Implementing Sustainable Management of Muskmelon Diseases, Weeds, and Insect Pests in Partnership with Iowa Growers
The goal of our three-year, multidisciplinary research project is to enhance sustainability of muskmelon disease, weed, and insect pest management. Muskmelon, a key high-value crop in the North Central Region, relies heavily on costly synthetic pesticides and fertilizers that endanger grower and consumer health, kill non-target organisms, and pollute the environment. We will investigate innovative tactics that can reduce or replace these toxic inputs.
Field trials will use mass trapping augmented by trap crops, row covers, neem, and kaolin clay to deter bacterial wilt transmission and feeding injury by cucumber beetles. Another field experiment will evaluate control of anthracnose, gummy stem blight, and Alternaria leaf blight by Bacillus subtilis, potassium bicarbonate, and the Melcast disease-warning system, which allows growers to time fungicide sprays efficiently. We will also assess suitability of a hairy vetch-winter rye mulch to replace synthetic herbicides and fertilizers in managing weeds and N fertility. We will integrate the most promising new strategies into systems-level field trials and determine input costs, profitability, and risks associated with all strategies. Growers will learn about the new methods through eight on-farm trials per year, field days, a World Wide Web site, newsletter articles, press releases, and an Extension bulletin.
Muskmelon growers are full partners in the project from start to finish. Growers have provided feedback on the proposal, and a 10-grower Advisory Panel will help to design field trials, interpret findings, and prioritize education efforts.
Short- and intermediate-term outcomes include identifying practical management tactics that reduce reliance on synthetic chemical inputs, participation of 15 growers in on-farm trials, and (as determined by Year 1 and Year 3 surveys) adoption of one or more of the new methods by 50 growers and increased willingness of 100 growers to adopt these methods. Long-term outcomes include higher profits for growers, reduced environmental pollution, and healthier and more stable rural communities.
1) Suppress bacterial wilt and cucumber beetle feeding injury on muskmelons as effectively as conventional insecticide-based methods by integrating mass trapping, trap crops, row covers, neem, and kaolin clay.
2) Validate ability of a) the Melcast disease-warning system using site-specific weather estimates as inputs, b) Bacillus subtilis sprays, and c) potassium bicarbonate sprays to reduce or replace conventional fungicide sprays for control of muskmelon anthracnose, gummy stem blight, and Alternaria leaf blight.
3) Evaluate ability of a hairy vetch – rye cover crop to suppress weeds, reduce applications of conventional herbicides, and add nitrogen to soils.
4) Combine the most promising component strategies from Objectives 1 to 3 into systems-level strategies that integrate weed, insect, and disease management.
5) Document economic costs and profitability of these new management tactics in comparison to conventional practices.
6) Transfer the project’s findings to North Central Region muskmelon growers through on-farm demonstration trials with eight growers per year, a World Wide Web site, quarterly newsletter articles, presentations at regional grower meetings, annual field days, annual press releases, and an Extension bulletin.
Our advisory panel includes the following growers: Ray Jensen (Greenfield), Laura Krouse (Mt. Vernon), Greg Hoffman (Waterloo), Michael Nash and Solveig Hanson (Postville), Richard and Sharon Dix (Janesville), Gary Guthrie (Nevada), Dean Henry (Nevada), John Kiwala (Muscatine), Bob Furleigh (Clear Lake), and a group of Amish farmers from Davis County. We met in Ames on December 15 to discuss the results of last year’s experiments and plans for next year. The meeting allowed the growers to ask questions about our results, and brought to light several ideas for next year’s project. We also cooperated with several growers in Iowa on strategies for cucurbit pest management during the 2003 field season.
We also cooperated with several growers in Iowa on strategies for cucurbit pest management:
Phil and Nancy Funk (Dallas Center, IA) and Richard, Bill, and Sharon Dix (Janesville, IA) managed one row of muskmelons at each of their farms according to the Melcast system. In each of these trials, ISU personnel managed the weather data, and worked with each grower to determine spray timing. The Melcast system saved them at least one or two sprays this season, without an increase in anthracnose symptoms. Both growers were very happy with this strategy. They are interested in learning to manage the sprays on their own using the Melcast computer program.
Gary Guthrie (Nevada, IA) and Ray Jensen (Henry Wallace Country Life Center, Greenfield, IA) evaluated row covers for control of bacterial wilt. At both places, one of two 100-foot rows was covered with Reemay row cover until bloom. The treated row sections were compared with the non-covered row section. ISU scouts visited every week to assess wilt and count striped and spotted beetles on a sticky card in each plot. At the Guthrie farm, there were both striped and spotted beetles this summer. Beetle counts were taken five times this summer and averaged. There were more beetles found in the non-covered row (11.2 striped and 24.8 spotted) than the covered row (7.4 striped and 15.2 spotted), however, there was no bacterial wilt in either row. There were very few beetles found at the Henry Wallace Farm (an average of 0-3 per week in both rows), yet there were 13.3% wilted plants in the non-covered row and 6.7% wilted plants in the covered row.
Greg Hoffman (Waterloo, IA) treated one row of his muskmelon plants with a single application of the systemic insecticide Admire at planting. He also planted one row of melons that was treated with Gaucho (same active ingredient as Admire). We visited three times to look for beetles and signs of bacterial wilt. There were more beetles found in the ‘Gaucho’ row (average of 48.3 striped and 6.3 spotted) than the ‘Admire’ row (17.3 striped and 2.3 spotted). There were 7.3% of the plants wilted in the ‘Gaucho’ row compared to 7.3% in the ‘Admire’ row.
OBJECTIVE 1: Suppress bacterial wilt and cucumber beetle feeding injury on muskmelons as effectively as conventional insecticide-based method by integrating mass trapping, trap crops, row covers, neem, and kaolin clay.
Two sets of experiments evaluating several strategies for managing cucumber beetles were established in June 2004 at three university farms in Iowa. These locations include the ISU Horticulture Research Station in central Iowa, the Armstrong Research and Development Farm in Southwest Iowa, and the Muscatine Island Research Station in Southeast Iowa.
In one trial, five insecticide treatments were applied to 20-foot rows of ‘Athena’ in four replications. These treatments were applied to two fields at each farm, one of which was covered with ‘Reemay’ spun cotton row covers at the time of planting. The row covers were removed at bloom. Insecticide applications were made every two weeks. They began at the time of planting in the uncovered field and at bloom in the covered field. The evaluated insecticides were 1) Entrust (an organically approved spinosad insecticide), 2) Admire (a single application of the systemic insecticide Imidacloprid), 3) Admire applied at planting + Sevin or Capture applied after flowering, and 4) Invite (a cucurbitacin-impregnated compound that is highly attractive to cucumber beetles) + Sevin or Capture. A conventional control (Sevin alternated with Capture) and a non-treated plot were also included in the trial. Striped and spotted beetle populations on five plants per plot, as well as the number of plants with bacterial wilt, were recorded each week. The number and weight of melons harvested per plot was measured twice per week at maturity.
There was a clear benefit to covering the plots with Reemay, as it significantly increased both the number and weight of melons harvested compared to the uncovered. Some of the yield benefits of these covers may have been due to a warming effect in the spring, but they also decreased or delayed the incidence of bacterial wilt, indicating that they protected the plants from cucumber beetles.
Of the insecticide treatments evaluated, all resulted in yields greater than the non-treated control. This is mostly from the increase in marketable fruit. There was no correlation between yield and either beetle population. There was a significant negative correlation between bacterial wilt incidence (at the later dates) and marketable melon number and weight (P<0.001).
In a second trial, the feasibility of a trap crop strategy was investigated. The trap crop used was ‘Black Beauty’, a zucchini cultivar that that is highly attractive to cucumber beetles. It is hypothesized that the melons will be ignored by cucumber beetles when a more attractive alternative (‘Black Beauty’ zucchini) is available to the beetles. The trap crop selection is different from 2003 (Turk’s Turban). Problems from 2003 were that the trap crop wilted early in the season, leaving the fields without their trap crop. The ‘Black Beauty’ zucchini is still highly attractive to the beetles, yet has some resistance to bacterial wilt (according to a source from Connecticut).
At two locations in Iowa, three small fields of ‘Athena’ muskmelons were established at least 1000 feet from one another to evaluate the efficacy of ‘Black Beauty’ zucchini as a trap crop alone and in combination with an insecticide. In two of the fields, a row of zucchini was established between every five rows of muskmelons. Admire (imidacloprid) at planting, and Capture (bifenthrin) or Sevin (carbaryl) applied weekly were the insecticides for the treated zucchini field. As a control, only muskmelons were planted in the third field. Striped and spotted beetle populations on five plants per row, as well as the number of plants with bacterial wilt, were recorded each week. The number and weight of melons harvested per plot was measured at maturity.
Overall, more striped and spotted beetles visited fields with treated zucchini than fields with just melons or non-treated zucchini (P<0.05). This means that some of the trap crop fields are a regional attractant to beetles. However, having more beetles in a trap cropped field is only a problem if they move from zucchini plants to melon plants. The trap crop may attract more beetles to the field, and doesn’t distract all beetles from the melon plants. Consequently, there was the same average number of wilted plants in melon rows from fields with and without the trap crop.
We hoped to slow the dispersal of beetles from gourd plants to melon plants by making weekly sprays of insecticide in the trap crop in one field. These sprays did not reduce the average number of beetles or wilted plants in melon rows when compared to non-sprayed rows. It appears, from average beetle counts and wilt ratings, that a zucchini trap crop, sprayed or unsprayed, does not decrease the incidence of bacterial wilt in muskmelon.
Yield data from these three fields, however, show that the field without a trap crop and the insecticide-treated trap crop field produced a fewer melons and total weight of melons than did the non-sprayed trap crop field. This is exactly opposite of the data from 2003. Also, there was no correlation between bacterial wilt development and beetle population dynamics this season (P<0.05).
From the two years of data, most of the fields with a trap crop had more beetles in the trap than in the melons. However, this did not decrease the number of beetles in the melon rows. Furthermore, the trap crop did not have any influence on the number of wilted plants either year. As for marketable melons, there were conflicting results from year to year; suggesting that more than beetles, bacterial wilt, and trap crops influence the number and weight of melons.
OBJECTIVE 2: Test alternative strategies against the fungal disease complex of anthracnose, gummy stem blight, and Alternaria leaf blight.
This trial was located in central and southeast Iowa. Seedlings of ‘Athena’ muskmelon were transplanted into black plastic in a randomized complete block with eight treatments and four replications. Plots were 25-ft-long rows with 2-ft spacing between plants and 8 ft between row centers. Treatment rows alternated with guard rows. Treatments included Bravo Ultrex (chlorothalonil) applied at 7 and 14-day intervals, as well as Serenade (Bacillus subtilis) and Kaligreen (potassium bicarbonate) applied at 7-day intervals. In addition, three treatments were sprayed according to the Melcast disease-warning system: a treatment based on weather data collected from a CR10 weather data logger located on-site, another treatment based on remotely-estimated SkyBit weather data, and a final treatment based on SkyBit weather data that was corrected using models developed in previous years of the study.
All rows were inoculated on 13 July in Muscatine and 14 July in Ames with a virulent strain of Colletotrichum orbiculare isolated from melon plants in 2003. Immediately prior to inoculation, all treatments except the non-sprayed control were applied with a backpack sprayer for the first time when vines first touched between rows. Subsequent sprays were applied either on a calendar-based schedule or according to the Melcast system. Percentage of foliage covered with anthracnose lesions was measured using Horsfal-Barrett’s scale, and the weight and number of marketable and cull melons was determined at harvest. Disease data were converted to midpoints prior to analysis.
At Ames, the Melcast disease-warning system suggested only two sprays during the season for the on-site weather data, the SkyBit weather data, and the corrected SkyBit data. Scheduled sprays called for 3 to 6 sprays, indicating that the Melcast system used with any source of weather data saved several sprays. Compared to the non-sprayed control, all treatments had significantly lower disease. However, this did not result in an increase in the number or weight of marketable melons.
At Muscatine, the Melcast disease-warning system suggested only 3 sprays during the season for the SkyBit weather data, the on-site and the corrected SkyBit data. Scheduled spray treatments called for 3 to 5 sprays, indicating that the Melcast system used with any source of weather data saved two sprays compared to the 7-day spray schedule. The final percent of foliage infected was approximately equal in all of the treatments, and was extremely high throughout the field. The non-sprayed control and the plots treated with Kaligreen had significantly higher final percent disease than chlorothalonil sprayed every 7 days. The number of marketable melons was approximately equal among treatments, though the weight of melons was significantly lower in the non-treated plots compared to most of the treated plots.
Overall, all treatments had significantly less disease at both locations. However, at Muscatine, none of the fungicide treatments were effective at controlling disease. Many of the plants died prematurely. The Melcast system did save a few sprays at Ames, but it is unclear how often sprays were needed in Muscatine for adequate disease control. A few differences from 2003 were that a more virulent isolate was used for inoculations and all rows were inoculated, instead of guard rows only.
A second goal of objective two for summer 2003 and 2004 was to collect data for use in better estimating leaf wetness. Weather stations, including thermometers, pyranometers, relative humidity sensors, rain gauges, wind speed sensors, and leaf wetness sensors were established in central and southeastern Iowa. Leaf wetness sensors were placed at various heights in and outside of the muskmelon canopy. Data were measured hourly for all instruments, and files were downloaded from the weather station each week. Remotely-estimated SkyBit data were ordered from June 1 to Sept 15 for each of these locations. Leaf wetness sensors were also placed at 5 established CR-10 stations across Iowa. SkyBit data for each of these locations was collected as well. Cloud cover data were obtained from the ISU Mesonet, http://wepp.mesonet.agron.iastate.edu/GIS/rainfall.phtml. Data from all of the Iowa sites are being compiled and modeling efforts are underway.
OBJECTIVE 3: Evaluate the ability of a hairy vetch-rye cover crop to suppress weeds, reduce applications of conventional herbicides, and add nitrogen to soils.
Four weed management strategies were tested in Ames and Muscatine, Iowa in 2004. These management strategies are similar to those in 2003. These strategies included conventional mulch or cover crop mulch combined with either “weed-free” hand weeding or “weed-management” hand weeding. Treatments include all possible combinations of conventional mulch, cover crop mulch, “weed-free” hand weeding, and “weed-management” hand weeding. The objective of “weed management” hand weeding is to simulate the realistic (i.e. incomplete) on-farm weeding effort required to keep weed competition in check and to limit increases in the soil weed seed bank.
Hairy vetch and rye were seeded in eight 30’ x 32’ plots at each location in September 2003. In mid-June of 2004, when the hairy vetch was flowering, the plots were crushed using a cultipacker. Four rows were marked out in each plot and ‘Athena’ muskmelon seedlings were planted into the mulch. Another eight “conventional” plots were established at the same time using pre-emergence herbicide and black plastic mulch. The conventional plots received tillage as needed until the vines touched between the rows. One-m2 weed samples were collected at the time of planting in each plot, and weeds were separated by species and weighed. This type of sampling was repeated in early August and at the time of harvest. In “weed-free” plots, weeds were removed weekly, and in “weed-management” plots, weeds were removed at 3, 6, and 9 weeks after planting. The person-hours spent weeding and the species and weight of weeds removed during each weeding event were recorded. The number and weight of melons harvested from each plot was measured at maturity.
In Ames, we were quickly short of man-hours when it came time to weed the vetch/rye plots. Weed pressure was so high that it quickly became unrealistic to count and weigh weed samples from the whole plot. The high weed pressure resulted in very few melons and no marketable ones. This is very different from last year (a very dry year with less weed pressure) where there were no differences between cover crop and black plastic mulch plots. In Muscatine, we also found that cover cropped plots had lower yields than plastic mulched plots (P<0.0001; Table 6). The weed pressure was not as bad in Muscatine. However, the rye that was knocked down in June started to re-seed and became established.
Overall, half of the field locations had a significant decline in yield compared to the black plastic mulch. Even in the years with comparable yields, there was a huge delay in harvest time from delaying planting until mid-June.
OBJECTIVE 4: Combine the most promising component strategies from Objectives 1-3 into systems-level strategies that integrate weed, insect, and disease management.
Most of the trials in 2004 were repeated from 2003. Trials in 2005 will be combining the most promising strategies.
OBJECTIVE 5: Document economic costs and benefits of these new management tactics in comparison to conventional practices.
Economic analyses are ongoing.
OBJECTIVE 6: Transfer the project’s findings to muskmelon growers through on-farm demonstration trials, a WWW site, quarterly newsletter articles, presentations at regional grower meetings, annual field days, press releases, and an extension bulletin.
Although newsletter articles and press releases will not be written until our experiments are complete, we have prepared an introductory WWW site (http://www.public.iastate.edu/~cucurbitipm/). It includes background information on the biology of cucumber beetles, anthracnose, gummy stem blight, Alternaria leaf blight, and common weeds in Iowa, Minnesota, and Colorado, as well as some basic management strategies. We are currently adding the results of our 2003 and 2004 experiments.
We have also completed a web-based extension bulletin, “Melcast: A Weather-based Disease Warning System for Muskmelon Anthracnose”, that describes the system and publicizes a newly-developed Excel computer program that growers can download from our website to help them use the Melcast system. This computer program will be modified at the end of these experiments to incorporate an improved leaf wetness model developed from ongoing modeling efforts.
Furthermore, we presented our research to approximately 75 growers at an Iowa Fruit and Vegetable Grower’s field day in July, as well as at Practical Farmer’s of Iowa Field Days in July.
Impacts and Contributions/Outcomes
Our impacts and contributions are listed in the previous section.
To summarize: The cover crop for weed control did not preform as well as hoped. Rainfall in 2004 was consistent and led to great weed pressure. The rain also delayed planting so melons did not have time to mature.
The trap crop was changed from 2003 from Turks Turban to a zucchini. This did not change our results. The trap crop did attract more insects, but it did not lower beetle numbers in the melons. The trap crop did not positively affect melon yield either.
The row covers were considered a success. At all three locations, there was an increase in yield and a delay in wilt when using row covers. This increase in yield may be from early season warming as well as delaying and reducing bacterial wilt.
The Melcast disease model was successful in one location, but it had no effect in the other. The inoculum used probably overwhelmed any fungicides used at the unsuccessful location.
From the past two years of studying many different possible techniques, the two successful ones were row covers with insecticides (Admire) and fungicides applied using Melcast. In 2005, a combination of row covers, Admire fungicide, and fungicide applications timed using Melcast will be completed.