Mustard Cover Crops Are Ineffective in Suppressing Soilborne Disease or Improving Processing Tomato Yield

Mustard (Brassica spp.) cover crop residue has been reported to have signifi cant ‘biofumigant’ action when incorporated into soil, potentially providing disease suppression and yield improvement for the succeeding crop. The effects of growing over-winter mustard cover crops preceding processing tomato (Lycopersicon escultentum Mill.) production were investigated in six fi eld trials in the Sacramento Valley of California from 2002–04. A selection of mustard cover crops were compared to a legume cover crop mix, a fallow-bed treatment (the current grower practice in the region), and in two of the six trials, fumigation treatments using metam sodium. Mustard cover crops removed 115 to 350 kg·ha N from the soil profi le, reducing NO 3 -N leaching potential. Soil populations of Verticillium dahliae Kleb. and Fusarium spp. were unaffected by the cover crops, and there was no evidence of soilborne disease suppression on subsequent tomato crops. Mustard cover crops increased tomato yield in one fi eld, and reduced yield in two fi elds. In one of two fi elds, metam sodium fumigation signifi cantly increased tomato yield. We conclude that, while environmental benefi ts may be achieved, mustard cover cropping offers no immediate agronomic benefi t for processing tomato production. Processing tomato is an important crop in the Sacramento Valley of California, and the sustainability of tomato production from both an economic and environmental perspective has been the subject of considerable research. Between the Sustainable Agriculture Farming Systems (SAFS) (Clark et al., 1999) and the Long-term Research on Agricultural Systems (LTRAS) (Denison et al., 2004) projects there have been nearly 20 site-years comparing organic, reduced input, and conventional production practices. One feature of organic and reduced input systems that might be useful in conventional production is the use of over-winter cover crops; currently nearly all commercial fi elds in the region are maintained in a fallow condition before tomato production. In both the SAFS and LTRAS projects over-winter cover cropping improved soil physical and biological properties (Clark et al., 1998; Colla et al., 2000; Martini et al., 2003; Poudel et al., 2001); unfortunately, the direct effects of cover cropping on tomato yield were unclear, since other management differences (fertilizer, weed control, direct seeding vs. transplanting, etc.) among production systems were confounding factors. Miyao and Robbins (2000) evaluated cover cropping in conventionally managed Sacramento Valley tomato fi elds across three production seasons (1998–2000). In all years cover cropping signifi cantly reduced winter runoff and increased tomato yield up to 7% compared with a winter fallow treatment. Although the factors responsible for yield that the existing buried drip irrigation system could be reused the following season. The following cover crops were seeded on 22 Nov: 1) ‘Humus’ (rape, Brassica napus L.); 2) ‘Pacifi c Gold’ [indian mustard, B. juncea (L.) Czern.]; 3) ‘Ida Gold’ (white mustard, Sinapis alba L.); 4) ‘ISCI 20’ (S. alba); and 5) a mixture of ‘Lana’ woolypod vetch (Vicia dasycarpa Ten.), ‘Magnus’ pea (Pisum sativum L.), and ‘Cayuse’ oat (Avena sativa L.). The cover crops were sown and lightly incorporated on the top of the raised beds. A fallow bed control treatment was also included to represent the typical over-winter management practice. The experimental design was randomized complete block, with four replications; individual plots were one bed wide × 30 m long. Cover crops were germinated by precipitation and were allowed to grow undisturbed until 28 Mar. 2003, when above-ground biomass was determined. One representative 0.5-m section per plot was harvested, and cover crop fresh and dry biomass recorded. On 31 Mar. cover crops were fl ail-mowed and immediately incorporated to a depth of about 15 cm with two passes of a bed disk. The fi eld was left undisturbed until mid-May when the soil beds were fertilized at 20 and 25 kg·ha N and P, respectively. Tomato transplants (‘Halley’) were set in the fi eld on 23 May. The crop was sprinkler-irrigated for establishment, then dripirrigated for the remainder of the season. An additional 165 kg·ha N was applied through the drip system over the growing season. A destructive fruit harvest was made on 17 Sept., with total fruit yield determined; marketable yield was determined after the removal of cull (green, sunburned and rotten) fruit. Soluble solids concentration of marketable fruit (°Brix, by refractometer) was measured, and the soluble solids yield (marketable yield × °Brix, a measure of overall productivity) was calculated. In Fall 2003, a second experiment was initiated at UCD. A fi eld of Yolo silt loam in which processing tomatoes had been grown that summer was prepared for cover crop planting. A randomized complete block design was used, with four replications of six treatments; individual plots were one 1.5-m bed wide × 30 m long. The following cover crops were seeded on 6 Nov.: 1) ‘Pacifi c Gold’; 2) ‘Caliente’ (a commercial blend of two separate cultivars, one B. juncea and one S. alba); and 3) vetch–pea–oat mixture. Additional treatments were a winter fallow, winter fallow plus springapplied metam sodium, and ‘Pacifi c Gold’ plus spring-applied metam sodium. The cover crops were sown on the bed tops and germinated by precipitation. On 18 Mar. 2004, the cover crops were fl ail-mowed and immediately incorporated as described for the 2002–03 trial. The beds were then rolled and sprinkler-irrigated for 6 h to provide ideal conditions for hydrolysis of glucosinolates in the cover crop residue. Before incorporation, cover crop above-ground fresh and dry biomass was determined on a 0.6-m section of each plot. Metam sodium (Soil Prep; Wilbur Ellis Co., San Francisco, Calif.) was applied to designated plots 1 week after cover improvement were not identifi ed, possible mechanisms included improved in-season irrigation water infi ltration (Colla et al., 2000) or suppression of soilborne diseases (VanBruggen, 1995). While cover crop evaluation in the SAFS and LTRAS projects centered on legumes due to their contribution to N fertility, the use of other types of cover crops may be more appropriate in conventional rotations where N availability is seldom a production constraint. Given their widely documented biofumigation effects (Brown and Morra, 1997; Fenwick et al., 1989), mustard cover crops (family Brassicaceae) present a potentially useful alternative. Mustards contain glucosinolates, compounds that upon hydrolysis release isothiocyanates (ITCs) and related compounds with broad biocidal activity. Among the soil pathogens reported to be effectively suppressed by the action of mustard residues are Verticillium dahliae (Olivier et al., 1999; Subbarao et al., 1999), Sclerotinia minor (Subbarao, 1998), and Helminthosporium solani (Vaughn et al., 1993). This study was conducted to evaluate the effects of over-winter mustard cover crops on soil pathogen populations, disease expression and fruit yield in conventionally managed processing tomato rotations. Materials and Methods Field trials were conducted at the University of California–Davis (UCD) and in commercial fi elds in Yolo County, Calif., from 2002–04. At UCD in 2002 processing tomatoes were grown in a Yolo silt loam soil. Production practices typical of the local commercial industry were used. Following harvest in early September, crop residue was incorporated into the raised, 1.5-m-wide beds, which were maintained so HORTSCIENCE 40(7):2016–2019. 2005. Received for publication 15 July 2005. Accepted for publication 29 Aug. 2005. We are indebted to the California Tomato Research Institute for fi nancial support, and to J. Davis of the University of Idaho for technical support. University of California Cooperative Extension. Department of Plant Pathology. DecemberBook.indb 2016 10/19/05 5:22:58 PM 2017 HORTSCIENCE VOL. 40(7) DECEMBER 2005 crop incorporation, into soil near fi eld capacity moisture content. The fumigant was applied through the buried drip irrigation system at 89 kg·ha a.i. in 60 m·ha of water. Preplant fertilizer was applied at 20 and 25 kg·ha N and P, respectively. ‘Heinz 9665’ transplants were planted in single rows per bed on 6 May. Irrigation and additional N fertilization were as described for the previous season. A destructive fruit harvest was made on 31 Aug. Total and marketable fruit yield, and fruit soluble solids yield, were determined. Four commercial fi eld trials were initiated near Woodland, Calif., in Fall 2003; site characteristics and cultural details are given in Table 1. At one site (designated Woodland A) fi ve overwintering treatments were compared: 1) ‘Pacifi c Gold’; 2) ‘Caliente’; 3) vetch–pea–bell bean (Vicia faba L. var. equina) mixture; 4) winter fallow; and 5) winter fallow + springapplied metam sodium. The experimental design was randomized complete block with 6 replications; individual plots were three beds wide × 45 m long. Cover crops were seeded 6 Nov. and incorporated 16 Mar. 2004. Cover crop biomass was determined on 1-m sections of the middle bed of each plot before incorporation; biomass of the B. juncea and S. alba components of the ‘Caliente’ blend were evaluated separately. Cover crops were fl ail-mowed and immediately incorporated with a bed disc. One week after incorporation metam sodium was applied at 107 kg·ha a.i. in designated plots by a tractor-mounted applicator through four injection shanks per bed. In the remaining fi elds (designated Woodland B, C, and D), fall-seeded ‘Caliente’ blend was compared to fallow bed management. At these sites the experimental design was randomized complete block, with 10, 8, and 5 replications, respectively; individual plots were 3 beds wide by 30 to 45 m long. Mustard biomass was determined as at Woodland A. Mustard plots were fl ail-mowed and incorporated by disking within 30 min. At commercial maturity, all Woodland fi eld trials were machine harvested; total and marketable fruit yield and fruit soluble solids yield were determined. Glucosinolate concentration of aboveground mustard biomass was determined for all 2003–04 trials except Woodland C. Immediately before soil incorporation, fi ve to six representative whole plants were collected from each of three plots of each type of mustard cover crop; for the ‘Caliente’ blend, B. juncea and S. alba plants were collected and analyzed separately. The plants were freeze-dried and ground to pass through a 1-mm mesh screen. Concentrations of the primary glucosinolates (benzyl, 2-propenyl, 2-phenylethyl, 2-hydroxy-3-butenyl, and 4-hydroxybenzyl) were determined by gas chromatography using a modifi cation of the procedure of Daun and McGregor (1991). Net N mineralization rate from 2003–04 cover crop residues was evaluated in a laboratory incubation. Oven-dried cover crop samples were ground and analyzed for total N content by combustion (Carlo Erba 1500; Fisons Instruments, Beverly, Mass.). Soil from the UCD site was wetted to fi eld capacity. Ground residue from each cover crop type from each site was blended into the moist soil at 0.5% by dry weight. Three replicate subsamples of moist, unamended soil, and of each residue/soil blend, were extracted in 2 N KCl and analyzed for mineral N concentration (NO 3 -N + NH 4 -N) using a fl ow injection analyzer (Lachat Instruments, Milwaukee, Wis.). The remainder of the unamended moist soil and the soil–residue blends were incubated in sealed containers at 20 °C; the containers were vented weekly to maintain aerobic conditions. After four weeks triplicate samples of unamended soil and of each soil–residue blend were extracted in 2 N KCl for determination of mineral N concentration. The change in mineral N, adjusted for the change in unamended soil, represented net N mineralization or immobilization by the residue. The effect of cover crops on soilborne pathogens was evaluated at four of the 2003–04 trial sites. At UCD and Woodland A soil samples were collected in Fall 2003, before cover crop planting, and again in Spring 2004, before transplanting tomatoes. In each plot a composite sample of 12 soil cores (0 to 15 cm deep) was collected. After the sample was homogenized, a 3-g sample was added to 30 mL of 0.1% water agar. Following agitation for 5 min, two aliquots of 1 mL each were plated on selective media (pectate agar for Verticillium dahliae and Komada’s medium for Fusarium spp.). Depending on the concentration of propagules in the soil, dilutions were adjusted to achieve a visually manageable number of colonies on the plates. Colonies of V. dahliae and Fusarium spp. were counted with the aid of a dissecting microscope. Soil pathogen data were analyzed by the ANCOVA procedure (SAS Institute, Cary, N.C.) using soil pathogen population preceding cover crop production as the covariate in the analysis of the spring populations. During the 2004 tomato season signifi cant foliar expression of plant infection by V. dahliae (Woodland A), fusarium foot rot (Fusarium solani) (Woodland B), and fusarium wilt (F. oxysporum f.sp. lycopersici W.C. Snyder & H.N. Hans.) (Woodland C) were observed. At each site the number of plants exhibiting disease symptoms was recorded 4 to 6 weeks before harvest. The identity of each pathogen was confi rmed by plating symptomatic tissue on water agar or potato–dextrose agar.