment Station (33.256076° N, 88.553837° W) in Brooksville, MS, on an Okolona silty clay (fine, smectitic, thermic Oxyaquic Hapluderts) with 8% sand, 51% silt, 41% clay, 2% organic matter, and a pH of 6.8; and on a Brooksville silty clay (fine, smectitic, thermic Aquic Hapluderts); and at the R. R. Foil Plant Science Research Center (33.469066° N, 88.760782° W) in Starkville, MS, on a Marietta fine sandy loam (fine-loamy, siliceous, active, thermic Fluvaquentic Eutrudepts) with 71% sand, 17% silt, 13% clay, 1.03% organic matter, and a pH of 5.9. Differences from 2012, 2013, and 2015 involved the addition of an extra cleanout procedure and the addition of a rate titration followed by aqueous sample collection and analytical analysis. Planting date, planting populations, and seed variety varied among locations (Table 1).
Table 1 Planting year, location, date, population, and seed variety information for dicamba hose sequestration trials.
Field studies conducted in 2012 and 2013 involved five different types of agricultural spray hoses by two cleanout procedures (water and ammonia). Each hose measured 3 m in length and had an inside diameter of 1.3 cm, which is enough carrying capacity to deliver a sufficient volume to treat a plot-sized area of 2 by 12 m. Hose types included: John Deere PMK 4131-08 (yellow/PVC/high tensile–strength yarn/1 ply), John Deere PMA 4086-08 (blue/linear/low-density polyethylene blend), John Deere PMA 1687-08 (green/PVC/polyurethane/high tensile–strength yarn/2 ply), John Deere PMA 1628-08 (gray/PVC/polyurethane blend/high tensile–strength yarn/2 ply), and a Goodyear hose (black/Versigard synthetic rubber). Each hose end was fitted with a female pneumatic coupling to allow for sequestration of the solution within each hose and to prevent leakage. Field studies in 2015 involved the same hose types previously described and added a cleanout (water, ammonia, and no-cleanout) along with a rate titration of dicamba at 0.56, 0.140, 0.0087, and 0.0022 kg ae ha−1 to use for comparison. Samples were collected from each hose type by cleanout procedure and rate titration. Analysis was performed on high-performance liquid chromatography–mass spectrometry (HPLC-MS).
In 2013 and 2015, herbicide treatments consisted of dicamba (Engenia®, 600 g L−1, BASF, 26 Davis Drive, Research Triangle Park, NC 27709) at 0.56 kg ae ha−1. In all years, glyphosate (Roundup WeatherMax®, 540 g ae L−1, Monsanto, St. Louis, MO 63167) was applied at 1.1 kg ae ha−1.
For soybean analysis, spray lines were filled with dicamba and glyphosate at a rate of 0.56 and 1.1 kg ae ha−1, respectively and left to equilibrate for 48 h. The spray solution was then flushed out of the lines and the hose section cleaned with one of three cleanout procedures: no cleanout or water cleanout or ammonia cleanout at a rate of 11.35 L of water per line to simulate an actual in-field cleanout procedure with the hose then left to equilibrate in its designated cleaning solution for 24 h. For the ammonia cleanout, a 1% v/v solution was used. After 24 h, lines were flushed of the designated cleaning solution and left empty for 48 h. The spray lines were then filled with glyphosate at a rate of 1.1 kg ae ha−1. This solution was then equilibrated for 48 h to aid in the release of any sequestered herbicides before collection. The solution from each hose type by cleanout procedure was then collected using CO2 to push the solution from each hose to a collection bucket. A 10 ml aliquot was taken from each collection container for chemical analysis. The remaining solution was then applied to soybean at the R2 growth stage. Each hose type by cleanout combination was replicated three times; in essence, there was only one hose type per cleanout procedure per replication. Hoses were used for the same treatment from one year to the next throughout the entirety of the study.
Herbicide treatments from hose equilibrated solutions were applied with a CO2-pressurized backpack sprayer equipped with TTI110015 wide-angle, air induction, tapered flat spray tip (TeeJet Technologies, P.O. Box 7900, Wheaton, IL 60187) at an application volume of 140 l ha−1 and a pressure of 220 kPa. Visual estimates of soybean injury were recorded 7, 14, 21, and 28 DAT using a scale of 0 to 100%, where 0=no injury and 100=total plant death. Chlorosis, necrosis, stunting, leaf cupping, epinasty, and plant height reduction as compared with the untreated control were visually evaluated to estimate injury. Plant height and plant height reduction from the check were collected 7, 14, 21, and 28 DAT. Soybean was machine harvested and yield and yield reduction were calculated.
The experiment was arranged as a factorial arrangement of treatments in a randomized complete block with factor A consisting of hose type and factor B consisting of cleanout procedure. The rate titration was averaged separately and used as a comparison. Three replications for each treatment were used in the experiment with a plot-sized area of 2 by 12 m.
Treatments described in the 2015 field studies were also evaluated in the greenhouse in 2014. The trial was replicated in the greenhouse in October and November of 2014. Soybean seeds were planted approximately 2.5 cm deep in 9.8 L plastic pots (RM3R RootMaker Pot, Stuewe and Sons, 2290 SE Kiger Island Drive, Corvallis, OR 97333) containing commercial potting soil mix (Metro-Mix® 360, Sungro Horticulture, 770 Silver Street, Agawam, MA 01001). After planting, plastic containers were surface irrigated with tap water for the duration of the experiment. Plants were thinned to four plants per container within 1 wk of emergence and grown at 35/30 C day/night temperature. Natural light was supplemented with light from sodium vapor lamps (General Electric Sodium Vapor Lamps, Lucalox LU 400, General Electric Consumer and Industrial Lighting, 1975 Noble Road, Nela Park, Cleveland, OH 44112) to provide a 16 h photoperiod.
Approximately 2 wk after thinning, when plants had reached the V3 growth stage, hose cleanout-solution spray treatments were initiated using a compressed-air spray chamber equipped with a single 80015EVS flat-fan nozzle (TeeJet Technologies) at an application volume of 140 L ha−1 and a pressure of 220 kPa. Herbicide treatments consisted of dicamba at 0.56 kg ae ha−1 and glyphosate applied at 1.1 kg ae ha−1. For the greenhouse experiments, all spray lines were filled with dicamba and glyphosate at the same rate and cleaned in the same manner as in the field experiments in 2015. The solution from each hose type by cleanout procedure was then collected using CO2 to push the solution from each hose to a collection bucket for analysis. A 10 ml aliquot was then taken from each collection bucket for chemical analysis. The remaining solution was then added to 355 ml bottles and applied to soybean at the V3 growth stage in the spray chamber.
Visual estimates of soybean injury were recorded 3, 5, 7, and 14 DAT, using a scale of 0 to 100%, where 0=no injury and 100=total plant death. Chlorosis, necrosis, stunting, leaf cupping, epinasty, and regrowth were visually evaluated to estimate injury. Plants were cut at the soil line 21 DAT, dried, and weighed to calculate dry matter and dry matter reduction from the untreated check. Three replications for each treatment were used in the experiment with one pot representing one hose per hose type by cleanout procedure for each replication. Data were pooled across site years because experimental replication was considered a random variable. Untransformed and arcsine square-root-transformed data were subjected to analysis of variance, but interpretations were similar to untransformed data; therefore, untransformed data were used for analysis. Data were analyzed using PROC GLIMMIX in SAS v. 9.4, and means were separated using Fischer’s protected LSD test at P=0.05.
Samples from field and greenhouse studies were collected in 2014 and 2015 in 20 ml liquid scintillation vials (Sigma-Aldrich, 3050 Spruce Street, St. Louis, MO 63103) with a sample volume of 10 ml. Rinsates from field and greenhouse samples were taken at the time of the experiment and frozen for analytical analysis. Samples were collected using a 50 ml silicone pipette filler, three-way valve (Cole-Parmer Instrument Company, 625 East Bunker Court, Vernon Hills, IL 60061) attached to a 10 ml serological, sterile, individually wrapped pipette (Cole-Parmer Instrument Company). Samples were collected with one pipette per sample to eliminate cross contamination.
Dicamba analysis was performed at the University of Tennessee (Knoxville, TN 37996). Instrumentation used in the analysis began with the Agilent 1100 series, which included a quaternary pump, an auto sampler, a thermostated column compartment, and a 6120 quadruple single-quad MS (Agilent Technologies, 5301 Stevens Creek Boulevard, Santa Clara, CA 95051). The liquid phase of the analysis was acetonitrile +0.1% formic acid (70%) and water +0.1% formic acid (30%).
Samples collected from field and greenhouse studies were prepared by vortexing the aliquot solutions (Fisher Vortex Genie 2, Scientific Industries, 80 Orville Drive, Suite 102, Bohemia, NY 11716) for 30 s. A 1 ml aliquot was added to 19 ml of methanol to constitute a 0.05 dilution. For rate titration of dicamba at 0.56 and 0.14 kg ae ha−1, the dilution rate was increased to 0.00063. Dilution of the samples moved the dicamba concentration to within the linear range of the MS instrument by adding 1 ml of the aliquot solution to 19 ml of methanol and then extracting 250 µl of that solution into 19.75 ml of methanol. For the lower end of the rate titration of dicamba at 0.0087 and 0.0022 kg ae ha−1, the dilution rate of 0.05 was maintained. After dilutions were made, a final vortex of the solution was done for 30 s. A 2 ml extraction was made with a BD 10 ml syringe with Luer-LokTM (Becton, Dickinson, 1 Becton Drive Franklin Lakes, NJ 07417-1880) and passed through a 0.45 µm hydrophobic polytetrafluoroethylene (PTFE) membrane filter (09-719H, Thermo Fisher Scientific, 300 Industry Drive, Pittsburgh, PA 15275) directly into a clear glass vial with a polypropylene open top bounded with PTFE/silicone septum (Thermo Fisher Scientific).
The analysis began with an injection of methanol (to verify a lack of background carryover) followed by dicamba standards of 16.5, 30, 300, and 1000 ppb to establish linearity of MS response. A dicamba standard (30 ppb) was analyzed after every four unknown samples to verify consistency of MS detector response over time. The conservative lower limit of detection was 5 ppb, and all samples (with the exception of untreated samples) had dicamba concentrations above this amount. Three replications for each treatment were used in the experiment, with one sample representing one hose per hose type by cleanout procedure for each replication.
Hose Analysis Using Scanning Electron Microscopy (SEM)
For hose analysis using SEM, subsamples of hoses used throughout experiments were derived by randomly selecting hose types used and comparing them to hoses of the same type that were never used and have never had solution within them. The used hoses were used a total of eight times in the previous experiments. Three subsamples were cut from each hose type into 7.6 cm samples using a ratcheting hose and PVC cutting tool (Professional Ratcheting Hose and PVC Cutter 37100, Superior Tool Company, 100 Hayes Drive, Cleveland, OH 44131). Samples were then cut into smaller pieces roughly measuring 6.4 by 2.5 mm. Samples were then randomly chosen and glued to a 25.4 mm pin stub (16144, Ted Pella, 4595 Mountain Lakes Boulevard, Redding, CA 96003) using EPO-TEK® conductive Silver Epoxy and a liquid hardener (Ted Pella H-22) to affix four samples per pin stub with the outside of the hose attached to the stub for analysis of the inner tube. After 24 h the samples were coated; it was necessary to use a platinum coating to create a charge. The platinum coating was applied with an EMS 150T ES Coater (EMS, P.O. Box 550, 1560 Industry Road, Hatfield, PA 19440) using argon gas as the supply. Samples were coated in less than 1 min and left to cure for 24 h.
Samples were then loaded to a Zeiss Evo 60 EP-SEM (Zeiss International, Carl-Zeiss-Strasse 22, 73447 Oberkochen, Germany) connected to a Bruker AXS Quantax 4010 energy dispersive X-ray spectrometer (Bruker Corporation, Permoserstrasse 15, 04318 Leipzig, Germany). The Bruker software was used for graphing the elemental makeup of the sample and for creating a color-coded map of the sample in which different colors pertain to different elements. The Quantax 4010 was equipped with a silicon drift detector that provided a high-resolution and accurate map and/or graph of the sample.