On-Farm Use of a Hybrid Vetch Cover Crop to Reduce Fusarium Wilt in Seedless Watermelon

Final Report for OS07-035

Project Type: On-Farm Research
Funds awarded in 2007: $9,900.00
Projected End Date: 12/31/2009
Region: Southern
State: South Carolina
Principal Investigator:
Anthony Keinath
Clemson University
Expand All

Project Information


A hairy vetch winter cover crop was compared with rye, fallow, and vetch plus methyl bromide fumigation for managing Fusarium wilt of seedless watermelon on a private farm. Wilt was lower with vetch plus fumigation than in all other treatments. Early season yield of watermelon was similar in vetch alone and vetch plus fumigation. Cover cropping with vetch produced a net return of $2,675/acre, whereas fallow and rye lost an average of $410/acre. The poor stand of vetch in no-till corn stubble may have limited the potential benefits of hairy vetch.

Tables, figures or graphs mentioned in this report are on file in the Southern SARE office.

Contact Sue Blum at 770-229-3350 or
sueblum@uga.edu for a hard copy.


Fusarium wilt is the most wide-spread, damaging soilborne disease of watermelon in the South. Cropping the same fields again and again to watermelon without rotation or with short rotations selects the specialized form of the soilborne fungus Fusarium oxysporum, F. oxysporum f. sp. niveum (abbreviated FON), that attacks watermelon (Hopkins et al., 1992). Fusarium wilt kills individual vines or whole plants, directly reducing the number of fruit produced. In addition, this fungus lasts indefinitely in infested soil. It can reduce the yield, acreage, and economic return of watermelon crops, both crops grown now and those planted in the future.
As of 2006, there are four races of FON that are separated by the cultivars they attack. Race 0 is pathogenic only on older and heirloom watermelon cultivars that have no resistance. Many diploid, seeded hybrid cultivars have resistance to race 1. Previously, race 2 was limited to Texas, Oklahoma, and Florida. Since 2000, race 2 has been found in five new states: Indiana, Maryland, Delaware, Georgia, and South Carolina (Egel et al., 2005; Keinath, 2009).
Traditionally, 7-year rotations have been used to manage Fusarium wilt. However, today’s watermelon growers specialize in watermelon production and cannot afford to leave much acreage out of production for more than one or two years. In South Carolina, all seedless watermelons and a majority of seeded watermelons are irrigated. Growers are limited to planting in fields that they have equipped with wells, pumps, or center pivot irrigation. In addition, rotations with agronomic crops, such as wheat and soybean, do not reduce FON in soil (Zhou and Everts, 2004).
Although hybrid seeded cultivars resistant to race 1 were useful in the past, the watermelon market has shifted to seedless cultivars, which are susceptible to Fusarium wilt. Two seedless watermelon cultivars, ‘Matrix’ and ‘Revolution,’ have resistance to race 1, but they are not grown because their fruit types are not acceptable to produce buyers. Both cultivars are susceptible to FON race 2 and thus would perform like susceptible cultivars in fields infested with a mixture of races. In addition, the O’Neal’s cannot grow these cultivars because oblong fruit do not fit in their state-of-the-art packing line.
Methyl bromide was used in the past to manage Fusarium wilt. During the phase-out, some growers tried other fumigants, such as Telone-C35 (1, 3-dichloropropene plus 35% chloropicrin). However, restrictions on use of Telone have been imposed as a result of FQPA review, and few watermelon growers currently use Telone.
To summarize the problem, there is a lack of commercially viable control options for Fusarium wilt of watermelon. Therefore, watermelon growers in the South and other regions of the U. S. need sustainable controls that fit their production practices and are effective against all races of FON.

Project Objectives:
  • Determine the level of Fusarium wilt in treatments;
    Measure the yield (number) of marketable-sized fruit;
    Calculate the cost and economic return of each treatment;
    Determine the level of the Fusarium fungus in soil collected from treated plots;
    Communicate results to stakeholders


Click linked name(s) to expand
  • Bradley O'Neal


Materials and methods:

The experiment was done on a private watermelon farm in Hampton County, SC, in a portion of a 50-acre field in which Fusarium wilt was diagnosed in previous watermelon crops. The soil was a fine loamy sand with a pH of 6.2. The field was cropped to corn in 2007. Four treatments, fallow, hairy vetch (Vicia villosa), ‘Wren’s Abruzzi’ rye, and hairy vetch plus methyl bromide-chloropicrin, were arranged in a randomized complete block design in one field block (six planted rows between a drive row). All treatments were replicated three times. Plots were 100 ft long by 43.8 ft (6 rows) wide.
Hairy vetch at 65 lb/A and rye at 120 lb/A were seeded on 23 Oct 07. (Note that although ‘Cahaba White’ hybrid common vetch was supposed to have been used, ‘Cahaba White’ was not available locally in South Carolina because of a crop failure. The grower—without consulting the Project Director—purchased and seeded hairy vetch instead.) Cover crops were incorporated on 1 Feb 08 by cultivating with a disk. Watermelon rows were raised beds shaped 1.5 ft wide on 7.3-ft centers. Fertilizer was applied to all beds, and they were covered with black polyethylene mulch. Selected plots were fumigated with 200 lb/A methyl bromide-chloropicrin (50/50) as mulch was laid. Watermelon was transplanted into the prepared rows on 1 Apr. The seedless triploid cultivar ‘7167’was planted in rows 1, 3, 4, and 6. A mixture of the seedless cultivar and equal numbers of two seeded diploid pollenizers, ‘Mickey Lee’ and ‘Summer Flavor 800,’ was planted in rows 2 and 5. Fertilizer was injected through the drip irrigation system with N, P and K values adjusted weekly according to soil and plant analysis done by the grower.
Plants with any symptoms of wilting and dead plants were counted on 30 April and 20 May. Symptomatic plants of ‘7167’ were collected from rows with seedless watermelons in the fallow, hairy vetch, rye, and hairy vetch plus methyl bromide-chloropicrin treatments in block 1. Discolored vascular tissue was cultured on Komada’s medium. Fruit in rows 1, 4 and 6 were counted on 10 June. Counts of fruit per plot were divided by the number of surviving plants per plot to account for differences in survival.
Ten soil cores were collected from each plot in December 2007, February 2008, and June 2008. Soil was diluted in water and aliquots of soil suspensions were cultured on Komada’s medium, which is semi-selective for F. oxysporum. The number of colonies was counted 9 days later. Selected colonies were kept from the dilution plates and tested in the greenhouse for pathogenicity to ‘Sugar Baby’ watermelon to determine the number and percentage of F. oxysporum f. sp. niveum in soil. Isolates obtained from wilted plants were inoculated onto three differential cultivars in the greenhouse to determine which races were present.

Research results and discussion:

Hairy vetch stand was poor, due to being seeded into no-till corn stubble. Stand averaged 6.4 plants per square foot in December 2007. Based on a biomass of 0.26 kg dry weight per square meter, vetch contributed approximately 49 pounds of N per acre. Note that the stand of vetch and the amount of N were less than half of what was obtained in experiments at the Coastal Research and Education Center over several years. In fall 2008, the O’Neal’s seeded hairy vetch on 150 acres of land. Stands in 2008 in soil that had been disked and harrowed were 9.4 plants per square foot. The amount of N per acre calculated from a dry weight biomass of 0.52 kg/square meter was 127 pounds per acre. Thus, the potential effects of vetch on Fusarium wilt and watermelon yield are likely to be greater than what was realized in the 2007-2008 experiment.
Fusarium wilt was very severe in the area of the field where the experimental plots were located. All 23 plants cultured yielded F. oxysporum. Many symptomatic plants in the nonfumigated treatments were completely wilted and stunted, whereas most plants in the fumigated plots were normal size and only partially wilted. At the first rating on 30 April, mean wilt incidence was higher in the four rows with only seedless watermelon than in the two rows with a mixture of seedless and pollenizer watermelons (paired t test, P=0.0075). Overall wilt incidence ranged from 49% to 66%. Hairy vetch plus methyl bromide-chloropicrin reduced wilt only in the rows with a mixture of seedless and pollenizers. At the second rating on 20 May, wilt incidence ranged from 50% to 96%. It was lower in the hairy vetch plus methyl bromide-chloropicrin treatment than in all other treatments for rows with seedless watermelons and the mixture of seedless and pollenizers.
The number of fruit set per surviving plant on 10 June was double in the vetch plus methyl bromide treatment (1.4) than in the rye and fallow treatments (mean of 0.7 fruit/plant). Fruit number in the vetch alone treatment, 0.74 fruit/plant, was lower but statistically equivalent to the vetch plus methyl bromide treatment (P=0.10). The grower had seeded the remainder of the field to hairy vetch. Yields of watermelon in that part of the field were approximately 40,000 lb/A, which exceeded expectations for yields in a field infested with F. oxysporum f. sp. niveum.
The baseline population of F. oxysporum in soil in all treatments in December 2007 was approximately 8.0 colony-forming units (CFU) per gram of oven-dry soil. After incorporation of the cover crops and fumigation in February 2008, populations decreased to 2.9 CFU in the vetch plus methyl bromide treatment, whereas populations increased slightly in fallow (15 CFU) and vetch (11 CFU) treatments and increased greatly to 27 CFU in the rye treatment. At the end of the experiment, the population of F. oxysporum was still significantly greater in the rye plots (5.6 CFU) than in the vetch and fallow plots (2.9 CFU), which were greater than in the vetch plus methyl bromide plots (1.5 CFU).
Of 19 isolates of F. oxysporum recovered from diseased plants, 17 were pathogenic i.e., were F. oxysporum f. sp. niveum. All 17 isolates were race 2. Of 81 isolates from soil that were tested for pathogenicity, 21 (or 26%) were F. oxysporum f. sp. niveum.

Participation Summary

Educational & Outreach Activities

Participation Summary

Education/outreach description:

Keinath, A. P. and Hassell, R. L. 2009. On-farm evaluation of hairy vetch and fumigation for integrated control of Fusarium wilt on seedless watermelon. Plant Disease Management Reports 3:V035. (http://www.plantmanagementnetwork.org/pub/trial/pdmr/reports/2009/V035.pdf)

Two outreach events were held. On August 26, 2008, I met with Bradley O’Neal and Angela O’Neal Chappell to discuss the results of the project. On November 17, 2008, I met with the members of Board of Directors of the South Carolina Watermelon Association to present a summary of the project. There were seven leading South Carolina growers, one scout, and two employees of the SC Dept. of Agriculture present. Because of the problems with the vetch stand and the lack of a visual response of watermelon to vetch, based on number of wilted and dead plants, we decided not to hold a field day, since we concluded that this was not a representative field site.

Project Outcomes

Project outcomes:

Demonstrable impacts are that the O’Neal’s planted the remainder of the field used in 2007-08, 50 acres, to hairy vetch. This was the first year they used vetch as a winter cover crop. In the fall of 2008, they planted hairy vetch in three other fields for a total of 150 acres. If hairy vetch increases yields again in 2009, then I expect other growers to start using hairy vetch, based on favorable comments from the O’Neal’s.

Economic Analysis

Variable costs and fixed costs for irrigated, seedless watermelon are calculated to be $2,864/acre and $155/acre, respectively, for 2008-2009 in South Carolina (http://cherokee.agecon.clemson.edu/melndrp6.pdf). The cost of seeding vetch was $127/acre, plus $2 for incorporation, for an additional cost of $129/acre above the total cost for irrigated, seedless watermelon. Fumigation cost an additional $800/acre. Seeding and incorporating rye cost $43/acre. Yields and gross receipts were based on early season fruit counts, which were doubled to estimate total yields, an average weight of 14.5 pounds/melon, and an average price of $0.20/pound. (The grower estimated that his yields in the reminder of the field, which was cropped to vetch and fumigated, were 40,000 pounds/acre. Thus, doubling the crown set was used as an approximation for total yield in the field, as the estimate for the vetch plus fumigation treatment was 42,000 pounds/acre.) Using this formula, cover cropping with vetch produced a net return of $2,675/acre. This compared very favorably with fallow and rye, as these control treatments lost an average of $410/acre. Vetch plus methyl bromide fumigation yielded a net return of $4356/acre.

Farmer Adoption

The O’Neal’s planted the remainder of the field used in 2007-08, 50 acres, to hairy vetch. This was the first year they used vetch as a winter cover crop. In the fall of 2008, they planted hairy vetch in three other fields for a total of 150 acres.

Nine farmers were reached through outreach activities.


Areas needing additional study

The grower wanted to know what the benefit was to adding vetch to soil fumigation. Because there was no fumigation treatment without vetch included in the 2007-2008 experiment, this could not be directly determined. An additional experiment is in progress, in which plots with and without vetch are being compared in an area fumigated with chloropicrin. The grower also wanted to know if more plants showing partial wilting early in the season “recovered” in the vetch treatment than in the other treatment, as a partial explanation for why the plants in the vetch treatment yielded better than expected.

Any opinions, findings, conclusions, or recommendations expressed in this publication are those of the author(s) and do not necessarily reflect the view of the U.S. Department of Agriculture or SARE.