University of Nebraska - Lincoln

DigitalCommons@University of Nebraska - Lincoln

Faculty Publications: Department of Entomology, Department of

Entomology

2007

Economic Threshold for Soybean Aphid (Hemiptera: Aphididae)

D. W. Ragsdale
University of Minnesota, ragsd001@umn.edu

B. P. McCornack
University of Minnesota

R. C. Venette
U.S. Forest Service

B. D. Potter
University of Minnesota

I. V. Macrae
University of Minnesota

See next page for additional authors

Follow this and additional works at: https://digitalcommons.unl.edu/entomologyfacpub

Part of the Entomology Commons

Ragsdale, D. W.; McCornack, B. P.; Venette, R. C.; Potter, B. D.; Macrae, I. V.; Hodgson, E. W.; O'Neal, M. E.;
Johnson, K. D.; O'Neil, R. J.; Difonzo, C. D.; Hunt, Thomas E.; Glogoza, P. A.; and Cullen, E. M., "Economic
Threshold for Soybean Aphid (Hemiptera: Aphididae)" (2007). Faculty Publications: Department of
Entomology. 297.
https://digitalcommons.unl.edu/entomologyfacpub/297

This Article is brought to you for free and open access by the Entomology, Department of at
DigitalCommons@University of Nebraska - Lincoln. It has been accepted for inclusion in Faculty Publications:
Department of Entomology by an authorized administrator of DigitalCommons@University of Nebraska - Lincoln.
Authors
D. W. Ragsdale, B. P. McCornack, R. C. Venette, B. D. Potter, I. V. Macrae, E. W. Hodgson, M. E. O'Neal, K. D.
Johnson, R. J. O'Neil, C. D. Difonzo, Thomas E. Hunt, P. A. Glogoza, and E. M. Cullen

This article is available at DigitalCommons@University of Nebraska - Lincoln: https://digitalcommons.unl.edu/

entomologyfacpub/297
 FIELD AND FORAGE CROPS

Economic Threshold for Soybean Aphid (Hemiptera: Aphididae)

D. W. RAGSDALE,1 B. P. MCCORNACK, R. C. VENETTE,2 B. D. POTTER,3 I. V. MACRAE,
E. W. HODGSON, M. E. O’NEAL,4 K. D. JOHNSON,4 R. J. O’NEIL,5 C. D. DIFONZO,6
T. E. HUNT,7 P. A. GLOGOZA,8 AND E. M. CULLEN9
Department of Entomology, University of Minnesota, 219 Hodson Hall, 1980 Folwell Avenue, St. Paul, MN 55108

J. Econ. Entomol. 100(4): 1258Ð1267 (2007)

ABSTRACT Soybean aphid, Aphis glycines Matsumura (Hemiptera: Aphididae), reached damaging
levels in 2003 and 2005 in soybean, Glycine max (L.) Merrill, in most northern U.S. states and Canadian
provinces, and it has become one of the most important pests of soybean throughout the North Central
region. A common experimental protocol was adopted by participants in six states who provided data
from 19 yield-loss experiments conducted over a 3-yr period. Population doubling times for Þeld
populations of soybean aphid averaged 6.8 d ⫾ 0.8 d (mean ⫾ SEM). The average economic threshold
(ET) over all control costs, market values, and yield was 273 ⫾ 38 (mean ⫾ 95% conÞdence interval
[CI], range 111Ð567) aphids per plant. This ET provides a 7-d lead time before aphid populations are
expected to exceed the economic injury level (EIL) of 674 ⫾ 95 (mean ⫾ 95% CI, range 275Ð1,399)
aphids per plant. Peak aphid density in 18 of the 19 location-years occurred during soybean growth
stages R3 (beginning pod formation) to R5 (full size pod) with a single data set having aphid
populations peaking at R6 (full size green seed). The ET developed here is strongly supported through
soybean growth stage R5. Setting an ET at lower aphid densities increases the risk to producers by
treating an aphid population that is growing too slowly to exceed the EIL in 7 d, eliminates generalist
predators, and exposes a larger portion of the soybean aphid population to selection by insecticides,
which could lead to development of insecticide resistance.

KEY WORDS yield loss, population dynamics, invasive species

The soybean aphid, Aphis glycines Matsumura al. 1994). In Minnesota, soybean aphid outbreaks are
(Hemiptera: Aphididae), is native to Asia, and it has associated with a reduction in plant height, pod num-
caused substantial damage to soybean, Glycine max ber, seed size and quality, and yield (Ostlie 2001). The
(L.) Merr., in North America since its conÞrmed oc- damage potential at low-to-moderate aphid densities
currence in August 2000 (Ragsdale et al. 2004). At is less clear, but soybean aphid feeding is known to
present, the soybean aphid is the most signiÞcant in- disrupt the photosynthetic processes at relatively low
sect threat to soybean production in North America. aphid densities (Macedo et al. 2003). Soybean aphid is
In China and in other parts of Asia, this insect is only also a vector of numerous plant viruses (Clark and
an occasional pest of soybean, and when plants are Perry 2002, Davis et al. 2005), which can further limit
colonized by soybean aphid in early vegetative growth soybean yield and seed quality.
stage, yield loss in excess of 50% can occur (Wang et Aphid population declines in annual cropping sys-
tems are attributed to variable host plant quality (e.g.,
1 Corresponding author, e-mail: ragsd001@umn.edu. physiological age and antibiosis), increased activities
2 U.S. Forest Service, North Central Research Station, 1561 Lindig of natural enemies, and weather extremes (van den
Ave., St. Paul, MN 55108. Berg et al. 1997, Fox et al. 2004, Karley et al. 2004, Li
3 Southwest Research & Outreach Center, University of Minnesota,

Lamberton, MN 56152.
et al. 2004). In controlled environments, soybean
4 Department of Entomology, Iowa State University, Ames, IA aphid populations can double in 1.5 d (McCornack et
50011. al. 2004), but these high intrinsic rates of increase are
5 Department of Entomology, Purdue University, 901 W. State St.,
only obtainable under ideal conditions where popu-
West Lafayette, IN 47907.
6 Department of Entomology, Michigan State University, 243 Nat- lation growth is not constrained by host quality, effects
ural Science Bldg., East Lansing, MI 48824. of weather, or natural enemies. Soybean aphid biology
7 Department of Entomology, University of Nebraska-Lincoln, and the speciÞc conditions that trigger rapid increases
Haskell Agricultural Laboratory, 57905 866 Rd., Concord, NE 68728. in population densities that are associated with yield
8 University of Minnesota, Regional Extension Center-Moorhead,

715 11th St. N, Suite 107C Moorhead, MN 56560.

reductions are not well understood in North America
9 Department of Entomology, 536 Russell Labs, University of Wis- (Ragsdale et al. 2004). However, Þeld estimates of
consin-Madison, 1630 Linden Dr., Madison, WI 53706. soybean aphid population growth rates are less than

0022-0493/07/1258Ð1267$04.00/0 䉷 2007 Entomological Society of America

August 2007 RAGSDALE ET AL.: ECONOMIC THRESHOLD FOR SOYBEAN APHID 1259

the theoretical intrinsic rate of increase (Costamagna ing trifoliolate containing three to Þve aphids was
and Landis 2006). Therefore, basing an economic excised and placed on approximately one plant per 30
threshold (ET) on population doubling times derived cm of row within each plot on 23 July 2004. The timing
from laboratory experiments that occurred in the ab- of this artiÞcial infestation matched the general ap-
sence of any environmental resistance will result in an pearance and density of aphids in most Nebraska soy-
artiÞcially low economic threshold. Such an economic bean Þelds. In all location-years, a foliar insecticide,
threshold based on laboratory derived intrinsic rate of lambda-cyhalothrin at 16.8 Ð28.0 g (AI)/ha (Warrior
increase has been calculated (Olson and Badibanga with Zeon Technology, Syngenta Crop Protection,
2005a), resulting in a threshold of three aphids per Greensboro, NC), was applied to all plots in a given
plant, which in their model had the highest economic treatment by using ground equipment once a target
return. Such a threshold is not realistic, because it aphid density in a treatment in terms of CAD was
assumes that the multiple sources of environmental reached (averaged across all blocks). In all cases, in-
resistance would not prevent exponential growth of secticides were applied within 2 d after aphid counts
soybean aphid populations. were completed. If soybean aphid populations began
The objective of this study was to quantify the to increase after the initial insecticide application,
relationship between aphid densities and yield loss additional applications were applied to prevent aphid
under Þeld conditions in which biotic and abiotic populations from increasing.
factors were allowed to inßuence soybean aphid den- Aphid Sampling and Soybean Yield. Nondestruc-
sities. These data were used to estimate the aphid tive whole-plant samples were taken to enumerate the
density at which control measures should be applied total number of aphids per plant. To detect small
to prevent yield losses. ETs and economic injury levels populations early in the season, up to 20 plants per plot
(EIL) were developed based on current expected were inspected. As the season progressed and the
yields, control costs, and market values for U.S. soy- frequency of encountering plants with aphids in-
bean. creased to 50%, 10 plants per plot were sampled. When
⬎80% of plants were aphid infested, Þve plants were
Materials and Methods counted per plot at each sampling date. For analysis,
all data were converted to mean number of aphids per
Field Plot Design. In 2003, 2004, and 2005, a com- plant per plot. Soybean growth stages (Fehr and Cavi-
mon experimental protocol was used at sites located in ness 1977), whether vegetative or reproductive, were
six states (Iowa, Michigan, Minnesota, Nebraska, noted each week.
North Dakota, and Wisconsin), so that comparisons Yield was estimated by harvesting the entire middle
could be made across locations and years (19 location- two rows of each plot with a small-plot combine and
years). At each location, a soybean variety was se- adjusting seed moisture to 13%. Linear regression
lected that was adapted for that area, and it was (PROC REG; SAS Institute 2001) was then used to
planted from mid- to late May (Table 1). Plots were relate percentage yield reduction to CAD; slope and
3.0 m in width (four rows) by 12.3 m in length with intercept estimates were used in all EIL calculations.
a 76.2-cm (30-in.) row spacing. We used predeter- Values Used in Calculation of an Economic Injury
mined, targeted aphid population densities based on Level. Cost estimates for insecticide and application
cumulative aphid-days (CAD) of 0, 2,000, 4,000, 8,000, costs, market value, and expected yield were used to
12,000, 16,000, and an untreated control (⫽maximum calculate an EIL for soybean aphid. A gain threshold
CAD) as treatments. Cumulative aphid-days is a single (GT) expressed in percentage yield loss was calcu-
value that provides a measure of aphid abundance lated by estimating control costs (C) [$/ha] divided
over time, and it can be calculated weekly as sampling by estimated market value (V) [$/ton] by using var-
occurs. We calculated CAD by using the procedures ious yield potentials (Y) [tons/ha] (Pedigo et al.
outlined in HanaÞ et al. (1989). 1986), which is equivalent to
Insecticide treatments varied among locations and
years and depended largely upon the natural level of
aphid infestation in any given location-year. Each tar- C
GT 共% yield loss兲 ⫽ ⫻ 100 [1]
get aphid density was replicated a minimum of four V⫻Y
times within each location-year, and treatments were
arranged in a randomized block design. With the ex- Average retail price of representative insecticides reg-
ception of one location-year in Minnesota in 2003 istered for soybean aphid control and their associated
where the study was located in a commercial produc- application costs were obtained from an informal
tion Þeld, fallow ground of ⬇3 m surrounded each plot phone survey of multiple local elevators along with
to facilitate application of insecticide to individual published sources (Dobbins et al. 2004, WASS 2004,
plots, minimize spray drift among plots, and encourage Edwards and Smith 2005). Average soybean prices
uniform aphid colonization throughout the experi- from 2000 to 2005 were obtained from the National
ment (DiFonzo et al. 1996, Hodgson et al. 2005). Soy- Agriculture Statistical Services (NASS 2006). Finally,
bean aphids were allowed to naturally colonize the soybean yield potentials used in the calculation of the
Þeld except in Nebraska in 2004 where soybean aphids GT represent the range of long-term average soybean
were seeded into plots by using Þeld-collected aphids yield throughout the North Central growing region
from a nearby Þeld. In the Nebraska plots, an expand- (NASS 2006).
 1260

Table 1. Soybean aphid population growth rates for the untreated controls in each of the 19 location-years

Population Discrete daily

Planting Julian Julian Doubling
Location (city, state) Brand; variety growth ratec Interceptd ⫾ SEM R2 P value growth rate
date start datea end dateb time(d)e
(r) ⫾ SEM (␭)
Rosemount, MN 21 May 2003 NK; S19-V2(RR) 189 209 0.192 ⫾ 0.010 ⫺33.9 ⫾ 2.0 0.995 0.003 1.211 3.6
Rosemount, MN 3 June 2003 NK; S19-V2(RR) 189 209 0.229 ⫾ 0.024 ⫺40.1 ⫾ 4.7 0.979 0.010 1.257 3.0
Lamberton, MN, Þeld 1 10 May 2003 Thompson; 7227CR 200 234 0.198 ⫾ 0.016 ⫺38.1 ⫾ 3.5 0.974 ⬍0.001 1.219 3.5
Lamberton, MN, Þeld 2 16 May 2003 Cropplan; R1976 199 227 0.261 ⫾ 0.031 ⫺50.5 ⫾ 6.8 0.957 0.004 1.298 2.7
New Ulm, MN 5 May 2003 Stine; 1918-4 193 220 0.074 ⫾ 0.022 ⫺9.5 ⫾ 4.6 0.733 0.030 1.077 9.4
Prosper, ND 24 May 2003 Asgrow; AG0801(RR) 202 230 0.144 ⫾ 0.006 ⫺27.9 ⫾ 1.3 0.995 ⬍0.001 1.155 4.8
Rosemount, MN, Þeld 1 28 May 2004 Pioneer; 91B91(RR) 209 223 0.113 ⫾ 0.012 ⫺22.0 ⫾ 2.6 0.988 0.069 1.119 6.2
Rosemount, MN, Þeld 2 28 May 2004 Pioneer; 91B91(RR) 209 223 0.143 ⫾ 0.024 ⫺19.0 ⫾ 11.9 0.781 0.310 1.110 6.7
Rosemount, MN, Þeld 3 28 May 2004 NK; S19-R5(RR) 179 214 0.143 ⫾ 0.024 ⫺25.2 ⫾ 4.7 0.901 0.004 1.153 4.9
Concord, NE 29 May 2004 Asgrow; 2703 207 242 0.139 ⫾ 0.018 ⫺25.5 ⫾ 4.1 0.934 0.002 1.149 5.0
Chariton, IA 10 May 2005 Stine; 3532-4 209 249 0.071 ⫾ 0.022 ⫺11.9 ⫾ 5.0 0.783 0.046 1.073 9.8
24 May 2005 Stine; 3532-4 209 249 0.067 ⫾ 0.024 ⫺10.8 ⫾ 5.6 0.715 0.071 1.069 10.4
Ames, IA 23 May 2005 Prairie Brand; PB-2183RR 202 241 0.077 ⫾ 0.009 ⫺11.8 ⫾ 2.0 0.944 0.001 1.080 9.0
16 June 2005 Prairie Brand; PB-2183RR 202 241 0.085 ⫾ 0.014 ⫺14.1 ⫾ 3.0 0.906 0.003 1.089 8.2
Nashua, Iowa 5 May 2005 Crows; C 2133 R 206 243 0.053 ⫾ 0.011 ⫺6.5 ⫾ 2.5 0.887 0.017 1.055 13.1
21 May 2005 Crows; C 2133 R 206 243 0.052 ⫾ 0.014 ⫺6.3 ⫾ 3.3 0.810 0.038 1.053 13.4
East Lansing, MI 1 May 2005 Pioneer; 92B38(RR) 179 229 0.128 ⫾ 0.013 ⫺20.5 ⫾ 2.7 0.942 ⬍0.001 1.137 5.4
JOURNAL OF ECONOMIC ENTOMOLOGY

Rosemount, MN 24 May 2005 NK; S19-R5(RR) 192 214 0.161 ⫾ 0.016 ⫺28.4 ⫾ 3.3 0.980 0.010 1.175 4.3
Arlington, WI 18 May 2005 NK; S19-V2(RR) 164 194 0.129 ⫾ 0.015 ⫺20.4 ⫾ 2.6 0.950 0.001 1.137 5.4
Mean 197 228 0.127 1.138 6.8
SEM 3 4 0.014 0.016 0.8

a
Date when soybean plants in untreated control plots reached 80% infestation.
b
Date when soybean aphid populations peaked in untreated control plots.
c
The underlying model for population growth is Nt ⫽ N0ert , which is equivalent to ln(Nt) ⫽ ln(N0) ⫹ rt, where N0 is the initial density, r is the population growth rate (i.e., slope from the linear regression),
and t is expressed in Julian days.
d
The intercept equals the ln(N0), where N0 is the initial density.
e
Doubling time equals ln(2) divided by r.
Vol. 100, no. 4
August 2007 RAGSDALE ET AL.: ECONOMIC THRESHOLD FOR SOYBEAN APHID 1261

An EIL expressed in CAD was calculated using an the underlying model for population growth is Nt ⫽
adjusted percentage yield potential by subtracting the N0ert, which is equivalent to ln(Nt) ⫽ ln(N0) ⫹ rt,
GT calculated in equation 1 from the maximum po- where N0 is the initial aphid density, r is the population
tential yield set at 100% and incorporating the y-in- growth rate (i.e., slope from the linear regression), and
tercept (␤0) and slope (␤1) parameters estimated t is expressed in d (Julian days). Discrete daily growth
from the linear regression of CAD against percentage rate (␭) was calculated as er, averaged across all lo-
yield loss and can be written as follows: cation-years, and used in the calculation of all ETs.
␤ 0 ⫺ 共% yield potential)
EIL 共CAD兲 ⫽ Results and Discussion
␤1 [2]
To convert the EIL expressed as CAD to an EIL based Market Value and Control Costs Used in Calcula-
on aphid density (aphids per plant) at a particular tion of the EIL. Application cost of using a personally
point in time, we used the general formula for the owned, nonpropelled, boom sprayer was estimated at
summation of a geometric progression: $5.09/ha by Lazarus and Selley (2005). Their costs
included fuel, lubricants, repairs, maintenance, labor,
l␭ ⫺ a and power and implement depreciation, interest, in-
s⫽ [3]
␭⫺1 surance, and housing of equipment. Custom applica-
tion costs for ground application in 2005 averaged
where in our application, s is the CAD on a per plant
$12.23/ha (range, $8.65 to $21.00/ha) (Dobbins et al.
basis; ␭ is the discrete time population growth rate
2004, WASS 2004, Edwards and Smith 2005). Custom
(␭ ⫽ er with e being the base of the natural logarithm
aerial application costs in 2004 and 2005 averaged
and r being the population growth rate); a is aphid
$14.85/ha (ranged from $12.36 to $16.68/ha) (Dob-
density at the start of accumulation of aphid-days; and
bins et al. 2004; Edwards and Smith 2005). We ob-
l is aphid density on the last day of the series. In this
tained retail cost of commonly used insecticides for
case, l is also the EIL but expressed as aphids per plant.
soybean aphid control that included pyrethroids
For convenience we conservatively set a ⫽ 1 under
(lambda-cyhalothrin; zeta-cypermethrin, or Mustang,
the assumption that the accumulation of aphid-days is
FMC Corporation, Philadelphia, PA; and esfenvaler-
insigniÞcant until densities reach an average of one
ate or Asana XL, DuPont, Wilmington, DE) and an
aphid per plant. We rearrange this equation to solve
organophosphate (chlorpyrifos or Lorsban-4E, Dow
for l, which yields the following:
AgroSciences LLC, Indianapolis, IN). Total control
s共 ␭ ⫺ 1兲 ⫹ 1 costs included the insecticide and its application, and
l⫽ [4] we derived three estimates of soybean aphid control
␭
costs: 1) lowest application rate of $16.41/ha ($6.64/
In this equation, s is equivalent to the EIL calculated acre) by using the lowest cost insecticide applied with
from equation 2. Finally, to convert this EIL to an grower-owned equipment, 2) a mid-range control cost
economic threshold, expressed in aphids per plant that of $24.51/ha ($9.92/acre) that is representative of
crop managers will use, we calculated aphid densities custom ground application of a moderately priced
t days before reaching the EIL based on the equation insecticide, and 3) a high control cost option at
l␭-t. For our purposes we set t, or lead time, to 1, 3, 5, $32.94/ha ($13.33/acre), which represents a maxi-
and 7 d. This ET also assumes that the aphid population mum labeled rate of an expensive insecticide custom
is increasing and that crop managers will need any- applied by air (Table 2).
where from 1 to 7 d to make arrangements to have a Market values used in the calculation of the EIL
foliar insecticide applied to a Þeld. To solve for the represented three probable soybean prices for the
number of d (t) it takes for a given aphid density to North Central Region: 1) $202.09/ton soybean ($5.50/
reach a speciÞc EIL (l) we can rearrange the equation bu) as a conservative or lowest expected market price,
used to calculate lead time, 2) a mid-range market price of $220.46/ton ($6.00/
ln共l/a兲 bu), and 3) an optimistic soybean price of $238.83/ton
t⫽ ($6.50/bu). Our soybean price estimates are not a
ln共 ␭ 兲 [5] Þne-tuned forecast for soybean prices, but rather we
where a is the initial aphid density and ␭ is derived used these different values to assess the sensitivity of
from the estimated population growth rates. the soybean aphid EIL to ßuctuating soybean prices
To estimate ␭ used in equations 4 and 5, we averaged and application costs (Barrigossi et al. 2003). Yield
the population growth rate (r) from all 19 location- expectations used in calculating the EIL ranged from
years by using aphid population growth data from the 2.02 ton/ha (30 bu/acre) to 4.04 ton/ha (60 bu/acre).
untreated control plots where maximum CAD oc- These yield expectations represent a range of average
curred. Aphid densities were natural log transformed yields reported for various states or regions within the
and simple linear regression was used to estimate r for United States (NASS 2006). Other economic factors
each individual location-year. The population growth that may affect some production systems also should
rate was calculated using aphid densities between the be considered when estimating an ET, but they are too
periods when aphid populations Þrst reached 80% numerous to estimate here. For example, in narrow
plant infestation and the point in time when peak row soybean production systems, yield losses from 1 to
aphid densities were observed. For this application, 2.5% are caused by driving ground equipment through
1262 JOURNAL OF ECONOMIC ENTOMOLOGY Vol. 100, no. 4

Table 2. Economic thresholds (ET) and economic injury levels (EIL) for soybean aphid with various control costs, market prices, and
soybean yield potentials

Cost of control Market price Yield potential EIL: cumulative EIL: aphids ET with different lead times (d):
($/ha)a ($/ton)b (ton/ha)c aphid-days per plant 1 3 5 7
16.41 202.09 2.02 5,649 684 601 465 359 278
2.69 4,188 507 446 345 266 206
3.36 3,309 401 353 272 211 163
4.04 2,715 329 289 224 173 134
220.46 2.02 5,160 625 549 425 328 254
2.69 3,821 463 407 314 243 188
3.36 3,015 366 321 248 192 148
4.04 2,471 300 263 204 157 122
238.83 2.02 4,747 575 505 391 302 233
2.69 3,510 425 374 289 223 173
3.36 2,766 335 295 228 176 136
4.04 2,264 275 241 187 144 111
24.51 202.09 2.02 8,546 1,035 909 703 543 420
2.69 6,363 771 677 523 404 313
3.36 5,051 612 538 416 321 248
4.04 4,164 504 443 343 265 205
220.46 2.02 7,816 946 832 643 497 384
2.69 5,815 704 619 478 370 286
3.36 4,611 559 491 379 293 227
4.04 3,798 460 405 313 242 187
238.83 2.02 7,198 871 766 592 457 353
2.69 5,350 648 570 440 340 263
3.36 4,240 514 452 349 270 208
4.04 3,489 423 372 287 222 172
32.94 202.09 2.02 11,561 1,399 1,230 950 734 567
2.69 8,627 1,044 918 709 548 424
3.36 6,863 831 730 564 436 337
4.04 5,671 687 604 467 360 279
220.46 2.02 9,696 1,174 1,032 797 616 476
2.69 7,890 955 840 649 501 387
3.36 6,273 760 668 516 399 308
4.04 5,180 627 552 426 329 254
238.83 2.02 8,933 1,081 951 735 568 439
2.69 7,266 880 773 598 462 357
3.36 5,773 699 615 475 367 284
4.04 4,765 577 507 392 303 234
Mean 5,563 674 592 458 354 273

a
Cost of control in $/ac equivalents are $6.64, 9.92, and 13.33, respectively.
b
Market value estimates in $/bu equivalents are $5.50, 6.00, and 6.50, respectively.
b
Yield expectations in bu/ac equivalents are 30, 40, 50 and 60, respectively.

mature soybean Þelds without a nonplanted tram line with the x-axis (Fig. 1B), this represents few or no
to apply insecticides (Beuerlein et al. 2005). Such aphids per plant were found during subsequent sam-
mechanical losses caused by ground application are pling periods.
not included in our calculated ET. Conversely, avail- Initial soybean aphid colonization across all
ability of generic insecticides could reduce control location-years occurred between 1 June and 23 July,
costs and seed treatment does seem to slow the early corresponding with plant growth stages V1 (Þrst tri-
season population growth rate of soybean aphid. How- foliolate) to reproductive stage R2 (multiple ßowers).
ever, to obtain maximum economic yield during aphid Mean peak aphid density and standard error (⫾SEM)
outbreak years, a foliar application may still be war- in untreated control plots for all location-years was
ranted in Þelds where seed treatments were used 1,262 ⫾ 351 aphids per plant with peak densities rang-
(McCornack and Ragsdale 2006). Here, we focused on ing from 17 to 4,275 aphids per plant. Peak aphid
the major economic variables when calculating a densities typically occurred between late July and late
range of ETs and EILs. August (Fig. 2) when soybeans were in reproductive
Aphid Population Densities and Associated Yields. growth stages R3 (pod formation) through R5 (full
An example of the graphical relationship between size pod). Mean CAD across all location-years was
aphids per plant and CAD from one location-year is 10,573 ⫾ 1,338 and ranged from 1 to 70,771 across all
presented in Fig. 1A and B, respectively. It is difÞcult target aphid densities (Fig. 2). Mean population
to discern the population trends among various target growth rate, r, across all location-years was 0.127/d ⫾
aphid densities (Fig. 1A) when plotting aphid density 0.014 or ␭ ⫽ er ⫽ 1.138/d ⫾ 0.016 with R2 values for
on a per plant basis. Differences are apparent when all location-years ranging from 0.715 to 0.995 (Table
converting from aphids per plant to a CAD scale (Fig. 1). Recall, r was calculated using transformed weekly
1B). When an individual CAD line remains parallel aphid counts beginning when sampling data showed
August 2007 RAGSDALE ET AL.: ECONOMIC THRESHOLD FOR SOYBEAN APHID 1263

the absence of natural enemies will result in too many

Þelds requiring treatment without realizing an eco-
nomic beneÞt. In other work (McCornack and Rags-
dale 2006), treating soybean aphid populations that
did not exceed the ET calculated here resulted in
accrued control costs without a measurable yield ben-
eÞt, thus a net loss to producers. For other aphid pests,
frequent application of insecticide has resulted in high
levels of insecticide resistance (Radcliffe and Rags-
dale 2002, Wang et al. 2002). Care must always be
exercised when dealing with aphids that reproduce
parthenogenically during the growing season to avoid
repeated insecticide applications that could lead to
resistance.
Across all 19 location-years in plots where the target
aphid density was 0 CAD, the observed CAD averaged
1,567 ⫾ 446. Most of the CAD in these plots occurred
during vegetative growth stages V1 to V8, and subse-
quent insecticide application kept aphid density low
during the reproductive stages. Myers et al. (2005)
showed that application of insecticides to vegetative
growth stages for soybean aphid control had no mea-
surable impact on yield, so any plant injury caused by
aphids feeding during vegetative growth stage in this
Fig. 1. Soybean aphid population curves expressed as study was likely immeasurable. At the time of the Þrst
aphids per plant (A) and CAD (i.e., one aphid per plant per insecticide application to the plots with a target aphid
day equals one aphid-day) (B) over multiple sampling times density of 0 CAD, mean aphid density on a per plant
in one location-year (Rosemount, MN, in 2003). basis was 17 ⫾ 3 aphids per plant. Mean aphid density
on a per plant basis after application of lambda-cyha-
lothrin was 3 ⫾ 1 and ranged from 0 to 20. Only in four
⬎80% of plants were infested and ended when peak of the 19 location-years (two locations in Minnesota in
aphid density was attained. Average aphid density at 2003 and two locations in Iowa in 2005) was a second
80% infestation was 21 ⫾ 7 aphids per plant in the insecticide application made to plots with a target
untreated controls. From the estimates of r we calcu- aphid density of 0 CAD, and in two location-years
lated the average doubling time in days using the (one each in Iowa and Michigan in 2005), a third
equation ln(2)/r (Table 1) and for all 19 location- insecticide application was needed. In these six loca-
years the average doubling time for Þeld populations tion-years where additional insecticide application
was 6.8 ⫾ 0.8 d with observed doubling times ranging were deemed necessary, mean aphid density at the
from 2.7 to 13.4 d. McCornack et al. (2004) estimated time of the second application was 77 ⫾ 13 and for the
intrinsic rate of increase under controlled conditions third application mean aphid density was 12 ⫾ 9 aphids
at various constant temperatures and values ranged per plant. We intentionally avoided more frequent
from ⫺0.383Ð 0.474 at constant 35 and 20⬚C, respec- application of insecticides, e.g., weekly, to plots where
tively. Their estimate of population doubling time at the target aphid density was 0 CAD. By applying
the optimal temperature of 27.8⬚C was calculated as insecticides too frequently, especially pyrethroids,
1.3 d. These population growth estimates by McCor- twospotted spider mite, Tetranychus urticae Koch, out-
nack et al. (2004) represent the biotic potential for breaks can occur (Yang et al. 2002, Steffey et al. 2006),
soybean aphid in absence of environmental resistance. which would confound the yield loss relationship. No
Here, we estimated growth rates (r) of Þeld popula- spider mite outbreaks were noted from any of the 19
tions which accounted for impact of natural enemies, location-years. In all remaining location-years (13 of
weather, increase or loss of aphids from winged aphids 19) only a single application was needed to control
(immigration and emigration), and other factors (van aphids and population growth did not increase after an
den Berg et al. 1997, Li et al. 2004, Costamagna and insecticide application.
Landis 2006). Because the experiment was conducted in six states
Others have calculated an ET for soybean aphid by where the soybean maturity group typically planted in
using laboratory-derived intrinsic rate of increase (Ol- each state ranged from group 00 to group 3, we were
son and Badibanga 2005a) or from population growth unable to directly compare yield losses among loca-
rates of caged Þeld populations (Catangui 2006), tions. Average yield in plots that had the fewest CAD
which excluded natural enemies and other events was 3.38 ⫾ 0.23 ton/ha (50.2 ⫾ 3.4 bu/acre) and
such as rainfall that can signiÞcantly reduce popula- ranged from 1.96 to 4.8 ton/ha (29.2Ð71.4 bu/acre).
tion growth rates (Dixon 1976). Setting an ET too low Because weather conditions, soil type, maturity group,
by using population doubling times based on labora- and other agronomic factors were highly variable
tory-derived reproductive rates or those that occur in among location-years, we measured the change in
1264 JOURNAL OF ECONOMIC ENTOMOLOGY Vol. 100, no. 4

Fig. 2. Cumulative aphid-day curves for all 19 location-years during the 2003, 2004, and 2005 seasons. (AÐE) Minnesota
2003. (GÐI) Minnesota 2004. (O) Minnesota 2005. (F) North Dakota 2004. (J) Nebraska 2004. (KÐM) Iowa 2005 (each panel
for Iowa represents two planting dates at each of three unique locations for a total of six location-years). (N) Michigan 2005.
(P) Wisconsin 2005. Symbols represent different target CAD densities within each location-year.

yield (percentage of maximum yield) as aphid-days addition, the y-intercept [y ⫽ 0.9985 ⫺ 0.0688(CAD)]
accumulated. This allowed us to directly compare data from the linear regression passed through 100% of the
from all 19 location-years. Within a single location- proportion maximum yield (Fig. 2), indicating that the
year we measured yield where the target aphid density yield loss relationship was best explained by the linear
was 0 CAD and designated this yield as the maximum regression and that when CAD were near zero yield
yield (100%) obtainable. We then measured the per-
centage yield loss relative to this maximum yield in
treatments where the target aphid density was ⬎0
CAD. It is not possible to totally eliminate aphids even
with repeated insecticide application, but the goal was
to keep aphids as low as possible in the treatment
where the target aphid density was 0 CAD without
ßaring secondary pests. For these 19 location-years,
there was no evidence of bean leaf beetle, Cerotoma
trifurcata (Forster), feeding or injury caused from
other defoliating insects reported, so yield losses mea-
sured here are from plant damage solely attributed to
soybean aphid feeding injury.
We used linear regression to relate relative yield
obtained in the plots where the target aphid density
was 0 CAD to relative yield in plots where aphids were
allowed to accrue higher CAD. We observed that Fig. 3. Percentage of maximum yield comparing plots
CAD was negatively correlated with yield (F ⫽ 212.09; with the target aphid density of 0 CAD to plots with target
df ⫽ 1, 103; P ⬍ 0.0001; R2 ⫽ 0.665) (Fig. 3). In aphid densities ⬎0 CAD for all 19 location-years (n ⫽ 116).
August 2007 RAGSDALE ET AL.: ECONOMIC THRESHOLD FOR SOYBEAN APHID 1265

observed was not different from the yield potential. late the aphid density on a per plant basis that would
The 95% conÞdence interval (CI) for the y-intercept occur at selected intervals before reaching the EIL.
was 0.984 Ð1.016, and, therefore not signiÞcantly dif- We arbitrarily selected 1, 3, 5, and 7 d before reaching
ferent from 1.0 at ␣ ⫽ 0.05. Yield (tons per hectare) the EIL, and we estimated the ET in terms of aphids
was reduced by 6.88% (95% CI was 5.94 Ð7.82%) for per plant (Table 2). The mean ET across all yield
every 10,000 aphid-days accumulated (Fig. 3). expectations, market values, and control costs pre-
As an example of how long it takes to accumulate sented in Table 2 with a lead time of 7 d was 273 ⫾ 38
⬇5,563 aphid-days, the mean EIL listed in Table 2, for (95% CI) aphids per plant. The corresponding EIL was
a 7-d lead time, we provide the following hypothetical 674 ⫾ 95 (95% CI) aphids per plant.
calculation. If the starting aphid population was one In 18 of the 19 location-years, peak aphid abundance
aphid per plant (a) and we used the lowest and highest was reached between growth stages R3 to R5; at the
EILs from Table 2 of 275 and 1,399 for the values of (l) time of 80% infestation, when measurement of aphid
in equation 5 and the mean value for ␭ of 1.138 (Table population growth rates were initiated, plots were in
1), we can solve for the number of d (t) it will take this plant growth stages R1 (Þrst ßower) to R4 (interme-
aphid population to reach either EIL. For the EIL of diate pod formation). Only one location-year (Ne-
275, it will take 43 d and for the EIL of 1,399, it will take braska 2004) had peak aphid abundance that occurred
56 d or ⬇6 Ð 8 wk for aphids to go from one aphid per during soybean growth stage R6 (full size green seed).
plant until they reach the EIL. Although a signiÞcant yield loss was measured at the
A more relevant question might be how long will a Nebraska 2004 location, Ostlie (2001) showed with
starting population of 100 aphids per plant take to on-farm strip trials that producers were less likely to
reach the EIL? First, we must use equation 3 to esti- achieve an economic beneÞt by treating aphid pop-
mate how many CAD occurred between 1 and 100 ulations when plants were in reproductive stage R6 or
aphids per plant. Solving for s in equation 3, there were later. With the majority of the location-years having
817 CAD, and this value is subtracted from the average peak aphid abundance that exceeded 80% infestation
EIL in CAD (Table 2) of 5,563, leaving 4,746 CAD as early as R1 and the ET exceeded at the latest during
remaining. Using this remaining value in equation 4, R5, the ET developed here using a 7d lead time is valid
we can solve for l, which is 576 aphids per plant. The between R1 to R5. More research is needed to es-
number of days it will take to go from 100 aphids per timate a valid threshold for growth stages after R5.
plant to 576 aphids per plant can be estimated using Our data suggest that an ET for R6 and later growth
equation 5 by using the value of ␭ as 1.138 (Table 1), stages will exceed 273 aphids per plant, but we have
which when solved for (t) in equation 5, the EIL will too few data sets to accurately estimate the ET
be reached in 14 d. during R6 and no data for aphid populations that
In general, as aphid populations increase, the more might peak during R7.
accurate the prediction of when the EIL will be These data collected from 19 location-years over a
reached. To predict when the aphid density will reach 3-yr period and in six states represent a wide range of
the EIL starting with a density of one aphid per plant soybean production systems and aphid infestations
is nearly impossible. There is too much time (6 Ð 8 wk) with respect to the period between initial colonization
where weather, natural enemies, disease, host plant and peak population density (Fig. 2). EILs and the
quality, and other factors could inßuence the aphid calculated ET for any pest needs to be dynamic and
population growth rate. In the examples above, only respond to changing conditions in market value, ex-
15% of the CAD needed to reach the EIL were realized pected yield, and variable control costs (Barrigossi et
when aphid populations were between 1 and 100 al. 2003). Table 2 covers a wide variation in the asso-
aphids per plant. It would take ⬇4 Ð 6 wk for aphid ciated costs and market values using the yield-loss
densities to reach 100 aphids per plant. Conversely, regression equation (Fig. 3) to calculate an ET. The
85% of the CAD needed to achieve the EIL occurred aphid population data we used to derive the EIL and
after the aphid density reached 100 per plant and an corresponding ET were from Þeld populations that
estimated additional 14 d at the average aphid popu- increased in the presence of natural enemies (pred-
lation growth rates to reach the EIL once aphid den- ators, parasites, and pathogens), adverse weather con-
sities reached 100 per plant. ditions (heavy rainfall, drought stress, and low and
Economic Threshold. Although we estimated the high temperatures that are beyond optimal ranges),
yield response of soybean injury to soybean aphid and plant growth stages. We made the assumption that
feeding using CAD as a measure of season-long expo- all varieties used in the study were susceptible to
sure of plants to soybean aphid, we suggest that this soybean aphid, because no aphid-resistant soybean
method will not be as useful to most producers or their variety is currently available to growers in the North
advisors as an ET based on aphid density per plant. To Central region (Li et al. 2004). In the future, soybean
calculate CAD requires regular and multiple visits to varieties will be developed and released that are re-
the Þeld. If sampling occurs less frequently or at ir- sistant or tolerant to soybean aphid. Equally likely is
regular intervals that greatly exceed 7 d, such esti- the release and establishment of classical biological
mates of CAD are less reliable. Therefore, we con- control agents along with a greater understanding of
verted the EIL based on CAD to the average aphid the value of native natural enemies. This ET will there-
density per plant using the average, Þeld-based, Þnite fore need to be modiÞed to account for such changes
population doubling time of 6.8 d (Table 1) to calcu- to the soybean production system. For example, re-
1266 JOURNAL OF ECONOMIC ENTOMOLOGY Vol. 100, no. 4

sistant soybean varieties may delay aphid doubling References Cited

times and EILs for tolerant varieties would be higher
Barrigossi, J.A.F., G. L. Hein, and L. G. Higley. 2003. Eco-
than our reported values. Also, new yield loss rela- nomic injury level and sequential sampling plans for Mex-
tionships will need to be estimated for resistant or ican bean beetle (Coleoptera: Coccinellidae) on dry
tolerant soybean varieties. beans. J. Econ. Entomol. 96: 1160 Ð1167.
If the maximum soybean yield and highest soybean Beuerlein, J. E., R. B. Hammond, A. E. Dorrance, and D. R.
price is expected and the lowest possible treatment Mills. 2005. Using Skip rows for soybean pest man-
cost is assumed, the ET with a 7-d lead time is 111 agement increases proÞt. The Ohio State University,
aphids per plant with a corresponding EIL of 275 Agonomic Crops Network. (http://agcrops.osu.edu/
soybean/SkipRow5%20October%202005.pdf).
(Table 2). However, caution must be used in inter-
Catangui, M. A. 2006. Economic threshold of the soybean
preting these calculated thresholds that fall well below aphid, A. glycines, in South Dakota. (http://plantsci.
the average ET of 273 and EIL of 674 aphids per sdstate.edu/ent/entpubs/sa_economic_threshold.htm).
plant. The GT for the lowest treatment cost is approx- Clark, A. J., and K. L. Perry. 2002. Transmissibility of Þeld
imately the value of 0.07 ton/ha (⬇1 bu/acre), and isolates of soybean viruses by Aphis glycines. Plant Dis. 86:
signiÞcant yield differences this small were not mea- 1219 Ð1222.
surable from any of our 19 location-years. Essentially, Costamagna, A. C., and D. A. Landis. 2006. Predators exert
ET and EIL values can be calculated using the yield- top-down control of soybean aphid across a gradient of
agricultural management systems. Ecol. Appl. 16: 1619 Ð
loss equation from this study (Fig. 3), but we consider 1628.
these low ET values impractical, and the yield loss Davis, J. A., E. B. Radcliffe, and D. W. Ragsdale. 2005. Soy-
associated with the corresponding EILs is immeasur- bean aphid, Aphis glycines Matsumura, a new vector of
able. The value of Table 2 is to demonstrate the rel- Potato Virus Y in potato. Am. J. Potato Res. 81: 101Ð105.
ative sensitivity of the variables used in the EIL and ET DiFonzo, C. D., D. W. Ragsdale, E. B. Radcliffe, N. C.
calculations. Gudmestad, and G. A. Secor. 1996. Crop borders reduce
The ET calculated from this study of 273 ⫾ 38 (95% potato virus Y incidence in seed potato. Ann. Appl. Biol.
CI) overlaps a consensus action threshold that was 129: 289 Ð302.
Dixon, A.F.G. 1976. Timing of egg hatch and viability of the
promoted after a widespread soybean aphid outbreak sycamore aphid, Drepanosiphum platanoidis (Schr.), at
that occurred in 2003 of 250 aphids per plant with bud burst of sycamore. J. Anim. Ecol. 45: 593Ð 603.
⬎80% of plants being aphid-infested (NCSRP 2006). Dobbins, C. L., S. Wilson, and Z. Cain. 2004. Agricultural
This action threshold was derived from a subset of economics: Indiana custom rates 2004, pp 1Ð3. Purdue
location-years included in this study (six of the 19 University Extension Publication EC-130-W. Purdue Uni-
location-years). In a recent survey, 66% of respon- versity, West Lafayette, IN.
dents correctly identiÞed, this action threshold and Edwards, W., and D. Smith. 2005. Ag decision maker: 2005
84% indicated that scouting for aphids was critical for Iowa farm custom rate survey, pp 1Ð3. Iowa State Uni-
versity Extension, Ames, IA. (www.extension.iastate.
effective aphid management (Olson and Badibanga edu/agdm).
2005b). The long-term beneÞts of delaying treatment Fehr, W. R., and C. E. Caviness. 1977. Stages of soybean
for as long as practical are far reaching. Some natural development. Iowa State University Cooperative Ex-
enemies in soybean are known to follow soybean tension Service Special Rep. 80. Iowa State University,
aphid to its principal overwintering host, common Ames, IA.
buckthorn, Rhamnus cathartica L., (Yoo et al. 2005), Fox, T. B., D. A. Landis, F. F. Cardoso, and C. D. DiFonzo.
and these predators and pathogens (Nielsen and Ha- 2004. Predators suppress Aphis glycines Matsumura pop-
ulation growth in soybean. Environ. Entomol. 33: 608 Ð
jek 2005) continue to reduce aphid populations on the
618.
overwintering host well past soybean harvest. Con- Hanafi, A., E. B. Radcliffe, and D. W. Ragsdale. 1989. Spread
serving soybean aphid natural enemies is of utmost and control of potato leafroll virus in Minnesota. J. Econ.
importance and this ET, if widely adopted, will help Entomol. 82: 1201Ð1206.
preserve natural enemies on a landscape level and Hodgson, E. W., R. L. Koch, and D. W. Ragsdale. 2005. Pan
provide producers the means to make decisions that trapping for soybean aphid (Homoptera: Aphididae) in
avoid treating subeconomic aphid populations. Minnesota soybean. J. Entomol. Sci. 40: 409 Ð 419.
Karley, A. J., W. E. Parker, J. W. Pitchford, and A. E. Douglas.
2004. The mid-season crash in aphid populations: why
and how does it occur? Ecol. Entomol 29: 383Ð388.
Acknowledgments Lazarus, W., and R. Selley. 2005. Farm machinery economic
cost estimates for late 2005, pp 1Ð12. University of
We thank the North Central Soybean Research Program, Minnesota Extension Service, University of Minnesota,
the Iowa Soybean Association, the Minnesota Soybean Re- St. Paul, MN. (http://www.apec.umn.edu/faculty/
search and Promotion Council, the North Dakota Soybean wlazarus/mf2005late.pdf).
Council, the University of Minnesota Rapid Agricultural Re- Li, Y., C. B. Hill, and G. L. Hartman. 2004. Effect of three
sponse Fund, and the Agricultural Experiment Stations in resistant soybean genotypes on the fecundity, mortality,
Indiana, Iowa, Michigan, Minnesota, Nebraska, North Da- and maturation, of soybean aphid (Homoptera: Aphidi-
kota, and Wisconsin for Þnancial assistance and support for dae). J. Econ. Entomol. 97: 235Ð239.
Þeldwork conducted in the participating states. We also Macedo, T. B., C. S. Bastos, L. G. Higley, K. R. Ostlie, and S.
thank the many individuals in the six states that helped Madhavan. 2003. Photosynthetic responses of soybean
collect data, count aphids, manage Þeld plots and assisted in to soybean aphid (Homoptera: Aphididae) injury. J.
all aspects of this 3-yr study. Econ. Entomol. 96: 188 Ð193.
August 2007 RAGSDALE ET AL.: ECONOMIC THRESHOLD FOR SOYBEAN APHID 1267

McCornack, B. P., and D. W. Ragsdale. 2006. EfÞcacy of Ragsdale, D. W., D. J. Voegtlin, and R. J. O’Neil. 2004. Soy-
thiamethoxam to suppress soybean aphid populations in bean aphid biology in North America. Ann. Entomol. Soc.
Minnesota soybean. J. Crop Manage. (http://dx.doi.org/ Am. 97: 204 Ð208.
10.1094/CM-2006-0915-01-RS). SAS Institute. 2001. PROC userÕs manual, version 6th ed.
McCornack, B., D. W. Ragsdale, and R. C. Venette. 2004. SAS Institute, Cary, NC.
Demography of soybean aphid (Homoptera: Aphididae) Steffey, K., M. Gray, R. Estes, J. Schroeder, D. Bakken, D.
at summer temperatures. J. Econ. Entomol. 97: 854 Ð 861. Schaefer, and G. Roskamp. 2006. Twospotted spider
Myers, S. W., D. B. Hogg, and J. L. Wedberg. 2005. Deter- mites: case study in soybean pest management. Illinois
mining the optimal timing of foliar insecticide applica- Crop Protection Technology Conference, 2006
tions for control of soybean aphid (Hemiptera: Aphidi- Proceedings. (http://www.ipm.uiuc.edu/conferences/cptc/
dae) on soybean. J. Econ. Entomol. 98: 2006 Ð2012. proceedings.pdf).
[NASS] National Agricultural Statistics Service. 2006. van den Berg, H., D. Ancaza, A. Mamad, R. Rusli, H. A.
United States Department of Agriculture (http://www. Widayanto, H. B. Wirasto, and I. Yully. 1997. Evaluating
nass.usda.gov/index.asp). the role of predation in population ßuctuations of the
[NCSRP] North Central Soybean Research Program, Plant soybean aphid Aphis glycines in farmersÕ Þeld in Indone-
Health Initiative. 2006. Soybean aphid management. sia. J. Appl. Ecol. 34: 971Ð984.
(http://www.planthealth.info/aphids_mgmnt.htm). Wang, X. B., Y. H. Fang, S. Z. Lin, L. R. Zhang, and H. D.
Nielsen, C., and A. E. Hajek. 2005. Control of invasive soy- Wang. 1994. A study on the damage and economic
bean aphid, Aphis glycines (Hemiptera: Aphididae), pop- threshold of the soybean aphid at the seedling stage. Plant
ulations by existing natural enemies in New York State, Prot. 20: 12Ð13.
with emphasis on entomopathogenic fungi. Environ. En- Wang, K. Y., T. X. Liu, C. H. Yu, X. Y. Jiang, and M. Q. Yi.
tomol. 34: 1036 Ð1047. 2002. Resistance of Aphis gossypii (Homoptera: Aphidi-
Olson, K. and T. Badibanga. 2005a. A bioeconomic model of dae) to fenvalerate and imidacloprid and activities of
the soybean aphid treatment decision in soybeans. Selected detoxiÞcation enzymes on cotton and cucumber. J. Econ.
paper, American Agricultural Economics Association An- Entomol. 95: 407Ð 413.
nual Meeting, 24Ð27 July 2005, Providence, RI. (http:// [WASS] Wisconsin Agricultural Statistics Service. 2004.
agecon.lib.umn.edu/cgi-bin/pdf_view.pl?paperid⫽16358). WisconsinÕs 2004 custom rate guide (http://www.nass.
Olson, K., and T. Badibanga. 2005b. FarmersÕ awareness and usda.gov/Statistics_by_State/Wisconsin/Publications/
case of IPM for soybean aphid control: results from the 2005 custom_rate_2004.pdf).
survey. Staff Paper P05-13. (http://agecon.lib.umn.edu/ Yang, X., L. L. Buschman, K. Y. Zhu, and D. C. Margolies.
cgi-bin/pdf_view.pl?paperid⫽19374&ftype⫽.pdf). 2002. Susceptibility and detoxifying enzyme activity in
Ostlie, K. [ed.]. 2001. Soybean aphid reduces yields: har- two spider mite species (Acari: Tetranychidae) after se-
vest results from insecticide strip trials. University of lection with three insecticides. J. Econ. Entomol. 95: 399 Ð
Minnesota, St. Paul, MN. (http://www.soybeans.umn. 406.
edu/crop/insects/aphid/studyresults.htm). Yoo, H.J.S., R. J. O’Neil, D. J. Voegtlin, and W. R Graves.
Pedigo, L. P., S. H. Hutchins, and L. G. Higley. 1986. Eco- 2005. Host plant suitability of Rhamnaceae for soybean
nomic injury levels in theory and practice. Annu. Rev. aphid (Homoptera: Aphididae). Ann. Entomol. Soc. Am.
Entomol. 31: 341Ð368. 98: 926 Ð930.
Radcliffe, E. B., and D. W. Ragsdale. 2002. Aphid transmit-
ted potato viruses: the importance of understanding vec-
tor biology. Am. J. Pot. Res. 79: 353Ð386. Received 7 November 2006; accepted 16 April 2007.