Applied Engineering in Agriculture

Vol. 27(3): 317‐324 2011 American Society of Agricultural and Biological Engineers ISSN 0883-8542 317

DEVELOPMENT OF A DESICCATED CADAVER DELIVERY

SYSTEM TO APPLY ENTOMOPATHOGENIC NEMATODES FOR CONTROL OF SOIL PESTS

H. Zhu, P. S. Grewal, M. E. Reding

ABSTRACT. Entomopathogenic nematodes may be more capable of controlling soil pests when they are harbored by desiccatedcadavers. A small‐scale system was developed from a modified crop seed planter to effectively deliver desiccatednematode‐infected cadavers into the soil. The system mainly consists of a metering unit, an air pressure source, a cadaverscraper, a custom‐designed cadaver container, tension adjustment devices, double disk soil opener, a discharge tube, a packerwheel, and a press‐drive wheel. The metering unit intermittently discharges a cadaver to the discharge tube at a constant rate.The double disk opener creates a narrow 7.5‐cm deep slot for the placement of the discharged cadavers. The packer wheelthen covers the cadavers with loose and somewhat moist soil. The spring tension device and the press‐drive wheel controlthe slot depth. The number of cells on the metering plate controls the cadaver delivery rate regardless of the travel speed.The metering unit with modified double bean plates has delivery accuracy between 79% and 94% at 500‐Pa air pressure. Aslower travel speed and fewer cells on the metering plate improved the accuracy of delivery. When the travel speed was4.8 km/h or slower, the desiccated cadaver delivery system performed satisfactorily in both laboratory and field tests, anddelivered the cadavers at a rate of 1.6, 3.3, or 6.6 cadavers/m length in the soil.

Keywords. Biological insecticide, Farm machinery, Ornamental nursery, Turf application.

oot‐feeding larvae of insects such as the Japanesebeetle (Popillia japonica), European chafer(Rhizotrogus majalis), and Oriental beetle(Anomala orientalis) larvae are important

economical pests of floral and nursery crops, vegetables, turf,and other specialty crops, requiring conventional chemicalpesticides or natural pest control agents for crop protection.However, conventional agro‐chemicals have raised concernsover potential contamination to the environment anddisruption of sensitive ecosystems. Accordingly, biologicalcontrol agents, which are considered safer and cleaner andoffer environmental‐friendly advantages, have beenpromoted as an alternative.

Several species of entomopathogenic nematodes arecommercially available as biological predatory insecticidesto limit pest insect populations in nurseries, greenhouse,vegetables, and turf (Grewal et al., 2005). One of thesespecies is the Heterorhabditis bacteriophora Poinar strainGPS11 nematode. Approximately 500 to 1000 �m in lengthand 18 to 50 �m in width, the infective juvenile nematodes

Ssbmitted for review in March 2010 as manuscript number PM 8486;approved for publication by the Power & Machinery Division of ASABEin February 2011.

Mention of proprietary product or company is included for the reader'sconvenience and does not imply any endorsement or preferential treatmentby USDA‐ARS or The Ohio State University.

The authors are Heping Zhu, ASABE Member Engineer, AgriculturalEngineer, USDA‐ARS, Application Technology Research Unit (ATRU),Wooster, Ohio; Parwinder S. Grewal, Professor, Department ofEntomology, The Ohio State University/OARDC, Wooster, Ohio; andMichael E. Reding, Research Entomologist, USDA‐ARS ATRU, Wooster,Ohio. Corresponding author: Heping Zhu, USDA‐ARS ATRU, Ag. Eng.Bldg., OARDC, 1680 Madison Ave., Wooster, OH 44691; phone:330‐263‐3871; e‐mail: heping.zhu@ars.usda.gov.

harbor symbiotic bacteria (Xenorhabdus ssp.) and transmitthem into the insect body cavity. This bacterium kills theinsect within 48 h. H. bacteriophora have been reported toeffectively control scarab larvae in ornamental nurseries(Reding et al., 2008).

Conventional sprays (Klein and Georgis, 1992; Bateman,1999) and drip irrigation methods (Wennemann et al., 2003;Reding et al., 2008) have been used to apply nematodes asbiological insecticides. However, damages to the nematodesby spray application have been reported (Steinke and Giles,1995; Fife et al., 2004, 2005, and 2007). Insufficient moisturein the area where nematodes are delivered may kill themwithin 24 h after spray application. High soil moisturecontent and slow desiccation rates are required for nematodesto act effectively as biological insecticides (Bedding et al,1993). On the other hand, the concern of deliveringnematodes through drip irrigation system is the low recoveryrate and considerable distribution variations along drip lines(Wang et al., 2009).

Entomopathogenic nematodes can be applied usinginfected insect cadavers, which can extend theireffectiveness as biological insecticides (Shapiro and Lewis,1999; Shapiro‐Ilan et al., 2003). By using cadavers as acarrier, nematodes can be applied to less than perfect fieldconditions (Bruck et al., 2005). After placement of theinfected cadavers, the nematodes are expected to exit thecadavers to seek and infect insect larvae in the soil only whenthe conditions are suitable for their mobility. Depending onthe soil type and the availability of root insects, thenematodes can travel over 30 cm away from their dischargedpoint (Schroeder and Beavers, 1987). To optimize the insectcontrol, the cadavers should be discharged less than 60 cmapart. The most commonly used insect host to rear andmaintain nematodes is the larvae of the greater wax moth

R

318 APPLIED ENGINEERING IN AGRICULTURE

(Galleria mellonella) because they are very susceptible tonematode colonization and are commercially available.However, nematode infected hosts can stick together orrupture during transport or application (Shapiro‐Ilan et al.,2001). One technique used to alleviate this problem is toformulate the cadavers in a powdery substance such asMirasperse (a starch) + clay (calcium silicate) to preventcadavers sticking together (Shapiro‐Ilan et al., 2001). Furtherformulation can be done by desiccating cadavers, which canincrease the longevity of the nematodes by decreasing theirmetabolism (Shapiro‐Ilan et al., 2001). Nematode infectedcadavers have irregular shapes and are very fragile, thereforebreak easily. They must be buried 5 to 7.5 cm beneath the soilsurface to avoid damage from animals or exposure tosunlight. Because of these special physical limitations anddelivery requirements, no equipment is currently available toefficiently deliver desiccated cadavers into soils withoutdamaging them.

In this research, our hypothesis was that the mechanism ofcommonly available crop seed planters could be adapted forthe development of a system to deliver the desiccatedcadavers infected by nematodes into the soil laterally. Inessence, the placement of cadavers would resemble theprocess of planting crop seeds. Because of the frailty of thedesiccated cadavers, commercially available mechanicalseed metering units may not be able to discharge the cadaversat a constant and accurate rate, and could even potentiallydamage them. Consequently, the planters underconsideration must be modified to accommodate thedifferent physical properties of cadavers from those of solidseeds.

The objective of this research was to develop a workingdevice that could deliver desiccated cadavers, orG. mellonella larvae harboring nematodes at steady andpredictable rates for control of soil pests. A majorrequirement for the delivery system was to place the cadavers5 to 7.5 cm deep in the soil at a discharge rate of 6.6, 3.3, or1.6 cadavers/m length (or 15, 30, or 60 cm apart).

MATERIALS AND METHODSDELIVERY SYSTEM DESIGN

Modification of Seed Metering Unit

Before development of the delivery system, a preliminarytest was conducted to select metering units that would bepotentially used for discharging desiccated cadavers.Mechanical and pneumatic metering systems arecommercially available for seed planters. Since desiccatedcadavers have a brittle hard cover outside and soft bodyinside, the mechanical metering system could potentiallydamage the cadavers and was not considered for the study.The pneumatic system might be a feasible method to deliverthe cadavers without damage during operation. It uses eithervacuum or positive air pressure to place seeds on the meteringplate for accurate delivery. Due to its relative simplicity overa vacuum system, the positive pressure metering system wasselected for the tests.

A demonstration metering unit for White Planters�manufactured by Agco (Duluth, Ga.) was used to test whethera row crop seed metering unit would be able to actually meterdesiccated cadavers (fig. 1). The unit consisted of a sealedhopper to store cadavers, a seed metering device including ametering plate and a seed scraper, an air blower with an airpressure controller, and a metering plate speed controller.

Based on the shape and mass of desiccated cadavers, threedifferent metering plates, used for planting field corn seeds,sunflower seeds and double soybeans were chosen for thetest. The outside diameter of these plates was 30 cm. The fieldcorn plate had 30 cells, the sunflower plate 30 cells, anddouble bean plate 60 cells uniformly distributed along thecircumference of the plate. A single cell in the double beanplate was designed to hold two soybeans at a time while a cellin the other two plates was to hold one seed at a time.

The desiccated cadavers were approximately 20 mm long,5.5 mm wide, and 1.5 mm thick, weighing about 85 mg andwere covered with white clay powder (fig. 2). The powdermainly consisted of 1% Mirasperse (AE StaleyManufacturing Co., Decatur, Ill.) and Microcel clay(Manville Products Corp, Denver, Colo.). The cadavers hada layer of brittle skin to maintain their shape and cover the

Cadaver hopper Desiccated cadavers

Metering chamber

Metering plate

Air blower with airpressure controller

Cadaver collector

Scraper

Motor and speed controller

Figure 1. Schematic diagram of the demonstration seed metering unit used to test three different metering plates to discharge desiccated cadavers.

319Vol. 27(3): 317‐324

Figure 2. Desiccated cadavers infected with nematodes used in the test.

soft body. They were produced by exposing each larva of agreater wax moth consisting of 100 infective juvenilenematodes of H. bacteriophora GPS11. Three days later, theinfected cadavers were partially desiccated at 70% relativehumidity in desiccators for 6 days, and then were formulatedin Mirasperse and clay as described by Shapiro‐Ilan et al.(2001). Each desiccated cadaver carried approximately100,000 infective juveniles. They were applied into the seedcontainer before the test. In operation, the air pressure and thescraper in the seed chamber were supposed to place onecadaver into each cell on the revolving plate. The cadaverswere released outside when they reached the bottom of themetering chamber where the air pressure lost its effectivenessto hold the cadavers in the cells.

The test was conducted with the air pressure at 250 and500 Pa, and the plate rotation speed at 5, 7, and 10 rpm,respectively. Each test was run for 120 s, and cadaversdischarged from the metering unit with each plate werecounted and compared with predicted number of cadavers.The predicted number of cadavers was calculated from thenumber of cells on each plate, the plate rotation speed and themeasurement time (120 s). The comparison was made fordetermination of metering accuracy for each plate, and it waspresented as recovery rate which was the percentage of thecounted number of cadavers divided by the predicted numberof cadavers. The number of broken cadavers was alsocounted during the test to determine the percentage ofdamaged cadavers .

Development of Delivery System

After the preliminary test with the three different seedplates that supported the concept that the seed planter wouldpotentially be able to deliver desiccated cadavers as a carrierto apply nematodes into the ground, a core part of themetering unit used for White Planters� was modified tobecome a metering unit specifically for the cadaver. Themetering unit included a chamber, a modified double beanmetering plate, a cadaver scraper, and a cadaver hopper. Thedouble bean metering plate was used because in thepreliminary test it had the highest accuracy to meter cadaversamong the three plates tested. The hopper was fabricated witha custom‐built, cylinder shape, 4,250‐cm3 capacity,aluminum container to store infected cadavers. It used apolycarbonate lid with a seal to prevent pressurized air fromleaking.

The desiccated cadaver delivery system was developed byadapting the modified metering unit to a John Deere model71 Flexi Unit planter (Moline, Ill.). The original meteringunit on the planter was removed because it used vacuum to

hold seeds on the plate inside its metering chamber. Thesystem also included a double disk opener to cut a slot in thesoil, a rectangular‐shaped discharge tube to guide cadaversinto the soil slot, a spring tension device, a 2.5‐cm wide and17.5‐cm diameter packer wheel to close the slot and cover thedischarged cadavers with loose soil, a press‐drive wheel topress the covered slot and maintain the desired depth as wellas drive the metering unit, and a small air pressure source.

The air pressure source was a centrifugal blower (Model#35440‐0000, ITT Industries, Gloucester, Mass.) with anominal air flow of 7.0 m3/min. It produced adjustable airpressure inside the metering chamber to hold the cadavers inthe cells on the plate during operation. The air pressure wasadjusted by changing the blower speed which was operatedwith a motor controller.

A steel bracket base was designed to couple the modifiedmetering unit to the modified planter frame. The bracket basewas 30 cm long, 30 cm wide, and 0.6 cm thick. It had a squarehole for the cadavers to fall though, and also had two 1.6‐cmdiameter holes with a stud welded to anchor the metering unitto the top side.

Unlike solid seeds, the fragile desiccated cadavers couldbe easily broken when the metering speed was very high.Another modification to the planter was to reduce meteringplate speed to prevent or minimize damage to the cadaversduring operation. The modification was made including theground press‐drive wheel, a roller chain and two sprockets.The press‐drive wheel was 0.40 m in diameter and was usedto drive a 14‐tooth sprocket. With a roller chain the 14‐toothsprocket turned a 22‐tooth sprocket which was mounted onthe same shaft of a 13‐tooth sprocket. The 13‐tooth sprocketwas connected to a 60‐tooth sprocket with another rollerchain. The 60‐tooth sprocket was mounted on the same shaftas the metering plate. With this design, the metering unit wasdriven by the press‐drive wheel, and the discharge rate(defined as the number of discharged cadavers/m) wouldonly depend on the number of cells on the metering plateregardless of the travel speed. The turning ratio between thepress‐drive wheel and the metering plate was 7.25:1. For thisdesign, the metering plate speed would be calculated fromthe equation,

D

VW T

�60

1000

25.7

1×= (1)

where W is the metering plate speed (rpm), VT is the travelspeed (km/h), and D is the diameter of press‐drive wheelwhich is 0.40 m.

The predicted cadaver discharge rate could be calculatedas,

320 APPLIED ENGINEERING IN AGRICULTURE

D

M

V

MWQ

T π=×=

25.71000

60 (2)

where Q is the discharge rate (number of cadavers/m), and Mis the number of cells on the metering plate. For the 60‐cellplate the discharge rate would be 6.6 cadavers/m (or twocadavers/ft) regardless of the travel speed. Based onequation 2, two additional double bean metering plates weremodified to deliver the cadavers at the predicted dischargerate of 1.6 and 3.3 cadavers/m (0.5 and 1 cadaver/ft),respectively. There were 15 cells on the plate for the1.6 discharge rate, and 30 cells on the plate for the3.3 discharge rate. Propoxy 20� epoxy putty (HerculesChemical Co., Passaic, N.J.) was applied to the 60‐celldouble been plate (fig. 3a) to fill alternating cells that werenot used for holding cadavers to form the 15‐ and 30‐cellplates (figs. 3b and 3c).

A solid steel tool bar of nominal 6‐cm square cross sectionwas used to hold the entire delivery system. The tool bar waspositioned above the ground and was high enough that theparallel linkage to the delivery system was also parallel to theground. This design ensured the system to have an optimalrange of motion to float up and down during operation. Thefloating action of the delivery system helped maintain properdepth on the uneven ground.

Because the soils in nursery and turf fields are harder thanrow crop fields, the existing planter tension unit could not cutslots deep enough for placing the cadavers. A spring tensiondevice was modified that included adding two heavy springson the parallel linkage between the delivery system and thetool bar. The spring tension was adjustable to provide properslot depth while the delivery system was adjusted to thedesired depth by turning the depth adjustment accordingly.The springs were attached to two adjustment handles. Whenthe handles were placed closer to the ground, more springtension was created, resulting in more force downward anda deeper planting depth.

Except for the spring tension device, two additionaltension adjustments were also added into the system toincrease or decrease the depth of the double disc opener andultimately the depth of the cadavers as well. The firstinvolved a depth adjustment knob on the side of the deliverysystem. This knob adjusted the position of the press‐drivewheel in relation to the double disc opener. The secondadjustment was the top linkage of the three‐point hitch on thetractor. By adjusting the top linkage of the hitch, the deliverysystem could be tilted forward or backward in relation to thedirection of travel. This action transferred a portion of thetractor weight to the delivery system and increased ordecreased the aggressiveness of the double disc opener,resulting in a change of depth. Because of the light weight ofthe delivery system, this adjustment helped the double diskopener cut hard soil while still allowing the delivery systemto float.

The complete delivery system mounted to the solid steeltool bar was offset to one side of the tractor, which allowedthe cadaver dispenser to be used beside a row of trees (fig. 4).The system was small and light. The relatively small size andlow weight provided the versatility of adapting the system tosmall tractors or even small automated transfer vehicles(ATV) commonly used in nurseries or orchards.

EXPERIMENTSIndoor Bench Tests

After the delivery system was developed, two indoor testswere conducted before outdoor tests. The first test was todetermine discharge accuracy at various travel speeds withthe 60‐cell double bean plate. The delivery system wasmounted on a bench, and then a Dayton AC gear motor (W.W.Grainger, Inc., Lake Forest, Ill.) was used to control thepress‐drive wheel speed to simulate the travel speed. Thegear motor speed was controlled with a variableautotransformer. The simulated travel speeds were 1.6, 4.8,8.0, 11.3, and 14.5 km/h. From equation 1, these speedsyielded metering plate speeds from 2.9 to 26.5 rpm. The airpressure used in the test was 500 Pa based on results from theprevious test. Both cadavers and soybeans were usedseparately for the test. Soybeans were used for comparison asa reference to verify the cadaver test results becauseproducing the infected desiccated cadavers was timeconsuming. Also, the cadavers were relatively fragile andsome were damaged after being discharged from themetering unit, so they could not be used repeatedly.

For the bench test, the delivery system was operated for3 min at each travel speed. The numbers of cadavers or beansdischarged were then counted and presented as the number ofcadavers or beans per cell. Each test was repeated three times.For this test, it was anticipated that each cell should have onecadaver or two soybeans.

The second indoor test was to verify the accuracy ofmetering cadavers for the three modified metering platescontaining 15, 30, and 60 cells. The delivery system wasagain mounted on a bench with a Dayton AC gear motor tocontrol the press‐drive wheel speed and obtain a simulatedtravel speed. For this test, the simulated travel speed was4.8 km/h and the air pressure was 500 Pa. The numbers ofcadavers discharged during 3‐min time periods were countedand then averaged as the number of cadavers per cell. Tosimplify the comparison, recovery rate was used to presentthe metering accuracy for each plate. It was calculated fromthe counted number of cadavers per cell divided by thepredicted number of cadavers per cell. The predicted numberof cadavers was one per cell. The number of cadavers per cellwas then converted to the discharge rate for each meteringplate. The predicted discharge rate was 6.6, 3.3, and1.6 cadavers/m for the three plates, respectively.

Outdoor Tests

After the indoor tests, the delivery system was mounted ona tractor and tested on a hard surface in a parking lot to verifythe accuracy of discharge rates. A 15‐m straight line wasdrawn on the hard surface and then covered with 20 cm wideduct tape. The line was used for two purposes: to guide thetractor travel direction, and collect the cadavers discharged.The double disk opener moved on the middle of the tape whenthe tractor travelled along the line. The 30‐cell metering platewith the predicted discharge rate at 3.3 cadavers/m was usedfor the test. Air pressure was adjusted to 500 Pa. Travel speedwas 1.6, 3.2, 4.8, and 6.4 km/h, respectively. Due to the hardsurface, the discharged cadavers bounced away from theirlanding position, so spacing between cadavers was notmeasured. Total number of cadavers on the 15‐m long tapewas counted for each travel speed, and then was averaged asthe measured discharge rate.

321Vol. 27(3): 317‐324

(a) Original double bean plate with 60 cells.

(b) Modified plate with 30 cells.

(c) Modified plate with 15 cells.

(d) An enlarged cell

Figure 3. Modified metering plates with 60, 30, and 15 cells to meter desiccated cadavers at the discharge rate of 6.6, 3.3, and 1.6 cadavers/m (or 2,1, and 0.5 cadavers/ft), respectively.

Following the laboratory tests to verify performances ofthe delivery system, a field test was performed in a flat turfgrass area to verify the discharge depth of cadavers. Threestrips of the grass, each 2 m wide and 30 m long, wereselected. The grass in each strip was cut below 2.5 cm heightwith a mechanical mower before the test. The deliverysystem was mounted onto a John Deere 2640 tractor and was

Figure 4. The complete delivery system mounted offset to one side of thetractor. (1) air blower with air controller, (2) metering unit, (3) cadaverhopper, (4) roller chain, (5) nematode infected desiccated cadavers,(6) press‐drive wheel, (7) packer wheel, (8) cadaver discharging tube,(9) double disk opener.

tested at 4.8 km/h travel speed and 500‐Pa air pressure. The15‐cell metering plate was used to produce the predicteddischarge rate at 1.6 cadavers/m (or 1 cadaver per 2 ft). Thespring tension device was adjusted to make a 7‐cm deep slotin the soil.

Immediately after the cadavers were discharged, each30‐m long strip was evenly divided into five sections.Presence of cadavers under the soil in the second and fourthsections was examined through gently removing soil by handwith a screw driver. When a cadaver was found,measurements were taken for location and depth.

RESULTS AND DISCUSSIONTESTS WITH DIFFERENT METERING PLATES

With the demonstration metering unit for WhiterPlanters�, the recovery rate of cadavers discharged from thedouble soybean plate was much higher than that fromsunflower and field corn plates at both 250‐ and 500‐Pa airpressures (table 1). Data in table 1 also illustrate that the platespeed did not influence the recovery rate for any of the plates.At 250‐Pa air pressure, the average recovery rate of threespeeds was 61% from the double bean plate while the meanrecovery rate from both sunflower and corn plates was below20%. Comparatively, at 500‐Pa air pressure, the average

322 APPLIED ENGINEERING IN AGRICULTURE

Table 1. Comparison of accuracy of metering desiccated cadavers from double bean, sunflower and field corn seed plates with the

demonstration metering unit at 250‐ and 500‐Pa air pressures and 5‐, 7‐, and 10‐rpm plate speeds.[a]

Air Pressure(Pa)

Plate Speed(rpm)

Mean Recovery Rate (%)[b]

Double Bean Sunflower Field Corn

250 5 57 (11) 12 (11) 19 (10)

250 7 62 (7) 15 (17) 17 (11)

250 10 65 (13) 19 (11) 18 (12)

500 5 94 (4) 13 (20) 21 (8)

500 7 91 (4) 14 (8) 25 (10)

500 10 84 (3) 10 (8) 27 (5)[a] Coefficient of variation (%) is presented in parenthesis.[b] Recovery rate is percent of the ratio between the measured and

predicted numbers of cadavers per cell.

recovery rate of three speeds was 89.7% from the double beanplate while the mean recovery rate was below 15% from thesunflower plate and below 30% from the corn plate. The netrecovery rate increased nearly 30% for the double bean platewhen the air pressure increased from 250 to 500 Pa.Therefore, the double bean plate provided the highestmetering accuracy among the three plates, supporting theconcept that the seed planter would be able to deliver thedesiccated cadavers into the ground.

The rate of broken cadavers after they passed through themetering unit varied very little with the type of plate and airpressure, and slightly increased with the plate speed. Theaverage percentage of broken cadavers was about 13.5% forall three plates at two air pressures and three plate speeds.

INDOOR BENCH TESTSThe delivery system with the 60‐cell metering plate in the

bench test discharged 1.1, 0.8, 0.5, 0.6 and 0.5 cadavers percell at the simulated travel speed of 1.6, 4.8, 8.0, 11.3, and14.5 km/h, respectively (fig. 5). The number of cadaversdecreased as the simulated travel speed increased from 1.6 to8.0 km/h and then remained nearly constant for speedsbetween 8 and 14.5 km/h. The expected number of cadaverswas one cadaver per cell. However, at the 4.8 km/h speed andabove, the delivery system was not accurate enough toproduce one cadaver per cell. That is, some cells missedholding cadavers during each revolution when the travel

Figure 5. Number of desiccated cadavers and soybeans discharged fromthe metering unit with a 60‐cell double bean plate at different simulatedtravel speeds. Error bars present standard deviations.

speed was over 4.8 km/h. The soybean test showed that thenumber of beans was greater than two for all speeds, and thenumber of beans at 1.6 km/h speed was higher than those athigher speeds (fig. 5). The expected number of beans was twoper cell. Based on equation 1, the plate speed was 2.9, 8.8,14.6, 20.7, and 26.5 rpm when the travel speed was 1.6, 4.8,8.0, 11.3, and 14.5 km/h, respectively. Therefore, the numberof cadavers per cell was influenced by the plate speed. Themetering plate at slower speed led to a higher number ofcadavers per cell or higher metering accuracy. For safeoperations, the tractor travel speed in most nurseryproductions are limited below 6.5 km/h while in turf orvegetable production systems the speed may be higher.

Bench tests also illustrated that the measured number ofcadavers per cell increased as the number of cells on themodified metering plate decreased (table 2). Theoretically,the three metering plates were predicted to discharge onecadaver per cell. At 4.8‐km/h travel speed, the measurednumber of cadavers from the 30‐cell metering plate was closeto one per cell while the 60‐cell plate was 21% below thepredicted rate and the 15‐cell plate was 12% above thepredicted rate. The reason might be that as the number of cellson the plate decreased the time for cadavers to fill each cellincreased, thus increasing the probability of each cellcatching a cadaver as the number of cells decreased. Previoustest results for cadavers and soybeans also illustrated thatincreasing plate speed decreased the number of cadavers percell. The recovery rate, derived from percent of the ratiobetween the measured and calculated numbers of cadaversper meter, was 79%, 94%, and 112% for the plates containing60, 30, and 15 cells, respectively.

Corresponding to the number of cadavers per cell from thethree modified metering plates, the predicted discharge ratesfor the 60‐ and 30‐cell plates were higher than the measureddischarge rate while it was opposite for the 15‐cell plate(table 2). The 30‐cell plate produced the measured dischargerate most similar to the predicted rate versus the 60‐ and15‐cell plates. The 60‐cell plate produced lower and 15‐cellplate produced higher discharge rates than their predictedrates. Therefore, in the process of choosing the discharge rateand the amount of cadavers to be used in the field, under orover application of cadavers from the 60‐ or 15‐cell platesshould be considered. However, compared to the percentdistribution of root insects presented in the field, thedischarge accuracy from all three plates may still beadequate. For this reason, large scale efficacy tests fordifferent root insects under different planting conditions areneeded to further verify performance of the delivery system.

Table 2. Comparison of metering accuracy for three modified plates at 4.8 km/h simulated travel speed.[a][b]

PlateNo.

No. ofCells

MeasuredNumber[c]

RecoveryRate (%)[d]

Discharge Rate (cadavers/m)

Predicted Measured

1 60 0.79 (16.2) 79 6.6 5.2 (16.2)

2 30 0.94 (19.8) 94 3.3 3.1 (19.8)

3 15 1.12 (31.1) 112 1.6 1.8 (31.1)[a] Predicted number of desiccated cadavers was one cadaver per cell

from every plate.[b] Coefficient of variation (%) is presented in parenthesis.[c] Mean measured number of cadavers per cell.[d] Recovery rate is percent of the ratio between measured and predicted

numbers of cadavers per cell.

323Vol. 27(3): 317‐324

TESTS ON HARD SURFACEOutdoor tests on the hard surface in a parking lot showed

that the discharge rate decreased as the travel speed increased(table 3). For the 30‐cell plate, the predicted discharge ratewas 3.3 cadavers/m regardless of the travel speed; however,the average measured discharge rate from three replicationswas 3.7, 3.4, 3.1, and 2.6 cadavers/m when the travel speedwas 1.6, 3.2, 4.8, and 6.4 km/h, respectively. Variation in themeasured discharge rate for all speeds was relativelyconsistent. The coefficient of variation for the repeated testswas between 10% and 21% (table 3). A higher travel speedthat yielded a higher metering plate speed, resulted in lesstime for each cell to catch cadavers thus lowering thedischarge rate. To provide a desired discharge rate which issimilar to the predicted discharge rate, the travel speed shouldbe between 3.2 and 4.8 km/h.

The accuracy of metering cadavers with the deliverysystem varied with the number of cells on the metering plateand the travel speed. However, the system accuratelydelivered the cadavers with the 30‐cell plate at the 3.2 and 4.8km/h travel speeds. An ideal delivery system should be ableto deliver cadavers accurately for a wide range of dischargerates regardless of travel speeds because the desireddischarge rate would depend on the insect population andmany other conditions in the field. The metering plates weremodified based on the existing double bean plate for row cropplanters. The cell construction did not perfectly match theshape and physical properties of cadavers. During the tests,it was also noticed that some locations occasionally had morethan one cadaver while some did not have a cadaver,indicating slightly uneven distribution in a line. Accordingly,custom‐designed plates should be developed specifically formetering desiccated cadavers for different discharge rateswhen the market demand for the delivery system warrants itin the future.

DISCHARGE DEPTH TESTFor the field test to verify the cadaver discharge depth,

there were 9 cadavers recovered from the soil at an averagedepth of 5.9 cm in the first row, 5 cadavers at 7.6 cm in thesecond row, and 10 cadavers at 7.0 cm in the third row. Somecadavers were not recovered because of cadaver size,cadaver color, and human error. There were six peopleassigned to search for the cadavers in the soil under sunnyconditions. It was difficult to find all the cadavers dischargedinto the soil because of their size and color, which was similarto the soil. Some cadavers might have been moved from their

Table 3. Discharge rate of desiccated cadavers from the 30‐cellmetering plate at 1.6, 3.2, 4.8, and 6.4 km/h travel speeds

tested on the hard surface in a parking lot.[a][b]

TravelSpeed(km/h)

PlateSpeed(rpm)

Discharge Rate (cadavers/m)

Replication

AverageNo. 1 No. 2 No. 3

1.6 2.9 3.2 3.5 4.5 3.7 (18)

3.2 5.9 3.3 3.7 3.1 3.4 (12)

4.8 8.8 3.1 2.8 3.4 3.1 (10)

6.4 11.7 2.0 3.1 2.7 2.6 (21)[a] The predicted discharge rate was 3.3 cadavers/m for every speed.[b] Coefficient of variation (%) is presented in parenthesis.

original location during the search. Nonetheless, the averagedepth (6.8 cm) of the 24 cadavers recovered from the soil wasrelatively consistent with a range between 5 and 8 cm and a23.8% coefficient of variation.

SUMMARY AND CONCLUSIONSA lightweight system was developed to deliver infected

desiccated cadavers carrying entomopathogenic nematodes(H. bacteriophora GPS11) into the soil. The system consistedof a modified soybean metering unit and a row crop seedplanter. The process of operating the delivery system startedwhen a double disc opener makes a small slot or trench in thesoil with minimal disturbance of surrounding soil. The slot inthe soil was maintained at a constant depth under various soilcompactions by adjustment of three tension devices. Apress‐drive wheel rotated the metering plate at a 7.25:1 ratioand air pressure carried the cadavers in the metering unitcells. The cadavers were released from the metering unitwhen the cells moved to the bottom position within themetering chamber. Behind the double disc opener was arectangular shaped tube that discharged cadavers into the soilslot. A 17.5‐cm diameter packer wheel behind the tubecovered the cadavers with soil which was compressed by apress‐drive wheel. As the delivery system movedcontinuously, it discharged the cadavers at a desired rate anddepth in the soil.

Not all seed metering plates delivered desiccated cadaversaccurately or consistently. The double bean plate was bestamong the three row crop seed metering plates tested, andwas used to discharge the cadavers in this delivery system.

The accuracy of the metering plate for discharging thedesiccated cadavers was 90% and 61% when operated at 500‐and 250‐Pa air pressure, respectively.

Although the accuracy of the delivery system varied withthe number of cells on the metering plate and the travel speed,the system accurately delivered the cadavers with the 30‐cellplate at travel speeds of 3.2 and 4.8 km/h. In a grassy field,the slot depth in the soil was fairly consistent in the range of5 to 8 cm.

The accuracy of the system to discharge cadavers at therates of 6.6, 3.3, or 1.6 cadavers/m length was determined.However, the discharge rate of cadavers to effectively controlspecific insect pests of roots remains to be resolved. Also,should a demand for the system develop, special meteringplates must be designed to improve the accuracy of deliveryof desiccated cadavers at various discharge rates.

ACKNOWLEDGEMENTS

The authors acknowledge Adam Clark for his valuabletechnical assistance in development of the delivery system,Joshua Sun and Austin Beery for their assistance inpreliminary tests during the delivery system development,and Dr. Ruisheng An for preparation of desiccated cadavers.

REFERENCESBateman, R. P. 1999. Delivery systems and protocols for

biopesticides. In Methods in Biotechnology, Biopesticides: Useand Delivery, 509‐528. Vol. 5. F. R. Hall and J. J. Menn, eds.Totowa, N.J.: Humana Press.

324 APPLIED ENGINEERING IN AGRICULTURE

Bedding, R.A., R.J. Akhurst, and H.K. Kaya. eds. 1993. Nematodesand the Biological Control of Insect Pests. CSIRO Publications,East Melbourne, Australia.

Bruck, D. J., D. I. Shapiro‐Ilan, and E. E. Lewis. 2005. Evaluationof application technologies of entomopathogenic nematodes forcontrol of the black vine weevil. J. Econ. Entomol. 98(6):1884‐1889.

Fife, J. P., R. C. Derksen, H. E. Ozkan, P. S. Grewal, and J. J.Chalmers. 2004. Evaluation of a contraction flow field onhydrodynamic damage to entomopathogenic nematodes: Abiological pest control agent. Biotech. Bioeng. 86(1): 96‐107.

Fife, J. P., H. E. Ozkan, R. C. Derksen, P. S. Grewal, and C. R.Krause. 2005. Viability of a biological pest control agentthrough hydraulic nozzles. Trans. ASABE 48(1): 45‐54.

Fife, J. P., H. E. Ozkan, R. C. Derksen, and P. S. Grewal. 2007.Effects of pumping on entomopathogenic nematodes andtemperature increase within a spray system. Applied Eng. inAgric. 23(4): 405‐412.

Grewal, P. S., R. U. Ehlers, and D. I. Shapiro‐Ilan, eds. 2005.Nematodes as Biocontrol Agents. Wallingford, Oxfordshire, UK:CABI Publishing, CAB International.

Klein, M. G., and R. Georgis. 1992. Persistence of control ofJapanese beetle (Coleoptera: Scarabaeidae) larvae withsteinernematid and heterorhabditid nematodes. J. Econ.Entomol. 85: 727‐730.

Reding, M. E., H. Zhu, and R. C. Derksen. 2008. Drip‐chemigationwith imidacloprid and nematodes for control of scarab larvae innursery crops. J. Environmental Hort. 26(2): 93‐100.

Schroeder, W. J., and J. B. Beavers. 1987. Movement of theentomogenous nematodes of the families Heterorhabditidae andSteinernematidae in soil. J. Nematology 19(2): 257‐259.

Shapiro, D. I., and E. E. Lewis. 1999. Comparisons ofentomopathogenic nematode infectivity from infected hostsversus aqueous suspension. Environmental Entomology 28(5):907‐911.

Shapiro‐Ilan, D. I, E. E. Lewis, R. W. Behle, and M. R. McGuire.2001. Formulation of entomopathogenic nematode‐infectedcadavers. J. Invert. Pathol. 78: 17‐23.

Shapiro‐Ilan, D. I., E. E. Lewis, and W. L. Tedders. 2003. Superiorefficacy observed in entomopathogenic nematodes applied ininfected‐host cadavers compared with application in aqueoussuspensions. J. Invert. Pathol. 83: 270‐272.

Steinke, W. E., and D. K. Giles. 1995. Delivery systems forbiorational agents. In Biorational Pest Control Agents – Formulation and Delivery, 80‐94. F. R. Hall and J. W. Barry,eds. Washington, D.C.: American Chemical Society.

Wang, X., H. Zhu, M. E. Reding, J. C. Locke, J. E. Leland, A. L.Spongberg, R. C. Derksen, and C. R. Krause. 2009. Distributionof agri‐chemicals and microbial biopesticides applied by dripirrigation. Applied Eng. in Agric. 25(6): 883‐893.

Wennemann, L., W. W. Cone, L. C. Wright, J. Perez, and M. M.Conant. 2003. Distribution patterns of entomopathogenicnematodes applied through drip irrigation systems. J. Econ.Entomol. 96(2): 287‐291.