Effects of shelf architecture and parasitoid release height on biological control of Plodia interpunctella (Lepidoptera: Pyralidae) eggs by Trichogramma deion (Hymenoptera: Trichogrammatidae).

Abstract

The effects of shelving type, packaging, and release height on success of Trichogramma deion Pinto & Oatman (Hymenoptera: Trichogrammatidae) parasitizing Plodia interpunctella (Hübner) (Lepidoptera: Pyralidae) eggs was studied under laboratory conditions. In trials on multipletiered gondola-type or open shelving units, with or without packaging, foraging success was evaluated by comparing parasitism and total mortality rates of sentinel egg disks among shelves after a single point-release of T. deion. Results showed that T. deion parasitized more egg disks and killed more total eggs on open shelves than on gondola shelving. The presence of packaging had no effect on parasitoid foraging on open shelves; however, packaging did interfere with parasitism of P. interpunctella eggs on gondola shelving. Egg parasitism and mortality patterns among shelves were not as evenly distributed on gondola-type shelving compared with open shelving. On gondola shelves without packages, changing the release point of T. deion from the middle to the lowest shelf shifted the distribution of parasitism toward the floor. Gondola shelving, especially in the presence of packaging, reduced foraging efficiency of T. deion for P. interpunctella eggs. Thus, to attain adequate control of P. interpunctella, it may be necessary to use two release heights on gondola shelving.

Full text

Page 1

STORED-PRODUCT
Effects of Shelf Architecture and Parasitoid Release Height on
Biological Control of Plodia interpunctella (Lepidoptera: Pyralidae)
Eggs by Trichogramma deion (Hymenoptera: Trichogrammatidae)
MATTHEW J. GRIESHOP,1PAUL W. FLINN,2,3JAMES R. NECHOLS,4AND JAMES F. CAMPBELL2
J. Econ. Entomol. 99(6): 2202Ð2209 (2006)
Theeffectsofshelvingtype,packaging,andreleaseheightonsuccessofTrichogramma
deion Pinto & Oatman (Hymenoptera: Trichogrammatidae) parasitizing Plodia interpunctella (Hu ¨b-
ner) (Lepidoptera: Pyralidae) eggs was studied under laboratory conditions. In trials on multiple-
tieredgondola-typeoropenshelvingunits,withorwithoutpackaging,foragingsuccesswasevaluated
by comparing parasitism and total mortality rates of sentinel egg disks among shelves after a single
point-releaseofT.deion.ResultsshowedthatT.deionparasitizedmoreeggdisksandkilledmoretotal
eggsonopenshelvesthanongondolashelving.Thepresenceofpackaginghadnoeffectonparasitoid
foraging on open shelves; however, packaging did interfere with parasitism of P. interpunctella eggs
on gondola shelving. Egg parasitism and mortality patterns among shelves were not as evenly
distributed on gondola-type shelving compared with open shelving. On gondola shelves without
packages, changing the release point of T. deion from the middle to the lowest shelf shifted the
distribution of parasitism toward the ßoor. Gondola shelving, especially in the presence of packaging,
reduced foraging efÞciency of T. deion for P. interpunctella eggs. Thus, to attain adequate control of
P. interpunctella, it may be necessary to use two release heights on gondola shelving.
ABSTRACT
KEY WORDS
habitat complexity, stored products, biological control
Augmentative biological control by using trichogram-
matidwaspsoffersapromisingnewapproachforman-
agingstored-productmoths.PreviousstudieswithTri-
chogrammaspecieshave
reductions in moth captures and infestations in bulk
peanut, Arachis hypogaea L., storage (Brower 1988);
bulk wheat, Triticum aestivum L., storage (Scho ¨ller et
al. 1996); and bakeries (Prozell and Scho ¨ller 1998,
Steidle et al. 2001) as well as in warehouses and retail
stores (Prozell et al. 1996).
The Indianmeal moth, Plodia interpunctella (Hu ¨b-
ner) (Lepidoptera: Pyralidae), is one of the most se-
rious pests in retail stores and warehouses, infesting a
variety of products, including raw and processed ce-
reals,driedfruit,pulses,andgarlic,AlliumL.(Coxand
Bell 1991, Perez-Mendoza and Agulera-Pena 2004). A
major obstacle to the development of augmentative
biological control for P. interpunctella in retail stores
and warehouses is the absence of practical guidelines
demonstrated marked
about when, where, and how many Trichogramma to
release so that maximum pest control is achieved at
minimal expense. An important consideration when
developing release guidelines is to deÞne and under-
standhowhabitat complexity
chogramma. For example, a high level of habitat com-
plexityhas beenshown
Trichogramma in a variety of natural and artiÞcial
systems, including tree branches and crowns (Mc-
Cravy and Berisford 1998); maize, Zea mays L., and
waxpapermodels(AndowandProkrym1990,Andow
and Olson 2003); pine needles and paper models
(Lukianchuk and Smith 1997); and artiÞcial plant
modelsofvaryingstructuralcomplexity(Gingrasetal.
2002, Gingras and Boivin 2002). In stored grain,
Scho ¨ller et al. (1994) demonstrated that Tri-
chogramma were unable to penetrate more than sev-
eral centimeters into bulk wheat. A related study by
Grieshop (2005) showed that although good suppres-
sion of P. interpunctella by Trichogramma deion Pinto
& Oatman (Hymenoptera: Trichogrammatidae) oc-
curred on packaged cornmeal, virtually no suppres-
sion was evident on bulk cornmeal.
On a larger scale, variation in habitat complexity
pertaining to shelf architecture (i.e., size, type, and
presence and type of packaging) may be one of the
most relevant factors to consider when making deci-
sions about the location and magnitude of Tri-
chogramma releases. A previous study showed that
inßuences Tri-
to negatively affect
Mention of trade names or commercial products in this article is
solely for the purpose of providing speciÞc information and does not
imply recommendation or endorsement by The U.S. Department of
Agriculture or Kansas State University.
1Washington State University, Tree Fruit Research and Extension
Center, 1100 North Western Ave., Wenatchee, WA 98801.
2Corresponding author, e-mail: paul.ßinn@gmprc.ksu.edu.
3USDAÐARS, Grain Marketing and Research Production Center,
1515 College Ave., Manhattan, KS 66502.
4Department of Entomology, Kansas State University, 123 Waters
Hall, Manhattan, KS 66506.

Page 2

gondola shelving seems to negatively affect host-for-
agingandparasitismbyTrichogrammapretiosumRiley
comparedwithopenshelving(Grieshop2005).These
Þndings indicate that guidelines for releasing Tri-
chogrammaspeciesinstored-productfacilitiesneedto
take into account variability in the structural environ-
ment,whichconsistsofdifferenttypesofshelving,the
presence and type of packaging stocked on those
shelves, and patches of spilled products that may har-
bor concentrations of host eggs. In addition, tests to
determine how to choose the release point for Tri-
chogramma are needed because reduced foraging ef-
Þciency caused by structural characteristics of retail
stores and warehouses may be avoided or hindered
further depending on how parasitoids disperse from
the release point.
Ingeneral,Trichogrammaarereleasedonorinprox-
imity to the products to be protected, typically at 1-m
distances laterally (Scho ¨ller et al. 2006). However,
little is known about how rates of parasitism are af-
fectedbythetypeofshelvingonwhichTrichogramma
are released may affect rates of parasitism. Because
making multiple uniform releases is labor-intensive,
inconvenient, and costly, the optimization of release
rates on different types of shelving is of economic
importance. Thus, the objectives of this study were to
investigate the effects of shelf type, presence, or ab-
sence of packaging, and release point of T. deion on
percentage of parasitism and adjusted mortality of P.
interpunctella eggs.
Materials and Methods
Insects. Insect colonies were maintained in walk-in
growthchambersattheUSDAÐARS,GrainMarketing
and Production Research Center (GMPRC) in Man-
hattan, KS. The Mediterranean ßour moth, Ephestia
kuehniellaZeller(Lepidoptera:Pyralidae),servedasa
rearinghostforT.deion,whereasP.interpunctellawas
used in experiments. Larvae of both moths were
reared on a standard diet of cracked wheat, wheat
shorts, honey, and glycerin described in McGaughey
andBeeman(1988).T.deionwererearedonsterilized
eggs of E. kuehniella. Ephestia eggs were sterilized by
exposing them to UV light (6-W Spectronics BLE
6254s) for 1 min (eggs were placed ?2.5 cm from the
bulb). Colonies of T. deion and E. kuehniella were
originally obtained from BeneÞcial Insectaries (Red-
ding, CA) in April 2003 and February 2002, respec-
tively (Kansas State University Entomology voucher
no. 171). The P. interpunctella used in experiments
were obtained from a colony maintained in continu-
ousproductionformorethanadecadeattheGMPRC
(Manhattan, KS). Rearing conditions for T. deion and
P. interpunctella were ?26?C and 60% RH.
Experimental Design. Experiments were con-
ductedintwowalk-ingrowthchamberslocatedatthe
GMPRC. Experiments were conducted at 23 ? 1?C
and 45 ? 5% RH, with a photoperiod of 16:8 (L:D) h.
Temperature and relative humidity settings for the
chambers were based on means calculated from 6 mo
of data collected between April and November 2002
from four retail stores in Manhattan, KS. Open-shelf
trials were run between 1 July and 5 November 2003,
whereas gondola-shelf trials were run between 13
February and 7 December 2004. Sentinel egg disks
were used to map parasitism. Disks consisted of 1-cm-
diameter cardstock punch outs with four viable P.
interpunctella eggs (?18-h-old) attached using an in-
ert glue, Traganth gum (Merck, Whitehouse Station,
NJ). T. deion were released as late-stage pupae inside
eggs of E. kuehniella glued onto cardstock. The num-
ber and timing of each release was adjusted so that
?500 female wasps emerged within 1 h of the cards
being placed on the shelves.
Separate comparisons were made to test the inßu-
ence of shelving type, the presence or absence of
packages, and the shelf height at which parasitoids
werereleased,onthehost-foragingsuccessofT.deion
The Þrst trial consisted of four treatments made up of
twofactors:shelvingtypeandeitheremptyshelvesor
shelveswithpackaging.Eachtreatmentwasreplicated
Þve times for a total of 20 experimental units. In treat-
ments that included packaging, an assortment of
empty breakfast cereal boxes was used. These boxes
were placed on the shelves between sentinel disks in
four columns, with six boxes per column and boxes
spaced equidistant to Þll the width of the shelf. Total
surfaceareaofthepackagesoneachshelfwas3.97m2.
The second trial consisted of two treatments with the
release point either on the third (middle) or Þrst
(lowest) shelf of gondola shelves without packages.
The open shelving units (Fig. 1) consisted of 1.27-
cm-thick particleboard shelves and painted sheet
metal risers. A 3 by 5 grid of sentinel egg disks was
placedoneachshelfandontheßoordirectlybeneath
theshelvingunit,withtheexceptionofthethirdshelf
from the bottom, which had the center sentinel egg
disk replaced with the T. deion release point (Fig. 1).
The resulting three-dimensional grid consisted of six
tiers of 15 egg disks/tier with the third (central) shelf
consisting of 14 egg disks. The inter-disk spacing for
sentinel egg grids measured 33 by 28.6 cm. Figure 1
shows the layout of the open shelves, grids of sentinel
egg disks, and the T. deion release point.
The gondola units (Fig. 2) consisted of Þve painted
sheet metal shelves on either side of two pegboard
(Þberboard) dividers that extended from the lowest
shelf to the highest. A 2 by 4 grid of sentinel egg disks
was placed on each shelf on either side of the divider
and on the ßoor directly beneath the shelving unit.
The resulting three-dimensional grid consisted of six
tiers of 16 egg disks per tier (Fig. 2). The interdisk
spacingforsentinelegggridsongondolashelvesmea-
sured 38.1 by 22.9 cm per shelf with 0.1 m (the peg-
board divider) between the two sides of the shelf. T.
deionwerereleasedatthecenterofshelfthreeorshelf
one on two 375-egg cards, one on either side of the
divider. Packaging was laid out in four columns of
three boxes on either side of the central divider.
An additional 30Ð40 sentinel egg disks (120Ð160
eggs) were kept for each replicate. These disks were
kept in sealed petri dishes within the test chambers
December 2006GRIESHOP ET AL.: HOST-FORAGING SUCCESS OF T. deion ON SHELVING
2203

Page 3

and were used to calculate incidental P. interpunctella
eggmortality.Replicationswith?20%incidentalmor-
tality were excluded from the experiment. Replicates
were run at least 5 d apart to allow adequate time for
T. deion from the previous experiment to die.
Data Collection. Sentinel egg disks were collected
after 48 h and held for 7 d in a growth chamber set at
26?1?Cand60?7%RHtoallowanyliveparasitoids
to develop and emerge. Subsequently, the eggs on
each sentinel egg disk were individually inspected
usingastereomicroscope(originalmagniÞcation160Ð
240?), to count the number of eggs hatched, parasit-
ized, or dead. We calculated percentage of parasitism
per shelf, percentage of egg disks attacked, and cor-
rected total mortality per shelf. Corrected total mor-
tality(parasitizedeggs?deadeggs)wascalculatedas
(experimental percentage of mortality ? control per-
centage of mortality)/(1 ? control mortality) (Ab-
bott 1925).
DataAnalysis.Datawereanalyzedusingthemixed,
GLM,andfrequencyprocedures(SASInstitute2000).
All percentages were transformed by using arcsine
squareroot[arcsine(x0.5)]beforeanalysestostabilize
variance. Complete factorial mixed model analysis of
variance (ANOVA) were run for percentage of eggs
parasitized,percentageofdisksattacked,andpercent-
age of corrected total mortality for the shelving type
andpackagingtrial,andforthelowerreleasepointon
gondolashelvingtrial.Fortheshelvingtypeandpack-
agingexperiment,toaccountforthetimefactor,shelv-
ing type and packaging were nested within experi-
mental run as the random factor, and shelving type,
packaging, and shelf were analyzed as Þxed factors.
Forthereleasepointongondolashelvingexperiment,
toaccountforthetimefactor,releasepointwasnested
within experimental run as the random factor, and
release point and shelf level were analyzed as Þxed
factors. In addition, a series of mixed model t-test
comparisons were made within packaging treatment
and shelf level by using percentage of parasitism, per-
centageofdisksattacked,andpercentageofcorrected
totalmortalityasresponsevariables.Weusedshelving
typenestedwithinrunastherandomfactorandshelv-
ing type as the Þxed factor. Degrees of freedom were
adjusted using the KenwardÐRodgers correction for
mixed models to account for nesting in the designs.
Finally, Þve one-way ANOVA models by using the
GLM procedure were used to compare percentage of
parasitism, percentage of disks attacked, and percent-
age of corrected total mortality among shelf levels
within individual treatments with TukeyÐKramer ad-
justed least signiÞcant difference (LSD) multiple
comparisons used to separate means.
Further analysis consisted of chi-square goodness-
of-Þt-tests performed using the frequency procedure;
thedistributionofparasitismacrossthesixshelveswas
Fig.1.
release point. Sentinel egg disks were spaced in a 33- ? 28.6-cm grid.
Diagramofopenshelvingunitshowing3?5sentineleggdisk(eggpatches)layoutandthecentralTrichogramma
Fig. 2.
centralorlowestshelfTrichogrammareleasepoint.Sentineleggdisks(eggpatches)werespacedintwo2?438.1-?22.9-cm
grids.
Diagram of gondola shelving unit showing four x four sentinel egg disk layout across a pair of shelves and the split
2204JOURNAL OF ECONOMIC ENTOMOLOGY
Vol. 99, no. 6

Page 4

tested to determine whether it was signiÞcantly dif-
ferent from a uniform distribution.
Results
Shelving and Packaging. SigniÞcant treatment ef-
fects for percentage of parasitism, percentage of egg
disks parasitized, and percentage of corrected mor-
tality were detected for shelf level and shelf type, and
for the shelf level ? shelf type and shelf level ?
packaging interactions in the three-way mixed model
ANOVA (Table 1). For shelves without packages, the
overall mean ? SEM percentage of parasitism was
56.16 ? 2.97% on open shelves versus 40.94 ? 4.4% on
gondola shelves; overall mean ? SEM percentage of
egg disks attacked was 68.75 ? 4.21% on open shelves
versus 42.37 ? 5.23% on gondola shelves; and overall
mean percentage of corrected total mortality was
79.1 ? 2.85% on open shelves versus 69.38 ? 2.94% on
gondolashelves.Onshelveswithpackages,theoverall
mean percentage of parasitism was 60.06 ? 5.96% on
open shelves versus 37.03 ? 3.6% on gondola shelves;
overallmeanpercentageofdisksattackedwas71.45?
7.38%onopenshelvesversus40.05?3.48%ongondola
Fig. 3.
mortality of P. interpunctella eggs for open and gondola shelves with empty shelves presented in the left column and shelves
with packages presented in the right column. Asterisk (*) indicates signiÞcant difference between treatments at that shelf
level (t-test: df ? 8, P ? 0.05).
Mean ? SEM percentage of parasitism, percentage of egg disks parasitized, and percentage of corrected total
Table 1.
adjusted mortality between packaging treatments and shelving type
Mixed model ANOVA results for percentage of parasitism, percentage of egg disks attacked, and percentage of Abbott’s
Effectdf
% parasitism
% egg disks
parasitized
% total corrected
mortality
FPFPFP
Packages
Shelf level
Packages ? shelf level
Shelving type
Shelving type ? packages
Shelving type ? shelf level
Shelving type ? packages ? shelf level
1, 16
5, 80
5, 80
1, 16
1, 16
5, 80
5, 80
0.01
12.98
4.35
19.24
0.61
4.61
0.29
0.99
0.01
0.01
0.01
0.44
0.01
0.92
0.11
14.56
3.80
11.59
0.66
4.84
0.64
0.74
0.01
0.01
0.01
0.43
0.01
0.67
0.02
21.86
2.68
31.06
0.17
4.20
0.47
0.88
0.01
0.03
0.01
0.69
0.01
0.80
December 2006GRIESHOP ET AL.: HOST-FORAGING SUCCESS OF T. deion ON SHELVING
2205

Page 5

shelves; and overall mean percentage of corrected
total egg mortality was 82.02 ? 5.41% on open shelves
versus 62.17 ? 3.44% on gondola shelves. In almost all
cases, the highest rates of egg parasitism, percentage
of disks attacked, and egg mortality were observed on
the release shelf (shelf 3), with the exception of per-
centage of parasitism on open shelves without pack-
ages, where shelf 2 had the highest rate.
SigniÞcant differences in percentage of parasitism,
percentage of disks attacked, and percentage of cor-
rected total mortality were detected within shelves
with packages, among shelves, and between shelving
types(Fig.3).Forshelveswithoutpackages,percent-
age of parasitism, and percentage of disks attacked
weresigniÞcantlygreaterforopenshelvesontheßoor
and on shelf 2, and percentage of corrected total egg
mortalitywassigniÞcantlygreaterforopenshelveson
the ßoor and on shelves 1Ð3 (Fig. 3).
Forshelveswithpackages,percentageofparasitism
and percentage disks attacked were signiÞcantly
greaterforopenshelvesontheßoor,shelf1,andshelf
2; and percentage of corrected total mortality was
signiÞcantlygreaterforopenshelvesontheßoor,and
on shelves 1Ð4 (Fig. 3).
One-way ANOVA for shelf levels by shelving type
and presence or absence of packages showed a sig-
niÞcantdifferenceindistributionofparasitismamong
the treatments (Tables 2 and 4). On open shelving
units, no signiÞcant differences were found among
shelves for percentage of parasitism or percentage of
disksattacked,eitherwithorwithoutpackages(Table
2). However, percentage of corrected egg mortality
wassigniÞcantlygreateronshelf3comparedwiththe
ßoor and shelf 1 for open shelves without packages
(F ? 3.72; df ? 5, 80; P ? 0.01) (Table 2). On open
shelves with packages, egg mortality was signiÞcantly
greater on shelf 3 than on shelf 1 (F ? 2.64; df ? 5, 80;
P ? 0.05) (Table 2).
In contrast, on gondola shelving without packages,
percentageofparasitism,percentageofdisksattacked,
and percentage of total corrected egg mortality all
differed signiÞcantly among shelves (F ? 7.73, df ? 5,
80, P ? 0.01; F ? 14.46, df ? 5, 80, P ? 0.01; and F ?
8.45, df ? 5, 80, P ? 0.01, respectively). Rates of
parasitism on the ßoor were signiÞcantly lower than
those on the Þve shelves. In addition, the percentage
ofdisksattackedonshelf1wassigniÞcantlylowerthan
on shelf 3 (Table 3). For gondola shelves with pack-
ages, percentage of parasitism, percentage of disks
attacked, and percentage of total corrected egg mor-
tality were also signiÞcantly different among shelves
(F ? 6.68, df ? 5, 80, P ? 0.01; F ? 9.28, df ? 5, 80, P ?
0.01; and F ? 4.95, df ? 5, 80, P ? 0.01, respectively),
with rates on the ßoor lower than those on the upper
three shelves (Table 3).
The distribution of parasitism across shelf units for
open shelves without packages was signiÞcantly dif-
ferent from a uniform distribution (?2? 30.995, df ?
5, P ? 0.01). This was also true for gondola shelves
without packages (?2? 88.12, df ? 5, P ? 0.01), and
gondolashelveswithpackages(?2?67.02,df?5,P?
0.01). However, the distribution of parasitism on open
shelves with packages was not signiÞcantly different
fromauniformdistribution(?2?7.62,df?5,P?0.18).
Release Height. Two-way ANOVA comparing re-
lease points on gondola shelves showed that shelf
level, and the interaction between shelf level and
height of parasitoid release, signiÞcantly affected per-
centage of parasitism, percentage of disks attacked,
and percentage of corrected egg mortality (Table 4).
Whenparasitoidswerereleasedonshelf1,percentage
ofparasitismwassigniÞcantlygreaterontheßoor,and
signiÞcantly lower on shelves 3 and 4, than when
Table 2.
by T. deion on open-type shelving
Effect of packages on parasitism of P. interpunctella
Treatment% parasitism
% egg disks
parasitized
% total
corrected
mortality
Without packages
Floor
Shelf 1
Shelf 2
Shelf 3a
Shelf 4
Shelf 5
With packages
Floor
Shelf 1
Shelf 2
Shelf 3a
Shelf 4
Shelf 5
39.00 ? 10.08a 58.67 ? 13.47a 52.04 ? 16.00b
50.67 ? 3.89a73.33 ? 5.33a
67.67 ? 3.10a89.33 ? 3.40a
64.29 ? 5.84a95.71 ? 1.75a
64.00 ? 6.07a 82.67 ? 4.00a
54.00 ? 9.77a76.00 ? 10.19a 61.83 ? 10.33ab
57.46 ? 6.04b
78.16 ? 2.70ab
93.22 ? 2.43a
71.45 ? 6.85ab
61.67 ? 5.55a
54.33 ? 6.25a
63.33 ? 7.13a
68.93 ? 3.94a
58.00 ? 12.07a 80.00 ? 13.50a 72.19 ? 13.67ab
54.67 ? 6.46a77.33 ? 7.77a
86.67 ? 5.58a
72.00 ? 6.80a
86.67 ? 3.65a
90.00 ? 5.35a
69.71 ? 4.96ab
60.34 ? 8.85b
72.13 ? 7.93ab
92.81 ? 4.48a
62.97 ? 8.62ab
Values (mean ? SEM) within the same column and treatment
followed by a different letter are signiÞcantly different (LSD with
TukeyÐKramer adjustment [P ? 0.05]).
aParasitoid release shelf.
Table 3.
P. interpunctella by T. deion on gondola-type shelving
Effect of packages and release point on parasitism of
% parasitism
% egg disks
attacked
% total
corrected
mortality
Treatment 1: without packages, release on shelf 3
12.81 ? 3.26b21.25 ? 5.80c
Shelf 138.44 ? 5.20a
Shelf 240.31 ? 6.62a
Shelf 351.88 ? 5.33a
Shelf 4 49.69 ? 3.22a
Shelf 5 52.50 ? 8.92a
Floor8.60 ? 2.56b
35.40 ? 6.09a
40.17 ? 8.31a
64.58 ? 9.13a
52.66 ? 6.30a
53.87 ? 9.03a
67.50 ? 6.37b
70.00 ? 3.64ab
92.50 ? 2.34a
82.50 ? 5.38ab
82.50 ? 5.38ab
Treatment 2: with packages, release on shelf 3
25.94 ? 6.20c46.25 ? 8.97c
24.69 ? 2.00c46.25 ? 2.50c
30.00 ? 4.03bc55.00 ? 6.06c
54.06 ? 6.86a85.00 ? 5.08ab
37.81 ? 4.59ab63.75 ? 6.06bc
49.69 ? 3.18ab 88.75 ? 4.15a
Floor
Shelf 1
Shelf 2
Shelf 3
Shelf 4
Shelf 5
28.70 ? 7.56b
29.18 ? 3.73b
34.65 ? 5.13ab
57.93 ? 6.53a
38.50 ? 4.36ab
51.34 ? 2.77ab
Treatment 3: without packages, release on shelf 1
48.75 ? 5.98a86.25 ? 4.59ab
Shelf 155.63 ? 6.26a
Shelf 238.13 ? 6.68ab
Shelf 323.13 ? 2.06b
Shelf 422.50 ? 6.14b
Shelf 5 33.13 ? 4.57ab
Floor70.23 ? 6.38ab
89.93 ? 3.83a
51.21 ? 8.42bc
31.52 ? 3.11c
28.27 ? 8.75c
47.90 ? 3.39bc
88.75 ? 4.59a
55.00 ? 8.71c
45.00 ? 3.64c
38.75 ? 9.14c
58.75 ? 5.80bc
Values (mean ? SEM) within the same column and treatment
followed by a different letter are signiÞcantly different (LSD with
TukeyÐKramer adjustment [P ? 0.05]).
2206JOURNAL OF ECONOMIC ENTOMOLOGY
Vol. 99, no. 6

Page 6

parasitoids were released on shelf 3 (Fig. 4). In addi-
tion, a signiÞcantly greater percentage of disks was
attackedontheßoorandonshelf1,andasigniÞcantly
lowerpercentageofdiskswasattackedonshelves3Ð5
whenTrichogrammawerereleasedonshelf1(Fig.4).
Whenparasitoidswerereleasedonshelf1,percentage
of corrected total egg mortality was signiÞcantly
greaterontheßoorandonshelf1,andsigniÞcantlylower
onshelf3,comparedwithmortalityonthesameshelves
when parasitoids were released on shelf 3 (Fig. 4).
SigniÞcant differences were found among shelves
for both release heights, indicating different patterns
of parasitism between the two treatments. For shelf 1
releases, percentage of parasitism was greater on the
ßoor and on shelf 1 than on shelves 3 and 4 (F ? 5.93;
df ? 5, 80; P ? 0.01). Percentage of egg disks attacked
wasgreaterontheßoorandonshelf1thanonshelves
2Ð4 (F ? 10.88; df ? 5, 80; P ? 0.01). The percentage
ofcorrectedtotalmortalitywasgreateronshelf1than
on shelves 2Ð4; and greater on the ßoor than on
shelves3and4(F?15.21;df?5,80;P?0.01)(Table
3). In contrast, for shelf 3 releases, percentage of
parasitismandpercentageofcorrectedtotalmortality
were lower on the ßoor than on the Þve shelves (F ?
7.73, df ? 5, 80, P ? 0.01; and F ? 8.45, df ? 5, 80, P ?
0.01, respectively), and percentage of egg disks at-
tackedwaslowerontheßoorthanontheshelves,and
lower on shelf 1 than on shelf 3 (F ? 14.46; df ? 5, 80;
P ? 0.01) (Table 3).
For both parasitoid release points, the pattern of
parasitismamonggondolashelvesdidnotÞtauniform
distribution (?2? 78.43, df ? 5, P ? 0.01; and ?2?
88.12, df ? 5, P ? 0.01, respectively).
Discussion
Shelving and Packaging. The reduced levels of par-
asitism and T. deion-mediated mortality of P. inter-
punctellaeggsongondolashelvescomparedwithopen
shelves could be due to a variety of structural differ-
ences between the two shelving types. Perhaps the
most important difference is the pegboard divider
present in gondola shelves. The pegboard divider ex-
tends from the top of the Þrst shelf to the bottom of
the Þfth shelf and in addition to adding to the overall
surface area of the shelving unit, might reduce spatial
continuity between the two sides of a shelving tier.
ThatparasitismonthetoptierdidnotdiffersigniÞcantly
betweenthetwoshelvingtypessuggeststhatthecentral
partition may interfere with foraging (Fig. 3).
Additional differences between gondola and open
shelves are apparent in shelf continuity and construc-
tion. Gondola shelves are less continuous than open
shelves as only the two shelf ends are in contact with
the central partition, leaving a 0.5-cm gap running the
length of the shelf. Although T. deion is capable of
ßight, this gap may limit shelf-to-shelf movement.
AÞnalmajorstructuraldifferencebetweenthetwo
shelvingstylesisthepresenceofkickplatesongondola
shelves (Fig. 2), which extend from the bottom of the
lowest shelves to the ßoor. Percentage of eggs para-
sitized,percentageofeggdisksattacked,andpercent-
age of egg mortality on the lowest shelf did not differ
signiÞcantly between the two shelving types, which
suggests that few T. deion foraged underneath the
Table4.
release points.
MixedmodelANOVAresultsforpercentagesofparasitism,eggdisksattacked,andtotalmortalityongondolashelvesbetween
Effect df
% parasitism
% egg disks
parasitized
% total corrected
mortality
FPFPFP
Release point
Shelf level
Release point ? shelf level
1, 8
5, 40
5, 40
0.48
3.73
17.86
0.51
0.01
0.01
2.57
3.95
24.35
0.15
0.01
0.01
4.46
5.43
23.37
0.07
0.01
0.01
Fig. 4.
ageofeggdisksparasitized,andpercentageofcorrectedtotal
mortality of P. interpunctella eggs for either a Þrst or third
shelf release of T. deion on gondola shelves. Asterisk (*)
indicates signiÞcant difference between treatments at that
shelf level (t-test: df ? 8, P ? 0.05).
Mean ? SEM percentage of parasitism, percent-
December 2006GRIESHOP ET AL.: HOST-FORAGING SUCCESS OF T. deion ON SHELVING
2207

Page 7

kickplates. When the release point was shifted from
shelf 3 to the lowest gondola shelf, signiÞcantly more
eggs were parasitized or had increased mortality.
However, spillage typically accumulates on ßoors un-
derneath shelving (Roesli et al. 2003, Nansen et al.
2004), and other studies have demonstrated the neg-
ative inßuence of increased microhabitat complexity
onTrichogrammaforagingability(Scho ¨lleretal.1996,
Grieshop 2005). Thus, host eggs located in spillage
beneath shelves are likely to escape parasitism due to
increased microhabitat complexity.
A central release point would probably provide the
best overall coverage of products located on both
open and gondola shelving. However, to achieve ad-
equate control of P. interpunctella, managers should
make sanitation under shelves a priority. Where san-
itation is not feasible, releases of the late larval para-
sitoidHabrobraconhebetor(Say)(Hymenoptera:Bra-
conidae) might be used in conjunction with T. deion
to manage P. interpunctella populations located in
spillage under shelving (Grieshop 2005).
In contrast to shelving type, the presence or ab-
sence of packages seemed to have a limited effect on
the percentage of egg parasitism, percentage of egg
disks attacked, and egg mortality. An experiment on
the effects of millet and ßour on the host-foraging
success of three species of Trichogramma species
showed that increases in small-scale habitat complex-
ityhadanegativeeffectonTrichogrammaeggforaging
(Grieshop 2005). Differences in the distribution of
percentage of parasitism, percentage of egg disks par-
asitized, and egg mortality between gondola shelves
with and without packages (Table 3) suggests that
gondola shelves may pose a signiÞcant problem for
successful T. deion host foraging.
ReleaseHeight.Parasitismandmortalityofsentinel
eggdisksongondolashelvesweregreatestontheshelf
where T. deion was released. However, there were
differences in the vertical distribution of these rates
depending on the height of the release shelf. For
example, when parasitoids were released on shelf 3,
there was a more even distribution of parasitism and
mortalityonshelvesthanwhenreleasesweremadeon
shelf 1 (lowest). However, when released on shelf 3,
T. deion was not very effective in killing host eggs on
theßoor.Incontrast,parasitismandhosteggmortality
werehighontheßoorwhenparasitoidswerereleased
onshelf1(Fig.4).Furthermore,intheshelf1releases,
whereas the percentage of egg disks attacked and
percentageofhostmortalityseemsimilarforthelower
shelves and the ßoor, a greater percentage of parasit-
ized egg disks was found on the upper level shelves
than eggs parasitized or killed (Fig. 4). This may be
due to the early exploitation of eggs closest to the
releasepointfollowedbyaloweredlevelofparasitism
and/or successful larval development on older eggs
locatedfartherfromthereleasepoint.Variousstudies
have shown that host eggs become less attractive
and/or suitable for parasitism by Trichogramma spe-
cies over time as the caterpillar embryo develops
(Marston and Ertle 1969, Monje et al. 1999). Because
asingleinsectwithinaninfestedproductresultsinthe
loss of that product, a release pattern that results in a
more complete exploitation of host patches is of crit-
ical importance to successful pest management. Thus,
a central release point will likely provide better con-
trol than releases made on the Þrst shelf.
Interestingly,withingondolashelveswhereT.deion
was released on shelf 1, although percentage of par-
asitism, percentage of egg disks attacked, and cor-
rected mortality decreased progressively on most of
the shelves above the Þrst level, a slight increase was
observedontheuppermost(Þfth)shelf(Fig.4).Allen
and Gonzalez (1974) described a horizontal bimodal
dispersion of parasitism for T. pretiosum in cotton,
GossypiumhirsutumL.,Þelds,whichtheyexplainedas
eitherwinddriftorotherenvironmentalfactors,orthe
result of “parasitoid saturation,” where parasitoids
havehadadequatetimetoforageovertheentirestudy
area. We are unsure why T. deion parasitized eggs at
aslightlyhigherrateonshelf5comparedwithshelf4.
Thiseffectonlyoccurredwhenreleasesweremadeon
the Þrst shelf. The fact remains that releases made on
shelf 3 were superior for overall coverage of the
shelves compared with releases made on shelf 1.
Other studies that have examined vertical disper-
sion of Trichogramma species have shown a bias for
either the upper or lower strata of plants in Þeld and
greenhouse conditions. Thorpe and Dively (1985)
found that T. pretiosum foraged primarily at the bot-
tom of laboratory arenas. Smith (1988) showed a dif-
ferential bias in vertical dispersion of Choristoneura
fumiferana (Clemens) (Lepidoptera: Tortricidae)
eggs parasitized by Trichogramma minutum Riley,
with greater rates of parasitism at the tops of trees
whentheparasitoidwasreleasedat0.25m,andgreater
rates of parasitism on eggs located on lower branches
when parasitoids were released at ground level. In
contrast, Wang et al. (1997) found that T. ostriniae
Pang and Chen was more likely to parasitize sentinel
egg clusters on the lower two-thirds of sweet corn
plants in the Þeld when released at a height of 1 m.
The most important Þnding from this study for
stored product managers is that gondola shelves pre-
sented a less favorable foraging environment for T.
deionthanopenshelves.Onepotentialsolutiontothis
problemwouldbetousereleasepointsattwoheights,
ortoreleasegreaternumbersofparasitoidsongondola
shelving.Wealsofoundthatthepresenceofpackages
on shelves does not seem to affect the ability of T.
deiontolocatehostsonopenshelves,butitmayhinder
host foraging on gondola shelves. Finally, the vertical
location of the release point greatly inßuenced the
distribution of parasitism. We found that releasing T.
deion on shelf three provided the most even coverage
of shelves 1Ð5.
Acknowledgments
We thank Beverly Wilson (Kansas State University) and
Kenlee Friesen (USDAÐARS) for help with colony mainte-
nance and experimental setup; Matthias Scho ¨ller (Biolo-
gischeBeratungBerlin,Berlin,Germany)forhelpwithinitial
experimental design; James Higgins (Department of Statis-
2208JOURNAL OF ECONOMIC ENTOMOLOGY
Vol. 99, no. 6

Page 8

tics,KansasStateUniversity)forassistancewithdataanalysis,
andSinthyaPenn(BeneÞcialInsectaries),andMichaelHoff-
manandJefferyGardner(CornellUniversity,Ithaca,NY)for
providingT.deion.WealsothankRichardArbogast(USDAÐ
ARS, CMAVE, Gainesville, FL) and Gerald Wilde (Depart-
mentofEntomology,KansasStateUniversity)fortheirvalu-
able comments on an earlier version of the manuscript. This
research was funded by: USDAÐCSREES (RAMP) under
agreement00-51101-9674.Thismanuscripthasbeenassigned
contribution no. 06Ð255J from the Kansas Agricultural Ex-
periment Station.
References Cited
Abbott, W. S. 1925. A method of computing the effective-
ness of an insecticide. J. Econ. Entomol. 18: 265Ð267.
Allen,J.C.,andD.Gonzalez. 1974. Spatialattackpatternsof
Trichogramma pretiosum around release sites compared
with a random diffusion model. Environ. Entomol. 3:
647Ð652.
Andow, D. A., and D. M. Olson. 2003. Inheritance of host
Þnding ability on structurally complex surfaces. Oecolo-
gia (Berl.) 136: 324Ð328.
Andow, D. A., and D. R. Prokrym. 1990. Plant structural
complexity and host-Þnding by a parasitoid. Oecologia
(Berl.) 82: 162Ð165.
Brower, J. H. 1988. Population suppression of the almond
moth and the Indianmeal moth (Lepidoptera: Pyralidae)
by release of Trichogramma pretiosum (Hymenoptera:
Trichogrammatidae) into simulated peanut storages. J.
Econ. Entomol. 81: 944Ð948.
Cox, P. D., and C. H. Bell. 1991. Biology and ecology of
moth pests of stored foods, pp. 181Ð193. In J. R. Gorham
[ed.], Ecology and management of food-industry pests.
FDA Tech. Bull. 4, Arlington, VA.
Gingras, D., and G. Boivin. 2002. Effect of plant structure,
host density and foraging duration on host Þnding by
Trichogramma evanescens (Hymenoptera: Trichogram-
matidae). Environ. Entomol. 31: 1153Ð1157.
Gingras,D.,P.Dutilleul,andG.Boivin. 2002. Modelingthe
impact of plant structure on host-Þnding behavior of
parasitoids. Oecologia (Berl.) 130: 396Ð402.
Grieshop, M. J. 2005. Evaluation of three species of Tri-
chogramma egg parasitoids for the biological control of
theIndianmealmothinretailstoresandwarehouses.Ph.D.
dissertation, Kansas State University, Manhattan, KS.
Lukianchuk, J. L., and S. M. Smith. 1997. Inßuence of plant
structural complexity on the foraging success of Tri-
chogramma minutum: a comparison of search on artiÞcial
and foliage models. Entomol. Exp. Appl. 84: 221Ð228.
Marston, N., and L. E. Ertle. 1969. Host age and parasitism
by Trichogramma minutum (Hymenoptera: Trichogram-
matidae). Ann. Entomol. Soc. Am. 62: 1476Ð1482.
McCravy, K. W., and C. W. Berisford. 1998. Parasitism by
Trichogramma spp. (Hymenoptera: Trichogrammatidae)
in relation to Nantucket pine tip moth (Lepidoptera:
Tortricidae) egg density and location. Environ. Entomol.
27: 355Ð359.
McGaughey, W. H., and R. W. Beeman. 1988. Resistance to
Bacillus thuringiensis in colonies of Indianmeal moth and
almondmoth(Lepidoptera:Pyralidae).J.Econ.Entomol.
81: 28Ð33.
Monje, J. C., C.P.W. Zebitz, and B. Ohnesorge. 1999. Host
and host age preference of Trichogramma galloi and T.
pretiosum (Hymenoptera: Trichogrammatidae) reared
on different hosts. J. Econ. Entomol. 92: 97Ð103.
Nansen, C., B. Subramanyam, and R. Roesli. 2004. Charac-
terizing spatial distribution of trap captures of beetles in
retail pet stores using SADIE software. J. Stored Prod.
Res. 40: 471Ð483.
Perez-Mendoza, J., and M. Agulera-Pena. 2004. Develop-
ment,reproduction,andcontroloftheIndianmealmoth,
Plodiainterpunctella(Hu ¨bner)(Lepidoptera:Pyralidae),
in stored seed garlic in Mexico. J. Stored Prod. Res. 40:
409Ð421.
Prozell, S., and M. Scho ¨ller. 1998. Insect fauna of a bakery,
processingorganicgrainandapplyingTrichogrammaeva-
nescens Westwood. Integr. Prot. Stored Prod. (IOBC
Bull.) 21: 39Ð44.
Prozell, S., M. Scho ¨ller, S. A. Hassan, and Ch. Reichmuth.
1996. Release of Trichogramma evanescens as a compo-
nent of an integrated pest management programme in
organic food bakeries and stores (Hymenoptera: Tri-
chogrammatidae). In B. Baccetti [ed.], Proceedings of
theXXInternationalCongressofEntomology.TipograÞa
TAF, Florence, Italy.
Roesli, R., B. Subramanyam, J. F. Campbell, and K. Kemp.
2003. Stored-product insects associated with a retail pet
store chain in Kansas. J. Econ. Entomol. 96: 1958Ð1966.
SAS Institute. 2000. SAS release 8.01. SAS Institute, Cary, NC.
Scho ¨ller,M.,Ch.Reichmuth,andS.A.Hassan. 1994. Studies
on biological control of Ephestia kuehniella Zeller (Lep:
Pyralidae)withTrichogrammaevanescensWestwood(Hym:
Trichogrammatidae)Ðhost Þnding ability in wheat under
laboratory conditions, pp. 1142Ð1146. In E. Highley, E. J.
Wright,H.J.Banks,andB.R.Champ[eds.],StoredProduct
Protection: Proceedings of the 6th International Working
Conference on Stored-Product Protection, February 1994,
Canberra,Australia.CABInternational,Wallingford,United
Kingdom.
Scho ¨ller, M., S. A. Hassan, and Ch. Reichmuth. 1996. EfÞ-
cacy assessment of Trichogramma evanescens and T. em-
bryophagum (Hym: Trichogrammatidae) for control of
stored products moth pests in bulk wheat. Entomophaga
41: 125Ð132.
Scho ¨ller, M., P. W. Flinn, M. J. Grieshop, and E. Zˇd’a ´rkova ´.
2006. Biological control of stored product pests, pp. 67Ð
87. In J. Heaps [ed.], Insect management for food storage
and processing, 2nd ed. AACC International, St. Paul, MN.
Smith,S.M. 1988. Patternofattackonsprucebudwormegg
masses by Trichogramma minutum (Hymenoptera: Tri-
chogrammatidae) released in forest stands. Environ. En-
tomol. 17: 1009Ð1015.
Steidle, J.L.M., D. Rees, and E. J. Wright. 2001. Assessment
of Australian Trichogramma species (Hymenoptera: Tri-
chogrammatidae) as control agents of stored product
moths. J. Stored Prod. Res. 37: 263Ð275.
Thorpe, K. W., and G. P. Dively. 1985. Effects of arena size
on laboratory evaluations of the egg parasitoids Tri-
chogramma minutum, T. pretiosum, and T. exiguum (Hy-
menoptera: Trichogrammatidae). Environ. Entomol. 14:
762Ð767.
Wang,B.,D.N.Ferro,andD.W.Hosmer. 1997. Importance
of plant size, distribution of egg masses, and weather
conditions on egg parasitism of the European cornborer,
Ostrinia nubilalis by Trichogramma ostriniae in sweet
corn. Entomol. Exp. Appl. 83: 337Ð345.
Received 2 June 2006; accepted 29 August 2006.
December 2006GRIESHOP ET AL.: HOST-FORAGING SUCCESS OF T. deion ON SHELVING
2209