Introduction
Bacterial soft rot, caused by Erwinia carotovora and possibly by Pseudomonas species, often affects broccoli production in the Willamette Valley. The rot can affect floret or stem tissue and occurs mainly during periods of moderate temperatures and high humidity or rainfall. Chemical controls are ineffective and resistance has not been identified in varieties well-adapted to the local climate and acceptable to processors. While reductions in planting density may somewhat reduce disease incidence, the decreased disease incidence is not sufficient to offset the loss in number of heads harvested.
Previous work at NWREC (R. Ludy, M.S. thesis, Oregon State University, 1990) indicated that the amount of water applied to a broccoli crop did not affect floret soft rot (head rot) incidence or severity, at least within a range of applied water that allowed for economic yields. However, in non-replicated observations of the effect of irrigation frequency (once versus three times per week), head rot incidence more than doubled under the greater frequency. We propose that the incidence of rot is affected by the frequency and duration of water droplets on the floret and stem tissue. The purpose of these trials was to determine the effects of two or three frequencies of irrigation and the time of day of the irrigations, at equal amounts of applied water, on yield and head rot incidence in broccoli.
Methods
'Gem' broccoli was seeded to a Willamette silt loam, Ph 6.0, on 7 July, 1993 and 6 July, 1994. In 1993, the experimental design was a split-plot with a 2x2 factorial combination of irrigation frequency (every second and every eighth day) and timing (morning and evening) as main plots and harvest as subplot treatments. In 1994, the design was a factorial combination of three frequencies (every second, fourth, and eighth day) and the same timings. The main plots in 1993 consisted of 12 30-foot rows (18 in 1994), with three rows on a 60-inch bed. The subplots consisted of individual three-row beds. Main plot treatments were replicated four times in a completely random design. Plot preparation included a broadcast and incorporated application of 1000 pounds 10-20-20 fertilizer/acre, 0.75 pounds trifluralin, 1.3 pounds chlorpyrifos, and 10.0 pounds Solubor/acre. The area was then rototilled and cultimulched to form a seedbed. Plots were seeded with a gang of three tractor-mounted Planet Jr. planters. Broccoli was thinned to about 10 inches between plants during the last week of July. Diazinon was applied at 2.0 pounds/acre for maggot control in early August. In 1993, eighty pounds N/acre, as ammonium nitrate, was sidedressed on 17 August. In 1994, 100 pounds N/acre, as urea, was sidedressed on 2 August and an additional 50 pounds N/acre was sidedressed on 1 September.
Plots were irrigated uniformly as needed, about 4 cm/week, until 17 September, 1993 (8 September, 1994) when the irrigation treatments were initiated. Most plants had 1-inch diameter heads at this date. The irrigation system was constructed such that each of the treatment combinations was controlled by a separate valve with flexible hose used to connect the four plots making up the replicates of each treatment. Water was applied from a sprinkler located at each end of the main plots. Main plots were isolated from each other by 30 feet in all directions. Sprinklers were set to water a 180-degree pattern.
A spontaneous mutant of Erwinia carotovora subsp. carotovora (EccW3C105) was obtained from J.E. Loper, USDA, Corvallis, OR and selected for field trials. It was resistant to 100 g/mL nalidixic acid and phenotypically similar to the parental strain. Lawns of the Erwinia isolate were grown on Lauria Bertani (LBnal) media (10 g Bacto tryptone, 5 g Bacto yeast extract, 10 g NaCl, 20 g Bacto agar in 1.0 liter of distilled H2O with the addition of 100 g/mL nalidixic acid) for 3 days. The bacteria were washed from plates and the concentration of inoculum was adjusted to 107 cells/mL. On 16 and 23 September at 8 am and 7 pm, respectively, plots were sprayed with the bacterial suspension in 25 gal water/acre using a CO2-pressurized boom sprayer with three flat fan nozzles.
To determine the survivability of Erwinia over time, broccoli florets were sampled every other day at 9:00 am from 17 September to 8 October, 1993. Additionally, florets were sampled two hours after the first bacterial application and the morning following the second application. A five-foot section across each main plot (approximately 48 plants) was used for sampling. Five samples (approximately 0.2 g of floret tissue) per head were collected from three heads in each plot. Individual samples were placed in separate test tubes containing 2.24 mL of sterile potassium phosphate buffer (ph 6.5, 0.05 M). Test tubes were agitated and then racks of test tubes were sonicated for 5 min. After agitating again, a 0.01 ml aliquot of the wash and a 1:22.4 dilution of the wash were each plated on separate halves of plates filled with LBnal medium. Minimum detection limits for bacteria in the wash and the dilution were 2.24 x 102 and 5.0 x 103, respectively. Bacterial populations were counted three days after plating.
The duration of the irrigation treatments, one-half, one, or two hours, was intended to deliver 1 cm of water for the 2-day frequency, 2 cm for the 4-day, and 4 cm for the 8-day frequency, respectively. Rain gauges were located in the center of each main plot; actual water accumulation was measured immediately after each irrigation. Preliminary measurements at less than 1 knot winds indicated excellent uniformity of water application (less than 0.1 cm variation) both among plots of the same treatment and within main plots. The half-hour irrigations commenced at 8:30 a.m. (morning) or 6:30 p.m. (evening). The longer irrigations commenced at correspondingly earlier times, with all irrigations terminating at 9 a.m. or 7 p.m.
Soil moisture content of the surface 15 cm of the soil profile, as a percentage, was measured by time-domain reflectometry at about 2 p.m. each day from 17 September until 1 October, 1993 and on 19 and 21 September, 1994. Leaf wetness duration was measured with a Campbell Scientific 21X micrologger (Campbell Scientific, Logan, UT) in four plots representing each treatment. Four replicate sensors attached to ringstands were placed in each plot in the center of the canopy. Sensors were sampled at 30-minute intervals and wetness duration was computed as the percentage of the interval that the sensor was wet.
Visual disease ratings were made at four-day intervals in a three-row subplot in each plot. Heads were harvested from the appropriate subplots on 28 September, and 4 and 8 October, 1993 (19 and 29 September, 1994), counted, weighed, and checked for disease incidence.
Results and Discussion
In 1993, mean weight of broccoli heads at the first harvest was greater for the 2-day irrigation interval than for the 8-day interval and tended to be greater for the morning as opposed to evening irrigation (Table 44). At the second harvest mean head weight was nearly identical between frequencies, but was greater for the evening irrigation (Table 45). Most of these differences appeared to be due to delayed maturity and small heads for one replicate of the 8-day x morning treatment combination. The same trend was not observed at the third harvest (Table 46) and mean head weight did not vary with treatment for the sum of all harvests (Table 47). Since there were no significant interactions of irrigation frequency and timing affecting yield or rot incidence, only main effects have been presented in the tables.
The incidence of head rot in 1993 was very low, less than 10 percent. Irrigation treatment had no effect on rot incidence at the first harvest, but there was a greater incidence of rot with the evening irrigations at the second harvest. At the third harvest, the 2-day interval produced more rotten heads and, although not significant, the evening irrigation still tended to produce more rot than did the morning irrigation. For the sum of all harvests, only the effect of irrigation interval was statistically significant. In the observational sub/plots that were not harvested, only three heads were observed with head rot on 4 October. Four days later, on 8 October, the number and percentage of heads with rot was significantly greater in the 2-day irrigation interval treatment compared with the 8-day interval. The incidence of head rot did not differ significantly between morning and evening irrigations, although incidence was greater in the evening-irrigated plots.
Populations of the Erwinia mutant were recovered from all of the heads sampled two hours after the first inoculation and varied from 102 to 105 colony forming units (cfu)/g fresh weight of tissue. Two and four days later, colonies were recovered from only 2 and 3 plots, respectively. After the second inoculation, bacteria were recovered from most of the heads with populations ranging from 102 to 105. Recovered Erwinia populations immediately began to decline. Two, four and six days later, bacteria were recovered from 29, 12 and 10% of the heads, respectively. No bacteria were recovered after 30 September. Several possible reasons for the sharp decline are proposed. Environmental conditions were unfavorable to establishment and survival of the bacteria at the time of and immediately following application. Weather conditions prevalent during the sampling period were characterized by dry and warm days. Bacteria applied to aerial portions of plants are susceptible to desiccation. The ability of the Erwinia mutant to survive in situ has not been fully investigated. Due to the poor survival of the applied mutant, an analysis of treatment effect on survival of bacteria could not be made.
Several bacteria were isolated from diseased heads, including several Psuedomonads and undetermined isolates, but no Erwinia. These bacteria will be tested for their ability to cause head rot on broccoli.
Total precipitation between seeding and initiation of the irrigation treatments was 8 cm in 1993. During the irrigation treatments only 0.15 cm precipitation was recorded and probably had little influence on the treatments. The weather was unusually warm and dry during most of the treatment period. However, temperatures returned to normal seasonal averages during the last week of the experiment, with mostly cloudy skies.
Averaged over the length of the treatments, soil moisture did not vary with treatment if only the measurements taken on the no-irrigation days are used (Table 48). Since the soil moisture readings were taken midway between the morning and evening irrigations, the readings were greater for morning than for the evening-irrigated plots, if measured on the day of irrigation (data not shown). Soil moisture for the 2-day irrigation interval tended to be fairly constant over time, although there was an increase late in the experiment, perhaps related to greater cloud cover, reduced temperatures, and reduced evapotranspiration (Fig. 6). For the eight-day interval, soil moisture was much less constant. Soil moisture readings were terminated on 1 October because of instrument failure.
The lack of a consistent treatment effect on yield, and the lack of effect of treatment on soil moisture and average amount of water applied to each plot, indicate that the effects of treatment on head rot incidence are due to frequency or timing of irrigations rather than to the amount of water applied. Although timing and frequency did not affect the total amount of water applied or average soil moisture, it did affect leaf wetness duration. Mean hours of leaf wetness per day over the course of the study was greater in the two-day irrigation interval compared to the eight-day irrigation interval. The effect of timing was not significant, although plots irrigated in the evening tended to have greater average daily hours of leaf wetness (Table 49). Across all treatments, average daily hours of leaf wetness increased as the season progressed. The last three days of September, leaf wetness averaged 12.9 hours per day and by the final three days of the study leaf wetness averaged 19.9 hours per day.
In 1994, the amount of water (0.68 cm/day) applied to the plots during the treatment period slightly exceeded the goal of 0.5 cm/day. However, this included two occurrences of precipitation, averaging 0.07 cm/day. The amount of water applied did not vary significantly with treatment. Mean soil moisture also did not vary with treatment (Table 50).
Mean weight of broccoli heads and yield per unit area in 1994 did not vary with either frequency or timing of irrigation at either harvest (Tables 51 and 52). The first harvest occurred during a period of unusually hot, dry weather and no rot was observed. Since there were no significant interactions of irrigation frequency and timing affecting yield or rot incidence, only main effects have been presented in the tables.
The incidence of head rot was fairly high at the second harvest, averaging 22 percent for all treatments (Table 52). As in 1993, incidence of rot decreased with increasing interval between irrigations but timing of the irrigation had no effect.
Taken together, the 1993 and 1994 results support the hypothesis that development of disease is related to the availability and duration of free moisture. The data also provide evidence that irrigation practices may be useful in disease control. Avoiding frequent irrigation and timing irrigations to allow rapid drying of plants may reduce rot incidence.
Table 44. Main effects of frequency and timing of irrigation on yield and head rot incidence in broccoli at first harvest, NWREC, 1993 Treatment No. heads harvested Yield Mean head No. rotten per plot kg/plot wt. (g) heads per plot Frequency 2-day interval 50.0 11.5 231 0.2 8-day interval 45.7 8.2 182 0.0 Significance NS NS * NS Timing morning 46.5 10.2 220 0.0 evening 49.2 9.4 194 0.2 Significance NS NS NS NS Table 45. Main effects of frequency and timing of irrigation on yield and head rot incidence in broccoli at second harvest, NWREC, 1993 Treatment No. heads harvested Yield Mean head No. rotten per plot kg/plot wt. (g) heads per plot Frequency 2-day interval 61.1 18.1 296 1.9 8-day interval 60.8 18.2 299 0.5 Significance NS NS NS NS Timing morning 62.9 17.1 272 0.3 evening 59.0 19.2 325 2.2 Significance NS NS * * Table 46. Main effects of frequency and timing of irrigation on yield and head rot incidence in broccoli at third harvest, NWREC, 1993 Treatment No. heads harvested Yield Mean head No. rotten per plot kg/plot wt. (g) heads per plot Frequency 2-day interval 56.0 23.2 415 12.3 8-day interval 54.5 22.1 403 5.5 Significance NS NS NS * Timing morning 59.0 23.9 409 8.1 evening 51.5 21.4 410 9.7 Significance NS NS NS NS Table 47. Main effects of frequency and timing of irrigation on yield and head rot incidence in broccoli, sum of all harvests, NWREC, 1993 Treatment No. heads harvested Yield Mean head No. rotten per plot kg/plot wt. (g) heads per plot Frequency 2-day interval 167.1 52.7 317 14.4 8-day interval 161.0 48.5 300 6.2 Significance NS NS NS ** Timing morning 168.4 51.2 303 8.6 evening 159.7 50.1 313 12.0 Significance NS NS NS NS Table 48. Main effects of frequency and timing of irrigation on the mean amount of water applied per day and on the mean percent soil moisture during the course of the experiment, NWREC, 1993 Treatment Water applied, cm/day Soil moisture (%) Frequency 2-day interval 0.51 14.9 8-day interval 0.51 14.3 Significance NS NS Timing morning 0.53 14.0 evening 0.50 15.1 Significance NS NS Table 49. Main effects of frequency and timing of irrigation on mean daily hours of leaf wetness and the number and percentage of heads with rot on 8 Oct, 1993, NWREC Treatment Rotten headsz Rotten heads Leaf wetness Number/plot % mean daily hours Frequency 2-day interval 5.7 11.4 15.1 8-day interval 2.3 4.4 13.0 Significance ** ** * Timing morning 3.2 6.5 13.1 evening 5.4 10.4 14.9 Significance NS NS NS zNon-harvested heads from observation strips in each main plot. Table 50. Main effects of frequency and timing of irrigation on the mean amount of water applied per day and on the mean percent soil moisture during the course of the experiment, NWREC, 1994 Treatment Water appliedz Soil moisture cm/day (%) Frequency 2-day interval 0.66 17.4 4-day 0.70 15.7 8-day 0.70 18.9 Significance NS NS Timing morning 0.72 18.6 evening 0.63 16.0 Significance NS NS zIncludes 1.5 cm precipitation during the course of the experiment. Table 51. Main effects of frequency and timing of irrigation on yield and head rot incidence in broccoli at first harvest, NWREC, 1994 Treatment No. heads harvested Yield Mean head No. rotten per plot kg/plot wt. (g) heads per plot Frequency 2-day interval 24.5 6.4 268 0 4-day 26.7 7.5 279 0 8-day 24.0 6.4 273 0 Significance NS NS NS NS Timing morning 27.1 7.4 278 0 evening 23.1 6.1 269 0 Significance NS NS NS NS Table 52. Main effects of frequency and timing of irrigation on yield and head rot incidence in broccoli at second harvest, NWREC, 1994 Treatment No. heads harvested Yield Mean head No. (%) rotten per plot kg/plot wt. (g) heads per plot Frequency 2-day interval 62.1 20.4 330 18.9 (30.4) 4-day 60.4 22.7 373 11.1 (18.3) 8-day 58.0 19.8 346 8.9 (15.3) Significance NS NS NS ** ** Timing morning 59.9 20.9 354 13.0 (21.7) evening 60.4 21.0 346 12.9 (21.4) Significance NS NS NS NS NS