Integrating Perimeter Trap Crops and Row Covers into Cucurbit-crop Farming Systems

Final Report for LNC13-350

Project Type: Research and Education
Funds awarded in 2013: $199,250.00
Projected End Date: 12/31/2016
Grant Recipient: Iowa State University
Region: North Central
State: Iowa
Project Coordinator:
Dr. Mark Gleason
Iowa State University
Expand All

Project Information


This project’s goal was to modify perimeter trap cropping and row covers for managing bacterial wilt and cucumber beetles in conventional and organic muskmelon production, respectively. Field experiments in Iowa and Ohio focused on perimeter trap cropping (PTC) in conventional systems and delayed row-cover removal (DRCR) – that is, delaying the removal of row covers by 10 days after the start of flowering – in organic systems. In replicated PTC experiments in each state during 2014 and 2015, muskmelon (cv. Athena) was the main crop; the experiment compared performance with vs. without a 2-row perimeter of Buttercup squash (cv. Space Station). Results were similar in both states and between years: PTC did not consistently reduce the number of insecticide sprays on muskmelon, nor did it consistently suppress incidence of bacterial wilt or improve marketable yield compared to the non-PTC control treatment. These results suggest that PTC is not a promising option for conventional muskmelon growers in the Midwest. Spunbond polypropylene row covers, deployed over ‘Athena’ muskmelon on wire hoops at 18-inch height in single rows from transplanting until 10 days after female flowers began to bloom, did not result in significantly higher marketable yield or lower incidence of bacterial wilt than non-covered plots, and in some site-years the row cover treatment even had significantly more non-marketable fruit, although it did require fewer insecticide sprays. In on-farm demonstration trials, row covers neither consistently increased marketable yield nor decreased bacterial wilt incidence compared to a non-covered control treatment. Our findings suggest that the row cover strategy used in this project does not offer an advantage for Midwest organic growers of muskmelon. Project activities were shared with growers during field days and winter meetings in each state.


Bacterial wilt devastates cucurbit crops and threatens profitability of North Central Region vegetable farms. To control the cucumber beetles that spread the bacterium, growers rely mainly on insecticides. Alternative tactics are urgently needed because the widely used neonicotinoid insecticides can harm pollinators and organic insecticides are ineffective. Our two-state project sought to develop practical strategies to minimize insecticide use and boost profits. We based this project on the encouraging results of our previous NCR-SARE project (LNC10—323) by scaling up our alternative management strategies – perimeter trap cropping and delayed removal of spunbond row covers – to NC Region farming systems. In two years of field experiments (2014 and 2015) on muskmelon in Iowa and Ohio, we scaled up perimeter trap cropping in conventional production to realistic farm size, and make delayed removal of row covers more reliable for organic growers by validating strategies to protect the crop after the row covers come off. We also evaluated these modified approaches with collaborating growers during on-farm demonstration trials. We shared our findings with NC Region growers through field days and regional grower-conference talks.

Project Objectives:
  1. In a conventional muskmelon production system, evaluate perimeter trap cropping for bacterial wilt suppression and yield at a spatial scale that is representative of North Central Region vegetable farms, and without neonicotinoid insecticides.
  1. In an organic muskmelon production system using delayed row cover removal, evaluate alternative strategies for managing bacterial wilt and cucumber beetles during the period between row cover removal and harvest.
  1. Estimate costs and profits of the modified alternative IPM strategies in Objectives 1 and 2.
  1. Share project results with cucurbit growers throughout the region by means of on-farm demonstration trials, virtual and on-site field days, extension bulletins, webinars, and regional meeting presentations.



Click linked name(s) to expand/collapse or show everyone's info
  • Jean Batzer
  • Mary Gardiner
  • Sally and Luke Gran
  • Laura Jesse
  • Susan Jutz
  • Donald Lewis
  • Jan Libbey
  • Sally Miller
  • Ben Saunders
  • Celeste Welty


Materials and methods:

Objective 1: In a conventional muskmelon production system, evaluate perimeter trap cropping (PTC) for bacterial wilt suppression and yield at a spatial scale that is representative of North Central Region vegetable farms, and without neonicotinoid insecticides.

On university research farms in Iowa and Ohio, four replications of each of two subplots (PTC vs. No PTC) were separated by at least 500 ft to avoid interplot interference.  PTC subplots (200 x 42 ft) were surrounded by two rows of buttercup squash (cv. Space Station), whereas non-PTC plots had annual rye planted as a border. Paired PTC and non-PTC subplots were separated by at least 500 ft to minimize interplot interference by pest insects. Populations of cucumber beetles were monitored weekly in both border rows and main-crop rows along four transects within each plot.  Synthetic pyrethroid insecticides (Asana XL or Pounce) were sprayed on the squash border rows or main crop muskmelons when threshold numbers of cucumber beetles were reached. Threshold numbers for cucumber beetles varied according to melon plant size as follows: pre-flowering = 0.5/plant; during fruit pollination = 1.0/plant; and at vine touch = 3.0/plant. Bacterial wilt incidence (% wilted plants) was recorded one week before harvest. Harvested fruit were assessed for marketability, then counted and weighed.

Objective 2: In an organic muskmelon production system using delayed row cover removal, evaluate alternative strategies for managing bacterial wilt and cucumber beetles during the period between row cover removal and harvest.

Iowa. Five-week-old organic transplants of ‘Athena’ muskmelon were planted 2 ft apart in black plastic mulch with drip irrigation on 7-ft centers. Subplots consisted of 30-ft-long rows of 15 plants. Spunbond polypropylene row covers (Agribon AG-30) were installed on wire hoops immediately after transplanting. Weed management was achieved with 6 inches of corn stalk mulch between rows. A 3x2 factorial experiment (3 row cover treatments x 2 insecticide treatments) was conducted in a randomized complete block design with 4 replicate subplots per treatment. Row cover treatments included:  1) No row covers (NRC) 2) Row covers applied at transplanting and removed at anthesis (when female flowers start to open) (RC) and  3) Row covers applied at transplanting with the ends opened at anthesis and removed 10 days later (DRC).  The two insecticide regimes were: a) Surround (kaolin clay) applied to plants and reapplied after removal by rainfall, and b) Surround was applied as the previous treatment, but Pyganic EC (pyrethrin) and Trilogy (neem oil) were also applied when cucumber beetle thresholds exceeded the following thresholds: 0.5 beetle/plant before anthesis, 3 beetles/plant from anthesis to vine touch (vines in adjacent rows start to grow together), and 10 beetles/ plant from vine touch to harvest.  Insecticides were not applied while plants were under row covers. Striped and spotted cucumber beetle adults were counted weekly from transplant through the beginning of harvest using a) yellow sticky cards and b) visual monitoring of 3 randomly chosen plants per subplot.  Disease incidence was monitored weekly, and the number and weight of marketable and cull melons harvested from each subplot were recorded. Ohio. The 2014 and 2015 trials with ‘Athena’ muskmelon at an OSU research farm in Columbus had 6 treatments and 4 replications; treatments were a 3 x 2 factorial design combining row cover treatments (no row cover, and extended-duration row cover) and insecticides (spinosad plus cucurbitacin (CideTrak-D); kaolin (Surround WP); and no insecticide). The cucumber beetle population was assessed by examining three randomly chosen plants per plot, once per week. If the beetle threshold (1 beetle per plant) was exceeded, insecticides were applied. Fruit were counted, weighed, and rated for marketability at harvest. Marketable fruit were those with size greater than 2 lb, uniform netting, and absence of physical damage or rots.

Objective 3. Estimate costs and profits of the modified alternative IPM strategies in Objectives 1 and 2.We planned to calculate a partial budget (Calkins and Di Pietre, 1983) annually from results of the perimeter trap cropping and delayed row cover removal experiments. Returns were to be calculated based on local retail prices for conventional muskmelon and ‘Buttercup’ squash (Objective 1) and organic muskmelon (Objective 2) in Iowa and Ohio.

Objective 4. Iowa. Cooperators in on-farm trials were Jan Libbey (One Step at a Time Farm, Kanawha), Susan Jutz (ZJ Farm, Solon), Ben Saunders (Wabi Sabi Farm, Grimes), and Tony Thompson (New Family Farm, Elkhart). With these cooperators, we tested the row cover strategy with removal at anthesis (RC) and/or the delayed-removal row cover (DRC) strategy, using ‘Athena’ muskmelon and ‘Marketmore’ cucumber in 30-ft-long rows in non-replicated trials. Sites were visited weekly. A constant covering of Surround was maintained on plants, and applications of Pyganic EC (pyrethrin) and Trilogy (neem oil) were applied when cucumber beetle thresholds were reached. Thresholds for melons and cucumbers were three beetles/plant. Thresholds were determined by weekly visual monitoring of three randomly chosen plants in each subplot. Incidence of bacterial wilt was monitored weekly with a final count immediately before harvest. Fungicides were applied on an as-needed basis.Cooperators harvested and recorded data on marketable yield.  Ohio. Demonstration trials with extended-duration row covers were conducted in Ohio in 2014 and 2015 with commercial grower Guy Ashmore at That Guy’s Family Farm in Clinton County, in 2014 with commercial grower Allison Davis at the Ohio State University Student Farm in Franklin County, and in 2015 with commercial grower Mary Bridgman and Bridgman Farm in Fayette County. In all cases, the trial consisted of three treatments each with two replicates at That Guy’s Farm and three replicates at the OSU Student Farm and Bridgman Farm. The three treatments were: 1) no row cover; 2) row cover removed when the first female flower appeared; and 3) extended-duration row cover, removed 10 days after first female flower appeared. Each plot was a single row. Plot length was 15 plants at That Guy’s Farm and 7 plants at the OSU Student Farm, as determined by space available at each site. Both growers produced ‘Athena’ muskmelons grown from untreated seeds, and Agribon-19 row covers. Plots were scouted once per week for beetles by OSU’s entomology technicians, and insecticide was applied if the action threshold was exceeded. The threshold was one beetle per plant. Insecticide used was spinosad (Entrust SC, 8 fl oz per acre) plus cucurbitacin (CideTrak-D, 3.1 oz per acre); if more than two sprays were needed, then kaolin (Surround WP, 25 lb per acre) was used before returning to Entrust plus CideTrak-D. Fruit were counted, weighed, and rated for marketability at harvest. Marketable fruit were those with size greater than 2 lb, uniform netting, and absence of physical damage or rots.

Research results and discussion:

Objective 1. Iowa results. Incidence of bacterial wilt was significantly (P<0.0001) less in PTC than non-PTC plots in 2014, but not in 2015. However, both PTC and non-PTC plots received approximately the same number of insecticide sprays, so the Butternut squash barrier apparently did not reduce numbers of cucumber beetles in the muskmelon main crop, and did not reduce the use of insecticides. Muskmelon yield was not statistically different in PTC vs. non-PTC plots. Ohio results. Results were generally similar, although incidence of bacterial wilt was only numerically rather than statistically less in PTC compared to non-PTC plots, and the PTC plots received one fewer insecticide spray than the PTC plots. As in Iowa, yield did not differ significantly between the treatments. Discussion. These findings suggest that the somewhat encouraging results from smaller-plot PTC trials in an earlier NCR-SARE project (LNC-323) were not borne out by the larger-scale trials in the present project. Insecticide-spray savings with PTC were minimal to nonexistent, bacterial wilt suppression was erratic, and marketable yield of the muskmelon main crop was not impacted by PTC. Realistically, these results will not and should not make Midwest growers interested in using PTC with a muskmelon main crop, because PTC is inherently complicated by having to manage and market two crops rather than one in the same field. The advantage seen in New England with a butternut squash main crop – a 90% reduction in insecticide use while safeguarding marketable yield – did not materialize in our trials. A possible reason for this disparity is our use of muskmelon as a main crop (in contrast to butternut squash as the main crop in New England trials). PTC depends on a sharp difference in main crop (less attractive to cucumber beetles) and perimeter crop (more attractive) in order to keep beetles largely confined to the perimeter crop. However, by using muskmelon as our main crop, we may have had a too-attractive main crop, thus minimizing the value of the PTC strategy. We conclude that, based on our results, PTC is not a viable strategy for muskmelon production in the Midwest.

Objective 2. Iowa results. In 2014, using row covers resulted in twice the marketable and total yield compared to the non-covered control. Row cover treatments required an average of 4.5 fewer insecticide sprays than the non-covered control. Row cover treatments significantly (P<0.05) suppressed incidence of bacterial wilt compared to the non-covered control. In 2015, however, there were no significant differences in marketable yield or bacterial wilt incidence between any row-covered treatment and the non-covered control, although non-covered treatments required three more insecticide sprays than the row-covered treatments. Ohio results. The 2014 row cover treatments did not impact either cucumber beetle counts or marketable yield. However, bacterial wilt incidence was significantly lower for the row-cover treatments than the non-covered treatments (about 10% vs. about 70% incidence). In 2015, in contrast, treatments did not differ significantly in bacterial wilt incidence, nor in marketable yield. Interestingly, weight and number of unmarketable fruit were significantly higher in the row-cover treatments than in the non-covered control treatments. Discussion. Neither the delayed-removal row cover strategy nor removing row covers at anthesis paid off in our trials in that they usually failed to deliver higher marketable yield than the non-covered control, nor did they consistently suppress bacterial wilt more effectively. These outputs suggest that the use of row covers, as configured in these trials, in unlikely to be a viable strategy for organic growers of muskmelon on the Midwest. Despite the high cost and labor requirements of row covers, they are somewhat attractive to organic melon growers due to the absence of effective organic insecticides against cucumber beetles and bacterial wilt. In our trials, neither removing the covers at anthesis nor waiting for 10 additional days to remove them was effective against this insect/disease complex or in delivering higher marketable yield. Given the expense and labor requirements of row covers, and their failure to consistently enhance marketable yield, our results suggest that they should not be used by Midwest organic muskmelon growers – unless the row cover strategy is revised radically (see below under Areas Needing Additional Study).

Objective 3. Economic analysis was not done, because results of Objective 1 and Objective 2 field experiments showed convincingly that the strategies being trialed were highly unlikely to be attractive to growers, since they offered no advantages in yield or pest management over simpler systems that are currently in wide use.

Objective 4. Iowa. In both 2014 and 2015, cucumber beetle numbers and bacterial wilt pressure were generally low. At New Family Farm and ZJ Farm during 2014, row cover strategies resulted in considerably higher yield than the non-row cover control treatment. At Wabi Sabi Farm, however, in the absence of significant cucumber beetle and bacterial wilt pressure, an outbreak of aphids under the row covers suppressed yield in the RC treatment, resulting in higher yield for the control treatment. At One Step at a Time Farm, miscommunication with the grower resulted in harvest data being combined for the treatments, so evaluation of the impact of row cover treatments on yield was not possible. In the 2015 trials, there was no consistent impact of row cover strategies on marketable yield or bacterial wilt incidence in comparison to the non-covered control treatment. Ohio.  At Guys Farm in 2014, weed pressure under the row covers required temporary removal of the covers for manual weeding, after which cucumber beetles were noted inside the row covers. No data for yield or bacterial wilt were recorded, although yield in all trials was low. At the OSU Student Farm, incidence of bacterial wilt was 80-100% at harvest for all treatments; yields did not differ significantly among treatments, but number of marketable fruit was highest for the delayed-removal row cover treatment. In the 2015 trial at Guy’s Farm, there were no significant differences among treatments in bacterial wilt incidence or in number or weight of marketable fruit. At Bridgman Farm, cucumber beetle pressure was high, and the delayed-removal row cover treatment required two fewer insecticide sprays than the non-covered control. However, there were no significant differences among treatments in incidence of bacterial wilt, and all treatments showed a very high (77-100%) incidence by the time of first harvest. Marketable yield was not significantly different among treatments, although the mean weight or marketable fruit in the delayed-removal treatment was twice that in the non-covered control treatment; a similar trend had occurred at the OSU Student Farm in 2014. Discussion. The large range of variability in outcomes of on-farm demonstration trials in both states reinforces our conclusions from Objective 2 field experiments: row covers, as configured in these trials, are not a recommended option for Midwest muskmelon growers.

Research conclusions:

Results of our field trials indicate that, under the conditions of our field trials, neither perimeter trap cropping (PTC) nor delayed removal of row covers has strong potential to reduce dependency on insecticides for management of bacterial wilt on muskmelon in the Midwest. Given the substantial additional labor and cost requirements for both strategies, they would have to provide clear advantages in enhancing marketable yield in order to be recommended to growers as alternatives to current management practices. In fact, no consistent enhancement of marketable yield was noted for either PTC or delayed removal of row covers.

Economic Analysis

Due to an absence of consistently encouraging results in field trials, economic analysis was not done, since the practices under trial proved to be clearly disadvantageous to growers.


Farmer Adoption

Due to the negative nature of the findings, as described above, we did not pursue farmer adoption.

Participation Summary

Educational & Outreach Activities

Participation Summary:

Education/outreach description:

The project’s findings were shared with growers at annual field days in Ohio and Iowa during 2014 and 2015, and in presentations during grower conferences in both states.  In addition, a total of 12 on-farm demonstration trials were held in cooperation with commercial growers in the two states during these years.

Project Outcomes


Areas needing additional study

If failure is the best teacher, we ought to have learned quite a bit from this study. We think that we did.

Perimeter trap cropping is of interest primarily to conventional (non-organic) growers because it requires potent insecticides to kill cucumber beetles that congregate in the perimeter trap crop (organic insecticides are only minimally effective in killing cucumber beetles). For bacterial wilt management, the strategy may be least effective when the main crop (muskmelon in our study) is highly susceptible to the disease and highly attractive to cucumber beetles. It may be more effective when the main crop is butternut squash, as in New England, where the PTC strategy was developed and validated. However, use of PTC for bacterial wilt control on butternut squash may be less useful in the Midwest than in New England. The reason for this speculation is that recent (2013-2016) collaborative research at Iowa State University and Harvard University revealed two interesting patterns. First, there are two putative subspecies of the bacterial wilt pathogen, Erwinia tracheiphila, which differ from each other in host selectivity: one subspecies prefers hosts in the genus Cucumis (including muskmelon) whereas the other subspecies prefers hosts in the genus Cucurbita (squashes). Second, mapping the occurrence of these subspecies of E. tracheiphila from a large collection of strains revealed that both strains are common in New England, whereas only the melon strain occurs commonly in the Midwest. The upshot of this research is that the risk of bacterial wilt on squash in the Midwest is relatively low compared to the risk in New England. Therefore, there is little incentive currently for Midwest squash growers to design alternative management strategies, such as PTC, against bacterial wilt. These discoveries underline our conclusion that PTC may be unsuitable for controlling cucurbit bacterial wilt on Midwest farms.

Our present results notwithstanding, we continue to be optimistic about the future of row covers as a viable strategy for Midwest organic growers to use against bacterial wilt in muskmelon. Effectiveness of current options (organic insecticides all season, or row covers followed by insecticides after flowering) is so erratic that many Midwest organic growers will not grow muskmelon. The fundamental shortcoming of our row cover strategy may have been that we ultimately had to remove the covers, even with a 10-day delay after bloom started. This removal left the muskmelon plants exposed for many weeks to cucumber beetles and the bacterial wilt pathogen they carry, resulting in mid- to late-season wilting and death of plants. This line of reasoning implies that full-season (transplant to harvest) row covers could be a more viable strategy. But how can this be done? In a current (2015-2018) field study funded by USDA-NIFA Organic Transitions Program, the ISU PIs in the present study, with collaborators at University of Kentucky, are evaluating “mesotunnels” as a new strategy for full-season protection against cucumber beetles, bacterial wilt, and other pests and diseases. Mesotunnels reconfigure row tunnels with a more air-permeable row cover material (nylon mesh vs. spunbond polypropylene) and more vertical and horizontal space within each tunnel (twice the height of conventional row tunnels, and triple-row plots under each cover). Pollination is provided by purchased boxes of bumble bees that are inserted under the covers when female flowers begin to open. Initial (2016) results in Iowa were encouraging: the full-season mesotunnels significantly raised marketable yield in comparison to conventional spunbond polypropylene tunnels that were removed at the start of bloom. However, depressed yield under full-season mesotunnels in Kentucky during 2016 suggested that the bees had been inserted under the row covers too late in the bloom period. By the end of 3 years of field trials, we hope to have a blueprint for a consistently effective mesotunnel strategy for Midwest muskmelon growers.



Information Products

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.