Materials and methods

The use of biofumigant crops as viable options in rotation with commercial onions was studied at two field locations on the University of Idaho's Parma Research and Extension Center. The first field location was a high-lime calcareous soil—a Nyssaton silt loam (coarse-silty, mixed, superactive, mesic Xeric Haplocalcids). The second location was a low-lime calcareous soil—a Greenleaf-Owyhee silt loam (fine-silty, mixed, superactive, mesic Xeric Calciargids). The two field locations will hereafter be termed “high lime” and “low lime” to distinguish where different components of the project took place.

Treatments at the high-lime location were arranged in a randomized complete block design, and were fallow, nontreated control; metam sodium at 30 gal/acre; ‘Colonel’ oil seed radish seeds at 25 lb/acre; and ‘Idagold’ mustard seeds at 25 lb/acre. Plots were three beds wide by 40 ft long. Onions were planted 21 Mar. 2002, 17 Mar. 2003, and 12 Mar. 2004, and plant samples were collected on 16 Sept. 2002, 28 Aug. 2003, and 23 Aug. 2004. Onion yield was measured at the end of the season from the three rows (minus a 2.5-ft border on each plot end) by harvesting after the onions were lifted and field cured. Only pink root and yield data were recorded from the high-lime soil.

Treatments at the low-lime location were arranged in a split-plot design. The main plots were fallow, nontreated control; metam sodium at 30 gal/acre; ‘Colonel’ oil seed radish seeds at 25 lb/acre; ‘Idagold’ mustard seeds at 25 lb/acre; and ‘Sunrise’ canola (Brassica napus) seeds at 25 lb/acre. The subplots were weedy control; low herbicide inputs; high herbicide inputs; and weed-free control, hand weeded. Low herbicide inputs for two-leaf onions were bromoxynil (0.19 lb/acre a.i.; Buctril; Bayer CropScience, Research Triangle Park, NC) plus oxyfluorfen (0.094 lb/acre a.i.; Goal; Dow AgroSciences, Indianapolis) applied on 23 May 2002, 16 May 2003, and 3 May 2004. For three- to four-leaf onions, the herbicide application consisted of bromoxynil (0.25 lb/acre a.i.) plus oxyfluorfen (0.125 lb/acre a.i.) plus sethoxydim (0.19 lb/acre a.i.; Poast; BASF, Florham Park, NJ) applied on 31 May 2002, 27 May 2003, and 20 May 2004. For five- to six-leaf onions, the herbicide application consisted of oxyfluorfen (0.25 lb/acre a.i.) applied on 12 June 2002, 9 June 2003, and 3 June 2004. High herbicide inputs included all of the low input applications listed above, plus a pre-emergence application of glyphosate (0.375 lb/acre a.i.; Roundup; Monsanto, St. Louis) and pendimethalin (1.0 lb/acre a.i.; Prowl, BASF) on 2 Apr. 2002, 7 Apr. 2003, and 26 Mar. 2004, and another application of pendimethalin (0.5 lb/acre a.i.) on 12 June 2002, 9 June 2003, and 3 June 2004.

At the high-lime and low-lime locations, onions were harvested in 2002, 2003, and 2004. Before planting onions, rotational crops of wheat (Triticum aestivum) or barley (Hordeum vulgare) were harvested in July, and in August, the soil was tilled and seeded with the biofumigant crops at 25 lb/acre. Nitrogen (100 lb/acre) was applied for each biofumigant crop but not to the noncropped treatments. In November, the biofumigant crops were mechanically chopped (flailed) and the green tissue was rototilled 6 inches into the soil. The chemical fumigant metam sodium was shank applied at this time at 30 gal/acre to one of two plots in which biofumigants had not been planted. Two weeks after the biofumigant and metam sodium applications, the soil was disked and bedded on 22-inch spacings for onion planting the following spring.

High- and low-lime fields were planted with ‘Vaquero’ yellow sweet spanish onions in March. Onions were seeded on the bed centers in a double row, 4 inches between seeds within a row, at a rate of 14 lb/acre of pelleted seed. Both fields were lifted early in September. After 2 weeks of curing, the onions were topped and harvested.

Herbicide applications in 2002 and 2004 were made using a pressurized carbon dioxide backpack sprayer calibrated to deliver 20 gal/acre for pre-emergence and 40 gal/acre for postemergence applications at a pressure of 30 psi. In 2003, herbicides were applied with a four-wheeler mounted sprayer calibrated to deliver 30 gal/acre of water at 30 psi for pre- and postemergence applications. Plots were six beds that were 22 inches wide and 25 ft long. Weed control was evaluated using weed counts, visual weed control ratings, and weed biomass measurements. Weed counts were taken from 10 ft of rows 2 and 4 in each plot on 6 May, 3 June, 11 June, 25 June, and 24 July 2002; 13 May, 8 June, and 23 June 2003; and 20 Apr. and 1 June 2004. Visual weed control ratings were conducted on 29 July 2002, 11 July 2003, and 20 July 2004. Weed biomass samples were harvested from 5 ft of one row of each plot, separated by species, dried, and weighed on 29 July 2002, 16 July 2003, and 19 July 2004. After sampling, weeds were removed from all plots to allow for harvesting in September. Onion yield was determined by harvesting 25 ft of the center four beds of each plot. A reduction in onion stands in the first replication of 2003 was unrelated to treatments and required the first replication to be removed from the analysis of onion yield data.

Pink root data were collected from 20 onion bulbs for each main-plot treatment; samples were collected from the weed-free control subplots to eliminate variation from weeds and herbicides. The percent area of the roots infected was visually estimated by scoring onion samples from the middle row of each plot into six classes based on the infection of the roots according to the following scale: class 1 = 0% infected, class 2 = 1% to 3%, class 3 = 4% to 6%, class 4 = 7% to 10%, class 5 = 10% to 15%, and class 6 = 16% to 30%. Pink root samples were collected and evaluated on 31 July 2002, 5 Aug. 2003, and 26 July 2004.

Nematode populations were determined by collecting and mixing 20 soil cores (1.9 cm diameter × 30 cm deep) from the weed-free control subplots in each main plot of the low-lime soil. Samples were collected on 7 Dec. 2001, 14 May 2002, 25 Nov. 2002, 9 June 2003, 27 Nov. 2003, and 18 May 2004. Populations of root lesion nematode, stunt nematode (Tylenchorhynchus spp.), spiral nematode (Helicotylenchus variocaudatus), and ring nematode (Criconemella xenoplax) were estimated by extracting them from soil samples by Cobb's wet sieve technique (Cobb, 1918) followed by the modified centrifugal floatation technique (Jenkins, 1964).

Weed control data were analyzed using NCSS software (version 2001; NCSS Statistical Software, Kaysville, UT). When year by treatment interactions were absent, data were combined across years. Because herbicide by biofumigant interactions were not significant, weed control and biomass data are presented separately for herbicide and biofumigant main effects. Treatment means were compared using Fisher's protected least significant difference (lsd) with a critical value of P ≤ 0.05. Because the untreated control and the hand-weeded plots were not included in the analysis of herbicide programs for visual weed control, significant differences between the low- and high-input herbicide treatments were determined using a paired t test. Pink root, nematode, and high-lime yield data were analyzed by analysis of variance using the GLM procedure of SAS (version 8.2; SAS Institute, Cary, NC). Treatment means were compared using Fisher's protected lsd with a critical value of P ≤ 0.05. Pink root data were transformed using square root transformation .

Results

There were few effects of biofumigant treatments on weed counts, visual weed control (data not shown), or weed biomass (Table 1). Weed counts were collected throughout the growing season with no significant differences among treatments (data not shown). Visual weed control estimates indicate that fall-applied metam sodium provided partial control of some species [redroot pigweed (Amaranthus retroflexus), common lambsquarters (Chenopodium berlandieri), barnyard grass (Echinochloa crus-galli), and green foxtail (Setaria glauca)] in 2002 only. Also in 2002, redroot pigweed control was significantly better with the oil seed radish than the fallow treatment, and lambsquarters control was significantly better with the mustard than with the fallow treatment (data not shown).

Table 1.

Influence of fumigant (metam sodium) and weed control program main effects on dry weight weed biomass. Fumigant main effects are averaged over no-input, low-input, and high herbicide-input treatments for commercial onions. Weed control through herbicides main effects are averaged overall fumigant treatments.

View Table

Weed biomass data were combined across all 3 years for redroot pigweed and common lambsquarters. Hairy nightshade (Solanum sarrachoides), kochia, and total weed biomass were combined across 2002 and 2004. Barnyard grass and green foxtail biomass could not be combined across years. Main effects were not significantly different for all 3 years across all fumigant and biofumigant treatments (Table 1). Weed biomass was significantly reduced by the high-input herbicide treatment in all years when compared with the untreated control (Table 1). In most cases, the low-input herbicide treatment also reduced weed biomass when compared with the untreated control. The exceptions were in 2003 where kochia, barnyard grass, green foxtail, and total weed biomass in the low-input herbicide plots was equal to those in the untreated plots (Table 1).

Onion yield in the low-lime soil was unaffected by biofumigant crops in 2002 and 2003. In 2004, yield was significantly less after mustard and canola when compared with fallow and metam sodium treatments, and yield was significantly less following radish when compared with the fallow (Table 2). In the high-lime soil, onion yield was affected by biofumigants only in 2003 when yields were lower with the radish treatment than with the fallow treatment. Onion yields from metam sodium-treated plots were significantly higher than in the fallow in 2003 and were significantly higher in the radish and mustard treatments in 2004.

Table 2.

Onion yield as influenced by main effects of fumigant and weed control in the low- and high-lime soil in 2003 and 2004.

View Table

Herbicide-treated (low- and high-input weed control) and weed-free (hand-weeded) plots yielded greater than untreated plots all 3 years. In 2002, the high-input and weed-free treatments had significantly higher yields than the low-input treatment. In 2003, onion yields were similar between both herbicide treatments and the hand-weeded control. In 2004, the weed-free treatment was significantly better than the low- and high-input treatments (Table 2).

In 2002, ‘Colonel’ oil seed radish and ‘Sunrise’ canola volunteered, averaging 1.0 and 0.25 plants/yard², respectively. In 2003, oil seed radish volunteered, producing a significant number of plants (35 plants/yard²). Canola also volunteered, but not in significant numbers. In 2004, oil seed radish again volunteered. ‘Idagold’ mustard did not volunteer in any year.

Pink root severity in the low-lime soil was significantly (P ≤ 0.05) lower in metam sodium-treated plots than all other treatments in 2002 (Table 3). There were no significant differences among the treatments in 2003, but in 2004, metam sodium and mustard treatments had significantly lower (P ≤ 0.05) pink root severity when compared with radish and canola. Fallow plots were not significantly different from any other treatment in 2004. In the high-lime soil, metam sodium had significantly lower (P ≤ 0.05) pink root severity than all other treatments in 2002 and 2004. In 2003, metam sodium had significantly less pink root than the oil seed radish biofumigant, but was not significantly different from the mustard or fallow treatments (Table 3).

Table 3.

Pink root severity on ‘Vaquero’ onion roots when treated with biofumigants and the fumigant metam sodium in low- and high-lime soils.

View Table

Root lesion, stunt, spiral, and ring nematodes were recovered from soil samples. Mean counts per treatment for 2002 and 2003 are given in Table 4, but there were no significant differences (P ≤ 0.05) in nematode populations among the treatments at any sampling period. Bulb and root samples were not evaluated for nematode damage because there were no apparent nematode symptoms in the onions during all 3 years of this study.

Table 4.

Mean counts of root lesion nematode, stunt nematode, spiral nematode, and ring nematode when treated with biofumigants and the fumigant metam sodium in low-lime soils in 2002 and 2003.

View Table

Discussion

Fewer weeds were expected through the suppression of weed germination and emergence by biofumigants as noted by Al-Khatib et al. (1997), Bialy et al. (1990), Boydston and Hang (1995), Brown and Morra (1996, 1997), and Krishnan et al. (1998). However, our study had only limited control of a few weed species based on weed counts, visual estimates, and biomass data. Control was not consistent nor did it show any trends over 3 years. It is possible that the biofumigants and metam sodium did reduce weed germination and emergence because they had an effect on the stand of onions (Geary et al., 2006). However, the number of seeds in the soil was unknown and may have been large, thus prohibiting detection of differences in germination and emergence through weed counts, visual estimates, and biomass data.

The high-input herbicide program frequently used by onion growers in the Treasure Valley increased control of all weed species in all 3 years except for barnyard and green foxtail grass in 2003. The low-input program also increased control when compared with the untreated control in 8 of the 11 weed biomass columns (Table 1). Herbicide control of weeds was good, and it is understandable why growers are hesitant to use other options such as biofumigants until these alternatives are proven to provide a benefit for controlling weeds. Onion yields with herbicide inputs were at least six times higher than the weedy control and close to, or as good as, the hand-weeded, weed-free plots.

Yield was reduced by volunteers from oil seed radish and canola due to early competition with the onions. Each year, volunteer biofumigants were removed by hand, but without the use of hand labor, biofumigants would not have been effectively controlled by herbicide applications and would have been as competitive with the onion crop as many other weeds. Although the mustard did not volunteer in any year, onion yield was reduced by the mustard in 2002 and 2004. The results suggest that biofumigants are affecting seedbed conditions in a way that is reducing yield. Reductions in onion stand almost always resulted in reduced yields. Reduced yields occurred from biofumigant plants competing for resources as volunteers and by reducing stand numbers through inhibition of seed germination and emergence, either from residual isothiocyanate activity, poorer seed soil contact, or other means due to incorporation of organic residues. Herbicide treatments did not affect onion stand in any year (Geary et al., 2006).

Overall, the use of biofumigant crops provided inadequate weed control in all 3 years and reduced onion yield in 1 of 3 years. Additionally, oil seed radish and canola volunteered at varying levels in all 3 years and were difficult to control with herbicides and required removal by hand. Onion growers cannot accept the negative effects caused by the biofumigant crops without some significant increase in positive weed suppression.

Pink root severity on onion roots was significantly lower in metam sodium treatments than all other treatments in five of six samplings (Table 3), consistent with a previous report (Hartz et al., 1989). Pink root severity after biofumigant crops did not differ significantly from the fallow treatment, except in 2004 when oil seed radish and canola appeared to exacerbate pink root in the low-lime soil. There was no evidence for pink root control with the biofumigant crops in this study.

Glucosinolates from the biofumigant crops, which break down to ionic thiocyanate and isothiocyanate, averaged 0.0018 fl oz/acre and 0.0028 fl oz/acre for radish and mustard, respectively, in the high-lime soil (Brown et al., 2006). Although metam sodium breaks down to methyl isothiocyanate, when applied at 30 gal/acre it becomes ≈1328 fl oz/acre, depending on conversion efficiency from metam sodium to methyl isothiocyanate (Draper and Wakeham, 1993; Geddes et al., 1995). The difference in the amount of a.i. applied per acre of biofumigant plants and metam sodium is large and is likely the contributing factor for metam sodium reducing the amount of pink root in comparison with the biofumigant crops.

The canola biofumigant treatment was included in the low-lime soil study as a treatment involving incorporated biomass containing little if any glucosinolates. The intention was to isolate the effects of tissue-derived glucosinolates from vegetative biomass effects with little glucosinolate. Although canola and oil seed radish never differed in pink root severity, their pink root severities were greater than mustard or fallow (no biomass and no glucosinolates) in one comparison. Mustard and fallow treatments never differed in pink root severity in all 3 years. Thus, we were unable to distinguish between biomass effects due to plant glucosinolates or other biomass-induced conditions.

In summary, from a pest control perspective, oil seed radish, mustard, and canola biofumigants provided little advantage in commercial onion production. They also reduced onion yield in some instances. Therefore, their use is not recommended for Treasure Valley onion production.