Soil amended with green cover crops of Brassica napus or B. juncea generally had higher populations of Fusarium oxysporum and Pythium spp. than the methyl bromide treatment and the fallow control. Laying black polyethylene mulch at incorporation or one month after incorporation did not consistently influence the amount of isothiocynates detected in amended soils. Damping-off and Fusarium wilt on seedless watermelon were not consistently lower in brassica-amended soils compared to the nontreated control or methyl bromide.
Biofumigation refers to the suppression of soilborne pests and pathogens by biocidal compounds released in soil when glucosinolates (GSL), thioglucoside compounds in Brassica green manure or rotation crops, are hydrolyzed (Kirkegaard and Sarwar, 1998). Glucosinolates are sulfur compounds composed of a thioglucose group, a variable carbon side chain (R-group), and a sulphonated oxime (Mayton et al., 1996). There are about 20 different types of GSLs commonly found in brassicas that vary in their structure, depending on the type of organic side chain (aliphatic, aromatic, indolyl) (Kirkegaard and Sarwar, 1998).
Glucosinolates are hydrolyzed by myrosinase, an enzyme present endogenously in brassica tissues, to release a range of hydrolysis products including oxazolidinethiones, nitriles, thiocyanates, and various forms of volatile isothiocyanates (Gardiner et al., 1999; Kirkegaard and Sarwar, 1998; Mayton et al., 1996; VanEtten et al., 1969). The specific hydrolysis products formed depend on the R group of the parent glucosinolate and pH (Gardiner et al., 1999). Myrosinase is located in special cells from which it is released when the leaf tissue is damaged (e.g. when it is rubbed). Reports in the literature document the types of glucosinolates and their quantitative and qualitative differences in plant parts, ontogeny, and season of growth for many Brassica spp (Gardiner et al., 1999; Kirkegaard and Sarwar, 1998; Mayton et al., 1996; Sarwar and Kirkegaard, 1998; Sarwar et al., 1998; Smolinska and Horbowicz, 1999; Smolinska et al., 2003).
Mayton et al. (1996) found that volatile compounds from B. juncea cv. Cutlass were fungicidal in vitro to plant pathogenic fungi, including Fusarium sambucinum. Smolinska et al. (2003) evaluated the sensitivity of four F. oxysporum isolates to different isothiocyanates (ITCs) in vitro. They found that conidial and chlamydospore germination were highly susceptible to inactivation by isothiocyanates leading them to conclude that these two stages in the life cycle of Fusarium are most susceptible. A similar conclusion was reached by Smolinska and Horbowicz (1999), in an experiment in which Fusarium oxysporum chlamydospores exposed to the volatiles from B. juncea lost their viability completely. Soil amended with B. juncea residues had significantly fewer chlamydospores of F. oxysporum (Smolinska, 2000).
Gardiner et al. (1999) tested two cultivars of B. napus, Dwarf Essex and Humus, which are winter hardy. They report that the abundance of most GLS compounds peaked 30 h after incorporation or a little later, and trailed erratically to 20 days after incorporation by which time they were generally below the limit of detection. They also found that the dominant glucosinolate in the roots of both cultivars was 2-phenylethyl (a minor constituent in the shoots) and when they sampled the soil after incorporation, again 2-phenlyethly ITC was the most abundant, followed by benzenepropanitrile. Kirkegaard and Sarwar (1998) also report finding significant amounts of 2-phenylethyl in the roots of brassicas. Therefore, roots might play a more prominent role than shoots in contributing allelochemicals in soil. Because 2-phenylethyl is aromatic–and thus less volatile–it may persist for longer periods in the soil, and be released prior to incorporation (Kirkegaard and Sarwar, 1998). Soil residence times vary among compounds, because the ITC functional groups are reactive and sorption occurs to soil constituents (Gardiner at al., 1999). In addition to the 2-phenylethyl and benzenepropanitrile, Gardiner et al. (1999) also found 3-butenyl, 4-pentenyl, 4-methylthiobutyl, and 5-methylthiopentyl ITCs in small quantities. Detection of 5-methylthiopentanenitrile and 6-methylthiohexanenitrile in soil indicates that glucosinolate hydrolysis in green manures may produce nitriles at the expense of the respective ITCs. Relatively little is known about the activity of 2-phenylethyl ITC in the soil since many previous studies have concentrated on aliphatic types such as methyl ITC (a commercial soil fumigant) or 2-propenyl ITC (allyl) due to its early recognition as the active constituent of mustard oils (Kirkegaard and Sarwar, 1998). Therefore, all isothiocyanates released in the soil following biofumigation will be monitored.
Fusarium wilt of watermelon is a problem where watermelon is grown on short rotation sequences. The pathogen accumulates in soils and remains there indefinitely. Many diploid cultivars of watermelon have resistance to race 1 of F. oxysporum f. sp. niveum. However, only a few cultivars of triploid watermelon have resistance, and these cultivars produce elongated fruit, which is unacceptable to produce buyers. In addition, race 2 of F. oxysporum f. sp. niveum is present in several southern states, and there is no resistance to this race in cultivated varieties (Martyn and Bruton, 1989; Zhou and Everts, 2001). Therefore, biofumigation may be useful to control Fusarium wilt in susceptible cultivars of triploid watermelon and in diploid cultivars in areas where race 2 is present (Martyn, 1987).
Control of other soilborne pathogens also is important. Damping-off caused by Rhizoctonia solani and Pythium spp can be a problem on transplanted watermelon. Watermelon fruits are subject to fruit rot caused by R. solani, Pythium spp, and Sclerotium rolfsii. Therefore, the study will also look at the effect of biofumigation on these fungi with the goal of getting more information out of the study.
- Evaluate the effectiveness of biofumigation in the control of Fusarium wilt of watermelon, compared to control with methyl bromide.
Determine the best time to incorporate green manure prior to laying plastic (i.e., let the biofumigant decompose before tarping, or tarping immediately after plowing under).
Quantify inoculum density of Fusarium oxysporum, Rhizoctonia solani, Pythium spp, Sclerotium rolfsii, and fluorescent Pseudomonas in the soil before and after biofumigation.
Determine glucosinolate concentration in roots and shoots of the brassicas at the time of incorporation.
Quantify glucosinolate breakdown products in the soil after brassica incorporation
The experiment was conducted at the Clemson University Coastal Research and Education Center, Charleston, SC. The soil was Yonges loamy sand with a pH of 6.3 in 2004 and 6.1 in 2005. Previously, the field was planted with muskmelon (Cucumis melo L.) and watermelon (Citrullus lanatus) in the spring of 2003. To increase the inoculum density of F. oxysporum f. sp. niveum, watermelon cv. Black Diamond was direct-seeded 15 September 2003 and 11 and 25 August 2004 in all treatment plots. At the time of seeding, 280 kg/ha ammonium sulfate 21-0-0 and 560 kg/ha of 10-10-10 N-P-K was applied in 2003 and 2004, respectively. In 2003, Black Diamond plants were mowed and the field was disked twice on 19 November. After seeding in 2004, emergence was poor due to excessive rainfall from Hurricane Charley, Tropical Depression Bonnie, and Tropical Storm Gaston. The field was not replanted and it was disked twice.
The experimental design was a randomized complete block with six replications. Treatment plots measured 24.4 by 0.9 m with a 3-m fallow space between plots. Six treatments were evaluated: 1) non-amended control soil covered with black polyethylene mulch at transplanting; 2) soil fumigated with methyl bromide at 448 kg/ha and black polyethylene mulch applied a month before transplanting watermelon; 3) canola (Brassica napus L. cv. Dwarf Essex) (Columbia Grain, Clarkston, WA) incorporated into soil and black polyethylene mulch applied 1 month later when watermelons were transplanted; 4) canola incorporated into soil and soil covered immediately with black polyethylene mulch at incorporation time; 5) mustard (B. juncea [L.] Czern cv. Cutlass) (Dr. K. Downey, Saskatoon Research Center, Saskatoon, Canada) incorporated into soil and black polyethylene mulch applied 1 month later at transplanting; and 6) mustard incorporated into soil and black polyethylene mulch applied immediately at incorporation time.
On 24 November 2003 and 3 November 2004, 672 kg/ha of 10-10-10 N-P-K fertilizer was applied followed by direct seeding of canola and mustard at 2.24 and 4.48 kg/ha, respectively (Hair, 2001) using a Brillion turfmaker junior seeder (Brillion Iron Works Inc., Brillion, WI). The field was watered as needed with overhead irrigation. In 2005, 672 kg/ha ammonium sulfate mixed with 1 kg/ha boron (Solubor®) was applied 3 months after seeding (Hair, 2001).
To determine the biomass of canola and mustard, plants were dug from two randomly chosen 0.25-m2 areas per plot (Gardiner et al., 1999), approximately 1 week before incorporation on 15 March 2004 and 30 March 2005. The plants were counted, washed, and oven-dried at 50oC for 1 week. The plants were then cut at the crown to separate roots and shoots, and dry weights of each were measured.
A month after the incorporation of mustard and canola, seedless watermelon cv. Tri-X® 313, susceptible to Fusarium wilt, was transplanted into single-row plots. Transplants were spaced 30 cm apart within rows. ‘Black Diamond’ pollenizer plants were transplanted into rows between plots. Plots were watered and fertigated using drip irrigation. During the growing seasons, recommended preventative fungicides were applied to manage foliage diseases. On 28 June to 13 July 2004, and 20 July to 5 August 2005, fruits were harvested from a 12-m length in the middle of each plot and weighed.
The number of plants in each plot was counted weekly until canopies of adjacent plants joined. To identify pathogens causing damping-off and wilt, diseased watermelon plants were collected on 25 May 2005 from all treatments and washed under running tap water and portions of necrotic lesions on stems and roots were cultured on semi-selective media. The recovered fungi were identified.
To rate the percentage of wilted foliage in each plot, plots were divided into quarters and disease severity assessments were made in each quarter plot on the canopy just before the first harvest and then on weekly intervals using a modified Horsfall-Barratt 15-point rating scale (Keinath and DuBose, 2004).
In 2005, 20 hyphal tips from colonies of Pythium spp. randomly selected from soil dilution plates to represent all treatment plots were transferred to cornmeal agar (CMA). Nineteen Pythium isolates recovered from watermelon roots in 2005 also were transferred to CMA. Isolates were identified to species using a grass-leaf blade culture technique (Van der Plaats-Niterink, 1981). At desired time intervals (2 to 4 days for asexual structures and 3 to 10 days for sexual structures) leaf samples were placed on a slide, stained with 0.05% trypan blue with lactophenol, and examined with a compound microscope at 400× and 1000× (Martin, 1992).
On July 15 2004, four vine pieces were collected from each treatment plot. The vines were cut into 1-cm sections, disinfected in 0.6% NaOCl for 1 minute, rinsed in sterile distilled H2O for 1 minute, and placed on Komada’s media with five segments from one vine on each plate (four plates per plot). From these plates, 43 isolates of F. oxysporum were recovered. On May 29 2005, 36 isolates were obtained from wilted 2-month-old watermelon plants isolated as previously described. Three agar plugs from 1-week-old single-spore F. oxysporum isolates were placed in 50 mL of potato dextrose broth in a 250-mL flask and shaken at 150 rpm at ambient temperature (22 to 24 oC) for 4 days. Cultures were decanted through sterile cheesecloth in a sterile funnel into sterile 50-mL centrifuge tubes and the fungal mats were washed with 5 to10 mL sterile distilled water to dislodge conidia. Culture filtrates were centrifuged at 5,000 rpm for 10 min to pellet spores, and spores were resuspended in 10 mL of sterile distilled water. Spore concentration was determined by counting with a hemocytometer and concentration was then adjusted to 106 microconidia/mL. Isolates with low concentrations of conidia were not used in pathogenicity and race tests.
Pathogenicity tests were set up in a completely randomized experiment with five and three replications, respectively. A replication consisted of four seedlings in a 15-cm diameter pot. Black Diamond was used for all pathogenicity tests. Watermelon was seeded in flats containing 1:1 v/v sand: vermiculite. Two-week-old seedlings were removed and their roots were washed under running tap water. To inoculate plants, roots were dipped into the conidium suspension and transplanted into 15-cm diameter pots containing a 4:1:1 (v/v/v) mix of sand: peat: vermiculite (Egel et al., 2004). Roots of negative-control plants were dipped in sterile distilled water. Two isolates of F. oxysporum f. sp. niveum race 2, FON 997 and FON 998 (isolated from watermelon cv. Stars and Stripes, Colleton County, SC), were included as positive controls. Wilted plants were counted weekly starting 2 weeks after inoculation for a total of 6 weeks. Temperature was set at 28oC in the greenhouse during experiments, but ranged from 16 to 40C over the course of the experiment.
Analysis of variance was performed with PROC GLM of SAS (SAS Institute, Cary, NC; release 8.2 for personal computers). All data sets were checked for normality, and Hartley’s test for equality of variance was calculated. In both years colony forming units (CFUs) of F. oxysporum, Pythium spp., and fluorescent Pseudomonas spp. were transformed to logarithm values. Pre-planned comparisons using single-degree-of-freedom contrast statements were used to compare effects of treatment on microbial populations. For pre-treatment microbial densities, means of similar treatments (i.e., two non-planted, two canola, and two mustard treatments) were pooled before comparisons were made. Treatment means for final incidence of damping-off, Fusarium wilt severity, and yield were compared using Fisher’s Protected LSD (P=0.05). For glucosinolate concentrations, means were pooled within species and standard errors calculated. Single-degree-of-freedom contrast statements were used to compare total glucosinolate concentrations.
To determine pre-incorporation inoculum density of F. oxysporum, R. solani, Pythium spp., and S. rolfsii and population density of flourescent pseudomonads, soil samples were taken on 9 March 2004 and 25 and 26 March 2005. Post-incorporation inoculum densities were assayed on 27 April 2004 and 18 April 2005. In 2005, F. oxysporum was also quantified at the end of the season on 16 August. Using a 2.5-cm diameter soil probe, 30 soil cores were taken to a depth of 15 cm from each plot. Soil samples from each plot were thoroughly mixed before assaying for microorganisms.
To estimate the inoculum density of Pythium spp., 10 g of soil from each plot was added to 200 mL of 0.3% water agar and shaken for 30 seconds. Three aliquots (0.5 mL on each plate) were spread on plates of pimaricin-ampicilin-rifampicin-pentachloronitrobenzene (PCNB) medium prepared with pimaricin at 5 mg/L (P5ARP) (Jeffers and Martin, 1986). Plates were incubated at 20oC for 20 h and then adhering soil was washed away. Plates were reincubated at 20oC and large-sized colonies (ca. > 1 cm) were counted at 36 to 48 h after plating (randomly selected colonies were later identified to species level using morphological techniques, described later). Colony counts were expressed as colony forming units (CFU) per gram of soil dried for 24 h at 100oC.
To estimate the inoculum density of Rhizoctonia solani, 300 g of soil from each plot was wet-sieved through a # 18 mesh sieve (1-mm pore openings) to recover organic matter (van Bruggen and Arneson, 1986). The organic matter was air-dried at 22 to 24oC overnight. Dried organic matter from each sample was placed in 10 small heaps of equivalent size on ethanol-potassium-nitrate agar, prepared with 2% ethanol and 20 µL/L prochloraz (Trujillo et al., 1987). Total number of heaps with and without colonies of R. solani growing from them was recorded after incubating 3 to 4 days in the dark at 23 to 25oC. Hyphae of selected colonies were stained in situ with 3% KOH and alkaline safranin O stain then examined at 200× to 400× using a compound microscope to determine the number of nuclei per cell (Bandoni, 1979).
To estimate the population density of flourescent Pseudomonas spp., 10 g of soil from each plot was added to 90 mL sterile distilled water and shaken for 30 seconds. Further ten-fold dilutions were made and three 0.1-mL aliquots of 10-1 and 10-2 dilutions were spread on S-1 medium (Gould et al., 1985). S-1 plates were incubated in the dark at 23 to 25oC for 3 to 4 days and colonies of flourescent Pseudomonas spp. were counted under long-wave ultraviolet light. All colony counts were expressed as colony forming units (CFU) per gram of soil dried for 24 h at 100oC.
To estimate the inoculum density of Sclerotium rolfsii, 300-g aliquots of soil from each plot were air dried overnight in aluminum pans (30 × 24 × 4 cm) at ambient temperatures (22 to 24 oC). Dried soil was spread evenly in pans and moistened with 75 mL of 1.33% (v/v) methanol (Rodriguez-Kabana et al., 1980). Aluminum pans then were put into large Ziploc® plastic bags and incubated at 30oC, 3 to 4 days. Germinated sclerotia were counted by closely examining the soil surface visually for discrete colonies of S. rolfsii.
To estimate the inoculum density of F. oxysporum, 10-g of soil from each plot was added to 90 mL of sterile distilled water and shaken for 1 minute. Further ten-fold dilutions were made and three 0.1-mL aliquots of 10-1 and 10-2 (an aliquot per plate) were spread on Komada’s medium (Komada, 1975). The plates were held under diurnal light (16-hr photoperiod) at ambient temperature (22 to 24 oC). Colony counts were expressed as colony forming units per gram of soil dried for 24 h at 100oC.
To quantify glucosinolate content in the canola and mustard crops, five additional plants were dug from each plot 1 week prior to incorporation, put into plastic bags, and held on ice in a cooler chest. In 2004, the canola and mustard plants were immediately taken to the laboratory where roots and shoots were cut separately into small pieces. Two 5-g subsamples from each plot were frozen in liquid nitrogen before being freeze-dried. In 2005, the plants were stored at -80oF for 12 months before freeze drying. Freeze-dried plant parts were ground using a Wiley mill equipped with a 1-mm mesh sieve. The powder was kept in a screw-cap test tube and stored at 5oC in a refrigerator until it was assayed for glucosinolates 2 weeks later.
Glucosinolates were extracted from 300 mg of freeze-dried root and shoot tissues using the procedure described by Magrath et al. (1993) with modifications employed by Kirkegaard and Sarwar (1998). To quantify desulphoglucosinolates, 2-propenyl and benzyl glucosinolates were used as internal standards. Levels of other glucosinolates were determined using response factors for desulphoglucosinolates published by the European Economic Community (Commission Regulation [EEC] No 1864/90) (Kirkegaard and Sarwar, 1998). Desulphoglucosinolates were analyzed by injecting a 30 µL sample into Hewlett Packard 1090 high pressure liquid chromatograph (Agilent Technologies, Wilmington, DE). Peaks were identified using pure standards. Forty-eight and 24 samples were analyzed in 2004 and 2005, one sample per replication and one sample from half of the replications, respectively. To calculate glucosinolate concentration incorporated per square meter, glucosinolate concentrations per gram were multiplied by biomass per square meter (Morra and Kirkegaard, 2002; Sarwar and Kirkegaard, 1998).
Canola and mustard plants were mowed and then rototilled (Ferguson Tilrovator, Ferguson Mfg. Co., Suffolk, VA) 15 cm deep into the soil on 17 March 2004 and 5 April 2005. In 2004, 50% of canola and 10% of mustard plants had flowered whereas approximately 90% of canola and 50% of mustard plants had flowered in 2005. Most of the non-flowering plants had unopened flower buds.
After incorporation, soil samples were taken 0 to 15 cm deep using a 2.5-cm diameter soil probe at 6 time intervals: 0 h and 6 h, and 1, 2, 12, and 26 days after incorporation in 2004 and 0 h and 6 h, and 1, 2, 4, and 12 days in 2005 . Soil cores were placed in a 50-mL Pyrex centrifuge tube containing 5 mL of 0.2 M CaCl2 and 12 mL of dichloromethane with 0.1% cyclohexane as an internal standard (Gardiner et al., 1999). A total of 216 samples (6 treatments x 6 replications x 6 sampling times) collected in 2004 and 90 samples (6 treatments ×3 replications x 5 sampling times) collected in 2005 were analyzed. Fewer samples were analyzed from 2005 because isothiocyantes were not detected at the later sampling times in 2004. Gardiner’s et al. (1999) methods were used for extraction. Samples were analyzed on a Hewlett Packard 5890 gas chromatograph (GC). External standards used for quantification included allyl isothiocyante (Lancaster Synthesis, Inc., Pelham, NH), N-propyl, phenyl, benzyl, methyl, and 2-phenylethyl isothiocyanate and goitrin (oxazolidinethione) (Alfa Aesar, Ward Hill, MA).
In 2004, damping-off was lower in canola-amended soils than in mustard amended-soils, but these treatments did not differ from methyl-bromide or the fallow control. In 2005, damping-off was lower in canola-amended soil than in the control. Likewise, in 2005, severity of Fusarium wilt was lower in canola-amended plots than in plots treated with methyl bromide or mustard. In 2004, there were no differences among treatments for severity of Fusarium wilt in watermelon. Yields did not differ significantly in either year.
Post-treatment in 2004, populations of F. oxysporum were significantly higher in canola-amended soils covered with polyethylene mulch immediately after incorporation (P=0.01) or later at transplanting (P≤0.01), and in mustard-amended soils covered with mulch later (P=0.003) than in control soils. Compared to densities at the second sampling, F. oxysporum population increased significantly (P<0.0001) by the third sampling in methyl bromide-treated soils in 2005. At the end of the season in August 2005, compared to the control, F. oxysporum densities were significantly higher in methyl bromide-treated soils, in canola-amended soils and in mustard-amended soils covered with polyethylene mulch immediately or later (P<0.01).
Pythium spp. identified from soil collected in 2004 were P. irregulare and P. spinosum.
After brassicas were incorporated in 2004, Pythium spp. inoculum densities were significantly lower in methyl bromide-treated soils compared to those in mustard-amended soils covered with polyethylene mulch immediately (P=0.006) or later (P<0.003).
In 2005, before brassicas were incorporated, populations of Pythium spp. were not significantly different between those in soils cropped to canola and mustard. However, Pythium spp. densities were significantly higher in soils planted to mustard (P=0.002) compared to non-planted soils. Comparing post-treatment to pre-treatment densities in 2005, Pythium spp. decreased significantly in methyl bromide treated soils (P=0.0001), and in mustard-amended soils covered with polyethylene mulch immediately after incorporation (P=0.004).
Population densities of Rhizoctonia solani and Sclerotium rolfsii were not significantly different among treatments in either year.
Both mustard and canola produced glucosinolates when grown as a winter cover crop in coastal South Carolina, although total glucosinolate concentrations for both species were higher in 2005 than in 2004 (P<0.0001).. In both 2004 and 2005, five glucosinolates were identified in mustard and three glucosinolates were identified in canola. The predominant glucosinolate in the roots of both canola and mustard was 2-phenylethyl. Root tissues of both cultivars had significantly higher concentrations of 2-phenylethyl than shoot tissues (P<0.0001). Total glucosinolate concentration incorporated per square meter was significantly higher for mustard than for canola in both years, although the biomass of canola was numerically greater than that of mustard.
Although concentrations of glucosinolates were higher in 2005 than in 2004, isothiocyantes (ITC)were detected less frequently in 2005 than in 2004. Isothiocyanates were not detected in soil samples taken at zero sampling time prior to plant tissue incorporation either year.
In 2004, 2-phenylethyl isothiocyante was detected in 10 of 144 samples from brassica-amended soils with equal numbers from covered or uncovered plots. In 2004, 2-phenylethyl ITC was detected in samples from amended soil taken 6 hours to 12 days after incorporation. Benzyl ITC was detected in 2004 from samples taken 6 hours after incorporation.
Of the 72 samples obtained from amended soil in 2005, 2-phenylethyl and benzyl ITC were detected in 4 and 1 samples, respectively. 2-phenylethyl ITC was detected in soil samples taken 6 hours and 1 day after incorporation. Benzyl ITC was only detected in 1 sample from soil 12 days after incorporation in 2005.
Educational & Outreach Activities
Njoroge, S. M. C., M. B. Riley and A. P. Keinath. 2006. Effect of Biofumigation with Brassica Crops on Population Densities of Soilborne Pathogens in South Carolina. Phytopathology 96:S85.
S.M.C. NJOROGE, J.E. Toler, and A.P. Keinath. 2006. Single- and multiple-year effects of soil solarization on population density of soilborne microorganisms in South Carolina. Phytopathology 96:S187.
S.M.C. NJOROGE, J.E. Toler, and A.P. Keinath. 2006. Incidence of damping-off in soils solarized and planted single or multiple years in South Carolina. Phytopathology 96:S187.
S.M.C. NJOROGE, J.E. Toler, and A.P. Keinath. 2005. Pythium spp. inoculum density and disease development under different soil solarization strategies in South Carolina. Phytopathology 95:S76.
Njoroge, S. M. 2006. Assessment of the effects of soil solarization and incorporation of Brassica spp. on populations of soilborne microorganisms in South Carolina. Ph.D. Dissertation, Clemson University. 144 p.
Canola ‘Dwarf Essex’ appeared to be a better candidate for biofumigation in coastal South Carolina than ‘Cutlass’ mustard, although mustard produced three to four times more glucosinolates. We learned that there is more to choosing a biofumigant cover crop than the amount of glucosinolates it is reported to produce.
No growers have adopted biofumigation, as the results of this study were not particularly encouraging. However, growers of brassica crops, such as collard, cabbage, and broccoli, may be affecting soil microoganisms in their fields when they plow down crop residue.
Areas needing additional study
Future research should evaluate other Brassica crops for biofumigation potential and investigate factors that affect the conversion of glucosinolates to ITC after incorporation of the cover crop into soil.