Title: Florida Entomologist
Full Citation
Permanent Link: http://ufdc.ufl.edu/UF00098813/00012
 Material Information
Title: Florida Entomologist
Physical Description: Serial
Creator: Florida Entomological Society
Publisher: Florida Entomological Society
Place of Publication: Winter Haven, Fla.
Publication Date: 2006
Copyright Date: 1917
Subject: Florida Entomological Society
Entomology -- Periodicals
Insects -- Florida
Insects -- Florida -- Periodicals
Insects -- Periodicals
General Note: Eigenfactor: Florida Entomologist: http://www.bioone.org/doi/full/10.1653/024.092.0401
 Record Information
Bibliographic ID: UF00098813
Volume ID: VID00012
Source Institution: University of Florida
Holding Location: University of Florida
Rights Management: Open Access

Full Text

Cummings et al.: Fipronil Impact on Mole Cricket Behavior and Mortality 293


'Department of Agribusiness, Agronomy, Horticulture, and Range Management, Tarleton State University,
Stephenville, TX 76402

2Department of Entomology, North Carolina State University, Raleigh, NC 27695-7613

3Department of Molecular and Environmental Toxicology, North Carolina State University,
Raleigh, NC 27695-7633

4Crop Science Department, North Carolina State University, Raleigh, NC 27695-7620


In a greenhouse experiment, fipronil was applied at 0.014 kg ai/ha to bermudagrass, Cyn-
odon dactylon L., in plastic 5-liter containers 120, 90, 60, 30, and 0 days before adding one
tawny mole cricket nymph, Scapteriscus vicinus Scudder to the container. After the exposure
period, soil in the containers was divided into depth increments of 0-4, 4-8, and 8-18 cm, and
cricket status was recorded as dead, absent, or alive by thoroughly examining soil. Soil in
the 0-4 cm-increment was analyzed for fipronil and four fipronil metabolite residues.
Fipronil residue concentrations decreased with time (C = 0.00002x 0.0053x + 0.3675, R2 =
0.9998 where C = fipronil concentration (igg/g of soil) and x = days after treatment). Concen-
trations of two metabolites, fipronil sulfone and fipronil sulfide, increased as fipronil resi-
dues decreased. Each treatment's affect on late instar mole crickets was significantly
different from the non-treated; however, there were no significant differences in nymph sta-
tus among fipronil-treated containers. Fipronil and residues of its metabolites 120 days after
application were 0.047 ig/g of soil and were high enough to kill or repel mole crickets to the
same extent as the 0-day treatment, 0.368 uig/g of soil. Repellency of fipronil and its metab-
olites was significant as the majority of nymphs evacuated or died in the treated containers,
but 35 of 37 nymphs were found alive in the non-treated containers.

Key Words: Scapteriscus vicinus, fipronil metabolite, fipronil sulfide, fipronil sulfone, repel-


En un experiment de invernadero, fipronil fue aplicado a raz6n de 0.014 kg ia/ha sobre
past Bermuda, Cynodon dactylon L., en recipients plasticos de 5-litros 120, 90, 60, 30, y 0
dias antes de poner una ninfa del grillo topo, Scapteriscus vicinus Scudder en el recipient.
Despu6s del period de exposici6n, el suelo en los recipients fue dividido en incrementos de
0-4, 4-8, y 8-18 cm de profundidad, y el estado del grillo fue registrado como muerto, ausente,
o vivo por la examinaci6n complete del suelo. El incremento de suelo de 0-4 cm fue analizado
por residues de fipronil y cuatro metabolitos de fipronil. La concentraci6n de fipronil en los
residues diminuyeron con el tiempo (C = 0.00002x 0.0053x + 0.3675, R2 = 0.9998 donde C=
la concentraci6n de fipronil (ig/g de suelo) y x = dias de tratamiento). Las concentraciones
de dos metabolitos, sulfona de fipronil y sulfito de fipronil, aumentaron con la diminuiaci6n
de residues de fipronil. El efecto de cada tratamiento sobre los mayores estadios de grillo to-
pos fue significativamente diferente que en las pruebas no tratadas; sin embargo, no habian
diferencias significativas en el estado de las ninfas entire los recipients tratados con fipronil.
Los residues de fipronil y sus metabolitos 120 dias despu6s de la aplicaci6n, 0.047 ig/g de
suelo, fueron suficientemente altos para matar o repeler los grillos topos al mismo grado que
en el tratamiento de 0-dias, 0.368 ig/g de suelo. La repelencia de fipronil y sus metabolitos
fue significativa, ya que la mayoria de las ninfas emigraron o morian en los recipients tra-
tados, pero 35 de las 37 ninfas se encuentraron vivas en los recipients no tratados.

Tawny mole crickets, Scapteriscus vicinus rean lawngrass (Zoysia japonica Steud.), centi-
Scudder (Orthoptera: Gryllotalpidae), are one of pedegrass (Eremochloa ophiuroides (Munro)
the most injurious turfgrass pests of bermuda- Hack.), St. Augustinegrass (Stenotaphrum secun-
grass (Cynodon dactylon L.), zoysiagrass or Ko- datum (Walt.) Kuntze), bahiagrass (Paspalum no-

Florida Entomologist 89(3)

tatum Fluegge), and pasture grasses on the
Coastal Plain regions of the US. From North
Carolina to Texas injury to the turf and cost of
control are high (Frank & Parkman 1999). Mole
crickets spend most of their lives tunneling
through soil feeding on turfgrass roots and soil or-
ganisms. As a consequence, they uproot turf
plants which dry out and die (Leslie 1994). Young
nymphs are difficult to monitor because of their
subterranean nature. As a result, mole crickets
are among the most difficult turfgrass insect to
control and account for hundreds of millions of
dollars in damage and control each year (Potter
1998). Very severe turf damage may require re-
seeding or resodding; and reinfestation is a com-
mon problem (Frank & Parkman 1999). The turf
may be further damaged by predators trying to
dig up the mole crickets (Frank & Parkman 1999).
Furthermore, thinned turf may be colonized by
opportunistic weeds (Frank & Parkman 1999).
Fipronil (5-amino-l-[2,6-dichloro-4-(trifluo-
1H-pyrazole-3-carbonitrile), a member of the phe-
nyl pyrazole class of pesticides, is an insecticide
that provides excellent mole cricket control.
Fipronil and its metabolites effectively control
home pests, termites, fire ants, mole crickets, wa-
ter rice weevil, and field corn pest (USGS 2003).
In turf, fipronil is applied as a granular product at
very low use rates (g ai (active ingredient)/ha).
Fipronil has four major metabolites: desulfi-
nylfipronil (5-amino-l-[2,6-dichloro-4-(trifluorom-
carbonitrile), fipronil amide (5-amino-1-[2,6-
acid), fipronil sulfide (5-amino-l-[2,6-dichloro-4-
thyl)thio]-l-H-pyrazole-3-carbonitrile), and
fipronil sulfone (5-amino-l-[2,6-dichloro-4-[(trif-
nyl]-l-H-pyrazole-3-carbonitrile). Desulfi-
nylfipronil is formed through photodegradation in
water and on soil. Fipronil amide is the product of
alkaline hydrolysis in water and soil. Fipronil sul-
fide forms slowly during anaerobic metabolism.
Fipronil sulfone is the product of oxidation in soil.
The objectives of this study were to determine
the length of influence of fipronil and its metabo-
lites on tawny mole cricket nymphs in bermuda-
grass and to measure degradation of fipronil and
appearance of fipronil metabolites.


Bermudagrass Cultivation

Dormant 'Tifway' bermudagrass from the
Sandhills Research Station, Jackson Springs, NC
was sectioned with a shovel into semi-cube shapes
15-cm wide and 8-cm deep on 22 Mar., 2003. The

150 bermudagrass cubes were placed in plastic 5-
liter containers (19 cm in diameter x 18 cm deep
(Classic 400, Nursery Supplies, Inc., Chambers-
burg, PA)). The bottom five cm of each container
was fitted with fiberglass screen (1.4-mm (New
York Wire, Mt. Wolf, PA)) to prevent nymph es-
cape through the drain holes as nymphs have the
ability to tunnel deep in the container.
The containers of soil and bermudagrass were
placed in a greenhouse with a supplemental pho-
toperiod of 14:10 (L:D). Sufficient top soil from the
research station (Candor sand (sandy, siliceous,
thermic, Arenic Paleudult)) was added to the bot-
tom of each container to ensure that the screen
was correctly positioned over the drain holes and
to make certain that when the bermudagrass
cube was added, the surface of the cube would be
2.0 cm below the rim of the container. The ber-
mudagrass was maintained at 2.0 cm mowing
height with battery-powered, reciprocating
shears. Turf in the containers was cut twice
weekly. Containers were turned /4 turn after each
mowing and rotated across the bench once a
week. Automatic overhead irrigation provided 0.6
cm water/day (2x annual mean). Peters 20-20-
20 fertilizer was applied at 24 kg N/ha on 4 Apr,
30 Apr, 23 May, 25 Jun, and 18 Aug 2003.
Fipronil (Chipco Choice 0.1 granular (G),
Bayer Environmental Science, Research Triangle
Park, NC) was applied with a saltshaker to the
surface of the bermudagrass containers at 0.014
kg ai/haand agedl20, 90, 60, 30, or 0 days before
adding one late-instar tawny mole cricket to each
container. All mole crickets were added on the
same day in both experiments. The five treat-
ments were arranged in a randomized complete
block with 11 replicates, and the experiment was
conducted twice. The number of replicates in the
untreated checks varied but was greater than 11.
Fipronil was applied in Experiment 1 on 7 May, 6
Jun, 8 Jul, 5 Aug, and 5 Sep 2003 to bermuda-
grass in unique containers. For Experiment 2,
fipronil was applied 21 May, 25 Jun, 22 Jul, 19
Aug, and 19 Sep 2003. Immediately following
each fipronil application, 0.3 cm of irrigation wa-
ter was applied. Experiment 1 included 11 non-
treated containers, and Experiment 2 included 26
non-treated containers (total of 147 containers
with one mole cricket nymph in each). Only one
cricket was applied to each test container because
tawny mole crickets are cannibalistic.
Tawny mole cricket nymphs were collected
from Brunswick County in southeastern NC on 14
Aug 2003 from the Oyster Bay Golf Links (33
52.91' N 78 32.10' W) for Experiment 1 and on 2
Sep 2003 from Sea Trails Golf Links (33 54.28' N
78 30.60' W) for Experiment 2. Nymphs were
brought to the soil surface by applying a 0.4% so-
lution of lemon Joy brand liquid dishwashing
detergent in six liters of water/m2 of turf (Short &
Koehler 1979). The collected nymphs were rinsed

September 2006

Cummings et al.: Fipronil Impact on Mole Cricket Behavior and Mortality

in lake water to remove the soapy residue and
placed in a 19-liter bucket half filled with moist
soil from the surrounding area. The day after col-
lection, each nymph was placed in its own 473-ml
plastic container half filled with moist soil and
half an earthworm (Lumbriculus ssp.) as a food
source. Nymphs were monitored for at least two
weeks before being added to the fipronil-treated
containers to ensure that the nymphs had recov-
ered from the stresses of collection.
One nymph was placed in a 3-4 cm deep hole
made by a rod 1 cm in diameter in the thatch layer
of the bermudagrass in the center of each con-
tainer on 5 and 19 Sep 2003 for Experiments-1
and -2, respectively. Nymphs were added to the
containers three h after "watering in" the 0-day
treatment. Nymphs were added to containers in
the order of expected lowest fipronil concentration
to highest (non-treated, 120, 90, 60, 30, or 0 days).
For Experiment 1, on 15 Sep, 10 days after adding
the nymphs, the containers were divided into
three sections (0-4, 4-8, 8-18 cm). The thatch layer
was approximately 0-4 cm thick. The status of the
nymphs was recorded as live, dead, or absent.
Nymphs found alive were placed in individual
473-ml plastic container with moist soil and mon-
itored for four days to ensure the recorded status
was accurate because fipronil has been observed
to require several days to control mole crickets in
the field (R. Brandenburg, Department of Ento-
mology, North Carolina State University, personal
communication 20 Jul 2004). The depth increment
where the nymph was located was noted. Because
more nymphs were recorded as absent in Experi-
ment 1 than expected, containers in Experiment 2
were processed four d after adding nymphs on 23
Sep 2003. The decreased nymph residence time in
treated containers facilitated locating dead
nymphs. In addition, the number of non-treated
containers was increased from 11 to 26.

Sample Preparation

All soil sections were bagged individually in
polyethylene bags, weighed, and stored at -18C.
All sections from containers with the same
fipronil treatment were simultaneously thawed,
sub sampled to determine percent moisture by
weight, and placed in 30 x 25 x 6-cm disposable
aluminum pans for two days to air dry at room
temperature. Then, the soil sections were sieved
(1.4 mm) and sub sampled. In addition to prepar-
ing the soil for fipronil and its metabolites extrac-
tion, sieving was used to locate fragments of
nymphs in Experiment 1 before recording them
as absent.

Extraction Procedure

To determine if the concentrations of fipronil
and fipronil metabolites were different in contain-

ers where nymphs were found absent or dead, soil
samples from the 0-4 cm section (thatch layer)
from each of two replicates where the nymphs
were reported as absent and where the nymphs
were reported as dead at the end of the experi-
ment were randomly selected for analysis from
each treatment in both Experiments 1 and 2. Be-
cause only four nymphs were found alive in all
fipronil-treated containers, only two samples
(120-day and 0-day) were analyzed where
nymphs were found alive.
Soil samples were extracted by sonication
(EPA Method 525.2 modified as described below).
Ten g of soil and 150 ml of acetone and n-hexane
(1:1) were placed in a 250-ml beaker and soni-
cated (Branson Sonifier 450, Branson Ultrason-
ics, Danbury, CT) for five min at 450 watts and 40
min duty cycle. The supernatant was filtered
through anhydrous sodium sulfate and glass wool
into a 500-ml boiling flask. Another 150 ml of ex-
traction solvent was added to the beaker and son-
icated for three min. The supernatant was filtered
into the same boiling flask. The volume was re-
duced to 2.0 ml by rotary evaporation under vac-
uum at 35C; then transferred quantitatively to a
10-ml test tube by using n-hexane for rinsing. In
order to remove the acetone from the solution, the
rinsate was reduced to 0.5 ml with a stream of dry
nitrogen and diluted to 8.0 ml with n-hexane. The
volume was reduced to 0.2 ml with dry nitrogen,
and then diluted to 3.0 ml with n-hexane. A flori-
sil solid phase extraction (SPE) Sep-Pak cartridge
(Waters, Milford, MA) was prepared with 10 ml of
hexane. The fipronil in 3.0 ml of hexane was then
injected into the cartridge very slowly. The
elutent was concentrated and analyzed to make
certain all fipronil was retained in the cartridge.
The fipronil and its metabolites were removed
from the cartridge with 5.0 ml of acetone mixed
with 5.0 ml of n-hexane. This elutent was reduced
to 1.0 ml with dry nitrogen and transferred to a 2-
ml GC auto sampler vial.
Samples were analyzed for fipronil, desulfi-
nylfipronil, fipronil amide, fipronil sulfide, and
fipronil sulfone with an HP 6890 GC coupled to an
HP 5973 MSD and a ZB 50 (Phenomenex, Tor-
rance, CA) column (30 m x 0.32 mm x 0.25pm film
thickness) with the following temperature pro-
gram: Injection port: 175C, initial temp: 80C,
initial hold: 1.0 min, ramp rate: 20C/min to
250C, ramp rate 6C/min to 287C and hold 1.0
min, ramp 25C/min to 300C and hold 5.0 min.
The electron capture detector (ECD) temperature
was 300C. Helium was used as the carrier gas at
a mean flow of 1.0 ml/min. Nitrogen was used as
the detector makeup gas. One pl of sample was in-
jected. Calibration standards at concentrations of
100, 10, and 5.0 pg/ml were used for quantifica-
tion. Separate soil samples treated with known
concentrations of fipronil and its metabolites
were run after every six samples. Every other

Florida Entomologist 89(3)

sample was injected into a mass spectrometer to
verify the concentration and molecule.

Data Analysis

The status of the nymphs was analyzed by
Pearson Chi Square analysis and by logistic re-
gression fitting experiment and treatment effects
with the GENMOD procedure of SAS version 8.0
(SAS Institute 2001). GENMOD requires a binary
response; thus to accommodate this procedure,
numbers of absent and dead nymphs (impacted
by fipronil or metabolite) were combined to com-
pare against the number of live nymphs. Likeli-
hood ratio chi square-tests (P < 0.05) were used to
determine significant differences. The most ap-
propriate degradation model for fipronil and
fipronil metabolites residual concentrations was
determined with SAS.


Only four nymphs of 147 were recorded as live
in fipronil-treated containers at the end of both
experiments even after 120 d of summer degrada-
tion in a greenhouse. The numbers of dead and
absent nymphs in the 110 fipronil-treated con-
tainers were 51 and 55, respectively. The data
suggest that fipronil and its metabolites modified
the behavior of the nymphs and indicate repel-
lency (Table 1). Of 37 nymphs in the non-treated
containers, 35 (94.6%) were found alive; one was
absent and one was dead. Therefore, the experi-
mental apparatus was not inherently lethal or re-
pellent to the nymphs. When fipronil concentra-
tions were greatest (0-day), the number of absent
nymphs was greatest. As fipronil concentrations
decreased and fipronil metabolite (fipronil sulfone
and fipronil sulfide) concentrations increased,
nymph mortality also tended to increase and re-
pellency tended to decrease (Fig. 1). Therefore,
fipronil or its metabolites were impacting the
mole crickets by either causing mortality or
avoidance behavior.
After noting the higher than expected number
of absent nymphs in Experiment 1, the nymph
residence time in Experiment 2 was reduced from
10 to four d. In Experiment 1, the absence of a
nymph in a container was attributed to quick
death followed by rapid mole cricket decomposi-
tion. While field observations have suggested
nymph avoidance of treated areas, fipronil and its
metabolites have not been documented to cause
nymph repellency. Villani et al. (2002) noted that
nymphs avoided biocontrol control agents (patho-
genic fungi) in the treated layer of soil at the sur-
face by remaining deep in the soil. If nymphs were
to detect the fipronil and its metabolites, they
were not expected to pass through the treated
layer at the surface and walk off the edge of the
container, but rather that they would tunnel

deeper into the soil (Villani et al. 2002). However,
Villani et al. (2002) conducted experiments in con-
tainers with high side walls and lids where mole
crickets could not leave the system; tunneling
deeper was the only option for avoidance.
Fipronil targets the y-aminobutyric acid type A
(GABA) receptor system which disrupts nerve
function in insects by blocking the GABA-gated
choride channels of neurons (California Environ-
mental Protection Agency (CEPA) 2001). Thus at
sufficient doses, fipronil causes excessive neu-
ronal excitation. At the conclusion of the 4-d
nymph exposure period of Experiment 2, 13
nymphs were found alive. Nine of these nymphs
were supine with rapid movement of the legs. All
nine of these nymphs died within four d and were
recorded as dead. Therefore, based on observation
of these nymphs and the mode of action of fipronil,
it is likely that some of the nymphs reported as
absent, received a high enough dose to be affected
by fipronil and its metabolites. The affected
nymphs may have walked over the edge of the
container. Seven nymphs were found dead below
the containers, but could not be assigned to a spe-
cific container. Perhaps some of the nymphs,
which were absent but not found, were consumed
by rodents known to inhabit the greenhouse.
When Experiments 1 and 2 are combined, the
number of nymphs recorded as absent is similar
to the number of dead (Table 1). During the statis-
tical analysis, the numbers of dead and absent
nymphs (impacted nymphs) were combined for
comparison against the number found live. All
fipronil treatment timings were different from the
non-treated (Chi-square = 124, P = <0.001). When
the results of the non-treated containers were re-
moved from the statistical analysis, there were no

0, 30, 60, 90, OR 120 D BEFORE ADDING ONE

Fipronil Live Dead Absent
Application % % %
Time (number) (number) (number)

0 Day 9.1(2) 27.3 (6) 63.6(14)
30 Day 4.5 (1) 54.5 (12) 40.9(9)
60 Day 0 (0) 45.5 (10) 54.5 (12)
90 Day 4.5 (1) 45.5 (10) 50 (11)
120 Day 0 (0) 59.1(13) 40.1(9)
Not Treated 94.6 (35)b 2.7 (1) 2.7 (1)

"Application Time 0, 30, 60, 90 and 120 d are not different
with regard the number of live insects recovered (P = >0.15).
bApplication times 0, 30 60, 90 and 120 d are all different
from the Not Treated (Chi-Square P = <0.001).

September 2006

Cummings et al.: Fipronil Impact on Mole Cricket Behavior and Mortality

significant differences among fipronil treatment
timings. Therefore, the 120-d application had the
same effect as the 0-d application in a greenhouse
during the summer where two times the normal
annual rainfall was applied (Chi-square > 2.095
with P > 0.15). Fipronil was expected to degrade
more rapidly as its reported disappearance time
in turf is 12-15 d vs. 33-75 d in bare soil (CEPA
2001). In this experiment, the half-life of fipronil
was approximately 40 d.
The mean 0-d-treatment concentration for Ex-
periments 1 and 2 for fipronil in the soil in the 0-
4-increment was 0.368 pg/g of soil, and mean 120-
d-treatment concentration was 0.047 pg/g of soil
(Fig. 1). The degradation of fipronil is best de-
scribed by the quadratic equation C = 0.00002x2 -
0.0053x + 0.3672, R2 = 0.9998 where C = fipronil
concentration (pg/g of soil) and x = days after
treatment. As the concentration of fipronil de-
creased, the concentrations of fipronil sulfone and
fipronil sulfide increased. Fipronil sulfone is the
major fipronil metabolite detected in this study.
Fipronil sulfone is formed through aerobic soil
metabolism and was expected to be the major deg-
radation product. The appearance of fipronil sul-
fone is best described by the quadratic equation C
= -0.00003x2 + 0.0039x + 0.1228, R2 = 0.8629

0.40 -

where C = fipronil sulfone concentration (pg/g of
soil) and x = days after treatment. The concentra-
tion of fipronil sulfone was greatest in the 60-d-
soil samples which indicates that this molecule is
also degrading. Fipronil sulfide concentrations
also increased as fipronil concentrations de-
creased with a maximum concentration reported
in the 60-d-samples. Fipronil sulfide is formed
through degradation in soil and water under
anaerobic conditions. The heavy irrigation rate
and screen in front of the drain holes may have al-
lowed anaerobic conditions to form temporarily.
There were no significant differences in fipronil
concentrations in containers where nymphs were
dead or absent at the end of the experiment. The
same relationship is true for fipronil sulfone and
fipronil sulfide. Metabolite residues were detected
in the 0-d-samples because nymphs were exposed
for 10 and four d after application for Experi-
ments 1 and 2, respectively.
Desulfinylfipronil and fipronil amide were de-
tected in all fipronil treated containers at minor
concentrations (data not shown). The mean con-
centrations were similar across treatments and
did not show a relation to fipronil residue concen-
trations. Desulfinylfipronil is formed during pho-
tolysis and was not expected to be present in high

-*- Fipronl
---- Fipronil Sulfone
--- Fipronil Sulfide
-- -- Fipronil (Quadratic)
------- Fipronil Sulfone (Quadratic)
- -- Fipronil Sulfide (Quadratic)
Fipronil Concentration = 2 x 10x' 0.0053x + 0.367
..R2=0.9998 P<0.0187

-Fipronil Sulfone Concentration = -3 x 10"4x2 + 0.0039x + 0.123
R2=0.8629 P<0.0102
- Fipronil Sulfide Concentration= -1 x 10x2 +0.0015x + 0.041
SR2 -0.8026 P<0.0002







0.00 -


30 60 90 120

Days after Treatment (x)

Fig. 1. Degradation of parent material and appearance of metabolites with time.

concentrations due to shading of the soil surface
by the bermudagrass canopy. Fipronil amide is
formed during alkaline hydrolysis; the conditions
necessary for fipronil amide to be produced were
not expected (USGS 2003).


Fipronil at a mean concentration of 0.047 pg/g
of soil in the 120-d-treatment impacted tawny
mole cricket nymphs as much as fipronil at a
mean concentration of 0.368 pg/g of soil in the 0-
d-treatment (0-4 cm). Fipronil or fipronil metabo-
lites impacted the nymphs similarly by either
causing death or repellency. As fipronil concentra-
tion decreased, fipronil sulfone and fipronil sul-
fide concentrations increased. Fipronil and its
metabolites are highly effective tawny mole
cricket nymph management tools and have the
potential to provide season-long control or pre-
vent reinvasion in the same season (120 d).


We acknowledge Bayer Environmental Science for
providing products; S. R. Thompson, R. M. Heltsley, and
Dr. G. E. Mahnken for technical assistance; North Caro-
lina Turfgrass Council and Center for Turfgrass Envi-
ronmental Research and Education for financial
support; and Dr. Cavell Brownie for assistance with sta-
tistical analysis. This work was completed as part of a
post graduate degree requirement.

September 2006


(CEPA). 2001. Environmental fate of fipronil. (On-
line.) Available at http://www.pw.ucr.edu/textfiles/
fipronil.pdf (accessed 13 May, 2004; verified 13 No-
vember, 2004). CEPA, Sacramento, CA.
FRANK, J. H., AND J. P. PARKMAN. 1999. Integrated pest
management of pest mole crickets with emphasis on
the southeastern USA. Int. Pest Manag. Rev. 4: 39-
LESLIE, A. R. (Ed.) 1994. Integrated Pest Management
for Turf and Ornamentals. Lewis Publishers, Boca
Raton, FL. pp. 234-235
POTTER, D. A. 1998. Destructive Turfgrass Insects: Bi-
ology, Diagnosis, and Control. Ann Arbor Press,
Chelsea, Michigan.
SHORT, D. E., AND P. G. KOEHLER. 1979. A sampling
technique for mole crickets and other pest in turf-
grass and pasture. Florida Entomol. 62: 282-283.
SAS INSTITUTE. 2001. User's Guide, Version 8.2. SAS
Institute, Cary, NC.
U.S. GEOLOGICAL SURVEY (USGS). 2003. Method of
analysis by the U.S. Geological Survey National Wa-
ter Quality Laboratory-A method supplement for
the determination of fipronil and degrades in water
by gas chromatography/mass spectrometry. Open-
File Report 02-462. U.S. Geological Survey Method
0-1126-02. Denver, Colorado.
Use of radiography and tunnel castings for observ-
ing mole cricket (Orthoptera: Gryllotalpidae) behav-
ior in soil. American Entomol. 48: 42-50.

Florida Entomologist 89(3)

Szalanski et al. Genetic Evidence for new Termite Species


1Department of Entomology, University of Arkansas, Insect Genetics Research Laboratory, Fayetteville, AR 72701

2Center for Urban & Structural Entomology, Department of Entomology, Texas A&M University,
College Station, TX 77843

3Dow AgroSciences LLC, 9330 Zionsville Road, Indianapolis, IN 46268


Genetic evidence for a new subterranean termite species herein named Reticulitermes
okanaganensis is provided based on DNA sequence analysis. Partial sequences of the mito-
chondrial DNA rRNA 16S gene were obtained from 27 samples of R. okanaganensis from
British Columbia, Idaho, Oregon, Nevada and California. Five nucleotide sites were variable
among the four observed haplotypes. One haplotype occurred only once, while the most com-
mon haplotype, 03, occurred in 37% of the samples. Molecular phylogenetic analysis of R.
okanaganensis relative to five other North American Reticulitermes species has clarified its
distinct position within the genus.

Key Words: 16S gene, DNA sequence, Reticulitermes, termite


Se provee evidencia gen6tica para una nueva especie de termita aqui nombrada Reticulitermes
okanaganensis basada sobre el analisis de la secuencia de ADN. Secuencias parciales del gen
16s de rARN del ADN mitocondrial fueron obtenidas de 27 muestras de R. okanaganensis de
Columbia Britanica, Idaho, Oregon, Nevada, y California. Cinco sitios de nucle6tidos fueron
variables entire los cuatro haplotipos observados. Un haplotipo ocurrio solamente una vez,
mientras que el haplotipo mas comun, "03", ocurrio en 37% de las muestras. El analisis filoge-
n6tica molecular de R. okanaganensis en relacion con las otras cinco species de Reticulitermes
norteamericanas ha clarificado su posici6n distinta dentro del g6nero.

There is a general consensus that the genus
Reticulitermes is in need of revision (Weesner 1970;
Nutting 1990; Scheffrahn & Su 1994; Forschler &
Jenkins 1999). This is an especially difficult be-
cause of the lack of discrete morphological charac-
ters which provide for accurate identification of
specimens to the specifies level. For this reason,
the application of non-morphological identifica-
tion methods such as the evaluation of cuticular
hydrocarbon analysis and mtDNA markers have
been used.
Most recently, the application of the mitochon-
drial rRNA 16S gene has been applied to identify
Reticulitermes populations from the south central
United States (Austin et al. 2004a, b, c) for species
across North America (Austin et al. 2005b), and
for clarification of exotic introductions of nearctic
Reticulitermes around the world (Austin et al.
2005a; Su et al. 2006). The application of this
marker has tremendous potential for molecular
diagnostics of Reticulitermes, with increased accu-
racy of positive species identifications (Szalanski
et al. 2003), and clarifying the identities of exotic

introductions around the world (Austin et al.
2005 a) and from North America (Austin et al.
2005 b). Recently, Copren et al. (2005) found evi-
dence for as many as seven new species of Reticu-
litermes from the western United States based
upon cuticular hydrocarbon phenotypes, but re-
solved to designate them as putative new species
after attempting to corroborate their relationship
with molecular phylogenetics and reproductive
flight dates. These discrepancies likely are attrib-
uted to the environmental plasticity of cuticular
hydrocarbons and stresses the need for fixed char-
acter states for species identification such as
mtDNA sequences.
We provide the first genetic evidence of a new
species, named herein as R. okanaganensis after
its first collection site in the Okanagan region of
British Columbia, Canada, and provide a phylo-
genetic analysis of Reticulitermes by applying the
16S mtDNA gene. In addition, the geographic
distribution of Reticulitermes species and haplo-
types of species throughout the region are dis-

Florida Entomologist 89(3)


Termites were collected from various locations
in British Columbia, Idaho, Oregon, Nevada, and
California, both from our own collecting efforts
and from the 2002 National Termite Survey (Ta-
ble 1). Samples were preserved in 100% ethanol.
Alcohol-preserved specimens were allowed to
dry on filter paper, and DNA was extracted ac-
cording to Liu & Beckenbach (1992) on individual
whole worker termites with the Puregene DNA
isolation kit D-5000A (Gentra, Minneapolis, MN).
Extracted DNA was resuspended in 50 pl of
Tris:EDTA and stored at -20C. Polymerase chain
reaction was conducted with the primers LR-J-
13007 (5'-TTACGCTGTTATCCCTAA-3') (Kamb-
hampati & Smith 1995) and LR-N-13398 (5'-
1994). These PCR primers amplify an approxi-
mately 428-bp-region of the mtDNA 16S rRNA
gene. The PCR reactions were conducted with 2 pl
of the extracted DNA (Szalanski et al. 2000), with
a profile consisting of 35 cycles of 94C for 45 s,
46C for 45 s, and 72C for 60 s. Amplified DNA
from individual termites was purified and concen-
trated with minicolumns (Wizard PCRpreps,
Promega) according to the manufacturer's in-
structions. Samples were sent to The University
of Arkansas Medical Center DNA Sequencing Fa-
cility (Little Rock, AR) for direct sequencing in
both directions. GenBank accession numbers
were DQ389178-DQ389180 and DQ438936 for
the Reticulitermes okanaganensis haplotypes found
in this study. Consensus sequences for each sam-

ple were obtained with Bioedit 5.09 (Hall 1999).
Mitochondrial DNA haplotypes were aligned with
MacClade v4 (Sinauer Associates, Sunderland,
The distance matrix option of PAUP* 4.0b10
(Swofford 2001) was used to calculate genetic
distances according to the Kimura 2-parameter
model of sequence evolution (Kimura 1980). Mi-
tochondrial 16S sequences from R. flavipes, R. hes-
perus, R. tibialis, R. hageni, R. flavipes, and R. vir-
ginicus (Szalanski et al. 2003; Austin et al.
2004a,b,c) were added to the DNA sequence
dataset for comparison. DNA sequences from the
Formosan termite, Coptotermes formosanus
Shiraki (GenBank AY558910), and Heterotermes
aureus (Snyder) (GenBank AY280399), were
added to act as outgroup taxa. DNA sequences
were aligned by CLUSTAL W (Thompson et al.
1994). Maximum likelihood and unweighted par-
simony analysis on the alignments were con-
ducted by PAUP* 4.0b10 (Swofford 2001). Gaps
were treated as missing characters for all analy-
sis. The reliability of trees was tested with a
bootstrap test (Felsenstein 1985). Parsimony
bootstrap analysis included 1,000 resamplings
with the Branch and Bound algorithm of PAUP*.
For maximum likelihood analysis, the default
likelihood parameters were used (HKY85 six-pa-
rameter model of nucleotide substitution, empir-
ical base frequencies with the exception of the
transition/transversion ratio, which was set to
2.596954:1). These parameters were used to
carry out a heuristic search by PAUP* with a
neighbor joining tree as the starting tree.


State/Prov City County Lat/Long Hap n

BC Osoyoos 49:02:09 N 119:27:51 W 01 2
CA Placerville El Dorado 38:43:47 N 120:47:51 W 01 1
NV Reno Washoe 39:31:47 N 119:48:46 W 01 3
ID Boise Ada 43:36:49 N 116:12:09 W 01 1
ID Lewiston Nez Perce 46:25:00 N 117:01:00 W 01 1
CA Chino San Bernardino 34:00:44 N 117:41:17 W 02 1
CA Grass Valley Nevada 39:13:09 N 121:03:36 W 02 1
CA Irvine Orange 33:40:10 N 117:49:20 W 02 1
CA Napa Napa 38:17:50 N 122:17:04 W 02 1
NV Reno Washoe 39:31:47 N 119:48:46 W 02 2
NV Carson City Carson City 39:09:50 N 119:45:59 W 02 1
CA Auburn Placer 38:53:48 N 121:04:33 W 03 2
CA Bakersfield Kern 35:22:24 N 119:01:04 W 03 1
CA Dinuba Tulare 36:32:36 N 119:23:10 W 03 1
CA Lafayette Contra Costa 37:53:09 N 122:07:01 W 03 1
CA Lake Arrowhead San Bernardino 34:15:52 N 117:11:04 W 03 1
CA Napa Napa 38:17:50 N 122:17:04 W 03 1
CA Strathmore Tulare 36:08:44 N 119:03:35 W 03 1
CA Walnut Creek Contra Costa 37:54:23 N 122:03:50 W 03 1
OR Klamath Falls Klamath 42:13:30 N 121:46:50 W 03 1
NV Reno Washoe 39:31:47 N 119:48:46 W 04 2

September 2006

Szalanski et al. Genetic Evidence for new Termite Species


DNA sequencing of the 16S rRNA amplicon
from R. okanaganensis revealed an average size of
428 bp. The average base frequencies were A =
0.41, C = 0.23, G = 0.13, and T = 0.23. Among the
27 R. okanaganensis mtDNA rRNA 16S DNA se-
quences, a total of 5 nucleotide sites were variable
(Table 2). Four distinct haplotypes (lineages) were
observed (Table 1), and genetic divergence among
these haplotypes ranged from 0.2 to 0.5 percent.
One haplotype, 04, occurred only once, while the
most common haplotype, 03, occurred in 37% of
the R. okanaganensis samples. While haplotype 04
was only found in Nevada, haplotype 01 was
found over the largest geographical area
(Table 1).
We conducted a phylogenetic analysis on R.
okanaganensis relative to all described North
American Reticulitermes species to clarify the phy-
logenetic relationships of R. okanaganensis within
the genus. Parsimony analysis of the aligned Reti-
culitermes spp. and the outgroup taxa used 436
characters, of which 90 were variable (20%) and
59 (14%) were parsimony informative. This anal-
ysis had a single consensus tree with a length =
154, and a CI value of 0.695 and verified the dis-
tinct monophyly of'R. okanaganensis' in the genus
(Fig. 1). Maximum likelihood analysis (-ln L =
1363.74728) also supported a distinct clade for R.
okanaganensis and was identical to the maximum
parsimony analysis. Reticulitermes okanaganensis
formed a sister group to R. flavipes, while R. hespe-
rus grouped with R. tibialis, and R. hageni formed a
distinct clade with R. virginicus.


Phylogenetic analysis of North American Reti-
culitermes revealed a distinct clade for R. okanagan-
ensis. There is clear genetic evidence for R. okana-
ganensis being identified as a discrete species, pos-
sessing multiple haplotypes over a broad geo-
graphical range including California, Idaho,
Oregon, Nevada, and British Columbia (Fig. 2).
The genetic isolation observed demonstrates that
this group represents a distinct taxonomic entity,
albeit cryptic due to overlapping morphological
characters with other congeners from the region
that merits further investigation.


haplotype 75 160 364 369 375

01 C G T T
02 A A
03 A
04 T

The identification of new species from north-
ern California has been reported by Haverty &
Nelson (1997), Haverty et al. (1999) applying cu-
ticular hydrocarbons, and further investigated
with ethological data (Getty et al. 2000a,b). It has
been proposed that extensive collecting from
western states may produce as many as 6 new
Reticulitermes species, and perhaps at least three
Reticulitermes species in Mexico (Myles 2000). We
have already identified other morphologically and
genetically distinct nearctic Reticulitermes from
this region and we are attempting to collect addi-
tional material in the near future so that subse-
quent investigations can be performed and the
overall relationships of western nearctic Reticuli-
termes can be better understood.
Reticulitermes okanaganensis has become a prob-
lem in British Columbia, where there is an un-
precedented number of attacks to structures, par-
ticularly around Kelowna, BC (Hugh Philip un-
published). Reports of difficulties controlling Reti-
culitermes infestations from urban structures in
northern California (Kistner & Sbragia 2001)
prompts us to consider whether these difficult
control scenarios are due to the structures them-
selves, or the possibility of infestations from this
newly identified species, that may have slightly
different nutritional, ecological, and behavioral
The western subterranean termite, Reticuliter-
mes hesperus, is the most destructive termite found
in California (Lewis 2001), but it is highly proba-
ble that R. okanaganensis accounts for a significant
portion of damage previously thought to be asso-
ciated with R. hesperus. Snyder (1949) originally
described R. hesperus based on material from Lit-
tle Bear Lake, San Bernadino mountains, Califor-
nia and samples previously collected from Van-
couver, British Columbia, and suggests the range
to extend from British Columbia through Wash-
ington, Oregon, California, and Nevada, extend-
ing down to Baja, California (Spencer 1937). The
northern limit of its range occurs in the upper
Frazer valley from Lytton to Kamloops, British
Columbia (Spencer 1937, 1945). Our genetic eval-
uation of populations from along the Pacific and
throughout the western states suggest that R. hes-
perus does not generally occur away from the pa-
cific coast region, as is described by Weesner
(1965), and is isolated west of the Sierra Nevada/
Cascades mountain ranges (unpublished). Reticu-
litermes tibialis overlaps entirely with the range of
R. okanaganensis, and extends east as far as Indi-
ana, but is not commonly found east of Texas and
Oklahoma (Banks & Snyder 1920; Austin et al.
2004a). Light and Pickens (1934) indicate collec-
tions of R. tibialis from the northern tip of Idaho,
while the U.S. Dept. of Agriculture (1959a,b) have
reports from Twin Falls and Lewiston to Lapwari,
Idaho. Weesner (1970) concedes that collections
from western California to as far east as Elko and

Florida Entomologist 89(3)

'R okanaganensis'CA hap 01
93 'R okanaganensis'CA hap 02
'R okanaganensis'CA hap 03
'R okanaganensis'NV hap 04
59 -R flavipes hap D
65 R flavipes hap F
R flavipes hap LL
80 R flavipes hap J
R flavipes hap FF
R flavipes hap M
R flavipes hap P
84 R flavipes hap X
R flavipes hap Z
R hesperus CA hap HE1
86 R hesperus CA hap HE2
R hesperus OR hap HE3
R hesperus OR hap HE4
100 ,61 R tibialis TX hap T1
R tibialis TX hap T2
R tibialis TX hap T3
91 R tibialis CA hap T9
66 R tibialis UT hap T10
R tibialis TX hap T4
R tibialis TX hap T5
R tibialis TX hap T6
R tibialis OK hap T7
R tibialis OK hap T8
R hageniTX hap H1
81 RhageniAR hap H2
R hageniAR hap H3
66 R virginicus TX hap V1
R virginicus AR hap V3
56 R virginicus LA hap V4
92 R virginicus AR hap V5
R virginicus MS hap V8
R virginicus GA hap V6
R virginicus MO hap V7
R virginicus VA hap V9
C formosanus
H aureus
Fig. 1. Maximum parsimony cladogram of Reticulitermes okanaganensis and related taxa.

Reno, Nevada appear to be R. tibialis, but are too
difficult to assign to either R. hesperus or R. tibialis.
Reticulitermes okanaganensis haplotype 01,
which has been reported in Osoyoos, BC, Canada,
is also found in El Dorado county California,
Washoe county Nevada, and Ada and Nez Perce

counties in Idaho. Osoyoos, B.C. is located in the
southern interior ecoprovince of Canada, a region
characterized as a desert climatic zone. We are
currently investigating the distribution of Reticu-
litermes species throughout the western United
States. From this study and ongoing research, we

September 2006

Szalanski et al. Genetic Evidence for new Termite Species

Fig. 2. Distribution of Reticulitermes okanaganensis
haplotypes in British Columbia, Idaho, Oregon, Ne-
vada, and California.

expect that R. okanaganensis will be found in the
drier regions north of the Cascades and East of
the Sierra Nevada Ranges. While the possible im-
portation of this species either to the United States
or to Canada should be considered, in all probabil-
ity we are dealing with a cryptic endemic species.
Establishment of this species in Canada may be at-
tributed to trade between the US and Canada, and
may account for the unprecedented number of
structures being attacked in British Columbia. It
may be equally plausible that urban sprawl and
encroachment into previously undeveloped areas
for housing needs has contributed to the increase
frequency of attacks to structures there.
An important criterion for determining the ex-
tent of genetic variation for a species lies in the
ability to sample from populations evenly distrib-
uted within the species range (Mayr & Ashlock

1991). For this reason, future studies of this un-
known species, including a proper species descrip-
tion, are required. This study represents an im-
portant first step towards this endeavor.


We thank M. Rust for critical review of this manu-
script. Thanks are extended to R. Saran Univ. of CA Riv-
erside, P. Pachamuthu, and S. Vega of Western
Exterminating (Anaheim, CA), numerous pest manage-
ment professionals, and especially to H. Philip of the Ca-
nadian Ministry of Agriculture, Food and Fisheries, Food
Safety and Quality Branch Plant Health Unit, Kelowna,
B. C. Canada, for providing samples. This research was
supported in part by the University of Arkansas, Arkan-
sas Agricultural Experiment Station, and the City of
New Orleans Mosquito and Termite Control Board.


2004a. Genetic variation and distribution of the sub-
terranean termite genus Reticulitermes (Isoptera:
Rhinotermitidae) in Arkansas and Louisiana. Flor-
ida Entomol. 87: 473-480.
FOSTER 2004b. Genetic variation and geographical
distribution of the subterranean termite genus Reti-
culitermes in Texas. Southwest. Entomol. 29: 1-11.
Distribution and genetic variation of Reticulitermes
(Isoptera: Rhinotermitidae) in Oklahoma. Florida
Entomol. 87: 145-151.
2005a. Genetic evidence for the synonymy of two
Reticulitermes Species: Reticulitermes flavipes (Kol-
lar) and Reticulitermes santonensis (Feytaud). Ann.
Entomol. Soc. Amer. 98: 395-401.
M. T. MESSENGER 2005b. Genetic variation of Reti-
culitermes flavipes (Isoptera: Rhinotermitidae) in
North America using the mitochondrial rRNA 16S
gene. Ann. Entomol. Soc. Amer. 98: (in press).
BANKS, N. A., AND T. E. SNYDER 1920. A Revision of the
Nearctic Termites. U.S. Nat. Mus., Bull. 108: 228 pp.
BANKS, N. A. 1946. Species Distinction in Reticuliter-
mes (Isoptera: Rhinotermitidae) MS thesis. Univer-
sity of Chicago, Chicago, Illiniois. 29 pp.
HAVERTY. 2005. Phylogenetic analyses ofmtDNA se-
quences corroborate taxonomic designations based
on cuticular hydrocarbons in subterranean termites.
Mol. Phylogen. Evol. 35: 689-700.
LETON. 2000. Morphological phylogenetics of ter-
mites (Isoptera). Biol. J. Linnean Soc. 70: 467-513.
EMERSON, A. E., AND E. M. MILLER 1943. A key to the
termites of Florida. Entomol. News 54: 184-187.
FELSENSTEIN, J. 1985. Confidence limits on phyloge-
nies: An approach using the bootstrap. Evolution 39:
FORSCHLER, B. T., AND T. M. JENKINS. 1999. Evaluation
of subterranean termite biology using genetic,
chemotaxonomic, and morphometric markers and

ecological data: a testimonial for multi-disciplinary
efforts. Trends Entomol. 2: 71-79.
GETTY, G. M., M. I. HAVERTY, K. A. COPREN, and V. R.
LEWIS. 2000a. Response of Reticulitermes spp.
(Isoptera: Rhinotermitidae) in northern California
to baiting with hexaflumuron with Sentricon Ter-
mite Colony Elimination System: J. Econ. Entomol.
93: 1498-1507.
GETTY, G. M., M. I. HAVERTY, AND V. R. LEWIS. 2000b.
Agonistic behavior between recently collected and
laboratory cultured Reticulitermes spp. (Isoptera:
Rhinotermitidae) from northern California: Pan-Pa-
cific Entomol. 76: 243-250.
HALL, T. A. 1999. BioEdit: a user-friendly biological se-
quence alignment editor and analysis program for
Windows 95/98/NT. Nucl. Acids. Symp. Ser. 41: 95-98.
HAVERTY, M. I., AND L. J. NELSON. 1997. Cuticular hy-
drocarbons of Reticulitermes (Isoptera: Rhinotermiti-
dae) from northern California indicate undescribed
species. Comp. Biochem. Physiol. B118: 868-880.
1999. New cuticular hydrocarbon phenotypes of
Reticulitermes (Isoptera: Rhinotermitidae) from the
United States. Sociobiol. 34: 1-21.
FRAHN. 1995. Intracolony morphometric variation
and labral shape in Florida Reticulitermes (Isoptera:
Rhinotermitidae) soldiers significance for identifica-
tion: Flor. Entomol. 78: 119-129.
KAMBHAMPATI, S., AND P. T. SMITH. 1995. PCR primers
for the amplification of four insect mitochondrial
gene fragments. Insect Molec. Biol. 4: 233-236.
KIMURA, M. 1980. A simple method for estimating evo-
lutionary rate of base substitutions through compar-
ative study of nucleotide sequences. J. Molec. Evol.
16: 111-120.
KISTNER, D. H., AND R. J. SBRAGIA. 2001. The use of the
Sentricon Termite Colony Elimination System for
controlling termites in difficult control sites in
northern California. Sociobiol. 37: 265-280.
LEWIS, V. R. July 2001. Pest Notes: Termites. UC ANR
Publication 7415. 6 pp. Also available online at
LIU, H., AND A. T. BECKENBACH. 1992. Evolution of the
mitochondrial cytochrome oxidase II gene among 10
orders of insects. Mol. Phylogen. Evol. 41: 41-52.
LIGHT, S. F. 1934. The termite fauna of North America
with special reference to the United States, pp. 127-
135 In C. A. Kofoid, S. F. Light, A. C. Horner, M.
Randall, W. B. Herms, and E. E. Bowe [eds.], Ter-
mites and Termite Control, 2nd ed., University of
California Press, Berkeley, CA.
LIGHT, S. F., AND A. L. PICKENS. 1934. American subter-
ranean termites and their distribution, pp. 150-156
In C. A. Kofoid [ed.], Termites and Termite Control,
Univ. of California Press, Berkeley, California.
MAYR, E., AND P. D. ASHLOCK. 1991. Principles of Sys-
tematic Zoology, 2nd ed. McGraw-Hill Book Com-
pany, New York.
MILLER, E. M. 1949. A Handbook on Florida Termites.
University of Miami Press, Coral Gables, Florida. 30
MYLES, T. G., AND M. M. HOOTEN. 2000. Evaluation of
the Occurrence of Termite Species and Potential
Termite Burrowing Depths in the Area of the Ne-
vada Test Site: Present Day and for the Next 10,000
Years. Neptune and Co., Inc. Document Number
05100-01. 17 pp.

September 2006

NUTTING, W. L. 1990. Insecta, Isoptera, pp. 997-1032 In
D. L. Dindal, D. L. [ed.], Soil Biology Guide. Wiley
and Sons, New York.
PICKENS, A. L. 1934a. The biology and economic signif-
icance of the western subterranean termite, Reticu-
litermes hesperus, pp. 157-183 In C. A. Kofoid, S. F.
Light, A. C. Horner, M. Randall, W. B. Herms, and
E. E. Bowe [eds.], Termites and Termite Control,
2nd ed. University of California Press, Berkeley, CA.
PICKENS, A. L. 1934b. The barren-lands subterranean
termite, Reticulitermes tibialis, pp. 184-186 In C. A.
Kofoid, S. F. Light, A. C. Horner, M. Randall, W. B.
Herms, and E. E. Bowe [eds.], Termites and Termite
Control, 2nd ed. University of California Press, Ber-
keley, CA.
SCHEFFRAHN R. H., AND N-Y Su. 1994. Keys to soldier
and winged adult termites (Isoptera) of Florida.
Florida Entomol. 77: 460-474.
AND P. FLOOK. 1994. Evolution, weighting, and phy-
logenetic utility of mitochondrial gene sequences and
a compilation of conserved polymerase chain reaction
primers. Ann. Entomol. Soc. Amer. 87: 651-701.
SNYDER, T. E. 1949. Catalog of the Termites (Isoptera)
of the World: Smithsonian Misc. Coll. 112: 1-490.
SOKAL, R. R., AND T. J. CROVELLO. 1970. The biological
species concept: a critical evaluation. Amer. Natur.
104: 127-153.
SPENCER, G. J. 1937. The termite situation in British
Columbia in 1936. Proc. Entomol. Soc. Brit. Colum-
bia 33: 42-43.
SPENCER, G. J. 1945. On the incidence, density and de-
cline of certain insects in British Columbia. Proc.
Entomol. Soc. Brit. Columbia 42: 21.
GIBLIN-DAVIS. 2006. Identification of Chilean Reti-
culitermes (Isoptera: Rhinotermitidae) Inferred from
Three Mitochondrial Gene DNA Sequences and Sol-
dier Morphology. Ann. Entomol. Soc. Am. 99: 352-
SWOFFORD, D. L. 2001. PAUP*: Phylogenetic analysis
using parsimony (*and other methods), ver. 4.0b10.
Sinauer, Sunderland, Massachusetts.
FRITZ. 2000. Population genetics and phylogenetics
of the endangered American burying beetle, Nicro-
phorus americanus (Coleoptera: Silphidae). Ann.
Entomol. Soc. Amer. 93: 589-594.
Identification of Reticulitermes Spp. (Isoptera: Rhi-
notermitidae) from the South Central United States
by PCR-RFLP. J. Econ. Entomol. 96: 1514-1519.
1994. CLUSTAL W: improving the sensitivity of pro-
gressive multiple sequence alignments through se-
quence weighting, position-specific gap penalties
and weight matrix choice. Nucleic Acids Res. 22:
1959a. Coop. Econ. Insect Rept. 9, 362.
1959b. Coop. Econ. Insect Rept. 9, 393.
WEESNER, F. M. 1970. Termites of the Nearctic Region,
pp. 477-525 In K. Krishna and F. M. Weesner [eds.],
Biology of Termites, vol. II. Academic, New York.
WEESNER, F. M. 1965. The Termites of the United
States: A Handbook, Printed for the National Pest
Control Association, Elizabeth, NJ. 71 pp.

Florida Entomologist 89(3)

Seal et al.: Scirtothrips dorsalis distribution in pepper fields


'University of Florida-IFAS, Tropical Research and Education Center, Homestead, FL 33033

2USDA, APHIS, PPQ, CPHST, Pest Detection Diagnostics and Management Laboratory, 22675 N. Moorefield Rd.,
Bldg. 6414, Edinburg, TX 78541-9398

3Ministry of Agriculture and Fisheries, St. Vincent, Richmond Hill, Kingstown, St. Vincent and the Grenadines

Scirtothrips dorsalis Hood is a significant pest of various vegetable, ornamental, and fruit
crops. Its biology and management are little known in the agro-ecosystems in western hemi-
sphere. We investigated distribution patterns of S. dorsalis in fields and plants of "Scotch
Bonnet' pepper, Capsicum chinense Jacq., on St. Vincent in 2004 and 2005. Scirtothrips dor-
salis adults and larvae were abundant on top leaves of the pepper plants followed by middle
leaves, lower leaves, flowers and fruits. The spatial distribution of S. dorsalis adults and lar-
vae on pepper was analyzed by using Taylor's power law and Iwao's patchiness regression.
These results were compared with the Index of Dispersion, Mean Crowding, Green's Index
and Lloyd's Patchiness Index. In Oct 2004, the distributions of S. dorsalis adults on the top
leaves were aggregated in 24- and 48 m2-plots. In the smaller plots adults were distributed
in a regular pattern. The distribution of larvae on the top leaves was aggregated irrespective
of plot size. In Williams Farms on Mar 2005, the distribution of adults was aggregated in the
largest plots (48 m2). In all other plots, the distribution of adults was regular as described by
Taylor's power law and Iwao's patchiness regression. In Baptist Farms on March 2005, the
distribution of adults according to both models was aggregated irrespective of plot size. The
optimum number of samples from a 24 m2 plot was 9 with a precision of 40% when there
were 0.5 individuals per top leaf of'Scotch Bonnet' pepper. However if the estimated density
was 2 individuals per top leaf, 9 samples from a same sized area were sufficient at the 10%
precision level. This information is essential to the development of a scouting-based inte-
grated management program for S. dorsalis. Based on this information, incipient infesta-
tions of S. dorsalis easily can be detected by examining young top leaves.

Key Words: Scirtothrips dorsalis, spatial distribution, within plant distribution, pepper, in-
vasive alien species, Caribbean

Scirtothrips dorsalis Hood es una plaga important de varias species de hortalizas, plants
ornamentales y frutales, pero se conoce poco su biologia y manejo. En los aios 2004 y 2005
investigamos el patron de la distribuci6n en el campo de S. dorsalis en piment6n'Scotch Bon-
net" en un campo commercial en St. Vincent. Los adults y las larvas de S. dorsalis fueron mas
abundantes en las hojas terminales de'Scotch Bonnet' seguido el conteo de las hojas del cen-
tro, hojas inferior, y las parties reproductivas. El patron de distribuci6n del de adults y lar-
vas de S. dorsalis en el campo de piment6n'Scotch Bonnet' fu6 determinado usando la lay de
poder de Taylor y la regresi6n de Iwao's. Estos resultados fueron comparados con el Indice
de Dispersi6n, punto maximo significativo de afluencia o poblaci6n de estos (Mean
Crowding), Green's Index y Lloyd's Patchiness Index. La distribuci6n de adults y larvas de
S. dorsalis en las hojas terminales fu6 agregada patronn de agregaci6n) independientemente
del tamano de la parcela (6, 12, 24, y 48 m2). En las frutas, la distribuci6n de adults fue
agregada en pequenas parcelas (6 y 12 m2), pero en forma regular o al azar en las parcelas
grandes (24 y 48 m2). La distribuci6n larval fue agregada patronn de agregaci6n) en las frutas
en todas las parcelas. El numero 6ptimo de la muestra fu6 9 con un 40% de nivel de precision,
cuando hay 0.5 individuals por hoja terminales de piment6n 'Scotch Bonnet'. Entonces si el
estimado de la densidad fu6 de 2 individuos por hoja terminal, luego 9 muestras serian su-
ficientes a un nivel de precision del 10%. Fu6 observado dano econ6mico en plants con 0.5
to 2 individuos por hoja terminal.

Translation by the authors.

Florida Entomologist 89(3)

Scirtothrips dorsalis Hood is a major pest of
various vegetable crops, cotton, citrus, and other
fruit and ornamental crops in eastern Asia, Af-
rica, and Oceania (Ananthakrishnan 1993,
CABI/EPPO 1997, CABI 2003). This pest occurs
on all above-the-ground plant parts of its hosts,
and creates damaging feeding scars on them
(Chang 1995). In India, S. dorsalis is a severe
pest of chilli pepper and hence is known as the
chilli thrips (Thirumurthi et al. 1972), and in Ja-
pan as the yellow tea thrips (Toda & Komazake
2002). Among the economically important hosts
of this pest listed by Venette & Davis (2004) are
banana, bean, cashew, castor, corn, citrus, cocoa,
cotton, eggplant, grapes, kiwi, litchi, longan,
mango, melon, onion, passion fruit, peach, pea-
nut, pepper, poplar, rose, sacara, soybean, straw-
berry, sweet potato, tea, tobacco, tomato, and
wild yams (Dioscorea spp.). The Florida Nursery,
Landscape and Growers Association considers S.
dorsalis one of the thirteen most dangerous ex-
otic pest threats to their industry (FNGLA
2003). Venette & Davis (2004) indicated that the
potential geographic distribution of S. dorsalis
in North America would extend from southern
Florida north to the Canadian border, as well as
to Puerto Rico and the entire Caribbean region.
It appears that most of Latin America is suitable
for colonization by this alien invasive species.
Scirtothrips dorsalis is a vector of various viral
and bacterial diseases, including peanut bud ne-
crosis virus, chlorotic fan spot virus of peanuts,
and tomato spotted wilt virus (TSWV) (Amin et
al. 1981; Mound & Palmer 1981; Ananthakrish-
nan 1993).
Efficient detection and reliable identification
of S. dorsalis are key prerequisites for developing
practices for managing it. Various methods have
been employed by entomologists to determine the
presence ofS. dorsalis. Bagle (1993) and Gowda et
al. (1979) sampled for this pest by dislodging lar-
vae and adults from young shoots or inflores-
cences onto black cardboard and counting the re-
covered insects. Suwanbutr et al. (1992) rinsed
thrips from plant material with 70% ethanol and
counted individuals collected on a fine muslin
sieve. Takagi (1978) constructed a sticky suction
trap to monitor the flight of S. dorsalis and other
tea pests. Okada & Kudo (1982) used a similar
suction trap for monitoring flight behavior of S.
dorsalis and other thrips. Saxena et al. (1996) re-
ported that S. dorsalis adults were attracted to
white sticky traps. Adults also may be attracted to
yellowish-green, green, or yellow boards
(Tsuchiya et al. 1995). Chu et al. (2006) evaluated
the effectiveness of the non-sticky 'CC' trap illu-
minated with a light-emitting diode (Chu et al.
2003) for capturing S. dorsalis and other thrips.
Seal and Baranowski (1992) separated Thrips
palmi, a related species to S. dorsalis, from bean
leaves by washing with 70% ethanol.

During 1984-2002, USDA-APHIS inspectors at
various U.S. ports-of-entry reported S. dorsalis 89
times on imported plant materials belonging to 48
taxa (USDA 2003). Most commonly the pest was
associated with cut flowers, fruits, and vegeta-
bles. On Jul 16, 2003, T. L. Skarlinsky, a Plant
Protection and Quarantine officer, intercepted S.
dorsalis at the port of Miami, Florida, on Capsi-
cum spp. from St Vincent and the Grenadines,
West Indies. This was the first interception at a
U.S. port of this thrips on a shipment from a port
of origin in the Western Hemisphere. Skarlinsky
(2003) made a preliminary assessment of the dis-
tribution and abundance of S. dorsalis on St.Vin-
cent and found it on pepper at several sites.
St.Vincent is a volcanic island located at lati-
tude 13015' N and longitude 61012' W within the
Windward Islands in the eastern Caribbean.
Temperatures fluctuate between 18 and 32C,
the dry season extends from December through
Jun, and the rainy season lasts from Jul through
Nov. The island's average annual rainfall ranges
from about 1,500 mm on the southeast coast to
about 3,800 mm in the interior mountains. Vege-
table and fruit crops are produced year round for
domestic consumption and export.
There are no published reports on within-plant
and field distribution patterns of S. dorsalis. Such
information is essential in the development of tac-
tics and strategies for managing this pest. Begin-
ning in Oct 2004, we undertook studies on the
spatial distribution patterns of S. dorsalis adults
and larvae on St. Vincent, as part of a larger effort
to determine the pest's host range, geographical
distribution and natural enemies, and to develop
efficient methods of detection, monitoring, and
control. Here we report on the thrips' within plant
distribution on pepper and spatial distribution in
pepper fields.


Within plant and field distributions of S. dor-
salis were investigated in a field of "Scotch Bon-
net' (Habanero type) pepper, Capsicum chinense
Jacq., on Williams Farms in Oct during rainy sea-
son, 2004 (Field 1) and in Mar, 2005 (Field 2); and
on Baptist Farms (Field 3) in Mar during dry sea-
son, 2005. All fields were located at Georgetown,
St.Vincent, and each field was about 3,035 m2.
"Scotch Bonnet' pepper had been planted into the
deep soil in each field 2-3 months prior to these
studies. The plants were spaced 90 cm apart
within the row with 1.2 m between rows. Plants
were maintained by using standard cultural prac-
tices recommended for St. Vincent. The pepper
plants were not treated with insecticides but they
received the recommended fungicide and fertil-
izer applications. Plants were sprayed with man-
cozeb and chlorothalonil at 7-10 d intervals and
irrigated weekly or as necessary through drip

September 2006

Seal et al.: Scirtothrips dorsalis distribution in pepper fields

tubing. For the purpose of studying distribution
patterns of S. dorsalis, an area of 332 m2 of each
field was divided into 60 equal plots, each 4.6 m
long and 1.2 m wide, and each plot contained 5
pepper plants.

Within "Scotch Bonnet' Pepper Plant Distribution

Five plants were randomly selected from each
of five plots at different locations in all three
fields. From each plant, a set of 3-4 leaves was col-
lected from each of the top, middle, and bottom
strata. In addition, three flowers and three fruits
were collected from each plant. Thus, from each
plot 1 -20 leaves were collected from each pepper
plant stratum, and 15 flowers and 15 fruits were
collected per plot. All samples of each category
from a plot were placed in zip-lock bags each la-
beled to indicate the field, plot, plant stratum,
and plant part. Samples were transported to the
laboratory for further processing. Adults and lar-
vae in each sample were washed off the plant
parts with 70% ethanol and collected by pouring
the ethanol through a sieve (6.35 cm dia., 500
Tyler equivalent mesh; 25 micrometer opening;
USA Standard Testing Sieve; ASTME-11 Specifi-
cation, W. S. Tyler, Inc., made in USA) (Seal &
Baranowski 1992). Identifications of adult and
larval thrips were based on the morphology of
adult and larval forms and their identities were
confirmed with recent taxonomic keys (Mound &
Kibby 1998). Adults of S. dorsalis were distin-
guished from other thrips based on body trans-
parency and color, and the presence of a dark cu-
ticular thickening medially on tergites III to VII.
Tergites ofS. dorsalis adults have three discal se-
tae in the lateral microtrichial fields (Mound &
Kibby 1998). Also the forewing cilia are straight.
The larvae of S. dorsalis were separated from
those of other thrips species based on color and
size, and confirmed by observing the funnel
shaped setae on the head and abdominal segment

Within "Scotch Bonnet' Pepper Field Distribution

The within field distribution ofS. dorsalis was
studied in plots of four different sizes- 6, 12, 24,
and 48 m2. Spatial distribution of S. dorsalis in
'Scotch Bonnet' pepper fields was studied in two
years (Oct, 2004 on Williams Farms and Mar,
2005 on both Williams and Baptist Farms) by col-
lecting the terminal leaf contained in a group of 3-
4 leaves at the tip of a branch. From each of five
randomly selected plants/plot the terminal leaf
was excised and placed separately in a ziplock bag
to prevent escape of S. dorsalis. The bags were
marked with the date, plot number, and plant
number. All bags were transported to the labora-
tory for further processing by the method de-
scribed above to record S. dorsalis adults and lar-

vae in each leaf sample. Each year each field was
sampled six times following the same procedure.
The spatial distribution patterns ofS. dorsalis
were determined by using Taylor's power law
(Taylor 1961) and Iwao's patchiness regression
(Iwao 1968). Taylor's power law parameters were
obtained by the regression of log0,-transformed
variances, s2, on log0,-transformed mean number
of S. dorsalis adults and larvae per sample, i.e., by
means of the linear regression model: log s2 = log
a + b log x (Taylor 1961). According to this model
a b value > 1 denotes a population with an aggre-
gated distribution, a b value significantly < 1 de-
notes a regular distribution, and a b value not sig-
nificantly different from 1 denotes a random dis-
tribution. The fit of each data set to the linear re-
gression model was evaluated by calculating the
r2, F, df, and P values. Student's t-test was used to
determine if the slopes (b values) obtained by
means of the linear regression procedure were
equal to 1, significantly < 1, or significantly >1
(Neter & Wasserman 1974). Separate regression
equations were calculated for different sample
types and plot sizes. Variation in plot size was
achieved by pooling data from adjacent plots to
obtain a range of sizes from the smallest (6 m2) to
the total field size. In a similar manner, Iwao's
patchiness regression was calculated for each
data set. Iwao's patchiness regression (x* = a +
bx), which may be seen as parallel to Taylor's
power law, is the regression of mean crowding, x',
on the mean x (Lloyd 1967; Iwao 1968). The factor
a depends on the size of the sampling unit and P
is the index of aggregation in the population. The
fit of each model to data from various plot sizes
was determined based on the values of r2d F, df,
and P as calculated by using the General linear
model (GLM) procedure of SAS (SAS 1988). GLM
procedures were also used to perform analyses of
variance of dependent variables (log of the vari-
ance of adults in Taylor's power law, and mean
number of adults in Iwao's patchiness regression)
of data collected from the various plots.
An index of dispersion (ID) was calculated as

ID = -
Where x is the mean number of S. dorsalis indi-
viduals per sample and s2 is the sample variance.
Values of ID greater than 1.0 indicate an aggre-
gated distribution of samples. ID is distributed as
a chi-square variable with n-1 degrees of freedom
(Elliott 1977), and therefore provides a test of sig-
nificant departure from randomness (i.e., a test of
aggregation). A generalized pattern of S. dorsalis
distribution was determined from Taylor's power
law and Iwao's patchiness regression by combin-
ing the first set of data on S. dorsalis adults in ter-
minal leaves from three fields (one field in 2004;
two fields in 2005). Finally, Mean Crowding (m*),

Florida Entomologist 89(3)

Green's Index (Cx) and Lloyd's "Patchiness' Index
(m /m) were calculated by combining data on S.
dorsalis adults from terminal leaves collected in
the above mentioned three fields.

Mean Crowding, m = m + 2 -1 (i)

m = J (ii)
j = 1
In equation (ii), m = mean density; x3 = no. in-
dividuals/plot (Q); j = plots (1 Q).

x;- x, /
52 =

62 = the square of the variance.
Thus, in equation (i), mean crowding equals the
difference between the ratio of the variance to
mean density minus 1 plus the mean density it-
self. In a random distribution, the variance and
mean density are equal, so the quantity in paren-
theses disappears, and m* and m become equal.
In the instance of Lloyd's Patchiness Index, the
value of m* is divided by m (m'/m) or [m + (62/m -
1)]/m. Green's Index can be calculated as: (62/m -
1)/(x 1).
These parameters were compared with other
characterizations of the S. dorsalis distribution

Sample Size Requirements

In order to estimate the population density at
a given level of reliability, the number of samples,
n, required for a particular plot size was deter-
mined by the equation (Wilson & Room 1982):

N = c2taxb-2

Where c is the reliability (half of the width of the
confidence interval as a percentage of the mean),
a and b are the coefficients of Taylor's power law,
x is the mean density, and t is Student's t-value
determined with n 1 degrees of freedom. This t
value is approximately 2.0 when n is large. Sam-
ple sizes were determined at three levels of preci-
sion (0.10, 0.20, and 0.40) for densities of 0.5, 1.0,
and 2.0 adults or larvae per sample. These densi-
ties were selected based on the number of S. dor-
salis collected per sample during the two-year
study, and on our observation of economic damage
associated with this range of pest density.

Statistical Analysis

Data on the within plant distribution were
subjected to square root (x + 0.25) transformation
to stabilize error variance (Steel & Torrie 1980).
Transformed data were analyzed with software
provided by Statistical Analysis System (SAS
1988). General linear model procedures were
used to perform the analysis of variance. The
Waller-Duncan K ratio t test was used to separate
treatment means where significant (P < 0.05) dif-
ferences occurred (Waller & Duncan 1969).


Within Plant Distribution of S. dorsalis

We found S. dorsalis on all above-ground pep-
per plant parts in three fields during Oct 2004
(rainy season), and Mar 2005 (dry season), re-
spectively (Table 1). In Field 1 (rainy season), the
mean number of S. dorsalis adults and larvae was
greatest on the top leaves, 2nd greatest on middle
leaves and least on bottom leaves, flowers, and
fruits (adults: F = 7.77; df= 4, 15;P < 0.05; larvae:
F = 13.93; df = 4, 15; P < 0.05; total: F = 16.88; df
= 4, 15; P < 0.05). The mean number of S. dorsalis
adults and larvae did not differ statistically
among those found on the bottom leaves, flowers,
and fruits. The lowest number of adults was found
on fruits, and the fewest larvae were found in
flowers, but these means were not significantly
different from the corresponding means for bot-
tom leaves and fruits.
In Field 2 (dry season), the mean numbers of S.
dorsalis adults and larvae were also larger on the
top leaves than on other plant parts; although not
significantly larger than those on the middle
leaves (adult: F = 3.36; df = 4, 15; P < 0.05; larva:
F = 7.61; df= 4, 15; P < 0.05) (Table 1). The mean
number of adults was greater on the middle
leaves than on bottom leaves, flowers, and fruits,
but these differences were not statistically signif-
icant. The mean number of larvae was lowest in
the flowers, although not significantly lower than
on bottom leaves and fruits. The cumulative mean
number of S. dorsalis was the greatest on the top
leaves followed by middle leaves and finally by
the other plant parts (total: F = 16.88; df = 4, 15;
P < 0.05).
In Field 3 (dry season), the mean number of S.
dorsalis adults on the top leaves was significantly
larger than on any other plant part (adult: F =
4.94; df = 4, 15; P < 0.05) (Table 1). Similarly, the
mean number of larvae was the largest on the top
leaves, but not significantly larger than on the
middle leaves (larva: F = 6.45; df = 4, 15; P < 0.05).
When adult and larval data were combined, the
mean number of S. dorsalis on the top leaves was
significantly greater than on any other plant part
(total: F = 12.93; df = 4, 15; P < 0.05). Although

September 2006

Seal et al.: Scirtothrips dorsalis distribution in pepper fields


Mean number of Scirtothrips dorsalis

Location on Pepper plant Adults Larvae Total

Field 1 (Oct 2004, rainy season)
Top leaf 4.50 a 5.50 a 10.00 a
Middle leaf 1.75 b 2.00 b 3.75 b
Bottom leaf 0.50 b 0.75 c 1.25 c
Flower 0.75 b 0.25 c 1.00 c
Fruit 0.25 b 1.00 bc 1.25 c
Field 2 (Mar 2005, dry season)
Top leaf 2.25 a 4.25 a 6.50 a
Middle leaf 1.00 ab 2.25 ab 3.25 b
Bottom leaf 0.25 b 0.75 bc 1.00 c
Flower 0.50 b 0.25 c 0.75 c
Fruit 0.50 b 0.75 bc 1.25 c
Field 3 (Mar 2005, dry season)
Top leaf 3.75 a 4.00 a 7.75 a
Middle leaf 1.25 b 1.75 ab 3.00 b
Bottom leaf 0.75 b 0.50 bc 1.25 bc
Flower 0.25 b 0.25 c 0.50 c
Fruit 0.50 b 1.00 bc 1.50 bc

Means within a column for each field followed by the same letter do not differ significantly (P > 0.05, Waller-Duncan k ratio pro-
cedure (Waller & Duncan 1969)).

the number found on middle leaves was larger
than either on fruits or bottom leaves, these dif-
ferences were not statistically significant. The
least number of life forms was recorded in flowers,
but this number was not statistically different
from the number either on the bottom leaves or on
the fruits.
The within plant distribution of S. dorsalis on
pepper differs from that of Thrips palmi Karny in

that the latter is very abundant in pepper flowers
(Seal 1996, 2001).

Within Field Distribution of S. dorsalis in 2004

In 2004, for adults (Table 2) the values ofr2 ob-
tained with both Taylor's power law (r2 = 0.54 -
0.99) and Iwao's patchiness distribution (r2 = 0.45
- 0.99) were moderate to large for all plot sizes.


Taylor's power law Iwao's patchiness regression

Plot size (m2) r2 a b r2 a P

6 0.54 0.07 0.93 REG 0.45 0.31 0.93 REG
12 0.63 0.03 0.94 REG 0.62 0.19 0.92 REG
24 0.76 0.02 1.08 AGG 0.73 0.008 1.05 AGG
48 0.99 0.009 1.20 AGG 0.99 -0.28 1.29 AGG

AGG, aggregated distribution, b significantly > 1; REG, regular distribution, b significantly < 1. These distributions are signifi-
cant at P < 0.05 based on Student's t-test. Numbers of plots (n) are 48, 24, 12, and 6 for the fields sized at 6, 12, 24, and 48 m2, re-
spectively. (Taylor's Power law. 6 m2: F =54.01, df = 1, 46, P = 0.001; 12 m2: F = 38.68, df= 1, 22, P = 0.001; 24 m2: F = 32.79, df = 1,
10, P = 0.001; 48 m2: F = 94.21, df = 1, 4, P = 0.001; Iwao's patchiness regression. 6 m2: F = 37.77 df = 1, 46, P = 0.0001; 12 m2: F=
36.12, df = 1, 22, P = 0.0001; 24 m2: F = 26.52, df = 1, 10, P = 0.0004; 48 m2: F = 112.23, df = 1,4, P = 0.0005).

Florida Entomologist 89(3)

This indicates a good fit of both models to the data
on adults with top leaves as the sampling unit. In
both models, the values ofF for various plot sizes
were significant (Table 2). The distribution of
adults in the two larger plots sizes (24 and 48 m2)
was aggregated. The slope values for these two
plot sizes in either model were significantly
greater than 1.00 (P > 1.00). For larval popula-
tions on top leaves (Table 3), the r2 values ob-
tained with both Taylor's power law and Iwao's
patchiness regression showed a good fit to the
data (r = 77 99) for all of the plot sizes. Simi-
larly, the F values in both models for all plot sizes
were significant. The slope in either model was
significantly greater than 1.00 (P > 0.05) indicat-
ing that the distribution of larval populations in
all plots, irrespective of size, was aggregated.
With respect to data collected on S. dorsalis
adults in March 2005 on Williams Farms, the
analyses with from both Taylor's power law and
Iwao's patchiness regression were in agreement
that the distribution of adults on top leaves was
regular irrespective of plot size (Table 4). The val-
ues ofr2 from Taylor's power law ranged from 0.21

to 0.99, indicating moderate to good fit to the data
collected from 6, 12, and 48 m2 plots. The values of
r2 and F (r2 = 0.09; F = 0.95, df = 1, 10, P = 0.35)
were low for data from 24 m2-plot. The value of r2
(r2 = 0.07) from Iwao's patchiness regression also
was low for 24 m2-plot (indicating poor fit to the
data). The F value (F = 0.74, df= 1, 10, P = 0.409)
for the corresponding data set was insignificant.
The values ofr2 ranged from 0.24 to 0.99 for the 6,
12, and 48 m2 plots (indicating a moderate to good
fit to the data). The values ofF calculated for the
data of these plots were significant.
The distribution patterns of S. dorsalis adults
in 2004 was regular in smaller plots (6- and 12 m2-
plots) and aggregated in larger plots (24- and 48
m2-plots). However, in 2005 the distribution of
adults was regular irrespective of plot size. The
distribution pattern of S. dorsalis larvae was ag-
gregated in all plot sizes. Both Taylor's power law
and Iwao's patchiness regression were in agree-
ment in describing the distribution of S. dorsalis
adults in pepper fields.
On Baptist Farms (Table 5) the adult distribu-
tion on top leaves based on Taylor's power law was


Taylor's power law Iwao's patchiness regression
Plot size (m2) 2 ab r2
r2 a b r2 a P

6 0.59 -1.49 2.52 AGG 0.89 -13.07 1.65 AGG
12 0.76 -2.63 3.33 AGG 0.92 -19.06 1.88 AGG
24 0.98 -3.74 4.11 AGG 0.99 -19.48 1.90 AGG
48 0.99 -5.64 5.49 AGG 0.99 -29.87 2.30 AGG

AGG, aggregated distribution, b significantly > 1. These distributions are significant atP < 0.05 based on Student's t-test. Num-
bers of plots (n) are 48, 24, 12, and 6 for fields sized at 6, 12, 24, and 48 m2, respectively. (Taylor's Power law. 6 m2: F =66.30, df = 1,
46, P = 0.001; 12 m2: F = 70.07, df= 1, 22, P = 0.001; 24 m2: F = 73.11, df = 1, 10, P = 0.001; 48 m2: F = 94.21, df = 1, 4, P = 0.001;
Iwao's patchiness regression. 6 m2: F = 2.77 df = 1, 46, P = 0.0001; 12 m2: F = 247.7, df = 1, 22, P = 0.0001; 24 m2: F = 390.90, df=
1,10, P = 0.0001; 48 m2: F = 332.23, df = 1, 4, P = 0.0001).


Taylor's power law Iwao's patchiness regression

Plot size (m2) r2 a b r2 a P

6 0.21 0.16 0.54 REG 0.24 1.00 0.55 REG
12 0.40 0.12 0.64 REG 0.51 0.65 0.70 REG
24 0.09 0.47 -1.47 REG 0.07 3.40 -1.21 REG
48 0.99 -0.55 5.18 AGG 0.99 -4.87 4.57 AGG

REG, regular distribution, b significantly < 1; RAN, random distribution, b not significantly different from 1. These distributions
are significant at P < 0.05 based on Student's t-test. Numbers of plots (n) are 48, 24, 12, and 6, and 48 for fields sized at 6, 12, 24,
and 48 m2, respectively. (Taylor's Power law. 6 m2: F =12.42, df = 1, 46, P = 0.0001; 12 m2: F = 14.65, df = 1, 22, P = 0.0001; 24 m2: F
= 0.95, df = 1, P = 0.35; 48 m':F = 99.21, df = 1, 4,P = 0.0001; Iwao's patchiness regression. 6 m2: F = 14.88, df= 1, 46, P = 0.0001;
12 m: F = 23.30, df = 1, 22, P = 0.0001; 24 m': F = 0.74, df = 1, 10, P = 0.409; 48 m': F = 211.30, df= 1, 4, P = 0.0001).

September 2006

Seal et al.: Scirtothrips dorsalis distribution in pepper fields


Taylor's power law Iwao's patchiness regression
r2 a b r2 a P

6 0.60 -0.25 2.10 AGG 0.64 -0.96 1.71 AGG
12 0.64 -0.32 2.57 AGG 0.57 -1.16 1.87 AGG
24 0.92 -0.50 3.33 AGG 0.98 -2.89 2.81 AGG
48 0.98 -0.63 4.00 AGG 0.99 -3.57 3.23 AGG

AGG, aggregated distribution, b significantly > 1; REG, regular distribution, b significantly < 1. These distributions are signifi-
cant at P < 0.05 based on Student's t-test. Numbers of plots (A) are 48, 24, 12, and 6 for fields sized at 6, 12, 24, and 48 m2, respec-
tively. (Taylor's Power law. 6 m2: F =68.72, df = 1, 46, P = 0.0001; 12 m2: F = 38.32, df = 1, 22, P = 0.0001; 24 m2: F = 109.18, df = 1,
10, P = 0.0001; 48 m2: F = 333.56, df = 1, 4, P = 0.0001; Iwao's patchiness regression. 6 m2: F = 80.67, df = 1, 46, P = 0.0001; 12 m2:
F = 28.83, df = 1, 22, P = 0.0001; 24 m2: F = 471.28, df = 1, 10, P = 0.0001; 48 m2: F = 615.17, df= 1, 4, P = 0.0001).

regular in the smallest plots (6 m2) and aggre-
gated (b > 1.00; P < 0.05) in all of the larger plots.
However Iwao's patchiness regression showed an
aggregated adult distribution in all plot sizes in-
cluding the smallest. The high r2 values (r2 = 0.61
- 0.99) indicated a good fit of the data from all plot
sizes to both Taylor's power law and Iwao's patch-
iness regression. The F values calculated from the
data of all plot sizes in both models were signifi-

General Pattern of Distribution of S. dorsalis Adults

When the cumulative data on adults from
three fields, two on Williams Farms (Field 1: Oct
2004, and Field 2: Mar 2005), and one on Baptist
Farms (Field 3: Mar 2005) were considered, the
distribution ofS. dorsalis adults on top leaves was
aggregated irrespective of plot size (Table 6). The

value of b ranged from 1.10 to 1.63 for Taylor's
power law and from 1.28 to 1.82 for Iwao's patch-
iness regression. Hence both methods consis-
tently gave slope values corresponding to an ag-
gregated distribution from cumulative data in-
volving one set of samples collected in the rainy
season and two collected in the dry season.
The values of Index of Dispersion, Mean
Crowding and Lloyd's Patchiness Index were >1
in all plots indicating an aggregated pattern of
distribution of S. dorsalis adults in top leaves (Ta-
ble 7). Green's Index did not fit the data on adults
on terminal leaves and hence showed negative
values for plots of all sizes. Both Taylor's power
law and Iwao's patchiness regression were in
agreement with Index of Dispersion, Mean
Crowding and Lloyd's Patchiness Index in denot-
ing an aggregated pattern of distribution of S.
dorsalis adults on top leaves.


Plot size (m2) n 2 Equation

Taylor's power law
48 6 0.83 log s2 = -0.02 + 1.63 log x
24 12 0.73 log s2= -0.01 + 1.51 logx
12 24 0.56 log s2 = 0.02 + 1.22 log x
6 48 0.43 log s2= 0.04 + 1.10 log x

Iwao's patchiness regression

48 6 0.89 x = -0.95 + 1.82 x
24 12 0.82 x = -0.88 + 1.77 x
12 24 0.59 x = -0.27 + 1.36 x
6 48 0.56 x"= -0.10 + 1.28 x

In the table n is the number of samples, r2 is the proportion of the sum of squares accounted for by the regression, x represents
the mean of the samples, s2 is the variance, and x' is the mean crowding index. Slopes (b values) of all equations are significantly
different from 1.0 (P > 0.05), indicating aggregated distributions of Scirtothrips dorsalis adults on terminal leaves.

Florida Entomologist 89(3)

FARMS (FIELD 3, MAR 2005).

Plot size (m2) Index of Dispersion Mean Crowding Green's Index Lloyd's Mean Crowding

6 1.23 1.30 0.02 1.24
12 1.18 1.25 0.01 1.19
24 1.16 1.24 0.004 1.17
48 1.23 1.30 0.02 1.01

Value > 1.0 indicates an aggregated pattern of distribution.

Southwood (1978) observed that when a popu-
lation in an area becomes sparse, the chances of
an individual occurring in any sample unit are so
low that the distribution is effectively random. In
the present study, the overall abundance ofS. dor-
salis was low but frequent occurrence in various
samples indicated an aggregated pattern. Such
aggregated distribution in the field is also typical
of T palmi (Seal 1996; Seal & Stansly 2000).
Southwood (1978) also reported that the disper-
sion of the initial insect invaders of a crop is often
random. We found that S. dorsalis adults were lo-
calized in certain parts of the crop field. The infes-
tation started along 6 m wide strip at one edge of
one field and proceeded with the prevailing wind
from south to north along a roughly 6 m-wide

Number of Leaf Samples Necessary for Reliably Esti-
mating S. dorsalis.

Mean densities of S. dorsalis adults ranged
from 0.5 to 1.0 per top leaf sample in fields that
suffered economic damage (Seal & Ciomperlik,
field observation). This information was used to

determine the optimum sample size (OSS) for es-
timating densities of S. dorsalis adults in pepper
fields based on the method of Wilson & Room
(1982). The OSS becomes large when the desired
level of precision is high (i.e., within 10% of the
mean) and when the average density of insects
per sample is low. We found that given an average
adult thrips density of 0.5 per sample and a 10%
precision level, the number of leaf samples re-
quired to estimate S. dorsalis abundance in 24 m2
is 140 (Table 8). It would be prohibitively labor in-
tensive to collect such a large number of samples
from a small pepper field. By relaxing the preci-
sion level to 40% in this example, the number of
samples required to obtain an estimate of the pop-
ulation density can be reduced to 9 for the same
size field (24 m2), and this is practical. The popu-
lation density of S. dorsalis strongly affects the
number of samples required. Thus, if the density
were assumed to be 2 adults per sample, then
number of samples needed for an estimate at the
10% precision level would be 9.
The following may be concluded: (1) S. dorsalis
populations maintained pest status in dry and
rainy seasons alike. (2) S. dorsalis populations


Sample size* at the following levels of precision

Field size (m2) Mean adults per sample 0.10 0.20 0.40

6 0.5 94 23 6
6 1.0 36 9 2
6 2.0 14 4 1

12 0.5 93 25 6
12 1.0 32 8 2
12 2.0 10 3 1

24 0.5 140 35 9
24 1.0 36 9 2
24 2.0 9 2 1

*Sample size values each rounded to the nearest whole number.

September 2006

Seal et al.: Scirtothrips dorsalis distribution in pepper fields

tend to be most abundant on top leaves; (3) S. dor-
salis populations tend to be aggregated irrespec-
tive of plot size; (4) the optimum sample size in
"Scotch Bonnet' pepper plots, when the estimated
population density is 2.0 per top leaf sample, is 9
with a 10% precision level. The above information
should be considered in developing management
programs against S. dorsalis for various host


We are very grateful to the Plant Quarantine Divi-
sion, Ministry of Agriculture, Industry and Labour,
Kingstown, Saint Vincent, and the Grenadines for the
use of laboratory facilities, local transportation, and ar-
rangements with growers. This study could not have
been accomplished without the facilitation and encour-
agement of Mr. Philmore Isaacs, Chief Agricultural Of-
ficer. We are grateful to Mr. Emil Williams and Mr.
Lauron Baptist for allowing us to conduct the studies on
their farms. Financial resources and guidance were pro-
vided by the Animal and Plant Health Inspection Ser-
vice, USDA through the leadership of Dr. Daniel A.
Fieselmann, National Science Program Leader and Ms.
Carolyn T Cohen, Caribbean Area Director. In addition,
financial support was provided by the Florida Agricul-
tural Experiment Station and the University of Flor-
ida's Center for Tropical Agriculture. We are grateful to
Dr. Jorge E. Pena for translating the abstract into Span-
ish. Ms. Catherine Sabines provided technical support.


1981. Transmission of tomato spotted wilt virus, the
causal agent of bud necrosis of peanut, by Scirto-
thrips dorsalis and Frankliniella schultzei. Plant
Disease 65: 663-665.
ANANTHAKRISHNAN, T. N. 1993. Bionomics of thrips.
Annu. Rev. Entomol. 38: 71-92.
BAGLE, B. G. 1993. Seasonal incidence and control of
Scirtothrips dorsalis Hood in pomegranate. Indian J.
Entomol. 55: 148-153.
CABI. 2003. Crop protection compendium: global mod-
ule. CAB International, Wallingford, UK.
CABI/EPPO. 1997. Quarantine Pests for Europe, 2nd Ed.
CAB International, Wallingford, UK.
CHANG, N. T. 1995. Major pest thrips in Taiwan. In T.
Lewis [ed.], Thrips Biology and Management. Ple-
num Press, New York.
AND T. J. HENNEBERRY. 2003. Plastic cup traps
equipped with light-emitting diodes for monitoring
adult Bemisia tabaci (Homoptera: Aleyrodidae). J.
Econ. Entomol. 96: 543-546.
ARDS, AND T. J. HENNEBERRY. 2006. Developing and
evaluating traps for monitoring Scirtothrips dorsalis
(Thysanoptera: Thripidae). Florida Entomol. 89 (1):
ELLIOTT, J. M. 1977. Some Methods for the Statistical
Analysis of Samples of Benthic Invertebrates, 2nd
ed. Freshwater Biol. Assoc. Sci. Publ. 25.
ASSOCIATION). 2003. The unlucky 13. Report of Ma-

jor Nursery Pest & Disease Identification Task
Force. Florida Nursery, Growers and landscape As-
sociation, Orlando, Florida, USDA. 1 p.
GOWDA, G., E. RAMAIAH, AND C. V. K. REDDY. 1979. Scir-
tothrips dorsalis (Hood) (Thysanoptera; Terebrantia:
Thripidae) a new pest on cashew (Anacardium occi-
dentale L). Current Res. 8: 116-117.
IWAO, S. 1968. A new regression model for analyzing the
aggregation pattern of animal populations. Res.
Popul. Ecol. (Kyoto) 4: 35-46.
LLOYD, M. 1967. Mean crowding. J. Anim. Ecol. 36: 1-30.
MOUND, L. A., AND J. M. PALMER. 1981. Identification,
distribution and host plants of the pest species of
Scirtothrips (Thysanoptera: Thripidae). Bull. Ento-
mol. Res. 71: 467-479.
MOUND, L. A., AND G. KIBBY. 1998. Thysanoptera, An
Identification Guide. 2nd ed. CAB International,
Wallingford, United Kingdom.
NETER, J., AND W. WASSERMAN. 1974. Applied Linear
Statistical Models. Irwin, Homewood, IL.
OKADA, T., AND I. KUDO. 1982. Relative abundance and
phenology of Thysanoptera in tea fields. Japanese J.
Appl. Entomol. Zool. 26: 96-102.
SAS INSTITUTE. 1988. SAS/STAT User's Guide, release
6.0. SAS Institute, Cary, NC.
AND U. RAIZADA. 1996. Pollination and gene flow in
chillies with Scirtothrips dorsalis as pollen vectors.
Phytomorphology 46: 317-327.
SEAL, D. R. 1996. Management and biology of Thrips
palmi Karny, pp. 161-181 In K. Bondari [ed.], New
Developments in Entomology. Research Signpost, T
C. 36/248 (1), Trivandrum-695 008, India.
SEAL, D. R. 2001. Seasonal abundance and distribution
of Thrips palmi Karny (Thysanoptera: Thripidae) in
southern Florida. Proc. Florida Hort. Soc. 114: 337-
SEAL, D. R., AND R. M. BARANOWSKI. 1992. Effective-
ness of different insecticides for the control of melon
thrips, Thrips palmi Karny (Thysanoptera: Thripi-
dae) affecting vegetables in south Florida. Proc.
Florida State Hort. Soc. 105: 315-319.
SEAL, D. R., AND P. A. STANSLY. 2000. Seasonal abun-
dance and within plant distribution of melon thrips
(Thysanoptera: Thripidae) on beans in southern
Florida. Proc. Florida Hort. Soc. 113: 201-205.
SKARLINSKY, T. L. 2003. Presence of Scirtothrips dorsa-
lis Hood in St. Vincent. A short report submitted to
SOUTHWOOD, T. R. E. 1978. Ecological Methods with
Particular Reference to the Study of Insect Popula-
tions, 2nd ed. Wiley, New York.
STEEL, R. G. D., AND J. H. TORRIE. 1980. Principles and
Procedures of Statistics: A Biometrical Approach.
McGraw-Hill, New York.
1992. A field trial on the efficacy of some insecticides
for controlling thrips attacking pummelo. Proceed-
ings of the International Symposium on Tropical
Fruit: Frontier in Tropical Fruit Research. Pattaya
City (Thailand); 20-24 May 1991. Acta Horticulturae
TAKAGI, K. 1978. Trap for monitoring adult parasites of
the tea pest. Japanese Agricultural Res. Quart. 12:
TAYLOR, L. R. 1961. Aggregation, variance and the
mean. Nature (London) 189: 732-735.

320 Florida Ento

1972. A note on the varietal incidence of grapevine
(Vitis vinifera) berry thrips (Scirtothrips dorsalis
Hood). South Indian Horticulture 20: 92-93.
TODA, S., AND S. KOMAZAKE. 2002. Identification of
thrips species (Thysanoptera: Thripidae) on Japanese
fruit trees by polymerase chain reaction and restric-
tion fragment length polymorphism of the ribosomal
ITS2 region. Bull. Entomol. Res. 92: 359-363.
attraction of yellow tea thrips (Scirtothrips dorsalis
Hood). Japanese J. Appl. Entomol. Zool. 39: 299-303.
USDA. 2003. Port Information Network (PIN-309):
Quarantine status database. US Department of Ag-


ologist 89(3) September 2006

riculture, Animal and Plant Health Inspection Ser-
vice, Plant Protection and Quarantine, Riverdale,
MD, USA. (Restricted access database.).
VENETTE, R. C., AND E. E. DAVIS. 2004. Chilli Thrips/
Yellow Thrips, Scirtothrips dorsalis Hood (Thysan-
optera: Thripidae) Mini Pest Risk Assessment. Univ.
of Minnesota, St. Paul, MN. 31 pp.
WALLER, R. A., AND D. B. DUNCAN. 1969. A Bayes rule
for the symmetric multiple comparison problem. J.
American Stat. Assoc. 64: 1484-1503.
WILSON, I. T., AND P. M. ROOM. 1982. The relative effi-
ciency and reliability of three methods for sampling
arthropods in Australian cotton fields. J. Australian.
Entomol. Soc. 21: 175-181.

Patel et al.: Rice Stink Bug Injury to Rice


Department of Entomology, Louisiana Agricultural Experiment Station, Louisiana State University AgCenter,
Baton Rouge, LA 70803
E-mail: mstout@agctr.lsu.edu


Greenhouse experiments were conducted to evaluate the effects of panicle age on quantita-
tive and qualitative injury caused by rice stink bug, Oebalus pugnax pugnax (Fab.), infesta-
tions on rice, Oryza sativa L. The effects were measured at two infestation levels (one and
two bugs per panicle) and compared with an undamaged control. Percentage of empty grains
and average weight of filled grains (quantitative injury) and percentage of pecky rice (qual-
itative injury) were evaluated at grain maturity. Regardless of infestation level, insect feed-
ing during anthesis and the early milk stage of grain development (first 8 d after anthesis)
caused substantially higher numbers of empty grains than feeding during later grain devel-
opment and the control. Average grain weights were lower in infestations during anthesis
and milk stage and higher in infestations during later grain development and the control.
Pecky rice was significantly higher during late milk and soft dough stages, 9-16 d after an-
thesis, compared with remaining stages of grain development and the control. Injury was
greater in the experiment in which panicles were infested with two bugs. Pecky rice was as-
sociated with highly significant reductions in germination of the grains. The data suggest
that rice is most vulnerable to rice stink bug injury during the first two weeks after anthesis,
and that the major effects of stink bug feeding change as panicles age.

Key Words: Oebalus pugnax pugnax, Oryza sativa, rice stink bug, injury, anthesis, rice pan-
icle age


Experimentos en invernadero fueron realizados para evaluar los efectos de la edad de la pa-
nicula en danos cualitativos y cuantitativos causados por la chinche hedionda del arroz., Oe-
balus pugnax (Fab.), en arroz, Oryza sativa L. Estos efectos fueron medidos a dos diferentes
niveles de infestaci6n (uno y dos insects por panicula) y comparados con un testigo sin dano.
El porcentaje de granos vacios y peso medio de los granos llenos (dano cuantitativo) y el por-
centaje de arroz picado (dano cualitativo) fueron evaluados durante la madurez del grano.
Independientemente del nivel de infestaci6n, danos causados por la alimentaci6n de los in-
sectos durante la antesis y el period inicial de desarrollo de grano lechoso primeross 8 dias
despu6s de la ocurrencia de la antesis) resultaron en un mayor numero de granos vacios que
durante estadios de desarrollo del grano mas avanzados y el testigo. El peso medio del grano
fue mas bajo cuando las infestaciones se produjeron durante la ocurrencia de la antesis y el
grano lechoso y mas alto en infestaciones durante estadios de desarrollo mas avanzados y en
el testigo. La ocurrencia de grano picado fue significativamente mas alta durante los esta-
dios de desarrollo de grano lechoso avanzado y masa blanda (9-16 dias despu6s de la ocu-
rrencia de la antesis), en comparaci6n con los otros estadios de desarrollo y el testigo. Danos
fueron mayores en el experiment en el cual las paniculas fueron infestadas por dos insects.
Arroz picado result en reducciones de la germinaci6n del grano muy significativas. Estos
datos sugieren que el arroz es mas vulnerable a danos causados por la chinche hedionda del
arroz durante las primeras dos semanas despu6s de la ocurrencia de la antesis, pero los efec-
tos de la alimentaci6n de las chinches cambian con la edad de las paniculas.

Translation by the authors.

The rice stink bug (Hemiptera: Pentatomidae), ductive phases of growth, in particular during
Oebalus pugnax pugnax (Fab.), is one of the most grain development (Douglas 1939; Rashid et al.
injurious pests of rice in the southern United 2006). Both adults and nymphs feed on develop-
States (McPherson & McPherson 2000). It is com- ing grains (Naresh & Smith 1983). Feeding re-
mon in the United States east of the Rocky Moun- sults in losses in yield and reduction in grain
tains and as far north as Minnesota and New York quality (Odglen & Warren 1962; Swanson & New-
(Sailer 1944). It is attracted to rice during repro- som 1962; Bowling 1963). The entire contents of

Florida Entomologist 89(3)

the rice grain may be removed during the milk
stage, resulting in false grains, or a portion of the
content may be sucked out, resulting in atrophied
grains (Swanson & Newsom 1962). Feeding dur-
ing soft and hard dough stages leaves a chalky
discolored area around the feeding site and rice so
affected is called pecky rice (Harper et al. 1993).
Fungi often enter the punctures made by rice
stink bug (Hollay et al. 1987; Lee et al. 1993).
Pecky rice easily breaks during milling, lowering
the percentage of whole grains and, thus, the
market value of the product (Odglen & Warren
1962). If pecky rice does not break during milling,
it will appear in head rice, resulting in inferior
quality of rice (Harper et al. 1993). For a brown
rice sample to qualify as US #1 or US #2, it should
contain no more than 1 or 2% pecky rice, respec-
tively (Fryar et al. 1986). Feeding also results in
losses due to reduced viability of the grain (Swan-
son & Newsom 1962).
Little effort has been made to develop
nonchemical controls for rice stink bug for several
reasons, including the short period of rice plant
susceptibility (heading to harvest, approximately
30 d for most varieties), the high mobility of the
bug, the low economic thresholds, and the rela-
tively low cost of chemical controls (Way 1990).
However, several of the chemical pesticides used
for controlling stink bugs may be removed in the
future due to label revision, to cancellation be-
cause of environmental and human safety con-
cerns, or to costs of the registration process (Todd
et al. 1994; McPherson & McPherson 2000). Thus,
investigations of alternate methods of control are
The susceptibility of a crop plant to injury from
pest insects usually varies with the stage of the
crop. In rice, although numerous attempts have
been made to quantify the relationship between
densities of rice stink bugs and injury (Odglen &
Warren 1962; Swanson & Newsom 1962; Bowling
1963; Robinson et al. 1980; Harper et al. 1993;
McPherson & McPherson 2000; Rashid 2003),
surprisingly few studies have attempted to quan-
tify the changes in rice susceptibility that occur as
rice panicles age and grains mature. Previous
studies that have investigated the influence of
panicle age or grain stage on susceptibility to
stink bug injury were conducted in the field,
where the presence of parasites, pathogens, and
weeds might have influenced results (Odglen &
Warren 1962; Rashid 2003; Tindall et al. 2004).
The objective of this study was to evaluate the ef-
fects of panicle age and grain maturity on the
quantitative and qualitative injury caused by rice
stink bug feeding in a controlled environment on
panicles of the rice variety 'Cocodrie', a recently-
released and widely-planted variety. Effects were
measured for two infestation levels of rice stink
bug. Effects on germination of infested grains also
were evaluated.


Qualitative and Quantitative Injury

Experiments were conducted during the sum-
mer of 2003 in a greenhouse on the campus of
Louisiana State University, Baton Rouge, LA.
Rice (Oryza sativa cv 'Cocodrie') was planted in
pots and grown in the greenhouse from Mar to
Jul. Rice for the first experiment was planted on
Mar 19 and for the second experiment on Mar 25.
Pots were 7" inches in height and 7" in diameter.
Growth medium was a mixture composed of 4
parts soil: 2 parts peat moss: 1 part sand: 1 part
vermiculite. Each pot was supplied with approxi-
mately 3.5 g of 23:12:12 NPK fertilizer at plant-
ing. Plants were watered as needed during the ex-
periments. Natural lighting was the only source
of light. Temperature ranged from 25 to 35C in
the greenhouse throughout these experiments.
Rice stink bugs were collected from heading or
headed rice as well as barnyard grass at the LSU
AgCenter Rice Research Station, Crowley, LA.
Bugs were maintained on panicles of barnyard
grass in the laboratory for approximately 2 d and
only bugs showing no signs of disease or damage
were used in experiments.
Experiments were initiated by tagging a large
number of panicles at the anthesis stage (approx-
imately 1 d after initial emergence of panicle) on
Jun 9 (experiment one) and Jun 13 (experiment
two). Panicles were randomly assigned to the fol-
lowing treatments: infestation at 1, 5, 9, 13, 17,
and 21 d after anthesis. In the first experiment,
each panicle was infested with one female rice
stink bug at the appropriate day for 4 d. In the
second experiment, panicles were infested with
two female rice stink bugs per panicle at the ap-
propriate day for 4 d. In both experiments, bugs
were placed inside muslin cloth sleeves enclosing
a rice panicle and tied at the bottom. Panicles
serving as controls were enclosed by muslin cloth
without stink bugs. Bugs were removed from the
muslin cloths after 4 d and the muslin cloth was
again put back on the panicle until harvest.
Treatments were arranged in a completely ran-
domized design with 18 replications in the first
experiment (one bug per panicle) and 10 replica-
tions in the second experiment (two bugs per pan-
Rice panicles were in the anthesis stage
(Counce et al. 2000) during approximately the
first 4 d after tagging (Patel, personal observa-
tion). Panicles then advanced into the milk stage
(Counce et al. 2000) approximately 5 to 12 d after
tagging. The soft dough (Counce et al. 2000) stage
ran approximately from 13 to 17 d after tagging
and then gradually progressed into the hard
dough stage.
Panicles were gently harvested by hand at ma-
turity and individually placed in plastic Ziploc

September 2006

Patel et al.: Rice Stink Bug Injury to Rice

bags. All panicles were taken out of the Ziploc
bags and air-dried on a lab bench at room temper-
ature for one week. Panicles were then individu-
ally threshed by hand. Filled and empty grains
were separated manually; partially filled grains
were counted as filled. The numbers of empty and
filled grains per panicle were counted and the
data were used to calculate the percentages of
empty and filled grains in each treatment. Total
weight of the filled grains also was determined.
The weight and number of the filled grains per
sample were used to determine the average
weight of a filled grain per treatment. Hulls were
removed mechanically from the rough rice sam-
ples by a McGill Sheller (H. T McGill Inc., Hous-
ton, TX). The resultant samples were separated
visually into pecky vs. nonpecky rice and then
weighed separately. All chalky discolored grains
were classified as "pecky." Weights of pecky and
pecky plus nonpecky rice were used to calculate
the percentage of pecky rice for each treatment
(time of infestation).

Effects on Germination

Pecky and nonpecky grains from the experi-
ment in which one bug per panicle was used to in-
jure rice were used for the germination experi-
ment. Grains were included from panicles in-
fested 1, 9, and 17 d after anthesis as well as those
from the control. The effects of rice quality (pecky
vs. nonpecky), time of infestation (1, 9, or 17 d af-
ter anthesis), and their interaction were tested in
this experiment. For each of the eight treatment x
time combinations, five replicates of 20 grains
were placed in a 5 x 4 matrix in 100 mm x 15 mm
sterile Petri dish (BD FalconTM, BD Biosciences,
Franklin Lakes, NJ), lined with three layers of
germination paper (Anchor Paper Co., St. Paul,
MN) saturated with 8 ml distilled water. Grains
were treated with Quadris 2.08 SC (Syngenta
Crop Protection, Greensboro, NC), a fungicide,
and covered with two layers of Kimwipe tissue pa-
per to ensure uniform hydration. Closed dishes
were incubated at 100% relative humidity for 14 d
at 30 C in darkness. Radical emergence was the
criterion for germination. The number of grains
germinated during the 14 d was recorded for each
Petri dish.

Data Analysis

Data from the two experiments with different
infestation levels were analyzed separately. Data
on quantitative (percentage of empty grains and
average weight of filled grains) as well as qualita-
tive injury (percentage of pecky rice) were ana-
lyzed by multivariate analysis of variance with
the MANOVA statement and Wilks' Lambda sta-
tistic in PROC GLM of SAS (SAS Institute 1996).
This statistic tested the null hypothesis of no

overall significant treatment (time of infestation)
effect on all three response variables. Correla-
tions among response variables were assessed
with Pearson correlation coefficients produced by
PROC CORR of SAS (SAS Institute 1996). Then,
each of these response variables was individually
subjected to analysis of variance by PROC GLM
and the Tukey HSD test for means separation
(SAS Institute 1996). Germination data were sub-
jected to two-way analysis of variance and were
analyzed with PROC GLM of SAS (SAS Institute


MANOVA Procedure and Pearson Correlation Coeffi-

The multivariate analysis suggested that
treatment (time of infestation) had an overall sig-
nificant effect on the response variables (percent-
age of empty grains, average weight of filled
grains, and percentage of pecky rice) in both ex-
periments (Wilks' Lambda, one rice stink bug per
panicle: F,,,332 = 36.47, P < 0.001; Wilks' Lambda,
two rice stink bugs per panicle: F,, 174 = 61.71, P <
0.001). Pearson correlation coefficients revealed
that only the percentage of empty grains and av-
erage weight of filled grains were significantly
correlated with each other. This correlation was
stronger at the higher infestation level (r = -
0.5245, P < 0.001 [one bug/panicle], r = 0.7548,P
< 0.001 [two bugs/panicle]).

Percentage of Empty Grains

The percentage of empty grains in panicles de-
creased as time of infestation after anthesis in-
creased in both experiments (one rice stink bug
per panicle: F, 119 = 31.25, P < 0.001, Fig. 1; two
rice stink bugs per panicle:F6,63= 81.11, P < 0.001,
Fig. 1). In both experiments, the percentage of
empty grains was statistically greater in panicles
infested 1 d after anthesis compared with that in
panicles infested during later grain development
and panicles in the control. Regardless of infesta-
tion level, the percentage of empty grains in pan-
icles infested 1 d after anthesis was approxi-
mately 2 times greater than the percentage in
panicles infested 9 d after anthesis. In both exper-
iments, infestation of panicles for 4 d beginning 1
and 5 d after anthesis produced greater percent-
ages of empty grains compared with panicles in-
fested 13, 17, and 21 d after anthesis and panicles
in the undamaged control. In the two bugs per
panicle experiment, panicles infested 9 d after an-
thesis also produced a greater percentage of
empty grains than panicles infested during later
grain development and panicles in the control. In-
festation of panicles 13, 17, and 21 d after anthe-
sis did not produce any significant reductions in

Florida Entomologist 89(3)



1 5 t 13 17
TIme f IArstMlon (Das ARn ArfifIwts
Fig. 1. Mean percentage (SE) of empt
rice panicles infested for a period of 4 d be
9, 13, 17, or 21 d after anthesis and in pani
untreated control (UTC). Two bars at eaci
time represent data from two experiments
tion levels of one or two rice stink bugs (RS
cle. Means within each infestation level
same lower or upper case letter did not d
cantly at a = 0.05 (Tukey, HSD).

the percentage of filled grains compare
control in either experiment. Panicles
trol averaged 6-7% empty grains in the
iments. Feeding by two rice stink bug
at least 1.5 times as many empty gral
ing by one bug in panicles infested 1, 5
ter anthesis.

Average Weight of Filled Grains

Treatment significantly affected t
weights of filled grains in rice panic
with one rice stink bug per panicle (F6,
< 0.001, Fig. 2) as well as two rice stir


panicle (F6,,, 6= 33.86, P < 0.001, Fig. 2). Average
weights generally increased with the time of in-
festation after anthesis in both experiments. In

the one rice stink bug per panicle experiment,
panicles infested 1 and 5 d after anthesis had
lower average weights compared with panicles in-
fested 21 d after anthesis and panicles in the con-
trol. In the two rice stink bugs per panicle exper-
iment, panicles infested 1 d after anthesis had
Slower average weights compared with panicles in-
fested 13, 17, and 21 d after anthesis and panicles
2' in the control. In the same experiment, panicles
infested 5 d after anthesis had lower average
;y kernels in weights compared with panicles infested during
ginning 1, 5, later grain development and panicles in the con-
les from the trol. When infested with one rice stink bug per
h infestation panicle, there were reductions of 8% and 10% in
with infesta-
wit infes pa- average weights in panicles infested 1 and 5 d af-
B) per pani-
followed by ter anthesis, respectively, compared with the con-
liffer signifi- trol; these reductions were 10% and 11% with two
rice stink bugs. This result suggests that feeding
during the anthesis, milk, and soft dough stages
of grain development reduced the average
ed with the weights of filled grains, with more injury during
in the con- early milk stage. Reductions in average weights
two exper- were high in panicles infested for 4 d beginning 1,
s produced 5, and 9 d after anthesis and low thereafter, dem-
ins as feed- onstrating that the first 12 d after anthesis were
and 9 d af- the most critical for injury in terms of reduced
,and 9 d af-
grain weight.

Percentage of Pecky Rice

he average Pecky rice as a percentage of the total weight of
es infested the de-hulled grains in each rice panicle is shown
119= 6.45,P in Fig. 3. In both experiments, controls had ap-
ik bugs per proximately 3% pecky rice. This result indicated

*t.--5 35

jo a. s

F scr
10t l b b
*r CI
ceO di~

1 I g 13 17 21 UTC
Tira ol In Minlt n (ODys ARe ntnl.)

Fig. 2. Average weight (g) of filled kernels (SE) in
rice panicles infested for a period of 4 d beginning 1, 5,
9, 13, 17, or 21 d after anthesis and in panicles from the
untreated control (UTC). Two bars at each infestation
time represent data from two experiments with infesta-
tion levels of one or two rice stink bugs (RSB) per pani-
cle. Means within each infestation level followed by
same lower or upper case letter did not differ signifi-
cantly at a = 0.05 (Tukey, HSD).

1 B 1 Is 47 12 UTC
Ti e of Itstallon (Days ARM AniteMst

Fig. 3. Mean percentage (SE) of pecky rice in rice
panicles infested for a period of 4 d beginning 1, 5, 9, 13,
17, or 21 d after anthesis and in panicles from the un-
treated control (UTC). Two bars at each infestation time
represent data from two experiments with infestation
levels of one or two rice stink bugs (RSB) per panicle.
Means within each infestation level followed by same
lower or upper case letter did not differ significantly at
a = 0.05 (Tukey, HSD).

September 2006

Patel et al.: Rice Stink Bug Injury to Rice

that pecky rice was caused by factors in addition
to rice stink bug. The percentage of pecky rice in
panicles differed with the time of infestation after
anthesis in both experiments: (one rice stink bug
per panicle: F6,119 = 138.92, P < 0.001, Fig. 3; two
rice stink bugs per panicle: F6, 3 = 200.23, P <
0.001, Fig. 3). In both experiments, the percent-
age pecky rice was greater in panicles infested 9
and 13 d after anthesis compared with that in
panicles in all other treatments and the control.
Similarly, the percentage pecky rice was greater
in panicles infested 5 and 17 d after anthesis com-
pared with that in panicles infested 21 d and 1 d
after anthesis as well as that in the control. The
percentage of pecky rice in panicles infested 1 and
21 d after anthesis did not differ, and infestation
at d 1 did not differ from the control. In both ex-
periments, the percentage pecky rice in panicles
infested 9 or 13 d after anthesis was at least 2
times greater than that in panicles infested 5 or
17 d after anthesis and approximately 4 times
greater than in panicles infested 1 or 21 d after
anthesis or that in the control. Thus, rice stink
bug caused pecky rice injury when rice panicles
were infested for 4 d at 5 to 21 d after anthesis,
with the most severe injury inflicted in panicles
infested on d nine and d 13. Incidence of pecky
rice was higher in the two bugs per panicle exper-

Percent Germination of Infested Grains

Peckiness was associated with reductions in
the germination of rice grains (F1, 3 = 935.03, P <
0.001), but the level of reduction did not differ
with time of infestation (F3,3,= 0.61, P < 0.6118).
There was no quality of rice x time of infestation
interaction (F,,32 = 1.05, P < 0.3860). This result
indicates that qualitative injury by rice stink bug
feeding reduced germination by nearly the same
amount at all times of infestation after anthesis
as well as in the control. Germination of nonpecky
grains averaged 89% while that in pecky grains
was 43%.


The data in these experiments showed that
rice grains became less susceptible to quantita-
tive injury (yield loss) by the rice stink bug as the
grains developed. Feeding during anthesis and
the milk stage produced significantly higher per-
centages of empty grains than did feeding during
later grain development. This finding supports
past field work by Pantoja et al. (2000) with a re-
lated species, Oebalus ornatus (Sailer), and by
Rashid (2003) and Swanson & Newsom (1962)
with the rice stink bug that showed severe losses
in rice yields resulting from rice stink bug feeding
during the flowering and the milk stage compared
with feeding during the soft dough stage. This re-

sult is partly explained by the feeding method of
this bug, which sucks out the contents of grains in
the milk stage (Odglen & Warren 1962). The exact
feeding mechanism of the rice stink bug on rice
grains at anthesis is not reported in the litera-
ture. Previous work by Every et al. (1990) indi-
cated that the wheat bug, Nysius huttoni White,
could suck sap rich in amino acids and sugars
from the ovary of wheat seeds at late anthesis.
Lee et al. (1993) found that rice stink bug feeding
during anthesis restricted further grain develop-
ment. Rice stink bug feeding also reduced the av-
erage weights of filled grains during anthesis and
the milk stage (first 12 d after anthesis). Feeding
during the milk stage has been shown to produce
atrophied grains (Swanson & Newsom 1962; Rob-
inson et al. 1980), which probably was a major
contributing factor to the reduced average
weights during the milk stage. Fryar et al. (1986)
stated that many pecky rice grains weigh sub-
stantially less because they are not fully devel-
oped. Therefore, it is likely that the higher per-
centages of pecky rice infested during the milk
and soft dough stages in our experiments signifi-
cantly contributed to the reduced average weights
during those stages. Previous work by Fuchs et al.
(1988) indicated that rice stink bug infestation
during grain development in sorghum reduced
the weight and size of the seeds.
There are at least two explanations for the de-
crease in quantitative injury to rice grains as the
grains matured. First, rice stink bugs may feed
less as grains develop and harden. Second, stink
bug feeding may be equal on grains of different
ages, but grains may become less susceptible to
injury from rice stink bug feeding as they mature.
Incidence of empty grains and reductions in
weights of filled grains were greater under the
higher infestation level, particularly during an-
thesis and the milk stage. Studies by Swanson &
Newsom (1962), Robinson et al. (1980), and
Rashid (2003) with the rice stink bug also found
significant reductions in the total weight per
grain at higher infestation levels compared with
lower infestation levels. This finding supports
previous greenhouse research with the rice bug
Leptocorisa oratorius (F.) that showed a negative
correlation of rice yield to bug density (Jahn et al.
2004). Yield losses in this latter study resulted
from increased numbers of empty and partially
filled grains under higher infestation levels.
The data for percentage of pecky rice (qualita-
tive injury) revealed two valuable pieces of infor-
mation. First, in contrast to the results for quan-
titative injury, the highest levels of pecky rice oc-
curred in grains infested during the soft dough
stage. Severe qualitative injury, at both infesta-
tion levels, occurred in panicles infested during
the soft dough stage (13 d after anthesis). Pani-
cles infested during the late milk stage (9 d after
anthesis), which had a significant number of

Florida Entomologist 89(3)

grains in the soft dough stage, and panicles in-
fested during the hard dough stage (17 and 21 d
after anthesis) also had considerable pecky rice.
The vulnerability of the soft and hard dough
stages is probably explained by the fact that this
bug removes a portion of the contents of grain,
leaving a discolored area around the site. Previous
studies have shown that grains attacked during
the soft and hard dough stages resulted in pecky
rice (McPherson & McPherson 2000; Harper et al.
1993); although not as common, pecky rice was
also reported in grains attacked during the milk
stage in the field (Odglen & Warren 1962).
Second, the presence of pecky rice in the con-
trols in both experiments suggests that peckiness
is caused by factors in addition to rice stink bug
feeding, perhaps fungi (McPherson & McPherson
2000). It is clear, however, that rice stink bug
feeding was a major factor contributing to pecky
rice in infested panicles in the current experi-
ments, either directly or indirectly by facilitating
the entry of microbes. The rice stink bug is known
to vector several pathogens through its stylets in
a transient manner (Hollay et al. 1987). Lee et al.
(1993) demonstrated that discoloration in pecky
rice resulted from fungi that were introduced
when rice stink bug was feeding. Maretti & Peter-
son (1984) demonstrated that rice stink bug feed-
ing was a major factor in grain discoloration, al-
though Bipolaris oryzae (Breda de Haan), a fun-
gus that causes brown spot, was a primary cause
of some grain discoloration and was one of several
microbes that colonized grains through feeding
punctures. Nematospora coryli Peglion, a fungus
capable of causing discolored areas, has also been
noted (Way 1990).
Pecky rice germinated at a significantly lower
rate than nonpecky rice, indicating that injury
due to rice stink bug feeding and/or microbes as-
sociated with pecky grains may have injured the
embryo of the attacked grains. It is also possible
that microbes present within the pecky grains in-
terrupted the germination process, although no
visible sign of differences in the microbial growth
between pecky and nonpecky grains were ob-
served during the germination test. A previous
study has documented reductions in viability of
grains because of rice stink bug feeding (Swanson
& Newsom 1962). In this study, grains that were
atrophied or injured at the proximal (germ) end
had reduced viability. Apparently, the embryo is
extremely sensitive to injury by the rice stink bug.
Rice stink bug attack during grain development
in sorghum reduced seed germination (Fuchs et
al. 1988). Although the seed cleaning process
would eliminate much of the seed severely atro-
phied by rice stink bug injury, observed reduc-
tions in germination were substantial enough to
prevent certification of seed for commercial sale,
which has an acceptable limit of 85% germination
(Douglas & Tullis 1950).

Rice producers have long relied on synthetic
insecticides to control rice stink bugs (McPherson
& McPherson 2000). Concerns about the toxicity
of insecticides to non-target organisms, continued
availability of currently registered insecticides,
and adverse effects of insecticides on the environ-
ment have prompted investigations of alternative
strategies for management of the rice stink bug.
The short window of vulnerability of the rice
plant to rice stink bug (approximately 30 d for
most varieties) has been an important factor in
restricting research in the development of
nonchemical control measures (Way 1990). The
current available action thresholds for rice stink
bug in rice (30 bugs per 100 sweeps for the first
two weeks of heading and 100 bugs per 100
sweeps from the dough stage until two weeks be-
fore harvest (Johnson et al. 1987) accounts to
some degree for age-related changes in grain sus-
ceptibility to injury. However, more precise infor-
mation such as that reported here on the suscep-
tibility of rice panicles may be important for the
refinement of the current thresholds and for the
development of alternative management strate-
gies for the rice stink bug.


We thank J. P. Geaghan (Department of Experimen-
tal Statistics, Louisiana State University AgCenter) for
help with statistical analyses, Mark Cohn (Department
of Plant Pathology and Crop Physiology, Louisiana
State University AgCenter) for providing helpful ad-
vice, and Boris Castro and Dennis Ring for helpful com-
ments on earlier drafts of the manuscript. This research
was partially funded by the Louisiana Rice Research
and Promotion Board. This paper was approved for pub-
lication by the Director of the Louisiana Agricultural
Experiment Station as manuscript no.05-26-0754.


BOWLING, C. C. 1963. Cage test to evaluate stink bug
damage to rice. J. Econ. Entomol. 56: 197-200.
2000. A uniform, objective, and adaptive system for
expressing rice development. Crop Sci. 40: 436-443.
DOUGLAS, W. A. 1939. Studies of rice stink bug popula-
tions with special reference to local migration. J.
Econ. Entomol. 32: 300-303.
DOUGLAS, W. A., AND E. C. TULLIS. 1950. Insects and
Fungi as Causes of Pecky Rice. U. S. Dept. Agric.
Tech. Bull. 1015.
1990. Wheat-bug damage in New Zealand wheats:
The feeding mechanism ofNysius huttoni and its ef-
fect on the morphological and physiological develop-
ment of wheat. J. Sci. Food Agric. 50: 297-309.
TUGWELL. 1986. The economics of controlling peck in
Arkansas rice. Arkansas Farm Res. 35(3): 7.
TEETES. 1988. Managing Insect and Mite Pests of
Texas Sorghum. Texas. Agric. Ext. Serv. B-1220: 1-16.

September 2006

Patel et al.: Rice Stink Bug Injury to Rice

AND J. W. MJELDE. 1993. Damage function analysis
for the rice stink bug (Hemiptera: Pentatomidae). J.
Econ. Entomol. 86: 1250-1258.
Structure and formation of feeding sheaths of rice
stink bug (Heteroptera: Pentatomidae) on rice
grains and their association with fungi. Ann. Ento-
mol. Soc. Am. 80: 212-216.
CIA. 2004. Effect of rice bug Leptocorisa oratorius
(Hemiptera: Alydidae) on rice yield, quality, and
seed viability. J. Econ. Entomol. 97(6): 1923-1927.
1987. Control of insects attacking rice. Arkansas
Coop. Ext. Serv. 330: 1-15.
WEIDEMANN. 1993. Role of fungi vectored by rice
stink bug (Heteroptera: Pentatomidae) in discolor-
ation of rice kernels. J. Econ. Entomol. 86: 549-
MARETTI, M. A., AND H. D. PETERSON. 1984. The role of
Bipolaris oryzae in floral abortion and kernel discol-
oration in rice. Plant Dis. 68: 288-291.
balus spp., pp. 141-158 In Stink bugs of Economic
Importance in America North of Mexico. CRC Press
LLC, Boca Raton, FL.
NARESH, J. S., AND C. M. SMITH. 1983. Development
and survival of rice stink bugs (Hemiptera: Pentato-
midae) reared on different host plants at four tem-
peratures. Environ. Entomol. 12: 1496-1499.
ODGLEN, G. E., AND L. O. WARREN. 1962. The Rice Stink
Bug, Oebalus pugnax F., in Arkansas. Arkansas Ag-
ric. Exp. Stn. Rep. Ser. 107: 1-23.
Population dynamics and effects of Oebalus ornatus
(Hemiptera: Pentatomidae) on rice yield and quality

in Southwestern Colombia. J. Econ. Entomol. 93(2):
RASHID, T. 2003. Rice Stink Bug Development, Feeding
Preference, Sampling Techniques and Damage Ef-
fects on Rice Yield. Ph.D. dissertation, University of
Arkansas, Fayetteville.
Sampling rice stink bug (Hemiptera: Pentatomidae)
in and around rice fields. Environ. Entomol. 35: 102-
HOLLAY. 1980. Rice stink bug: Relationship between
adult infestation levels and damage. Louisiana Ag-
ric. Ext. Stn., Rice Exp. Stn., Annu. Prog. Rep. 72:
SAILER, R. I. 1944. The genus Solubea (Heteroptera:
Pentatomidae). Proc. Entomol. Soc. Washington 46:
SAS INSTITUTE. 1996. SAS/STAT User's Guide: version
6, 4th Edition, Volume 1 and 2. SAS Institute Inc.
Cary, NC.
SWANSON, M. C., AND L. D. NEWSOM. 1962. Effect of in-
festation by the rice stink bug, Oebalus pugnax, on
yield and quality in rice. J. Econ. Entomol. 55:877-879.
Effects of the presence of barnyardgrass on rice wa-
ter weevil (Coleoptera: Curculionidae) and rice stink
bug (Hemiptera: Pentatomidae) populations on rice.
Environ. Entomol. 33: 720-726.
1994. Management tactics for soybean insects, pp.
115-117 In L. G. Higley and D. J. Boethel [eds.],
Handbook of Soybean Insect Pests. Entomol. Soc.
Am. Publ., Lanham, MD.
WAY, M. 0. 1990. Insect pest management in rice in the
United States, pp. 181-189 In B. T Grayson, M. B.
Green, and L. G. Copping [eds.] Pest Management in
Rice. Elsevier Applied Science, New York, NY.

Florida Entomologist 89(3)

September 2006


'University of Florida, Tropical REC, IFAS, Department of Entomology and Nematology 18905 SW 280ThSt.,
Homestead, FL 33031

2Centro Regional de Investigaciones Cientificas y Transferencia Tecnol6gica, Entre Rios y Mendoza s/n, CP 5301
Anillaco, La Rioja, Argentina


Seasonal abundance of the citrus leafminer, Phyllocnistis citrella Stainton (Lepidoptera:
Gracillariidae), was investigated between Nov 1999 and Apr 2003 in Taft Viejo (Tucuman
province). Phyllocnistis citrella populations increased during spring and summer, declined
during fall, and disappeared in the winter. Five species of parasitoids, one exotic and four in-
digenous, attacked citrus leafminer immature stages in commercial and experimental
lemon orchards. Ageniaspis citricola Logvinovskaya (Hymenoptera: Encyrtidae) was the
most abundant parasitoid. Cirrospilus neotropicus Diez & Fidalgo (Hymenoptera: Eu-
lophidae) was the most abundant indigenous species, followed by Galeopsomyia fausta La-
Salle (Hymenoptera: Eulophidae). The other indigenous species were not common and were
only occasionally collected from citrus leafminer larvae. Parasitoids and P. citrella exhibited
similar population fluctuations throughout the entire sampling period. A certain degree of
synchrony exists between the most abundant parasitoids (A. citricola, C. neotropicus, and G.
fausta) and the pest. The highest rates of parasitism were observed in the fall. Ageniaspis
citricola exhibited approximately 29.5% parasitism, whereas all the native species together
were only 8.2%. Data showed that a clear dependence existed between percentages of para-
sitism and citrus leafminer population density for the most frequent parasitoid populations.
The results of this study show that C. neotropicus has an important role among the native
species present in Argentina.

Key Words: pest, performance, parasitism, lemon, citrus leafminer


La provincia de Tucuman concentra el 90% de la actividad limonera de la Argentina, siendo
esta el primer productor e industrializador de lim6n del mundo. La presencia de Phyllocnis-
tis citrella Stainton, pone en riesgo la productividad de este cultivo, atacando a plants j6-
venes y de invernaderos. El objetivo del present trabajo fue conocer la fluctuaci6n
poblacional de P. citrella y sus parasitoides, datos de importancia al moment de implemen-
tar proyectos de manejo integrado de plagas en citricos. Se monitorearon cuatro campaias
citricolas entire Nov de 1999 y Jun de 2003; en dos quintas de limones ubicadas en el Depto
de Taft Viejo, Tucuman. Phyllocnistis citrella se registry en campo a principios de Noviem-
bre; aumenta su abundancia poblacional en sucesivas generaciones en el verano, 6poca en la
que muestra sus picos poblacionales maximos. A comienzos del otoio su abundancia pobla-
cional disminuye y desaparece en invierno. Cinco species de parasitoides, una ex6tica (En-
cytidae) y cuatro nativas (Eulophidae), atacaron los estadios inmaduros de P. citrella. El
parasitoide introducido, Ageniaspis citricola Logvinovskaya, fue el mas abundante con
29,5% de parasitoidismo. Cirrospilus neotropicus Diez & Fidalgo fue la especie native mas
abundante y frecuente con 7,8% de parasitoidismo, seguida por Galeopsomyia fausta LaSa-
lle. El resto de las species nativas no fueron muy comunes y s6lo ocasionalmente fueron co-
lectadas a partir de P. citrella. La especie native C. neotropicus demostr6 tener un rol
important, por su frecuencia y abundancia, sobre las poblaciones de la plaga, dando mues-
tras de la importancia de mantener la biodiversidad de la region.

Translation by the authors.

The citrus leafminer, Phyllocnistis citrella Phyllocnistis citrella was not reported in the
Stainton, originally from the east and west of western hemisphere until May 1993 when it was
Asia, is one of the main pests that affects Citrus discovered in citrus nurseries in Florida (Hepp-
and related species worldwide (Beattie 1993). ner 1993a). In Argentina, the citrus leafminer

Diez et al.: P citrella and its Parasitoids in Argentina

was reported for the first time in late 1995 in cit-
rus plantations in Northern Salta and Northwest-
ern Tucuman (Willink et al. 1996). At present, it
is found in all citrus areas of Argentina.
Heavy infestations of P. citrella can seriously
affect plants from nurseries and those recently
planted, although the damage is less significant
in mature trees. The presence of this pest makes
trees more susceptible to infestation by citrus
canker bacterium Xanthomonas axopodis pv. citri
(Sohi & Sandhu 1968). The serpentine mines un-
der the leaf cuticle caused by the larvae of citrus
leafminer, provide ample wounding on new
growth to greatly amplify citrus canker infection
(Gottwald et al. 2002). Larvae carve serpentine
mines on the adaxial and abaxial surfaces of
newly formed leaves, ingesting the sap and pro-
ducing a chlorotic leaf patch. This prevents young
leaves from expanding, making their margins
curl and sometimes even causing the leaves to
At present, P citrella control practices consist
of the application of chemical products as well as
biological control. Effective chemical control is
difficult because the larvae are protected from in-
secticides by the leaf cuticle (Legaspi et al. 1999).
It has been suggested that biological control can
become a successful tool for the regulation of pop-
ulations of this pest (Hoy & Nguyen 1994). Nearly
80 species of parasitoids have been obtained from
citrus leafminer worldwide. Of these, approxi-
mately 20 species occur in the Neotropical region
(Schauff et al. 1998). From P citrella larvae, 39
species of parasitoids from seven families, mainly
Chalcidoidea, have been identified (Heppner
1993b) and ten species occur in Thailand (Hoy &
Nguyen 1994). Of these, Ageniaspis citricola
Logvinovskaya is the most significant. LaSalle &
Peia (1997), Evans (1999), and Diez & Fidalgo
(2004) described three new species of Eulophidae
in the Neotropical region. Many of these species
are incidental and provide no effective control of
the pest. Others, however, play a significant role
in the regulation of the population levels, such as
species of Cirrospilus sp., and Galeopsomyia
fausta LaSalle (LaSalle & Pena 1997; Urbaneja et
al. 1998; Urbaneja et al. 2000).
Five species of parasitoids (Eulophidae) that
attack P citrella have been recorded in Argentina:
Cirrospilus neotropicus Diez & Fidalgo, Elasmus
sp. (a new species currently being described by
the authors), G. fausta, Elachertus sp., and Sym-
piesis sp. (Frias & Diez 1997; LaSalle & Peia
1997; Schauff et al. 1998; Fernandez et al. 1999a;
Diez 2001; Diez & Fidalgo 2004). The role of these
native parasitoids was studied by Diez et al.
(2000), and Diez (2001), who observed that C. neo-
tropicus is the most important native parasitoid
from P citrella due to its frequency and abun-
dance. Approximately 19% parasitism is due to C.
neotropicus, whereas the rest of these species are

not common and only occasionally collected from
citrus leafminer.
Classical biological control, through the intro-
duction of hymenopteran parasitoids, is a method
used in different parts of the world for the regula-
tion of the populations of P citrella (Hoy et al.
1995; Argov & R6ssler 1996; Smith & Beattie
1996; Hoy & Nguyen 1997; Siscaro et al. 1997;
Tsagarakis et al. 1999; Vercher et al. 2000;
Nogueira de Sa et al. 2000; Fernandez et al.
1999b; Diez et al. 2000; Willink et al. 2002). Two
Parasitoids, A. citricola and Citrostichus phylloc-
nistoides Narayanan, were introduced in Argen-
tina. The first was reported in Tucuman in 1997
as a case of ecesis in biological control (Fernandez
et al. 1999b; Diez et al. 2000). However, in 1998
this parasitoid was reintroduced in the citrus
plantations of Tucuman with specimens from
Peru (Figueroa et al. 1999). It was determined
that this parasitoid is widely established in the
northwestern region of Argentina, constituting
the most important species due to its abundance
and the frequency observed on P citrella (Diez et
al. 2000; Diez 2001). Citrostichus phyllocnistoides
was released in the Argentine provinces of Salta,
Jujuy, Tucuman and Catamarca in 2001. This
species was reared in large numbers and periodi-
cally released in citrus plantations throughout
the year 2002 (Willink et al. 2002).
The objective of this study was to describe the
population dynamics of P. citrella and its parasi-
toids in Tafi Viejo (Tucuman province) and evalu-
ate the importance of the latter for biological con-
trol of the pest.


The seasonal dynamics of citrus leafminer and
its parasitoids were studied between 1999 and
2003 in four sampling periods: Nov 1999 through
Apr 2000; Nov 2000 through July 2001; Aug 2001
through Jul 2002; and Aug 2002 through Apr
2003. The study was performed in plots from two
fields planted with lemon trees in Tafi Viejo, Tu-
cuman. One of the fields remained free of insecti-
cides unsprayedd, experimental orchard),
whereas the other was a commercial orchard and
was sprayed regularly. Both fields had approxi-
mately 500 lemon trees.
In the commercial orchard (sprayed), routine
agronomic practices for the region include fre-
quent use of agrochemicals, insecticides, fungi-
cides, and herbicides. Treatments applied to the
sprayed orchard were made to primarily control
mites and citrus leafminer. In all years, between
Sep and Jan, a copper and oil sprays were made
monthly for fungal pathogens and mite control. In
1999 and 2001 only, one application of mineral oil
was made in Sep and another in Jan to prevent
oviposition of citrus leafminer. To reinforce these
applications, a second application of mineral oil

Florida Entomologist 89(3)

was made within the first two weeks of Jan. In
2003 and 2004, abamectin was applied every 15
days between Jan and Mar, for P. citrella (Ing.
Agr. Fernando Carrera, personal communica-
To determine the density of P. citrella, 10 trees
were selected randomly from each plot every
week, and eight shoots measuring less than 10 cm
(9.32 + 2.6 leaves per shoot) were collected from
each tree (Ripolles et al. 1996). Occasionally,
fewer shoots were available in the field; in these
cases all the shoots present were collected. This is
reflected in the results as a decrease in the num-
ber of shoots monitored. Each shoot was placed in
a polyethylene-terephthalate bag (30 cm x 20
cm), with a sheet of absorbing paper to avoid con-
densation of humidity. The samples were trans-
ported to the laboratory and examined under the
microscope to determine the presence of all stages
of P. citrella. Leaves with larvae and pupae were
separated to observe the emergence of adults and
parasitoids of P citrella. They were placed in Petri
dishes (9 cm diameter) containing plaster of Paris
and covered with plastic film; distilled water was
added everyday to maintain humidity. The dishes
were kept under ambient conditions in the labora-
tory. All emerging adults of P citrella and parasi-
toids were identified and counted. These data al-
lowed us to describe the population fluctuation of
citrus leafminer and its parasitoids, the percent-
age of new flush damaged by citrus leafminer (Ur-
baneja Garcia 2000), the percentage of parasitism
(Van Driesch 1983), and species richness of para-
sitoids in both plots (Odum 1985).
Evaluation of differences between the popula-
tion density of P citrella and density populations
of parasitoids in both plots were made by Stu-
dent's t-test. A X2 test was used to evaluate differ-
ences between the abundance of the different spe-

cies of parasitoids. A stepwise multiple regression
analysis was performed in order to evaluate the
relationship between P citrella density, number
of monitored shoots, and different weather vari-
ables (daily high and low temperatures and daily
precipitation). Weather data were provided by the
personnel of the Park Sierras de San Javier.
These data were daily collected with a thermo-
graph and rain gauge, installed in the park,
within a meteorological square located at 700
m.s.n.m. The number of samples compared was
79. All the statistical analyses were made at sig-
nificance level a = 0.05 (Statistix, Analytical Soft-
ware, Tallahassee, FL).


Population Abundance

Between Nov 1999 and Apr 2003, 64,735
leaves were inspected to determine the presence
or absence of P citrella. In total, 45,335 citrus
leafminer specimens were recorded, including
eggs and larvae at different instars. The highest
population levels of P citrella were observed be-
tween Jan and Mar, while the lowest incidence oc-
curred in the months of Aug through Dec and Apr
through Jul (Fig. 1). The seasonal pattern ofRP ci-
trella was similar to observations in Florida (Pefia
et al. 1996) and in southern Texas (Legaspi et al.
1999), where population densities increase from
spring to fall and decline during the winter. In
general, the reason for the increase in density of
P citrlla in spring is the presence of new shoots
and the increase in temperature, which is favor-
able to the pest. In Florida (Peia et al. 1996) and
Mexico (Bautista-Martinez et al. 1998), as well as
in our working area, the development of some
flushes during winter, which were free of P. cit-

400 235
3.50 3

betweenen 1999 and 2003.3.00

:160 ?15

0.50 0\l I 5

0.0u p.l AM ..... .... ..-M .. F 0m
1999-2000 2000-001 aUl-A 2X)2-2003

a unsprayed -& sprayed a average temperature

Fig. 1. Average density of P. citrella in sprayed and unsprayed lemon orchard in Tafi Viejo, (Tucuman, Argentina)
between 1999 and 2003.

September 2006

Diez et al.: P citrella and its Parasitoids in Argentina

rella, were observed. According to Peia et al.
(1996), citrus leafminer oviposition declines dur-
ing the winter due to the low temperatures.
Hence, the increase in the population of citrus
leafminer in spring could be more related to the
increase in temperatures than to the presence of
shoots. In this study, the stepwise multiple re-
gression analysis between weather factors, moni-
tored shoots, and P citrella density showed that
the weather variables, and not the shoots, consis-
tently affected pest populations (Y = 1.88 + 2.84
minimum T deg + 0.55 Rainfall 2.4 maximum T(deg
,); R2= 0.99;F = 4836.7;P > 0.001; n = 79). The rise
in temperature and rainfall recorded during
spring contribute to this population increase.
Throughout all the sampling periods, a de-
crease in the population abundance of P. citrella
was observed, which fell an average 34% between
1999 and 2003. Climate variables play an impor-
tant role among the many factors that could con-
tribute to this situation. In that period a signifi-
cant decrease in precipitation and a small in-
crease in the average temperature was registered
in the study area. This decrease in rainfall could
have affected the quantity of shoots, bringing
their numbers down, which would be reflected as
a decrease in monitored shoots. This situation
could partially explain the decrease in the popu-
lation abundance of citrus leafminer in these
years. A similar situation was recorded in south-
ern Texas, where a general decrease in the popu-
lation levels of citrus leafminer was observed be-
tween 1995 and 1998 (Legaspi et al. 2001).
During the first sampling period four popula-
tion peaks were registered in Jan, Feb, and Mar of
2000, varying between 0.96 and 1.82 citrus leaf-
miner per leaf. In the second period, an important
population peak of 2.77 citrus leafminers per leaf
was recorded in the first week of Jan. In the third

period, two instances of higher population density
(0.96 and 1.27 citrus leafminers per leaf) were ob-
served in Jan 2002. In Dec 2002 and Feb 2003 (the
last monitored sampling period), two population
peaks were recorded, with maximum values of
2.61 and 2.87 citrus leafminers per leaf (Fig. 2).
These fluctuating patterns coincide with those
shown by the pest both in Florida (Peia et al.
1996) and southern Texas (Legaspi et al. 1999),
where populations are most abundant during the
summer and early autumn.
In general, there are no differences between
the P citrella population fluctuations in either
sprayed or unsprayed plots (t = 1.14; P = 0.14).
There were particular instances, however, where
the fluctuation curves were inverted (i.e.,when-
ever the density of P citrella increased in the com-
mercial plot, it decreased in the experimental
plot). This could be due to the action of other fac-
tors of mortality, such as predators, that could be
acting in this plot. In Jan, increases in citrus leaf-
miner population were observed in the experi-
mental plot whereas the population decreased in
the commercial plot (Fig. 1). These situations
could be partially related to the frequent use of
agrochemicals in the commercial plot.

Percentage Leaf Damage

The percentage leaf damage by the immature
stages of P. citrella was approximately 51%
throughout this study. Studies performed in
Spain during 1999 (Urbaneja Garcia 2000) re-
ported higher percentage leaf damage (86.7-
96.2%) than those observed in this investigation
for each monitored sampling period. In Spain, the
parasitoidA. citricola has not become established.
Except for the last sampling period, the percent-
age of total weekly attacks recorded in the exper-



19.W-200 2o-2001 2001-2002 202-20)03
paraito -. --P. citvla

Fig. 2. Average density ofP. citrella and its parasitoids in Tafi Viejo (Tucumin, Argentina) between 1999 and

Florida Entomologist 89(3)

mental plot was generally greater than in the
commercial plot. This is undoubtedly related to
the frequent use of agrochemicals in the commer-
cial plot. The highest percentage of citrus leaf-
miner attack to the crop was recorded in the un-
sprayed plot during the summer. Therefore,
chemical products used to control the pest in our
region should be applied starting in Nov and early
Dec, when the temperature and humidity condi-
tions are not yet adequate for development of the
pest, and its population density is low in the field.
This would thus limit the uncontrolled population
growth exhibited by P citrella around mid-Dec,
brought about in part by the absence in the field
of its parasitoid natural enemies.
Damage caused by P citrella larvae began in
late spring (Nov-Dec) amounting to 19.9% of the
total. The highest rates were observed in the sum-
mer ('!. I,, whereas in the fall damage decreased
to 13.4%. No damage occurred in winter. In Spain,
reported rates of damage were 30% in the first
months when citrus leafminer was observed and
90% during months exhibiting higher population
abundance (Urbaneja Garcia 2000).

Parasitoid Complex

The complex of parasitoids recorded on P cit-
rella in the four sampling periods was repre-
sented by one introduced exotic species, A. citri-
cola, and four native species: Cirrospilus (= Cir-
rospilus sp.) neotropicus, G. fausta, Elasmus sp.
and Elachertus sp., all members of the family Eu-
lophidae. This work confirms previous records ob-
tained for this region (Frias & Diez 1997; Fernan-
dez et al. 1999a; Diez 2001). The last species in-
troduced in our region (C. phyllocnistoides) has
failed to establish and overwinter successfully,
since no individual of this species was recovered
in the last sampling period.
During this study, five species, one introduced
(A. citricola) and four native species (C. neotropi-
cus, E. phyllocnistoides, G. fausta, and Elachertus
sp.), were recorded in the experimental plot,
whereas only three species, one introduced (A. cit-
ricola) and two native species (C. neotropicus and
G. fausta), were observed in the commercial plot.
According to Odum (1985), a harsh physical envi-
ronment, contamination, and other factors tend
to reduce the number of rare species and increase
the importance or degree of dominance of a few
common species (which can better tolerate the
pressure or are more adapted to it). This fact, in
conjunction with the use of chemical products in
the commercial plot, could explain at least in part
the differences found in the species richness in
both plots. Moreover, the presence of the native
species, C. neotropicus and G. fausta, in the com-
mercial plot would indicate that these species are
better able to withstand the pesticides used in
lemon plantations in the area. A similar situation

was described in Florida regarding the native
parasitoid Pnigalio minio (Walker) in crops where
insecticides were applied (Peia et al. 1996).
Ageniaspis citricola was the most abundant
parasitoid in our region, amounting to 62.3% of
the parasitoids observed in this study. Cirrospilus
neotropicus was the most abundant among the
native species (29.2%), followed by G. fausta
(1.3%), Elasmus sp. (0.1%), and Elachertus sp.
(0.05%) (X2 = 9.48). However, the structure of the
parasitoid complex of P citrella varies according
to the region under study. In Spain, Pnigalio pec-
tinicornis (Linnaeus) is the dominant species,
amounting to 57.1% of the recorded parasitoids
(Urbaneja et al. 2000). In Florida, the most impor-
tant native species was P minio (80%), followed
by Cirrospilus sp., Closterocerus sp., Zagrammo-
soma multilineatum (Ashmead), and Horismenus
sp. (Peia et al. 1996). In southern Texas, Z. mul-
tilineatum was the most important species,
amounting to 68-74% (Legaspi et al. 1999). In
Mexico, two species of Cirrospilus, and G. fausta
were observed as the most representative native
species (Bautista-Martinez et al. 1998). In Brazil,
G. fausta was recorded as the most important na-
tive species (' j. followed by Elasmus sp., and
Cirrospilus sp. (Montes et al. 2001). Cirrospilus
neotropicus exhibited a high frequency in the
commercial plot. This could be related to the pres-
ence of alternative hosts that are presumed to in-
habit nearby Citrus plantations and to act as res-
ervoirs for this native species in the area.
Parasitoids exhibited a population fluctuation
similar to that of the pest and also a similar be-
havior throughout the spring to winter period
(Fig. 2). Highest abundance of parasitoids coin-
cides with the highest pest density (summer) in
all sampling periods (R2 = 0.84). Legaspi et al.
(1999) reported the highest abundance of parasi-
toids in late summer and early fall in southern
Texas, corresponding to an increase in host popu-
The population density of different species of
parasitoids was the same in both plots (t = 1.48; P
= 0.13). In general, parasitoids were first ob-
served in both plots in Nov and Dec, exhibiting
the highest population levels in summer. Popula-
tion fluctuations recorded in the first, second, and
fourth sampling periods were very similar. Simi-
larly, the time of appearance of the parasitoids,
except for the population peak corresponding to
Jan 2000, was remarkably higher in the commer-
cial plot. In the third sampling period, parasitoids
in the commercial plot were first observed a
month later than in the experimental plot and ex-
hibited a marked asynchrony, showing a higher
density between Nov and Jan. The highest den-
sity of parasitoids in the experimental plot was
observed between Jan and Apr.
The most abundant populations,A. citricola, C.
neotropicus, and G. fausta exhibited similar popu-

September 2006

Diez et al.: P citrella and its Parasitoids in Argentina

lation fluctuations as P. citrella throughout the
entire sampling period. Two species, A. citricola,
C. neotropicus, were recorded in all sampling pe-
riods, whereas G. fausta was present only in the
first three sampling periods. No alternative hosts
ofA. citricola have been described. This suggests
that the parasitoid exhibits a good synchroniza-
tion with the pest, spending the winter in dia-
pause on P citrella, with delayed development in
the spring. This is probably due to the low relative
humidity, a known limiting factor in the rearing
of this species. The dry season in Tucuman coin-
cides with the winter and extends a few months
into spring.
Populations of Elasmus sp. were present spo-
radically, appearing only in Jan and Apr in the
first and third sampling periods, respectively.
Elachertus sp. was observed on P citrella only
during the first sampling period (Mar), and com-
pletely disappeared afterwards. Appearance in
the field varies for the different species of parasi-
toids according to the areas and species consid-
ered. In Florida, Cirrospilus species were found in
the field starting in the fall through the spring
(Pefa et al. 1996). In Brazil, G. fausta is found in
the field most of the year, Elasmus sp. between
Jan and Mar, and Cirrospilus sp. in Jun and Jul
(Montes et al. 2001).
In all but the last sampling periods, the native
species C. neotropicus was observed in the field
before the exotic species A. citricola. According to
Cornell and Hawkins (1993), parasitoids that suc-
cessfully colonize invading hosts are usually gen-
eralists, can change hosts more easily, and are
more successful initially than specialists. This
phenomenon could explain why C. neotropicus ap-
pears before the introduced parasitoid in all sam-
pling periods. Cirrospilus neotropicus remains
throughout the summer season exhibiting a more
or less stable frequency, which is suggestive of its
adaptation to its new host, P citrella.

Percentage of Parasitism

The average percentage of parasitism recorded
during the entire study was approximately 37.7%.
This is lower than those observed by others. Hoy
& Nguyen (1997) observed approximately 60% in
Florida. Bautista-Martinez et al. (1998) and Le-
gaspi et al. (2001) observed close to 70% in Mex-
ico. Percentages of parasitism recorded in both
plots were practically the same (commercial plot
39.9%, experimental plot 35%) (P = 0.75; t = 1.65).
Similar values were found in Spain in plots with
different exposure to insecticides (Vercher et al.
Throughout the study, the highest rates of par-
asitism were observed in the fall. This might be
partially due to a decline in the population ofRP ci-
trella during the fall season when all populations
of parasitoids, specialists or generalists, are al-

ready well established (Fig. 3). In addition, popu-
lations of parasitoids become synchronous with P.
citrella populations at this time of year. A similar
situation was observed in Veracruz, Mexico, with
high percentages of parasitism accompanied by
low abundance of collected P citrella (Bautista-
Martinez et al. 1998). Analysis of the data showed
that a clear dependence existed between percent-
ages of parasitism and citrus leafminer popula-
tion density for the most frequent parasitoid pop-

* Ageanasipn ro


p O *


, ,

- I..

0 0.5 1 1 5 2 2.5 3 3.5

)0 CmspikAsarwmpwc s
U .
a .
a .
S *
0 *

Q10 2
S* *

0 0.5 1 15

90 -
80 -
70 .

I 60-

a1 ^n -

25 3 35

Galopsomyia fasta


U ,UW a se ..
0 05 1 15 2 2.5 3 3.5
Number of circus leafinmer per leaf

Fig. 3. Relation between percentage parasitism of
Ageniaspis citricola, Cirrospilus neotropicus, and Gale-
opsomyia fausta, and average density of Phyllocnistis
citrella in Tafi Viejo (Tucuman, Argentina) between
1999 and 2003.

Florida Entomologist 89(3)

ulations (A. citricola, C. neotropicus, and G.
fausta). This means that the efficiency of the par-
asitoid increases with population density of the
pest (Fig. 3).

Native versus Introduced Parasitoids

Ageniaspis citricola exhibited approximately
29.5% parasitism, whereas all the native species
together showed a value of only 8.2%. Data re-
corded for A. citricola in this study were lower
than those recorded in other countries. For exam-
ple, studies conducted in Florida between 1994
and 1995 showed 60 and 80% parasitism for this
species (Hoy & Nguyen 1997). The first results on
A. citricola in Texas showed values of approxi-
mately 40% (Legaspi et al. 1999). As for native
parasitoids, 5-10% parasitism was observed in
Texas (Legaspi et al. 1999). The highest values re-
corded in Brazil were near 35% (Montes et al.
2001), but peaks of 85.7% were observed in Mex-
ico (Bautista-Martinez et al. 1998).
Cirrospilus neotropicus showed the highest
percentage of parasitism (7.8%) among the native
parasitoids, followed by G. fausta (0.43%), Elas-
mus sp. (0.08%) and Elachertus sp. (0.01%).
Elachertus sp. is an important species in China
with 40 to 54% parasitism (Hoy & Nguyen 1997).
In Mexico, species in the genus Cirrospilus and G.
fausta exhibit the highest levels (70%) of parasit-
ism (Bautista-Martinez et al. 1998). In Brazil, G.
fausta caused 53% of parasitism, followed by
Elasmus sp. with 38%, and Cirrospilus sp. with
9% (Montes et al. 2001).
According to LaSalle and Peia (1997), a large
complex of native parasitoids are now attackingP.
citrella, and in many cases are providing control
equal to, or greater than, that provided by A. cit-
ricola. The results of this study show that C. neo-
tropicus has an important role among the native
species present in our region due to the frequency
with which it is observed on the pest. Despite
their low number in the region, other native spe-
cies such as G. fausta and Elasmus sp. have also
been reported to be important control agents of P.
citrella in other countries.
Data obtained from the four monitoring sam-
pling periods in different times of the year deter-
mined that A. citricola (introduced species), C.
neotropicus, and G. fausta (native species), among
all the species that comprise the parasitoid com-
plex for P citrella in this region. These species
have characteristics that suggest they may be
good choices for future biological control projects.
These species were undoubtedly responsible, in
part, for the decrease in the population densities
of the pest since their introduction, and to current
levels. Ageniaspis citricola and C. neotropicus are
currently going through different processes i.e.,
the establishment and colonization ofA. citricola;
and the adaptation ofC. neotropicus to citrus leaf-

miner. More work is needed to better understand
the population dynamics of G. fausta.


Funds for this research were provided by the Na-
tional Research Council of Argentina (CONICET)
(Project # 0702 / 98) and are greatly appreciated. We
thank the authorities of the park Sierras de San Javier.


ARGOV, Y., AND Y. ROSSLER 1996. Introduction, release
and recovery of several exotic natural enemies for bi-
ological control of the citrus leafminer, Phyllocnistis
citrella, in Israel. Phytoparasitica 24: 33-38.
BRAVO MOJICA, AND S. D. KOCH. 1998. Natural par-
asitism ofPhyllocnistis citrella (Lepidoptera: Gracil-
lariidae) at Cuitlahuac, Veracruz, Mexico. Florida
Entomol. 81: 30-37.
BEATTIE, G. A. 1993. Integrated control of the citrus
leafminer. N. S. W. Agriculture, Rydalmere, N. S. W.,
Australia CAB, IIBC.
CORNELL, V. H., AND A. B. HAWKINS. 1993. Accumula-
tion of native parasitoids on introduced hosts: a com-
parison of hosts-as native and hosts-as-invaders.
American Natural. 114: 847-865.
DIEZ, P. A. 2001. Structure of the Complex of Parasi-
toides (Hymenoptera) ofPhyllocnistis citrella Stain-
ton (Lepidoptera: Gracillariidae) Attacking Lemon
Tree in Tafi Viejo, Tucuman Province, Argentina.
Thesis of Masters, National University of La Rioja -
Regional Centrro of Scientific researches and tech-
nological Transference, La Rioja, 98pp.
DIEZ, P. A., P. FIDALGO, AND E. FRIAS. 2000.Ageniaspis
citricola (Hymenoptera: Encyrtidae), specific parasi-
toid ofPhyllocnistis citrella (Lepidoptera: Gracillari-
idae): preliminary introduction and data on its
performance in Argentina. Acta Entomol6gica Chil-
ena. 24: 69-76.
DIEZ, P. A., AND P. FIDALGO. 2004. Cirrospilus neotropi-
cus sp. n. (Hymenoptera: Eulophidae): an indigenous
biocontrol agent of the citrus leafminer, Phyllocnistis
citrella (Lepidoptera: Gracillariidae). Entomol.
News 114: 76-81.
EVANS, G. A. 1999. A new species of Cirrospilus (Hy-
menoptera: Eulophidae) and two new synonymies of
parasitoids reared from the citrus leafminer, Phyl-
locnistis citrella (Lepidoptera: Gracillariidae). Flor-
ida Entomol. 82: 448-453.
HERRERO, P. DIEZ, AND E. WILLINK. 1999a. Parasi-
toids ofPhyllocnistis citrella Stainton (Lepidoptera-
Gracillariidae), and its distribution in ecosystems
citric of Tucuman, Argentina. Resumen X Jornadas
Fitosanitarias Argentinas, S. S. de Jujuy, abril de
1999, 248 p.
Ageniaspis citricola Logvinovskaya (Hymenoptera-
Encyrtidae) parasitoid ofPhyllocnistis citrella Stain-
ton (Lepidoptera-Gracillariidae) in Tucuman, Ar-
gentina. Resumen X Jornadas Fitosanitarias
Argentinas. S. S. de Jujuy, abril de 1999, 250 p.

September 2006

Diez et al.: P citrella and its Parasitoids in Argentina

1999. Biological Control of the citrus leafminer. Re-
sumen X Jornadas Fitosanitarias Argentinas. S. S.
de Jujuy, abril de 1999, 276 p.
FRIAS, E., AND P. A. DIEZ. 1997. Native parasitoids (Eu-
lophidae, Elasmidae) of citrus leafminer (Phyllocnis-
tis citrella Stainton) (Lep.: Gracillariidae) in the
Tucuman province. CIRPON Revista de Investiga-
ciones 10: 59-60.
2002. Citrus canker: The pathogen and its impact.
Feature Story, August/ Sep 2002. Online. Plant
Health Progress doi:10.1094/PHP-2002-0812-01-RV
HEPPNER, J. B. 1993a. Citrus leafminer, Phyllocnistis
citrella in Florida, (Lepidoptera: Gracillariidae:
Phyllocnistinae). Trop. Lepidop. 4: 49-64.
HEPPNER, J. B. 1993b. Citrus Leafminer, Phyllocnistis
citrella Stainton (Lepidoptera: Gracillariidae: Phyl-
locnistinae). Entomology Circular no. 359. Mayo/
Junio 1993 (Rev. 11/20/93).
HoY, M., AND R. NGUYEN. 1994. Classical biological con-
trol of the citrus leafminer in Florida. Citrus Indus-
try, Apr, 1994: 22-25.
Establishment of citrus leafminer parasitoids, Age-
niaspis citricola in Florida. Citrus Industry. Dec
1995: 12-17.
HoY, M., AND R. NGUYEN. 1997. Classical biological con-
trol of the citrus leafminer Phyllocnistis citrella
Stainton. Trop. Lepidop. 8: 1-19.
LASALLE, J., AND J. E. PENA. 1997. A new species of Ga-
leopsomyia (Hymenoptera: Eulophidae: Tetrastichi-
nae): a fortuitous parasitoid of the citrus leafminer,
Phyllocnistis citrella (Lepidoptera: Gracillariidae).
Florida Entomol. 80: 461-470.
WOOLLEY. 1999. The citrus leafminer Phyllocnistis
citrella (Lepidoptera: Gracillariidae) in South Texas:
incidence and parasitism. Florida Entomol. 82: 305-
C. LEGASPI, JR. 2001. Population dynamics of the cit-
rus leafminer, Phyllocnistis citrella (Lepidoptera:
Gracillariidae), and its natural enemies in Texas and
Mexico. Biol. Cont. 21: 84-90.
rencia de paratoides da larva minadora dos citros,
Phyllocnistis citrella Stainton, no Municipio de Pres-
idente Prudente, sp. Arquivos do Instituto Biologico
Sao Paulo, 68: 63-66.
IVEIRA, AND G. R. DE ALMEIDA. 2000. Parasitoids of
Phyllocnistis citrella in Jaguariuna, State of San
Pablo, Brazil, before and after of the introduction of
Ageniaspis citricola. Scientia Agricola, 57: 799-801.
ODUM, M. P. 1985. Ecologia. (Tercera edici6n). Editorial
sonal abundance of Phyllocnistis citrella (Lepi-
doptera: Gracillariidae) and its parasitoids in South
Florida citrus. Environ. Entomol. 25: 698-702.
FIBLA, AND R. CARULLA. 1996. M6todos de muestreo
para Phyllocnistis citrella, el minador de las hojas de
los citricos. Determinaci6n de la densidad de brot-
aci6n. Levante Agricola/4 Trimestre, 320-325.

1998. The genera of chalcid parasitoids (Hy-
menoptera: Chalcidoidea) of citrus leafminer Phyl-
locnistis citrella Stainton (Lepidoptera:
Gracillariidae). J. Nat. Hist. 32: 1001-1056.
1997. Prime acquisizioni sul controllo biologico e in-
tegrato della minatrice sepentina degli agrumi in
Italia. Informatore Fitopatologico, 7-8: 19-26.
SMITH, J. M., AND G. A. C. BEATTIE. 1996. Australian
citrus IPM and citrus leafminer. Proceedings from
the International Conference Managing the Citrus
Leafminer. Orlando (Florida): pp. 66-68.
SOHI, G. S., AND M. S. SANDHU. 1968. Relationship be-
tween Citrus leaf-miner (Phyllocnistis citrella Stain-
ton) injury and citrus canker (Xanthomonas citri
(Hasse) Dowson) incidence on Citrus leaves. Journal
Research. Punjab Agricultural University (Lu-
dhiana) 5: 66-69.
ence and impact of introduced and native parasi-
toids on Phyllocnistis citrella Stainton in Greece.
Evaluating indirect ecological effects of biological
control, global IOBC international symposium,
Montpellier, France, 17-20 October 1999. IOBC-
WPRS bulletin 22: 66 p.
GARRIDO. 1998. Dinamica e impact de los parasi-
toides aut6ctonos de Phyllocnistis citrella Stainton,
en la Comunidad Valenciana. Producci6n y Protec-
ci6n Vegetales, 13: 410-423.
URBANEJA-GARCIA, A. 2000. Biologia de Cirrospilus sp.
pr6ximo a lyncus (Hym.: Eulophidae), ectoparasi-
toide del minador de las hojas de los citricos, Phyl-
locnistis citrella Stainton (Lep.: Gracillariidae).
Dinamica e impact de los enemigos naturales del
minador. Tesis Doctoral, Universidad Polit6cnica
del Valencia, Escuela Superior de Ingenieros
Agr6nomos, Departamento de Producci6n Vegetal,
127 pp.
RIDO, AND J. A. JACAS. 2000. Indigenous natural en-
emies associated with Phyllocnistis citrella
(Lepidoptera: Gracillariidae) in Eastern Spain. Biol.
Cont. 18: 199-207.
VAN DRIESCHE, R. G. 1983. Meaning of percent parasit-
ism in studies of insect parasitoids. Environ. Ento-
mol. 12: 1611-1621.
CIA MARI. 1995. Parasitoides aut6ctonos del minador
de hojas de citricos Phyllocnistis citrella en las co-
marcas centrales valencianas. Levante Agricola/ 4
Trimestre, 306-312.
MARZAL, AND C. GRANDA. 2000. Importaci6n y es-
tablecimiento de parasitos del minador de hojas de
citricos Phyllocnistis citrella (Lepidoptera: Gracilla-
riidae). Boletin de Sanidad Vegetal, Plagas, 26: 577-
nador de la hoja de los citricos, Phyllocnistis citrella
en el NOA. Revista Avance Agroindustrial 16: 15-20.
2002. Introducci6n de Citrostichus phyllocnistoides
para el control biol6gico del minador de los citricos
en el NOA. Avance Agroindustrial, 23: 17.

Florida Entomologist 89(3)

September 2006


1Invasive Plant Research Laboratory, USDA-ARS, 3225 College Ave, Ft. Lauderdale, FL 33314

2Division of Plant Industry, Department of Agriculture and Consumer Services, 1911 34 St. Gainesville, FL 32614

3Entomology and Zoology Group, Plant Protection Research and Development Office, Department of Agriculture,
Chatuchak, Bangkok 10900, Thailand

4University of Florida, Institute of Agricultural Sciences, 3225 College Ave., Ft. Lauderdale, FL 33312


In an attempt to find potential biological control agents of the lobate lac scale (Paratachar-
dina lobata), an important pest in southern Florida and the Bahamas, we made collections
of the commercial lac scale (Kerria lacca) in northern Thailand. Four species of parasitoids
and two species of predaceous moths were reared from K. lacca infestations on twigs im-
ported into Florida quarantine. None of the parasitoids accepted P. lobata as a host. Parasi-
toids of P. lobata from India or Sri Lanka, the native home of this scale, probably have more
promise as potential biological control agents of this pest.

Key Words: biological control, Aphelinidae, Blastobasidae, Encyrtidae, Eulophidae, Eupelm-
idae, Noctuidae


En un intent de encontrar agents de control biol6gico potenciales para la escama lobada
de laca (Paratachardina lobata), una plaga important en la Florida meridional y las Baha-
mas, se obtuvieron colecciones de la escama de laca commercial (Kerria lacca) en el norte de
Tailandia. Cuatro species de parasitoides y dos species de polillas predadoras fueron cria-
das de K. lacca infestando ramas pequenas importadas a la cuarentena en la Florida. Nin-
guno de los parasitoides aceptaron P. lobata como hospedero. Parasitoides de P. lobata de la
India o Sri Lanka, la region native de la escama, probablemente tienen mas promesa como
agents de control biol6gico potenciales de esta plaga.

Translation by authors.

The lobate lac scale insect, Paratachardina lo-
bata (Chamberlin) (Hemiptera: Coccoidea: Kerri-
idae), native to India and Sri Lanka (Varshney
1976a), infests more than 300 economic and na-
tive plants in southern Florida and the Bahamas
(Howard et al. 2002; Pemberton 2003a, b; Howard
& Pemberton unpublished data). This recently in-
troduced pest continues to spread in Florida, is
present in the Bahamas, and could threaten culti-
vated and native vegetation in the West Indies
and Mexico as well as subtropical regions such as
Texas, California, and Hawaii should it make its
way to these areas (Pemberton 2003a; Howard &
Pemberton 2004).
Chemical control with imidacloprid and
bifenthrin can successfully control this pest but is
too expensive and/or inappropriate for large scale
usage in natural areas (Howard & Steinberg
2005). Other control methods are needed and bio-

logical control is a good possibility. Biological con-
trol has a high success rate against scale insects
(Kennett et al. 1999), and is thought to be suitable
for the lobate lac scale because this insect is non-
native and taxonomically isolated from native
scales and other Hemiptera in the region (Pem-
berton 2003a). Parasitoids and other enemies of
this pest are absent or rare in Florida (Pemberton
2003a; Schauff 2005; Howard unpublished data).
One of the approaches used in the program to de-
velop biological control for P lobata was to ac-
quire and evaluate the parasitoids of the commer-
cial lac scale, Kerria lacca (Kerr), which is also in
the Kerriidae. This approach was considered be-
cause the lobate lac scale has been difficult to find
and access in its native region. A two and a half
week survey in the type locality ofP. lobata in Sri
Lanka during Mar 2003 failed to locate the insect,
much less its enemies (Pemberton, unpublished

Pemberton et al.: Host Acceptance Trials of Kerria Lacca (Kerriidae) Parasitoids 337

data). Varshney, the Indian monographer of the
Kerriidae, has not seen the lobate lac scale in the
field in India, where it is assumed to be rare (Pem-
berton 2003a). In contrast, K. lacca and its rela-
tively well-known parasitoids (Narayanan 1962;
Bhattacharya 2002) were thought to be more ac-
cessible and some parasitoids of K. lacca also
have been recorded from P lobata (Varshney


Field Collection

Parasitoids of K lacca were first sought
through the Indian Lac Scale Institute in Ranchi,
Jarkhand, which has studied these pests of lac
scale culture for many years and was willing to co-
operate (Pemberton 2003a). However, this poten-
tial cooperation was stymied by the lack of en-
abling approvals from the Indian bureaucracy
above the Institute. Lac scale culture was subse-
quently sought outside of India and found in
northern Thailand. Preliminary surveys and a
collection ofK. lacca was made (by AW) in Ampur
Mae Tha in Lampang Province on Sep 20, 2003.
Two parasitoids emerged from the K lacca
brought to Bangkok and held. This indication of
the presence of parasitoids prompted additional
collections, which were made (by RWP and AW) at
the following five sites in Lampang Province on
Oct 13 and 14, 2003. Site 1, Oct 13, Moo 12, Baan
Look, Nakrou, Mae Tha (N1807' 69.1", E 99033'
57") lac cultivated on rain trees, Albizia saman
(Jacq.) F. Muell. (Fabaceae). Site 2, Oct 13, Moo 8,
Baan, Mae Tha (N 18007' 97.9", E 99033' 90.2"),
lac cultivated on rain trees. Site 3, 14 Oct, Wongk-
longsak farm near Ampur Jae Hom (N 18045' 27",
E 99034' 11") lac cultivated on rain trees. Site 4,
Oct14, Baan Sasobhok Tambon Bansa, Ampur
Jae Hom (N 18o 39' 14.1", E 99 32' 46.9") wild in-
festation of a longan tree (Dimocarpus longan
Lour., Sapindaceae) in yard of house in town. Site
5, Oct 14, Ban Toongkla, Tambon Maesook, Am-
pur Jae Hom (N 18048' 419", E 99034' 40.8") lac
scale was semiwild on large rain trees at edge of
town. At all sites, infested twigs were cut from the
trees and placed in cotton bags and hand-carried
to the Florida State Biological Quarantine in
Gainesville, and held for emergence. An addi-
tional collection was made again from the same
sites in Ampur Mae Tha (by AW) Mar 10-11, 2004
and shipped to the Florida quarantine.

Quarantine Emergence and Tests

Infested twigs were held in plastic boxes (35 x
22.5 x 15 cm) with screen ventilation ports.
Emerging parasitoids were sorted to species, held
in vials for a few hours and fed with honey, then
transferred to screen cages (0.6 x 0.6 x 0.9 m) con-

training 1- and 3-gal (3.8 andll.4 L) containers of
wax myrtle (Myrica cerifera L., Myricaceae) or
Inga edulis Mart. (Fabaceae) infested with P. lo-
bata. These plants were infested with P lobata
(by FWH) by placing them beneath and among
heavily infested trees at Secret Woods County
Park in Broward County, FL, usually for one-
month periods. Mixed stages, including crawlers,
young and mature scales were present at the time
of the tests. Honey and water were provided via
cotton wicks in each cage, which were in a glass-
house with mean ambient temperatures about
27C. Identification of parasitoids associated with
the K. lacca collections were made by Greg Evans
of the Florida Division of Plant Industry and
Michael Gates of the USDA-ARS Systematic En-
tomology Laboratory in Beltsville, MD, both aided
by specimens previously determined by the In-
dian Lac Research Institute (obtained by RWP).


Four parasitoid species of K. lacca emerged
from the collections (Table 1), including two pri-
mary parasitoids, Tachardiaephagus tachardiae
Howard (Encyrtidae) and Coccophagus tschirchii
Madhihassen (Aphelinidae) (Narayanan 1962;
Varshney 1976b). Reasonably large enough num-
bers of T tachardiae (119 from the Oct collections
and 77 from the Mar collection) emerged, which
enabled a good acceptance test against P lobata.
An adequate test probably did not occur with C.
tschirchii because only 8 individuals emerged
from the Oct collection and none from the Mar col-
lection. The other two parasitoids emerging from
K. lacca were Aprostocetus purpureus Cameron
(Eulophidae) and Eupelmus tachardiae Howard
(Eupelmidae), both of which can be either pri-
mary parasitoids of K. lacca or hyperparasitoids
(Narayanan 1962; Varshney 1976b).Aprostocetus
purpureus can be a secondary parasitoid of T ta-
chardiae or C. tschichii, while Eupelmus tachar-
diae can be a secondary parasitoid of braconid
parasitoids of the predaceous lac moths. Large
numbers ofA. purpureus emerged from both col-
lections (60 from the Oct collection and 866 from
the Mar collection) to have allowed for a good ac-
ceptance test. This was not true for E. tachardiae
because only 7 individuals were obtained from the
Oct collection and none from the Mar collection.
These species were tested after the first emer-
gence before the identifications were obtained
and their role as hyperparasitoids was recog-
nized. Aprostocetus purpureus from the Mar col-
lection was tested again to try to learn if it would
accept P lobata from which it has been recorded
(as Tetrastichus purpureus) (Varshney 1976b).
None of the parasitoids showed any behavioral
orientation (attraction, attenuation, host feeding,
or oviposition) to exposed lobate lac scales and no
parasitism occurred in the Florida quarantine. No

Florida Entomologist 89(3)

emergence holes were detected in P lobata and no
wasp adults were found. Random dissections of
more than 120 P lobata scales exposed to T ta-
chardiae and A. purpureus were made in late
May, after the second shipment and exposure pe-
riod, but no evidence of parasitoid attack was
found. Two predaceous moths of K lacca, Eub-
lemma roseoniuia Walker (Noctuidae) and Holco-
cera pulverea Meyr. (Blastobasidae), and their as-
sociated parasitoids, Brachymeria tachardiae
Cameron (Chalcididae) and Elasmus claripennis
Cameron (Elasmidae), also emerged from the col-
lected K lacca colonies (Table 1). Eupelmus ta-
chardiae may have been associated with the lac
moths as hyperparasitoids of their braconid wasp
parasitoids (Table 1).


It is unclear why neither T tachardiae nor A.
purpureus would accept P lobata in our tests.
Paratachardina lobata (subfamily Tachardini-
nae) and K lacca (subfamily Tachardiinae) are
actually not very closely related within the Kerri-
idae (Varshney 1976a) and may not be susceptible
to the same parasitoids. The records of T tachar-
diae and A. purpureus parasitizing P lobata may

be based on misidentifications of the parasitoids
or the hosts. Perhaps the form of P lobata adven-
tive in Florida differs in acceptability compared to
the forms of P lobata in India, or our rearing con-
ditions were in someway inadequate, such as the
number of suitable stages present in all cages.
The host plant species used by an herbivorous in-
sect can influence the level of attack by its parasi-
toids (Werren et al. 1992). Perhaps the host plants
used to culture P lobata for the quarantine test-
ing (Myrica cerifera and Inga edulis) may have
been factors inhibiting parasitism.
The predaceous moths, Eublemma roseoniuia
and Holcocera pulverea, which are known to be
among the most important mortality factors of K
lacca cultivation in India (Narayanan 1962), were
considered and rejected for biological control forP.
lobata (Pemberton 2003a). They lack specificity
and during their predation of K lacca destroy
large numbers of immature parasitoids (Bhatta-
charya 2002). The 25 moths (21 E. roseniuia and 4
H. pulverea) reared from our Oct collections prob-
ably greatly reduced the number of emerging par-
asitoids. All members of this K lacca enemy com-
plex are known from India but apparently not
previously recorded in Thailand where K. lacca is
cultivated on a small scale and not previously


Number emerged
Families and species Tropic roles' Total (Oct & Mar)

Coccophagus tschirchii Madhihassen Primary parasitoid of K. lacca 8 (8 & 0)
Brachymeria tachardiae Cameron Primary parasitoid of Eublemma and/or Holcocera 4 (4 & 0)
Elasmus claripennis (Cam.) Primary parasitoid of Eublemma and Holcocera 21 (21 & 0)
Tachardiaephagus tachardiae Howard Primary parasitoid of K. lacca 196 (119 & 77)
Aprostocetus purpureus Cameron Primary parasitoid of K. lacca
Secondary parasitoid of C. tsc92 (60 & 866)
Secondary parasitoid of C tschirchii
Secondary parasitoid of T. tachardiae
Eupelmus tachardiae Howard Primary parasitoid of K. lacca
Secondary parasitoids of braconids of Eublemma 7 (7 & 0)
and/or Holcocera
Holcocera pulverea Meyr. Predator of K. lacca 4 (4 & 0)
Eublemma rosenovia Walker Predator of K. lacca 25 (21 & 4)

Roles from literature (Narayanan 1962; Varshney 1976b; Battacharya 2002).

September 2006

Pemberton et al.: Host Acceptance Trials of Kerria Lacca (Kerriidae) Parasitoids 339

studied. Aprostocetus purpureus and T tachar-
diae were the most abundant of the parasitoids
attacking K lacca in our collections and are also
known to be the most abundant parasitoids of K
lacca in India (Bhattacharya 2002).
Additional collection and screening ofK. lacca
parasitoids against P. lobata could be worthwhile.
As mentioned above, the number of C. tschirchii
obtained were not thought be high enough to al-
low for an adequate acceptance test. This parasi-
toid was also one of the hoped for natural enemies
in the collections because it is also known only
from lac scales and is not known to be hyperpara-
sitic (Narayanan 1962; Varshney 1976b). Another
primary parasitoid of K lacca, Tachardiaphagus
somervilli Madhihassen (Encyrtidae), that may
have the ability to attack P lobata, could also oc-
cur in Thailand. Perhaps these parasitoids could
be obtained by collections at different times of the
year and from different sites. Our Oct and Mar
collections had large differences in parasitoid
richness even at the same site (Ampur Mae Tha),
although this may have been due to the very high
populations of the hyperparasitoid A. purpureus
during Mar (866 emerged from the collection). Ac-
quiring and evaluating the parasitoids of P lobata
itself will probably bring more success.


Greg Evans, Florida Division of Plant Industry
(Florida Department of Agriculture and Consumer Af-
fairs) and Michael W. Gates of USDA-ARS Systematic
Entomology Laboratory, Beltsville, MD identified the
parasitoids. The Indian Lac Scale Institute, Ranchi,
Jarkhand, India provided specimens which aided in the
identification of the parasitoids. John Hepner, Florida
Division of Plant Industry, determined the moths. This
research was supported in part by a grant from the Flor-
ida Division of Plant Industry to Robert Pemberton.

BHATTACHARYA, A. 2002. Lac insect and associated in-
sect fauna, pp. 97-103 In K. K. Kumar, R. Ramani
and K. K. Sharma [eds.], Recent Advances in Lac

Culture. Indian Lac Institute, Ranchi, Jarkhand, In-
dia. 290 pp.
FORD. 2002. Lobate lac scale, Paratachardina lobata
lobata (Chamberlin) (Hemiptera: Sternorrhyncha:
Coccidea: Kerriidae). University of Florida Featured
Creatures (1 Dec. 2002, http://creatures.ifas.ufl.edu/
orn/scales/lobate lac.htm.
HOWARD, F. W., AND R. W. PEMBERTON. 2004. The lo-
bate lac scale, a new insect pest of trees and shrubs
in Florida: implications of the Caribbean Region.
Proc. Caribbean Soc. Food Crops Soc. 38: 91-94.
drenches and topical insecticide treatments for con-
trol of the lobate lac scale, Paratachardina lobata
(Chamberlin). Proc. Florida State Hort. Soc. 118:
SELY. 1999. Biological control in subtropical and
tropical crops, pp. 713-742 In T S. Bellows and T. W.
Fisher [eds.], Handbook of Biological Control. Aca-
demic Press, San Diego. 1046 pp.
NARAYANAN, E. S. 1962. Pests of lac scales in India, pp.
90-113 In B. Mukhopadhay and M. S. Muthana
[eds.], A Monograph on Lac. Indian Lac Research In-
stitute, Ranchi, Bihar, India. 378 pp.
PEMBERTON, R. W. 2003a. Potential for biological con-
trol for control of the lobate lac scale, Paratachar-
dina lobata lobata (Chamberlin) (Hemiptera:
Kerridae). Florida Entomol. 86: 354-361.
PEMBERTON, R. W. 2003b. Invasion of Paratachardina
lobata lobata, (Hemiptera: Kerriidae) in South Flor-
ida: a snapshot sample of an infestation in a residen-
tial yard. Florida Entomol. 86: 374-378.
SCHAUFF, M. E. 2005. Ammonoencyrtus carolinensis, n.
comb. (Hymenoptera: Encyrtidae), a parasite of lo-
bate lac scale Paratachardina lobata (Chamberlin)
(Hemiptera: Kerriidae). Proc. Entomol. Soc. Wash-
ington 107: 115-118.
ODELL. 1992. Host plants used by gypsy moths affect
survival and development of the parasitoid Cotesia
melanoscela. Environ. Entomol. 21: 173-177.
VARSHNEY, R. K. 1976a. Taxonomic Studies on Lac In-
sects in India. Oriental Insects Supplement 5: 1-
VARSHNEY, R. K. 1976b. A check-list of insect parasites
associated with lac. Oriental Insects 10: 55-78.

Florida Entomologist 89(3)

September 2006


'Center for Biological Control, Florida A&M University, Tallahassee, FL

2USDA-APHIS-PPQ-CPHST, at Center for Biological Control, Florida A&M University, Tallahassee, FL

3USDA-ARS-CMAVE, at Center for Biological Control, Florida A&M University, Tallahassee, FL



The oligophagous cactus moth, Cactoblastis cactorum (Berg), has been recognized as a seri-
ous and immediate threat to Opuntia cacti in Florida and the southeastern United States.
The moth has successfully colonized new geographical ranges with lower annual tempera-
tures north of the Florida Keys where it was first detected in the continental United States
in 1989. This study evaluated the effect of temperature on egg development and egg hatch
of C. cactorum by utilizing various treatment temperatures, exposure times, and egg ages.
The temperatures used in this study ranged from a low of -20 C to a high of 50 C, thus en-
compassing the potential range of temperatures that eggsticks may be exposed to in poten-
tial new host areas. One-d-old eggs held at a constant temperature of 30 C resulted in the
highest percent hatch and shortest time to egg hatch. Eggs did not hatch when held at con-
stant temperatures <15 C or >35 C. Furthermore, one d of exposure at -10 C and 4 d of ex-
posure at -5 C were 100% lethal to one-d-old eggs. Eggs that were 7- and 14-d-old before
exposure to cold temperatures were generally more resistant to temperature effects than
one-d-old eggs.

Key Words: cactus moth, invasive species, insect development


La polifaga palomilla del cactus, Cactoblastis cactorum (Berg), ha sido reconocida como una
amenaza seria e inmediata a las cactaceas del g6nero Opuntia presents en el estado de Flo-
rida y en el sureste de los Estados Unidos de Am6rica. La palomilla ha sido exitosa en la co-
lonizaci6n de nuevas areas geograficas que poseen temperatures mas bajas que el area en
los Cayos de Florida donde la plaga fue detectada por primera vez en 1989. En este studio
se evaluaron los efectos de temperature sobre el desarrollo de huevecillos y la eclosi6n de
neonatos de la palomilla del cactus. Los huevecillos fueron expuestos a varias temperatures
por periods de tiempo diferentes y se realizaron experiment con huevecillos de distintas
edades. Las temperatures utilizadas en este studio variaron desde una temperature baja
de -20 C a una temperature alta de 50 C, rangos que abarcan las temperatures a las que los
huevecillos podrian estar expuestos si la palomilla coloniza nuevas areas. Los huevecillos de
un dia de edad que se mantuvieron a una temperature constant de 30 C mostraron el por-
centaje de eclosion mas alto y el tiempo de eclosion mas corto. Los huevecillos no tuvieron
emergencia de neonatos cuando fueron expuestos a temperatures de 35 C o 15 C. Asimismo,
no se obtuvo emergencia de neonatos cuando los huevecillos fueron expuestos a una tempe-
ratura de -10 C por un dia o a una temperature de -5 C por 4 dias. Los huevecillos de siete
y catorce dias de edad que fueron expuestos a temperatures bajas se mostraron mas resis-
tentes que los huevecillos de un dia de edad.

Translation by authors

The cactus moth, Cactoblastis cactorum (Berg) transported around the world to control exotic
(Lepidoptera: Pyralidae), was first recorded in pestiferous Opuntia spp. in Australia, South Af-
North America in the Florida Keys in 1989 (Ha- rica, Hawaii, and numerous other countries (Zim-
beck & Bennett 1990; Stiling & Moon 2001). This mermann et al. 2000). The moth was introduced
insect feeds on prickly pear cactus (Opuntia spp.) to Nevis in the Caribbean in 1957 to control na-
and is native to South America, but has been tive Opuntia spp. that were becoming weedy on

Mclean et al.: Temperature effect on C. cactorum Eggs

disturbed pasture land (Simmons & Bennett
1966). Heppner (2000) documents the spread of C.
cactorum through the Caribbean Islands aided by
deliberate introductions and natural dispersion.
Adult cactus moths mate in the early morning
(Hight et al. 2003) and oviposition on cactus usu-
ally begins the following evening. Eggs are laid on
top of one another forming eggsticks that are nor-
mally attached to cactus spines or directly to cac-
tus pads. Larvae emerge synchronously and to-
gether penetrate the thick outer layer of the host
plant, then move to feed inside cactus pads (Rob-
ertson 1989). After a pad has been completely con-
sumed, larvae tunnel to other pads through the
nodes or leave the destroyed pad to seek other un-
damaged host material. Larvae may inhabit all
portions of the cactus plant, including roots, and,
with the aid of associated rot organisms, the en-
tire cactus plant may be killed (Sweetman 1958).
Dodd (1940) and Pettey (1948) documented the ef-
fect of field temperatures on the phenology, fecun-
dity, and life cycle of C. cactorum in Australia and
South Africa, respectively. Dodd (1940) reported
that eggsticks and pupae have high mortality
when exposed to "unseasonably" high or low tem-
peratures. Larvae faired better because they were
able to move within the plant and thus find pro-
tection within the cactus.
Eggs appear to be the most vulnerable stage
in the C. cactorum life cycle. Eggsticks of C. cac-
torum take approximately one month to hatch in
the field, and are afforded little protection by the
cladodes and spines from prevailing environ-
mental conditions such as high or low tempera-
tures. Larvae are internal feeders and pupae are
found in leaf litter on the ground. The purpose of
this study was to evaluate the effects of high and
low temperatures and exposure times on C. cac-
torum egg development and hatch. Knowledge of
these temperature/time thresholds will become
increasingly important as C. cactorum continues
to expand its range in North America where av-
erage temperature variations are more extreme
than in its current range. Furthermore, this in-
formation will be helpful in streamlining control
strategies by enabling the focus of attention on
areas where the moth is most likely to colonize
and survive.


Eggsticks of C. cactorum were obtained from a
colony maintained at the USDA-ARS Crop Pro-
tection and Management Unit laboratory in Tif-
ton, GA. Larvae were reared on cladodes of Opun-
tia ficus-indica (L.) Miller inside rectangular
plastic boxes that were maintained at 26C + 1C,
a photoperiod of 14:10 (L:D), and 70% RH. As lar-
vae matured, cocoons were collected every 2-3 d
from the containers. Pupae were extracted from
the cocoons and placed in a screen cage (30.5 x

30.5 x 30.5 cm) where eclosion, mating, and ovipo-
sition occurred. Newly laid eggsticks (0-24 h old)
were collected daily and transported in a small
cooler (temperature approximately 4C) to the
USDA-ARS Center for Medical, Agricultural, and
Veterinary Entomology laboratory in Tallahassee,
FL, where the studies were conducted. The num-
ber of eggs per eggstick was determined under a
dissecting microscope. Individual eggsticks were
placed in a 30-ml plastic cup with a cardboard lid
(Jet Plastica, Hatfield, PA).

Low Temperature Effect Studies

Different aged eggsticks (1, 7, and 14 d) were
exposed to several treatment temperatures for
various lengths of time and assessed for egg
hatch. Eggsticks aged for seven and 14 d were
held at the control temperature prior to exposure
to treatment temperatures. Ten replicates (1 egg-
stick = 1 replicate) were completed at each treat-
ment temperature and exposure time. Treatment
temperatures were -10, -5, 0, and 5C, while expo-
sure times were 1, 2, 3, 4, 5, 10, 20, 40, and 60 d
(Table 1). Thermo-Forma growth chambers
(Thermo Electron Corporation, Marietta, OH)
used for the different temperature treatments
were programmed with a photoperiod of 0:24
(L:D) and 40-70% RH. Temperature and humidity
inside each chamber were verified with HOBO
data loggers (Onset Computer Co., Bourne, MA).
After the predetermined length of exposure time
at each treatment temperature, replicates were
returned to the control chamber where they were
checked daily for egg hatch. Date of first egg hatch
was recorded and newly emerged larvae were
carefully removed daily with a small brush. Egg-
sticks were inspected under a dissecting micro-
scope to determine when egg hatch was com-
pleted. The percent egg hatch, the time and dura-
tion of egg hatch, and the time to complete egg
hatch were calculated. Eggsticks that did not
hatch after 60 d were discarded.

Developmental Studies

Ten replicates of 1-d-old eggsticks (1 eggstick =
1 replicate) were completed at each constant
treatment temperature with a maximum expo-
sure time of 60 d or until egg hatch. Eggs that did
not hatch after 60 d were discarded. Treatment
temperatures were 10, 15, 20, 25, 35, 40, and 50C
(Table 1), photoperiod 0:24 (L:D), and 40-70% RH.
Replicates were inspected daily and the percent
egg hatch, time and duration of egg hatch, and
time to complete egg hatch were calculated as de-
scribed for the low temperature effect study
above. Data from a growth chamber set at 14:10
(L:D), 30C, and 40-70% RH served as the control
for both the low temperature effect study and the
developmental study.

Florida Entomologist 89(3)


temperature (C)

Egg age
at time of setup (d)

Number of days
at treatment temperature

Number of days at
30C control temperature

Development Studies


Effect of Low Temperature Exposure Studies



Data collected from developmental studies in-
volving 1-d-old eggsticks held at constant treat-
ment temperatures were analyzed by a two-factor
analysis of variance (ANOVA), with treatment
temperature and exposure time as sources of vari-
ation (PROC ANOVA and PROC GLM) (SAS In-
stitute 1989). Data collected from low tempera-
ture effect studies where 1-d-old eggsticks were
exposed to treatment temperatures for a pre-
scribed exposure time and then returned to con-
trol conditions were analyzed by ANOVA and re-

gression analysis, with exposure time and treat-
ment temperature as sources of variation (PROC
ANOVA and PROC GLM) (SAS Institute 1989).
Data collected from low temperature effect stud-
ies where eggsticks were aged for a prescribed pe-
riod under control conditions, exposed to treat-
ment temperatures for 24 h, and then returned to
control conditions, were analyzed by a three-fac-
tor ANOVA and regression analysis, with expo-
sure time, treatment temperature, and age as
sources of variation (PROC ANOVA and PROC
GLM) (SAS Institute 1989). Exposure time in the
statistical model denotes the length of time (in

30 (control)

September 2006

Mclean et al.: Temperature effect on C. cactorum Eggs

days) that the eggsticks were subjected to treat-
ment temperatures. The statistical model in- a
eluded the following dependent variables: percent I
hatch of eggsticks, time (in days) to first egg
hatch, mean days to egg hatch, and time (in days)
to last egg hatch. Percent egg hatch was calcu-
lated by comparing the total number of neonates -
with the total number of eggs in the eggstick. All g
other dependent variables were calculated from +
the date of oviposition. When the statistical model 0 <1
indicated significant treatment effects and signif-
icant interactions, differences among means were g
separated by the Tukey-Kramer statistic (P < 0
0.05) for multiple comparisons. Data met the as-
sumptions of ANOVA and were not transformed. + +1

Low Temperature Effect Studies P4 c -

Analysis of data obtained when 1-d-old egg-
sticks were exposed to treatment temperatures E H
for different exposure times revealed a significant
(F = 4.71; df = 6, 150; P = 0.0002) interaction be- 5 o
tween treatment temperature and exposure time 2 -a
for the mean number of days to first egg hatch, as g ,
well as a significant (F = 19.28; df = 10, 219; P <
0.0001) interaction between treatment tempera- 3
ture and exposure time for the mean percent &
hatch. Generally, with increasing exposure time
for each treatment temperature, time to egg a
hatch increased while percent hatch decreased c d
(Tables 2 and 3).
One-d-old eggsticks exposed to a treatment <1 H A
temperature of -10C for 24 h did not hatch when o
returned to control conditions, indicating that -10C E a
is lethal to young eggsticks (Table 3). In addition, 0 0
there was no egg hatch in 1-d-old eggsticks held at C1 0i
C +1 +1
-5C for 3 d. The percent egg hatch at all exposure E- r
time periods at treatment temperature -5C were z z6 c
significantly lower than the control. At this tem-
perature, exposure time sharply reduced the per- c
cent egg hatch as exposure time increased from 1 E .
to 3 d and resulted in 0% egg hatch of 1-d-old egg- E dd si
sticks exposed for 4 d. +1 +1 +1
At 0C and 5C, percent egg hatch was not signif- 1 4 zici + '$
icantly reduced until exposure times were greater q z 0
than 5 and 10 d, respectively. No egg hatch occurred C P E-
at 0C after 20 d exposure; no egg hatch occurred at P 0 H C
5C after 40 d (Table 3). It also appeared that no egg E 4 -
development occurred at O0C and 5C, and the eggs < P oc oc oo cc
that were exposed to these low temperatures essen- 0 00
tially took that many extra days to hatch when re- p P
turned to the control growth chamber, e.g., eggs took 9 J
approximately 21 d to hatch at 30C and 31 d to 0
hatch (10 d longer) when they were exposed for 10 d u
at 5C and then returned to the control chamber at o
30C (Table 2).
An analysis of exposure time on mean time to CN
egg hatch and mean hatch percent across treat-
ment temperatures showed that there were no sig- < H

Florida Entomologist 89(3)



la" ^

E0 --

+ +1 +1
8|^ 2

^|ig C00 i"
& Ez- -a .*

ul a)
m Ct

03 +1 +1
(3. 1 0 1 5
E-o oi

i Cr Q;

0 ,
0 o

E- C 1q,
S ,Q S B ^+1 +
^^ Q 1

E- A
E- E
piY I I

3 ~ ~ ~ - S -^ t
1^^O a B

1 a +1I
c t


0 + 1 +1 o

Swi i It J

ul cc cc _v

+1 +1 +1 +1

<1 ^^1

0cC~ C
h w S.01fl01f A
H g 0o H$
S^ m^` g
III+ +I +| | 1
^1 hj P

m a +1 +1 +1 g: m
2<^ TO ^-l 3

23 ('uo S
Wp ^<^ |
cr ^ ?"""? ^
pug ^ occco ^
w fcsri + + 1 1

a S s a;
^1^ 2 u u
H S-aS ,

September 2006

nificant differences between percent hatch of eggs
held at -5C, 0C, and 5C after 1 d of exposure.
There also were no significant differences in mean
percent hatch of eggs held at 0C and 5C after 5 d
of exposure. However, the percent egg hatch signif-
icantly increased when eggs were held at 0C
rather than -5C for both 2 and 3 d of exposure.
Percent egg hatch significantly decreased for egg-
sticks held for 10 d exposure at 0C when com-
pared to 5C. Eggs took significantly longer to
hatch at colder temperatures when exposed for 1 d
at treatment temperatures of -5C compared to
5C, and 10 d at 0C compared to 5C.
When the data for aged eggsticks were ana-
lyzed, there was a significant (F = 3.28; df = 2,
178; P = 0.0402) interaction between age (1, 7, or
14 d old) and exposure time (0 or 1 d) on the mean
number days to hatch. However, there was no sig-
nificant low temperature effect (-5, 0, or 5C),
therefore data from all temperature treatments
were pooled. For each eggstick age, exposure to
low temperature for 1 d significantly increased
the mean time to egg hatch when compared to the
control (exposure time of 0 d) (Table 4).
There were significant (F = 6.64; df= 2, 179; P
= 0.0017) interactions between temperature and
exposure time, and age and exposure time (F =
4.28; df = 2, 179; P = 0.0154) on the percent hatch
of eggsticks aged for 1, 7, and 14 d. Analysis of the
temperature/exposure time interaction reveals
that at -5C, 1 d of exposure significantly de-
creased percent egg hatch compared to the control
(0 d of exposure). Additionally, 1 d of exposure at
-5C significantly reduced egg hatch from that ob-
served at 0C and 5C for the same exposure time
(Table 5). Analysis of the egg age/exposure time

14:10 (L:D), AND 40-70% RH; AFTER EGGS

Time to egg hatch (d S.D.)'

24 h exposure to
Egg age (d) Control (30C)2 low temperature3

1 20.8 1.0 Aa 23.2 1.5 Ba
7 20.8 1.0 Aa 22.4 0.4 Bb
14 20.8 1.0 Aa 22.5 0.4 Bab

Means within each row followed by the same uppercase let-
ter are not different (P > 0.05); means within each column fol-
lowed by the same lowercase letter are not different (P > 0.05).
Data from a single control was used for comparison with all
'Statistical analysis indicated no effect due to different low
temperatures (-5C, 0C and 5C) and as a result data were
pooled across all temperature treatments.

Mclean et al.: Temperature effect on C. cactorum Eggs

Interaction (Table 5) shows that 1- and 7-d-old
P eggsticks had significantly lower percent hatch
When compared to the control, but 14-d-old egg-
+1 = sticks were not significantly different from the
cc 1 control. Additionally, 14-d-old eggsticks had a sig-
Soo U nificantly higher percent hatch when compared to
< o cc 1- and 7-d-old eggsticks after 1 d of exposure to
E low temperatures. However, at treatment temper-
S. atures of 0C and 5C there were no significant
difference in percent egg hatch when compared to
P E cthe control.
b o+o + Developmental studies
+1 +i
4 4 ^One-d-old eggsticks of C. cactorum exposed to
o I t constant treatment temperatures failed to hatch
Z at temperatures <15C or >35C and appeared to
be desiccated. The highest percent hatch and
shortest time to first egg hatch was obtained at
.- 30C (Table 2). At 30C, all eggs hatched within 4
Sm d of the initial hatch, with >50% occurring on the
c +1 +1 2 first d; while at 20C, hatch duration was pro-
+1 o longed over a 9 d period with the highest number
S- emerging on d 2 (Table 6).


STo predict how far an invasive pest species will
Extend its range or how an existing species range
o C1 will respond to climate change, it is important to
E +1 understand what factors limit the range of that
E- E particular species (Baskauf & McCauley 2001).
So According to Pemberton (1995), Carpenter et al.
S(2001a), and Mahr (2001), it is likely that the geo-
Sgraphical range of C. cactorum in the United
States will be limited by environmental factors,
Such as temperature and photoperiod, rather
S" than by host availability. North American Opun-
S+1 +1 + tia species have a broad geographical range and
0 o C S are tolerant of temperatures as low as -30C and
Soo as high as 45C (Carpenter et al. 2001b). Larvae
of C. cactorum are polyphagous, feeding on nearly
o a all Opuntia species tested to date (Stiling 2002).
Sa g According to Hoffmann & Blows (1994) and
E z 8 Baskauf & McCauley (2001), when climate di-
S0.6 t o rectly influences the extent of a species' range,
m 6 ^ A Alimiting factors such as temperature may be
So fairly easily identified by testing whether the re-
| distance of susceptible stages to abiotic stresses
S(such as extreme temperatures) matches the level
< of stress at the range limit.
<1 SThis study evaluated the effect of various
Treatment temperatures on the percent egg hatch
S 8 and time to first egg hatch of 1-d-old and aged C.
Scactorum eggsticks under laboratory conditions.
SThe results obtained from the constant treatment
a temperatures show that there was no egg devel-
CC opment at 15C and 35C. These results suggest
That the lower and upper temperature thresholds
Sfor C. cactorum egg development would be 20C
S and 30C, respectively. However, the validity of

Florida Entomologist 89(3)


Total number of days during which egg hatch occurred

1 2 3 4 5 6 7 8 9

Temperature ( C) Days to 1st hatch Percent hatch per day Total hatch (%)

15 NH 0
20 40 15 43 22 5 1 1 1 1 1 90
25 26 32 22 17 10 10 1 92
30 20 54 33 7 2 96
35 NH 0
40 NH 0
50 NH 0

such a conclusion would be suspect based on the
inability of the growth chambers utilized in this
study to maintain constant temperatures. Analy-
sis of temperature data obtained from data log-
gers placed inside the growth chambers revealed
that temperature within the chambers fluctuated
between 2.5C of a prescribed constant tempera-
ture. As a result, eggsticks held at constant treat-
ment temperatures of 15C and 35C were period-
ically exposed to lower and higher temperatures
(12.5C and 37.5C, respectively). It is possible
that the eggsticks were killed by periodic expo-
sure to the temperature extremes above and be-
low the treatment temperatures. In order to accu-
rately determine the developmental threshold of
C. cactorum eggs, this study should be repeated
utilizing growth chambers with a narrower range
of temperature fluctuations. The highest percent
egg hatch and shortest time to egg hatch obtained
in this study was observed at 30C, suggesting
that this temperature was closest to the optimal
temperature required for C. cactorum egg devel-
opment. At treatment temperatures of 20C, egg-
sticks required twice as much time to hatch and
had significantly lower percent egg hatch. The
longer period required for neonate emergence at
20C could be detrimental to neonate survival un-
der field conditions. Entry into the cactus pad is
thought to be achieved by the neonates chewing
gregariously on the same spot in order to pene-
trate the cuticle and/or overcome the plants de-
fenses (Pettey 1948). As such, when the period of
emergence is staggered, as occurs at 20C, there
may be too few neonates to successfully enter the
cactus pad, resulting in increased mortality due
to predation or eventual desiccation of the young
Percent egg hatch for 1-d-old eggsticks held at
low temperatures and returned to control condi-
tions showed an inverse relationship with the
length of exposure time at the low temperature.
The longer eggsticks were held at the low temper-
ature, the lower the percent egg hatch. At -5C,

percent egg hatch was low after 1 d and was fur-
ther reduced until no eggs hatched at 4 d of expo-
sure time. At -10C, 1-d-old eggsticks did not sur-
vive even after 1 d of exposure time. These results
suggest that 1-d-old eggsticks are highly suscepti-
ble to temperatures below 0C and that significant
mortality can be expected if 1-d-old eggsticks are
exposed to below freezing temperatures in the
field. At the 5C treatment temperature, percent
egg hatch was not significantly different from that
observed in the control until exposure time
reached 20 d. In addition, no egg hatch occurred
when eggsticks were exposed to 5C for 40 d. Time
to egg hatch showed a direct relationship with ex-
posure time. Longer times were required for egg
hatch as exposure time to low treatment tempera-
tures was increased. Eggsticks required twice the
number of days to hatch as compared to the control
when they experienced 20 d of exposure to 5C.
Eggsticks aged for 7 and 14 d before treatment
generally had higher percent egg hatch when
compared to 1-d-old eggsticks exposed to similar
temperature regimes. This would suggest that
older eggsticks are less susceptible to the effects
of low temperature under laboratory conditions.
However, additional studies are needed to deter-
mine if these results would be similar under field
The process of dispersal is a key feature in the
success of most invasive species. The ability of a
species to direct itself or its offspring to favorable
new habitats is a prerequisite for attaining wide-
spread ecological dominance (Krushelnycky et al.
2003). As such, understanding factors such as the
effect of temperature on rate of development that
ultimately influences the distribution and abun-
dance of an insect is a fundamental issue of insect
ecology (Andrewartha & Birch 1954) and is a
practical concern with insects like C. cactorum
that produce environmental and economic dam-
age (Baskauf 2003).
Data presented in this study offer valuable de-
velopmental information for C. cactorum egg-

September 2006

Mclean et al.: Temperature effect on C. cactorum Eggs

sticks under laboratory conditions and provide a
basis for future research projects designed to gain
a greater understanding of the development of C.
cactorum under laboratory and field conditions.
This information will become increasingly impor-
tant as C. cactorum continues to colonize new geo-
graphical regions. Many species with ranges that
extend over large geographic areas show regional
adaptations to climate and differentiate to the ex-
tent that separate populations are considered
ecotypes (Flint 1980). At this point it is unknown
whether such a scenario will occur with C. cac-
torum in North America, but, if it does, such a sit-
uation will have a significant impact on future
management strategies of this invasive species.


We thank the staff at USDA-ARS-CMAVE in Talla-
hassee, FL, especially Nathan Herrick, John Mass, and
Carla Evans, for assistance with monitoring experi-
ments. We also thank Susan Drawdy, Robert Caldwell,
and Robert Giddens at USDA-ARS-CPMRU, in Tifton,
GA, for providing technical assistance and eggsticks
from their colony of C. cactorum, and Richard Layton,
University of Georgia, for assistance with data analysis.
This research was supported by grants to Florida A&M
University from USDA-APHIS and USDA-CSREES.
Mention of trade names or commercial products in this
publication is solely for the purpose of providing specific
information and does not imply recommendation or en-
dorsement by the U.S. Department of Agriculture.


ANDREWARTHA, H. G., AND L. C. BIRCH. 1954. The dis-
tribution and abundance of animals. University of
Chicago. Press, Chicago, IL.
BASKAUF, S. J. 2003. Factors influencing population dy-
namics of the southwestern corn borer (Lepidoptera:
Crambidae): a reassessment. Environ. Entomol. 32:
BASKAUF, S. J., AND D. E. MCCAULEY. 2001. Evaluation
of low temperature mortality as a range-limiting fac-
tor for the southwestern corn borer (Lepidoptera:
Crambidae). Environ. Entomol. 30: 181-188.
Applications ofF1 sterility for research and manage-
ment of Cactoblastis cactorum (Lepidoptera: Pyral-
idae). Florida Entomol. 84: 531-536.
Inherited sterility in the cactus moth Cactoblastis
cactorum (Lepidoptera: Pyralidae). Florida Entomol.
84: 537-542.
DODD, A. P. 1940. The Biological Campaign against
Prickly Pear. Commonwealth Prickly Pear Board
Bulletin, Brisbane. 177 pp.
FLINT, M. L. 1980. Climatic ecotypes in Trioxys com-
planatus, a parasite of the spotted alfalfa aphid. En-
viron. Entomol. 9: 501-507.

HABECK, D. H., AND F. D. BENNETT. 1990. Cactoblastis
cactorum Berg (Lepidoptera: Pyralidae), a phycitine
new to Florida. Department of Agriculture & Con-
sumer Services Division of Plant Industries, Ento-
mology Circular 333.
HEPPNER, J. B. 2000. Tropical cactus borer, Cactoblastis
cactorum, intercepted in West Florida (Lepidoptera:
Pyralidae: Phycitinae). Lepidoptera News (Gaines-
ville) pp. 20-22.
TER 2003. Cactoblastis cactorum (Lepidoptera:
Pyralidae): Observations of courtship and mating
behaviors at two locations on the Gulf Coast of Flor-
ida. Florida Entomol. 86: 400-408.
HOFFMANN, A. A., AND M. W. BLOWS. 1994. Species bor-
ders: Ecological and evolutionary perspectives.
Trends Ecol. Evol. 6: 201-239.
ited spread ofunicolonial invasive insect and charac-
terization of seasonal patterns of range expansions.
Biological Invasions 6: 47-57.
MAHR, D. L. 2001. Cactoblastis cactorum (Lepidoptera:
Pyralidae) in North America: A workshop of assess-
ment and planning. Florida Entomol. 84: 465-473.
PEMBERTON, R. W. 1995. Cactoblastis cactorum (Lepi-
doptera: Pyralidae) in the United States: An immi-
grant biological control agent or an introduction of
the nursery industry? American Entomol. 41: 230-
PETTEY, F. W. 1948. The biological control of prickly
pear in South Africa. Science Bulletin (Department
of Agriculture of the Union of South Africa) 271: 1-
ROBERTSON, H. G. 1989. Seasonal temperature effects
on fecundity of Cactoblastis cactorum (Berg) (Lepi-
doptera: Pyralidae): differences between South Af-
rica and Australia. Entomol. Soc.of Southern Africa
52: 71-80.
SAS INSTITUTE. 1989. SAS User's Guide. SAS Institute,
Cary, NC.
SIMMONS, F. J., AND F. D. BENNETT. 1966. Biological
control of Opuntia spp. by Cactoblastis cactorum in
the Leeward Islands (West Indies). Entomophaga
11: 183-189.
STILING, P. 2002. Potential non-target effects of a bio-
logical control agent, prickly pear moth, Cactoblastis
cactorum (Berg) (Lepidoptera: Pyralidae), in North
America, and possible management actions. Biologi-
cal Invasions 4: 273-281.
STILING, P., AND D. C. MOON. 2001. Protecting rare
Florida cacti from attack by the exotic cactus moth,
Cactoblastis cactorum (Lepidoptera: Pyralidae).
Florida Entomol. 84: 506-509.
SWEETMAN, H. L. 1958. Interrelation of hosts and pests
and utilization in regulation of animal and plant
populations, pp. 429-499 In H.L. Sweetman, The
Principles of Biological Control. William C. Brown,
Dubuque, Iowa.
2000. The renowned cactus moth, Cactoblastis cac-
torum: Its natural history and threat to native
Opuntia floras in Mexico and the United States of
America. Diversity and Distributions 6: 259-269.

Florida Entomologist 89(3)

September 2006


'Entomology and Nematology Department, University of Florida, Gainesville, FL 32611-0630

2Escuela Agricola Panamericana, Apdo. 93, Tegucigalpa, Honduras

'Indian River Research and Education Center, University of Florida, Ft. Pierce, FL 34945

Lixadmontia franki Wood & Cave kills bromeliad-attacking weevils in the genus Metamasius.
This parasitic fly is being investigated for its potential as a classical biological control agent of
M. callizona (Chevrolat), which is devastating the native bromeliad flora in south Florida. A con-
tinuous rearing method was developed based on the fly's native host, M. quadrilineatus Cham-
pion, and a bromeliad species readily obtained from the field. Levels of parasitism of 3'd instars
on slices of Tillandsia standleyi L. B. Smith stems were equivalent to those on intact plants. Lar-
val feeding damage (>3 d) was necessary for successful parasitism. Average level of parasitism
per 30 weevil larvae exposed to 42 female flies for 14 d was 85%. Production of L. franki puparia
averaged 22.0 12.2 (SD) daily. Most (83%) final instars of the parasitoid exited their host 13-
16 d after initial exposure to the flies. Weekly inputs of 210 weevil larvae and 56 fly puparia pro-
duced, on average, 154 puparia per week for a net production rate of 2.75.
Key Words: parasitic fly, rearing, bromeliad weevils, biological control, parasitism

Lixadmontia franki Wood & Cave ataca a los picudos del g6nero Metamasius que viven en
bromeliaceas. Esta mosca esta siendo investigada por su potential como un agent de control
biol6gico clasico de M. callizona (Chevrolat), que esta destruyendo las poblaciones de brome-
liaceas nativas en el sur de Florida. Se desarroll6 un sistema de producci6n continue usando
el hospedero native, M. quadrilineatus Champion, de la mosca y una especie de bromeliacea
facilmente obtenida del campo. Niveles de parasitismo de larvas del tercer estadio del picudo
en pedazos de tallo de Tillandsia standleyi L. B. Smith fueron equivalentes a los que se ob-
servaron en plants enteras. Dano a la plant (>3 d) por la larva del picudo fue necesario
para parasitismo exitoso. Los niveles promedios de parasitismo por 30 larvas de picudo ex-
puestas a 42 moscas hembras por un period de 14 d fue de 85%. La producci6n diaria de pu-
parios fue en promedio de 22.0 12.2 (DE). La mayoria (83%) de las larvas del parasitoide
en el ultimo estadio emergi6 de su hospedero 13-16 d despu6s de la exposici6n inicial a las
moscas. Ingresos semanales de 210 larvas del picudo y 56 puparios de la mosca produjeron,
en promedio, 154 puparios por semana para una tasa neta de producci6n de 2.75.

Translation by authors.

The parasitic fly Lixadmontia franki Wood &
Cave (Diptera: Tachinidae) was discovered in
1993 parasitizing larvae of a bromeliad weevil,
Metamasius quadrilineatus Champion, collected
from native bromeliads in cloud forests of Hondu-
ras (Cave 1997; Wood & Cave 2006). Metamasius
quadrilineatus feeds primarily on bromeliads in
the genera Tillandsia, Catopsis, and Vriesea
growing at elevations of 1,600-2,000 m (Alvarez
del Hierro & Cave 1999). In its natural habitat,
M. quadrilineatus is not considered a pest due to
the presence of natural enemies such asL. franki,
predatory ants, and predatory beetles in the ge-
nus Platynus (Alvarez del Hierro & Cave 1999).
Moreover, its larval development is apparently re-
stricted to fallen bromeliads that are no longer

part of the breeding population. Its congener
Metamasius callizona (Chevrolat) was first de-
tected in Florida in 1989 (Frank & Thomas 1991,
1994; Frank 2000) and is now a serious pest.
Since its arrival, M. callizona has spread through
most of south Florida feeding on native species of
bromeliads (Frank & Thomas 1991, 1994; Frank
2000). The effects of M. callizona on the south
Florida bromeliad flora have been devastating
and 12 of the region's 16 species are now threat-
ened or endangered (Frank & Cave 2005). An ef-
fort to save these bromeliad species has been un-
dertaken by organizations including the Florida
Council of Bromeliad Societies and the Florida
Park Service through the collection of seeds from
these species to reproduce plants for future re-col-

Suazo et al.: Rearing of Lixadmontia frank

onization of affected areas, and through research
aimed at the implementation of classical biologi-
cal control to suppress weevil populations.
The suitability and feasibility of L. franki for
possible introduction into Florida to control M.
callizona are currently being studied by the Uni-
versity of Florida in conjunction with the Pan-
american School of Agriculture (Zamorano) in
Honduras. We have observed that L. franki can
parasitize larvae of M. callizona and, therefore,
could be used as a classical biological control
agent. However, lack of basic biological informa-
tion and an efficient method to produce sufficient
numbers of L. franki were limitations to imple-
menting a L. franki-based biological control pro-
gram for M. callizona. Large numbers of flies are
needed for research efforts oriented towards bet-
ter understanding the fly's biology and behavior.
Therefore, we developed a continuous and cost-ef-
fective procedure to multiply L. franki under lab-
oratory conditions as a first step towards the de-
velopment of a classical biological control pro-
gram for M. callizona.


Stock Material from the Field

Larvae of M. quadrilineatus were collected
from fallen bromeliads from three cloud forests
near El Zamorano, Honduras: Cerro Uyuca
(N1400' W8709', 1400-1800 m elevation), Cerro
Monserrat (N1356' W8654', 1780 m elevation),
and Cerro Apalagua (N1402' W8704', 1500 m el-
evation). Larvae were brought to the laboratory,
placed in 30-ml plastic cups, and provided with
portions of stems of Tillandsia standleyi L.B.
Smith as needed. The source of the stems was
fallen bromeliads in the same three cloud forests.
Flies emerging from parasitized weevil larvae
were used as stock material to start a fly colony in
the laboratory. About 300 flies constituted the
original stock material for initiation of the colony,
but variable numbers of flies were added at irreg-
ular intervals throughout the rearing period.

Parasitism of M. quadrilineatus by L. frank in the Lab-

All experiments were conducted in a room with
a constant temperature of 21C and 70% RH un-
der natural light conditions (approximately 12:12
L:D) supplemented with overhead fluorescent
lights from 0700 to 1600 h. In preliminary exper-
iments, adult fly emergence from puparia misted
twice daily with water from a hand sprayer was
seven times higher (90%) than from puparia left
dry (13%). Therefore, all puparia were misted at
this rate afterwards. Before conducting experi-
ments, newly-emerged flies, ten males and ten fe-
males, were placed in a screen cage (1.5 x 1.5 x 1.2

m) and left for 1 week to reach sexual maturity
and mate under the same conditions of tempera-
ture light and humidity. All plants used for these
experiments were healthy T standleyi (free of me-
chanical or bacterial damage) collected from the
same three cloud forests and checked to ensure no
weevil larvae were present before use. Weevil lar-
vae used in experiments were laboratory-reared
and all were in the 3rd instar at the time of expo-
sure to flies.
The first experiment was designed to investi-
gate fly preference for weevil larvae feeding on in-
tact plants versus excised stems. The treatments
were as follow: (1) weevil larvae were placed indi-
vidually in the stem of an intact plant by using a
15-cm nail to cut a hole (0.5 cm diameter, 1 cm
deep) into the side of the base of the plant and
gently inserting the larva into the hole; (2) weevil
larvae were placed individually in bare stems (no
leaves or roots) by using a nail to cut a hole (0.5
cm diameter, 1 cm deep) into one end of the stem
and gently inserting the larva into the hole
(Fig. 1). Twenty-eight weevil larvae in each treat-
ment were exposed to >8-day-old flies for 7 d im-
mediately after being put in the plant material.
After this exposure period, weevil larvae were re-
moved from the plants or stems and placed indi-
vidually in 30-ml clear plastic cups fitted with
screened lids. Larvae were fed portions (2 cm
thick) of T standleyi stem as needed (usually ev-
ery 3-4 d) to ensure larvae always had healthy
plant tissue on which to feed until parasitoid
emergence or weevil pupation. Differences in pro-
portions of parasitized and unparasitized weevil
larvae in each treatment were analyzed with a 2

Fig. 1. Third-instar Metamasius quadrilineatus in a
portion of Tillandsia standleyi stem in which a 0.5 cm
diam. hole was bored to accommodate the insect.

Florida Entomologist 89(3)

x 2 G-test of independence with Williams' correc-
tion (Sokal & Rohlf 1995).
A second experiment was designed to deter-
mine the influence of length of larval feeding time
in stems on level of parasitism. Insects, plants,
and cages were as described in the first experi-
ment. Three treatments were tested for this ex-
periment: 1) one third-instar M. quadrilineatus
within a freshly excised portion of T standleyi
stem and immediately exposed to flies for 1 d; (2)
one third-instar M. quadrilineatus allowed to feed
inside a freshly excised portion of T standleyi
stem for 3 d and then exposed to flies for 1 d; (3)
excised portion of T standleyi stem left without
weevil larva for 3 d, then one third-instar M.
quadrilineatus placed in the stem and immedi-
ately exposed to flies for 1 d. Sections of stems be-
gan to decompose from the time they were ex-
cised. Cutting of holes may have hastened the
rate of decomposition, and this may have been
compounded by feeding by weevil larvae. Thirty
larvae were exposed in each treatment. Each
treatment was exposed to a different set of flies to
avoid any effects of fly preconditioning or neonate
larva depletion. After exposure to flies, weevil lar-
vae were extracted from stems, placed individu-
ally in 30-ml clear plastic cups fitted with
screened lids and provided food as needed until
pupation or parasitoid emergence. Differences in
proportions of parasitized and unparasitized wee-
vil larvae in each treatment were analyzed with a
3 x 2 G-test of independence with Williams' cor-
rection (Sokal & Rohlf 1995). Larvae lost during
the experiment were not included in the analysis.

Continual-rearing Procedure for L. franki

Livestock for initiating the mass-rearing pro-
cedure originated from field-collected, parasitized
M. quadrilineatus larvae provided with T stand-
leyi stems as described above until mature fly lar-
vae emerged (Fig. 2A) and formed puparia
(Fig. 2B). Eight newly formed puparia were
placed in a Petri dish each day for 7 consecutive
days. Puparia in each dish were sandwiched be-
tween two layers of moistened paper towel, which
were misted twice daily to maintain the level of
humidity required for successful fly emergence.
This procedure was repeated weekly. Dishes with
puparia were kept in a room maintained at 21C
with 70% RH. Two weeks after the end of each
weekly collection of puparia, the Petri dish hold-
ing the puparia was introduced into a screened
exposition cage (1.5 x 1.5 x 1.2 m) (Fig. 3) where
adult flies emerged. The exposition cage was
maintained under the same environmental condi-
tions described for the puparia. Flies normally
emerged 21 d after they pupated. Considering an
average of 75% emergence and an adult fly life-
time of 2 weeks, six flies on average were expected
to emerge daily for a total of 42 in 1 week (21

males and 21 females assuming a 1:1 male:female
sex ratio). Flies were provided with commercially
available hummingbird instant nectar (Perky Pet
Brand, Denver, CO) and water.
One week after first adult fly emergence, a tray
containing 30 30-ml clear plastic cups, each with
a 5-cm piece of T standleyi stem containing one
third-instar M. quadrilineatus, was placed in the
exposition cage. A new tray of infested stems was
then added daily and each tray was exposed to
flies for 14 d. Fresh food was provided to larvae as
needed. Portions of old food were left in the cups
while the trays were inside the exposition cage,
but were removed at the end of the exposition pe-
After removal from the exposition cage, ex-
posed weevil larvae were maintained under the
same environmental conditions as the fly colony.
Weevil larvae were provided fresh food as needed
and checked for emergence ofL. franki larvae or
puparia every 3 d. To determine the efficacy of
this rearing method, percent parasitism of weevil
larvae and number of puparia recovered were re-
corded for three trays. Additionally, time in days
from first day of weevil exposure to emergence of
parasitoids was estimated.


When exposed for 7 d, parasitism ofM. quadri-
lineatus larvae feeding in intact plants (82%) was
not different (G = 0.41, 20051,, = 3.84) from larvae
feeding in only a portion of the stem (75%). This is
relevant because use of excised portions of stems
for weevil feeding substantially reduces the num-
ber of plants used, and therefore, the labor needed
to maintain the laboratory colony. A difference
was detected between levels of parasitism of lar-
vae in stems fed on for >3 d versus larvae in stems
fed on for 1 day (G = 87.37, 20,052 = 5.99).When
weevil larvae were allowed to feed on the stem for
>3 d and then exposed to the flies for 1 d, parasit-
ism was 100%, which contrasts sharply with no
parasitism of larvae in fresh stems and 8% para-
sitism of larvae in partially decomposed stems
with only 1 day of feeding damage. This suggests
that flies are possibly attracted to chemical cues
produced by larval feeding; the accumulation of
frass and/or masticated plant tissue may be re-
quired for the attraction to occur. Therefore, it is
essential to allow weevil larvae to feed in plant
stems for at least 3 d before exposing them to fe-
male L. franki to obtain a high level of parasitism.
An average of 6 adult flies emerged daily from
this rearing system. In this manner, a stable pop-
ulation of approximately 42 male and 42 female
flies could be maintained at all times by the regu-
lar addition of new puparia into the exposition
cage, considering an average lifespan of 2 weeks
for adult flies. Fly puparium production levels
were consistent when a stable population of flies

September 2006

Suazo et al.: Rearing of Lixadmontia frank

Fig. 2. Weevil carcass with mature larvae (A) and puparia (B) of Lixadmontia franki.

Florida Entomologist 89(3)

Fig. 3. Fly exposition cage holding rearing trays with 30-ml plastic cups containing a portion of Tillandsia stand-
leyi stem infested with third-instar Metamasius quadrilineatus for parasitism. The Petri dishes containing puparia
and cups with hummingbird food and water can be seen in the center of the cage.

was reached, averaging 22 12.2 fly puparia/day,
or approximately 154 puparia weekly. Levels of
parasitism averaged 85% (maximum = 90%, min-
imum = 75%) for 3 trays of 30 weevil larvae each.
Emergence of mature fly larvae from hosts was
observed as soon as 9 days and continued until 26
days after introduction into the exposition cage.
However, approximately 83% of the mature fly
larvae emerged 13-16 d after introduction of hosts
into the exposition cage. This continuous produc-
tion level is enough to supply flies for research
and field releases.
Large-scale production of puparia with this
rearing method is limited in Honduras by the re-
liance on weevil larvae and T standleyi (or other
suitable host plant) collected in the field. In order
to establish a large-scale production system in
Florida, this problem will need to be circum-
vented by the use ofM. callizona larvae reared in
pineapple crowns, which are available from some
grocery stores. Metamasius callizona can be
reared using pineapple crowns (Salas & Frank
2001), and a large-scale production system using
pineapple crowns to mass-multiply this insect has

been developed. Fortunately, L. franki parasitiz-
ing M. callizona larvae in pineapple stems has
been observed by R. D. Cave.
The method described here is effective for pro-
ducing sufficient numbers ofL. franki to facilitate
further investigation of this natural enemy to
control M. callizona in Florida. If L. franki is ap-
proved for release in the field, the techniques de-
scribed herein can be modified as needed to in-
crease the number of puparia produced to fit the
needs. The development of an artificial diet for M.
callizona would cut down the amount of labor and
plant material used. Furthermore, the isolation
and identification of fly attractants from stems
with 3-d-old feeding damage may also prove use-
ful for attracting and arresting the flies to the
sites of interest. These are all areas that need fur-
ther investigation.


We thank Rosa Ortega, Julio Torres, Marlon Godoy,
and Ana Samayoa (Zamorano, Honduras) for assistance
in rearing and collecting weevils and flies. We thank

September 2006

Suazo et al.: Rearing of Lixadmontia frank

Heather McAuslane and Norm Leppla for their review
of the draft manuscript. This research was supported by
a grant from the Florida Department of Environmental
Protection (RP509) and the Florida Department of Agri-
culture and Consumer Services (DACS 7276186-12).


Ecologia de Metamasius quadrilineatus (Coleoptera:
Curculionidae) yAdmontia (Diptera: Tachinidae) en
tres bosques montanos de Honduras. Ceiba 40: 43-
CAVE, R. D. 1997. Admontia sp., a potential biological
control agent of Metamasius callizona in Florida. J.
Bromeliad Soc. 47: 244-249.
FRANK, J. H. 2000. Florida's native bromeliads imper-
iled by exotic evil weevil. The Palmetto 19: 6-9, 12.
FRANK, J. H., AND R. D. CAVE. 2005. Metamasius calli-
zona is destroying Florida's native bromeliads. Vol 1,
pp. 91-101 In M. Hoddle [ed.], Second International

Symposium on Biological Control ofArthropods, Sep
2005, Davos, Switzerland. Publ. By USDA Fores Ser-
vice, Morgantown, WV, as FHTET-2005-08.
FRANK, J. H., AND M. C. THOMAS. 1991. Metamasius cal-
lizona kills bromeliads in southeastern Florida. J.
Bromeliad Soc. 41: 107-108.
FRANK, J. H., AND M. C. THOMAS. 1994. Metamasius cal-
lizona (Chevrolat) (Coleoptera: Curculionidae), an
immigrant pest, destroys bromeliads in Florida. Ca-
nadian Entomol. 126: 673-682.
SALAS, J., AND J. H. FRANK. 2001. Development of Meta-
masius callizona (Coleoptera: Curculionidae) on
pineapple stems. Florida Entomol. 84: 123-126.
SOKAL, R. R., AND F. J. ROHLF. 1995. Biometry: The
Principles and Practice of Statistics in Biological Re-
search, Third Edition. W.H. Freeman; New York,
New York.
WOOD, D. M., AND R. D. CAVE. 2006. Description of a
new genus and species of weevil parasitoid from
Honduras (Diptera: Tachinidae). Florida Entomol.
89: 239-244.

Florida Entomologist 89(3)

September 2006


Enns Entomology Museum, Division of Plant Sciences, University of Missouri, Columbia, Missouri 65211, U.S.A.
(e-mail: brett.landwer@mdc.mo.gov and SitesR@missouri.edu)


Despite widespread distributions and abundance, previously published diagnoses of the lar-
vae of the two species of the dragonfly genus Pantala often were contradictory or confusing.
Morphometric analysis of mensural characters and qualitative analysis of relative character
states were used to determine the ability of previously published characterizations to accu-
rately distinguish larvae of the two species. We found that many published characterizations
were inaccurate or insufficient, and their use in making species level determinations would
result in frequent misidentifications. In distinguishing between the two species, the most
useful and reliable characteristic was the palpal setal count. However, in specimens where
this count is intermediate, other characteristics may need to be evaluated.


A pesar de su distribuci6n amplia y su abundancia, las diagnosis publicadas anteriormente
de las larvas de dos species de liblulas del g6nero Pantala a menudo son contradictorias o
confusas. El andlisis morfom6trico de los caricteres mensurales y el andlisis cualitativo de
los caracteres de estado relatives fueron empleados para determinar la abilidad de las ca-
racterizaciones publicadas anteriormente para distinguir las larvas de las dos species con
exactitud. Nosotros encontramos que muchas de las caracterizaciones publicadas fueron
inexactas o insuficientes, y su uso para hacer determinaciones a nivel de especie resultarian
frecuentemente en identificaciones equivocadas. Para distinguir entire las dos species, la ca-
racteristica mas util y confiable es el conteo de las setas del palpo. Sin embargo, en especi-
menes donde el conteo es intermedio, se necesitan que otras caracteristicas sean evaluadas.

The genus Pantala Hagen is represented
worldwide by two species, both of which are wide-
spread and abundant in the United States. Many
characters have been used to diagnose the larva of
each species and distinguish between them. How-
ever, character states attributed to each species
frequently have been contradictory among au-
thors, raising questions as to the diagnostic reli-
ability of these characters.
Kennedy (1923) was the first to distinguish be-
tween P. hymenaea (Say) and P flavescens (Fabri-
cius). Needham (1901) had previously character-
ized the movable hook of P. flavescens as "hardly
longer than the teeth." Kennedy described the
larva of P hymenaea and distinguished it from
Needham's diagnosis of P. flavescens on the basis
of the moveable hook of the former as "twice as
long as the crenulations of the distal edge of the
lobe." Lamb (1929) investigated this character
and found that the mean ratio of the length of the
movable hook to length of the first palpal crenula-
tion was 2.2 in P hymenaea (3 specimens, 1
reared) and 2.6 in P flavescens (8 reared speci-
mens). Musser (1962) examined more specimens
than the previous authors (60 P hymenaea, 1 with
associated adult; 39 P flavescens, 3 with associ-
ated adults) and found the ratios to be 2.0 and 2.5-
3.0 for P hymenaea and P flavescens, respectively.

However, authors continued to characterize the
movable hook ofP. hymenaea as 2 times as long as
the crenulations of the labial palp while charac-
terizing that of P. flavescens as less (Byers 1930;
Smith & Pritchard 1956; Walker & Corbet 1975;
Huggins & Brigham 1982).
Two of the interspecific distinctions that appear
most commonly in the modern literature without
contradiction were first given by Musser (1962),
who stated that the lateral spines of abdominal seg-
ment IX were less than three times as long as the
basal width in P hymenaea, and at least three times
as long in P flavescens. Musser also mentioned a
"slight hump midway along" the dorsal margin of
the epiproct, and diagrammed the dorsal margin
abruptly decurved at a point somewhat beyond
midlength in P hymenaea, while characterizing and
depicting a straight dorsal margin of the epiproct in
P flavescens. Both of these distinctions subse-
quently have been repeated, and the accompanying
diagrams reproduced, in several faunistic treat-
ments (e.g., Walker & Corbet 1975; Huggins &
Brigham 1982; Needham et al. 2000). Paulson
(1966) also noted that the lateral spines of P flave-
scens were more elongate than those ofP hymenaea
but offered no quantification other than that the
spines of VIII reach or exceed the apical margin of
IX in the former and not the latter.

Landwer & Sites: Larval Characters to Distinguish Pantala species

Palpal setal counts also have been used to
characterize each species of Pantala and to distin-
guish between them. P flavescens is generally
characterized as bearing 12-14 palpal setae
(Needham 1901; Lamb 1929; Needham & Hey-
wood 1929; Byers 1930; Klots 1932; Smith & Prit-
chard 1956; Musser 1962; Walker & Corbet 1975;
Huggins & Brigham 1982) and P hymenaea as
bearing 15 (Kennedy 1923; Needham & Heywood
1929; Byers 1930; Klots 1932; Smith & Pritchard
1956; Musser 1962; Walker & Corbet 1975; Hug-
gins & Brigham 1982) or more (Lamb 1929; Klots
1932; Musser 1962). Paulson's (1966) table of pal-
pal setal counts is the only work documenting in-
traspecific variation and, in P. flavescens, he found
14 or fewer setae on at least one palp in all 16
specimens examined (4 reared), but found 15 on a
single palp in two specimens, and 16 and 17 on a
single palp in one specimen each. In P hymenaea,
he also found 16 or more palpal setae on at least
one palp in each of 13 specimens examined (1
reared), but found 15 on a single palp in one spec-
Needham & Westfall (1955) presented two
novel characters to distinguish between species of
Pantala. They characterized the epiproct of P fla-
vescens as longer than the paraprocts, and that of
P hymenaea as subequal to the paraprocts. This
distinction was adopted by Young & Bayer (1979).
Paulson (1966) believed this to be a fairly reliable
distinction, but found some specimens of each
species to be indistinguishable based on this dis-
tinction. Musser (1962) also agreed that this was
generally the case, but found measurement of
these characters too difficult to reliably quantify
and be useful. Needham & Westfall (1955) also
distinguished P hymenaea from P. flavescens on
the basis of a more marked color pattern in the
former, which also was adopted by Young & Bayer
(1979). Paulson (1966) and Musser (1962) both re-
futed this distinction as unreliable, yet it ap-
peared again in Needham et al. (2000).


A total of 46 specimens of Pantala was exam-
ined. Of these, 27 specimens were P hymenaea:
23 from Missouri (16 reared exuvial specimens, 7
final instars) and 4 from California (all reared ex-
uvial specimens). Nineteen specimens of P flave-
scens were examined, including 12 from Missouri
(1 reared exuvial specimen, 11 final instars) and 7
from Florida (all reared exuvial specimens). Mis-
souri specimens are deposited in the Enns Ento-
mology Museum, University of Missouri, Colum-
bia, Missouri; all other specimens are in the col-
lection of J. C. Abbott.
All measurements were performed with an oc-
ular micrometer on the strict dorsal aspect of the
specimen. Measurements were rounded to the
nearest 0.04 mm. The lengths of lateral spines

were measured on one side of the specimen, along
a line parallel to the long axis of the body, from the
posterior margin of the segment immediately ad-
jacent to the base of the spine to the level attained
by the tip of the spine. The basal width of a lateral
spine was measured along a tangent perpendicu-
lar to the long axis of the body, from the point on
the posterior margin of the segment immediately
adjacent to the base of the spine to the lateral
margin of the segment. Mid-dorsal segment
length was measured from the anterior to the pos-
terior margin of the tergite. Palpal setae were
counted on each palp.
Other characteristics were evaluated qualita-
tively. Specimens were examined in lateral view
to determine if the apex of the epiproct exceeded
the apices of the paraprocts. The color pattern
was considered distinct or indistinct. Due to non-
uniform telescoping reported in the abdomens of
larval and exuvial specimens (see Calvert 1934;
Huggins & Harp 1985), the posteriormost dis-
tance on segment IX attained by the lateral
spines of VIII was not evaluated. Also, precise
measurements of the relative lengths of the move-
able hook and crenulations of the labial palps
were very difficult to obtain without damaging
the specimens, and were evaluated comparatively
between species. The convexity of the dorsal mar-
gin of the epiproct in lateral view also was evalu-
ated comparatively.
Material examined.-CALIFORNIA: Fresno
Co., Enterprise Canal E. Clovis, 1 Nov 1976, S. W.
Dunkle exuviaee of 4 reared P hymenaea); FLOR-
IDA: Alachua Co., NE Gainesville, Nov 1975, S. W.
Dunkle exuviaee of 6 reared P. flavescens); same
data, Austin Cary Fishpond, coll. 22 Sep 1978,
emerged 23 Sep 1978, S. W. Dunkle exuviaee of 1
reared P flavescens); MISSOURI: Audrain Co., R.
M. White II Conservation Area, Sep 9, 1998,
BHPL & N. Whiteman (2 P. flavescens larvae);
Benton Co., Lost Valley Fish Hatchery, 6 Oct
2000, L. Trial exuviaee of 3 unreared, 1 larval P.
flavescens); Boone Co., Ditch near Vet. School,
University of Missouri campus, 30 Jul 2001,
BHPL exuviaee of 1 reared, 3 larval P hymenaea);
Christian Co., SW Nixa, ca. 200 gallon stock tank,
15 Aug 2001, BHPL exuviaee of 11 rearedP, hyme-
naea); Jackson Co., Jacomo Lake, no date, S.
Thewke (1 P flavescens larva); Pemiscot Co., Uni-
versity of Missouri Lee Farm, rice paddy, 25 Jul
2001, BHPL & C. Luppens, exuviaee of 1 reared,
larval P flavescens; exuviae of 3 reared, 4 larval
P hymenaea); Texas Co., 4 mi. S Simmons, 4 Oct
1972, S. Thewke (3 P flavescens larvae).


On average, the lateral spines were longer in P
flavescens and broader in P hymenaea (Table 1).
However, considerable interspecific overlap existed
in each of these measurements. Interestingly, less


Length Width Length Mid-dorsal Palpal Length/width Length spine 8/ length 8/
Species Specimen spine 9 spine 9 spine 8 length 8 setae spine 9 width spine 9 length spine 8

P. hymenaea



*Larval specimen or unreared exuviae.


Length Width Length Mid-dorsal Palpal Length/width Length spine 8/ length 8/
Species Specimen spine 9 spine 9 spine 8 length 8 setae spine 9 width spine 9 length spine 8

P. flavescens 1 2.40 0.56 1.28 1.36 13 4.29 2.29 1.06
2 2.20 0.56 1.28 1.28 13 3.93 2.29 1.00
3 2.24 0.60 1.32 1.36 13 3.73 2.20 1.03
4 2.04 0.56 1.16 1.24 13/14 3.64 2.07 1.07
5 2.24 0.64 1.20 1.32 13 3.50 1.88 1.10
6 2.44 0.72 1.44 1.28 13/14 3.39 2.00 0.89
7 2.80 0.68 1.56 1.44 14 4.12 2.29 0.92
8* 2.48 0.72 1.56 1.52 14 3.44 2.17 0.97
9* 2.72 0.72 1.48 1.44 14 3.78 2.06 0.97
10* 2.48 0.64 1.40 1.40 13 3.88 2.19 1.00
11* 2.92 0.72 1.80 1.44 13/14 4.06 2.50 0.80
12* 3.00 0.80 1.64 1.44 14/15 3.75 2.05 0.88
13* 2.88 0.80 1.76 1.44 14 3.60 2.20 0.82
14* 2.80 0.72 1.72 1.40 14 3.89 2.39 0.81
15* 2.80 0.68 1.84 1.40 14 4.12 2.71 0.76
16* 2.48 0.56 1.32 1.28 13 4.43 2.36 0.97
17* 2.52 0.68 1.44 1.36 14 3.71 2.12 0.94
18* 2.72 0.68 1.52 1.32 14 4.00 2.24 0.87
19 2.52 0.64 1.52 1.40 15 3.94 2.38 0.92

Mean 2.56 0.67 1.49 1.37 3.85 2.23 0.94

S.E. 0.27 0.08 0.20 0.07 0.28 0.19 0.10
*Larval specimen or unreared exuviae.

Florida Entomologist 89(3)

overlap existed in the mid-dorsal length
nal segment VIII. In fact, the smalles
ment in P hymenaea (1.44 mm) was e
largest in P flavescens, although this val
resented in 5 P flavescens specimens.
Length to width ratios of the lateral
hibited less overlap than did absolute
ments of these characters. The length to
of the spine of abdominal segment IX of
specimens of P flavescens overlapped t
(7%) of P hymenaea (Fig. 1). However, t
occurred at ratios considerably higher tl
fact, values greater than 3.00 were obs
(37%) specimens of P hymenaea and the
value observed in P flavescens was 3.39.
The ratio of the length of the later
abdominal segment VIII to the basal w
spine of IX yielded even less overlap tl
length/width ratio of spine IX. The va
character overlapped in only one specir
P flavescens (5%) andP hymenaea (4%
was less than 2 in all specimens of P
and greater than 2 in all but one spe
flavescens. The ratio of the length of t
abdominal segment VIII to the mid-do
of that segment showed greater over
tios in four P. flavescens specimens
those in three P hymenaea.
The larvae examined displayed a r
lack of symmetry in the number of palp
the left and right palp of individual
However, the total number of palpal sel
no overlap between species (Fig. 2).
each palp, the characteristic of 14 or fe
setae correctly identified 17 (89%) P.
larvae. Fifteen palpal setae on each pa
characteristic of any P. hymenaea spec
ever 15 or more correctly identified all
and 16 or more characterized all but a
of one specimen. No specimen of P flavu
more than 15 palpal setae on either
versely, all P. hymenaea specimens
more setae on at least one palp.

Al V llI lIIi ,ili
<2.7 2.8 3 3.2 3.4 3.6 3.8
Fig. 1. Frequency of expression of lengi
tios of spine 9 in specimens ofPantala

of abdomi-
t measure-
qual to the
ue was rep-

spines ex-
e measure-
width ratio
three (16%)
;hat of two
his overlap
ian 3.00. In
served in 10
e minimum

al spine of
idth of the
ian did the
lue of this
nen each of

O. hmelfwea
8 N P flarescens

4"--- -- --

26 27 28 29 30 31 32 33 34 35
Number of Setae (Both Palps)
Fig. 2. Frequency of expression of palpal setal counts
in specimens ofPantala


). The ratio This analysis clearly reveals the inadequacy of
hymenaea, existing interpretations of the distinction be-
cimen of P. tween species of larval Pantala. These results in-
he spine of dicate that the possession of 15 or fewer palpal se-
rsal length tae on each palp will distinguish P. flavescens lar-
ap; the ra- vae from P hymenaea, although two specimens of
overlapped P flavescens examined by Paulson (1966) vitiate
this characterization. If one includes the stipula-
emarkable tion that P. flavescens specimens bearing greater
lal setae on than 15 palpal setae on one palp bear fewer than
specimens. 15 on the other, this characterization is not vio-
tae showed lated. This distinction might be best expressed as
Applied to the sum of palpal setae from both palps (see Fig.
wer palpal 2). Thirty or fewer setae is characteristic ofP. fla-
flavescens vescens and 32 or more of R hymenaea. This char-
ilp was not acterization does not exclude Kennedy's (1923)
imen, how- original description of the larva of P hymenaea
specimens, and characterization based upon it. However,
single palp Lamb (1929) examined Kennedy's P hymenaea
escens bore material, and reported that the specimens bore at
palp. Con- least 17 palpal setae. It is not known if she exam-
bore 16 or ined the same specimens from which the descrip-
tion was prepared.
As noted by Musser (1962) and Paulson (1966),
the lateral spines of P hymenaea are generally
stouter than those of P flavescens. However, it is
OP h).mnlU clear from this analysis that the characterization
EPfla.,,I'sce of P hymenaea larvae as bearing lateral spines on
abdominal segment IX less than three times as
long as the basal width is inadequate. The maxi-
mum value of this ratio found in P hymenaea was
actually 3.50, slightly greater than the minimum
l fl value in P flavescens. Our results indicate that the
possession of lateral spines on segment VIII that
are less than two times as long as the basal width
I, of the lateral spines on segment IX in P hymenaea
4 >4.1 and greater than two times in P flavescens is a
more reliable quantification of the difference in the
form of lateral spines between the species.
th/width ra- The characterization of the moveable hook ofP.
flavescens as less than twice as long as the crenu-

September 2006

Landwer & Sites: Larval Characters to Distinguish Pantala species

lations of the labial palp is clearly erroneous. The
specimens examined generally agreed with the
assessments of Lamb (1929) and Musser (1962)
that the moveable hook of P flavescens is longer
than that ofP. hymenaea relative to the length of
the crenulations on the labial palp. It is likely
that all characterizations to the contrary are de-
rived from Kennedy's (1923) distinction of larvae
of P hymenaea from Needham's (1901) diagnosis
of the larvae P flavescens. However, Needham
probably offered the characterization of the move-
able hook as "hardly longer than the teeth" as an
obvious distinction from the long moveable hooks
of species of Tramea, and not as an absolute mea-
We found the decurvature of the dorsal margin
of the epiproct in lateral view to be a highly sub-
jective and unreliable character to distinguish be-
tween species. Although the dorsal margin of the
epiproct was generally more convex in P hyme-
naea (Fig. 3A), rarely was the decurvature as
abrupt or as near the apex as figured by Musser
(1962). Also, the dorsal margin of the epiproct ofP.
flavescens was often distinctly convex (Fig. 3B)
and in no specimen was it as straight as figured
by Musser (1962).

Fig. 3. Lateral view of posterior abdominal segments
of (A) Pantala hymenaea and (B) Pantala flavescens
(epiproct is uppermost appendage)

Our evaluation of relative lengths of epiprocts
and paraprocts is in accord with those of Musser
(1962) and Paulson (1966). Specifically, the dis-
tinction was obvious in many specimens, but un-
clear for several specimens of each species. We
also agree with these authors that there is no
clear distinction in the intensity of pigmentation.
In fact, the color pattern was distinct in all larvae
examined, and remained apparent in preserved
All specimens possessed spine like dorsal
hooks on abdominal segments II and III, and on
all but one P flavescens, IV as well. Our results
corroborate those ofWestfall & Tennessen (1996),
who also noted the presence of dorsal hooks on ab-
dominal segments II-IV in both species of Pan-
tala. This obviates the erroneous characterization
of the genus as lacking dorsal hooks (e.g.,
Needham & Heywood 1929; Byers 1930;
Needham & Westfall 1955; Smith & Pritchard
1956; Needham et al. 2000), the independent
characterization of P flavescens as lacking dorsal
hooks (Needham 1901; Byers 1930;Walker & Cor-
bet 1975), and the erroneous distinction of P. hy-
manaea from P. flavescens on this basis (Daigle
1992). This confusion may be traced to the origi-
nal description by Cabot (1890) who made no
mention of dorsal abdominal hooks.
It seems likely, due to intraspecific variation
and interspecific similarity, that no single charac-
teristic will reliably distinguish all larval speci-
mens of Pantala. We do agree, however, with the
assessment of Paulson (1966) that by evaluating
several distinguishing characteristics, almost all
specimens can be correctly determined. The sim-
plest characteristic to evaluate is the palpal setal
count. Evaluation of this character requires no
precise measurement and can be performed with-
out an ocular micrometer. Specimens with 14 or
fewer palpal setae on each palp are assuredly at-
tributable to P. flavescens, while those bearing 16
or more are attributable to P hymenaea. Speci-
mens bearing 15 or more palpal setae on one palp
and 15 or fewer on the other are best identified
with an alternate distinction. From this analysis,
it appears that lateral spines on segment VIII
greater than twice as long as the basal width of
the lateral spines on IX is a reliable characteriza-
tion of P flavescens, and a ratio less than 2 is char-
acteristic of P. hymenaea.
We hope that this evaluation of distinguishing
characteristics will prompt others to more care-
fully examine their specimens of Pantala. Dis-
crepancies between actual specimens and pub-
lished characteristics are unacceptable in a genus
comprising such ubiquitous, abundant, and easily
reared species. We urge researchers to report the
state of characters found in other regions so that
the extent of variation found throughout the
ranges of these widespread species can be docu-


We thank J. C. Abbott (University of Texas) for pro-
viding specimens and N. Whiteman, C. Luppens (Uni-
versity of Missouri), and L. Trial (Missouri Department
of Conservation) for assistance in collecting Missouri
specimens. Support for RWS was provided in part by
MU project number PSSL0232.


BYERS, C. F. 1930. A Contribution to the Knowledge of
Florida Odonata. Gainesville, University of Florida
Press. 327 pp.
CABOT, L. 1890. The immature state of the Odonata.
Part 3. Subfamily Cordulina. Memoirs of the Mu-
seum of Comparative Zoology at Harvard College
17(1): 1-52.
CALVERT, P. P. 1934. The rates of growth, larval devel-
opment and seasonal distribution of dragonflies of
the genus Anax (Odonata: Aeshnidae). Proc. Ameri-
can Philosophical Soc. 73: 1-70.
DAIGLE, J. J. 1992. Florida dragonflies (Anisoptera): a
species key to the aquatic larval stages. State of
Florida Department of Environmental Regulation
Technical Series 12(1): 1-29.
HUGGINS, D. G., AND W. U. BRIGHAM. 1982. Odonata,
Chapter 4, pp. 4.1-4.100 In A. R. Brigham, W. U.
Brigham, and A. Gnilka [eds.], Aquatic Insects and
Oligochaetes of North and South Carolina. Midwest
Aquatic Enterprises, Mahomet, Illinois.
HUGGINS, D. G., AND G. L. HARP. 1985. The nymph of
Gomphus (Gomphurus) ozarkensis Westfall (Odo-
nata: Gomphidae). J. Kansas Entomol. Soc. 58: 656-
KENNEDY, C. H. 1923. The naiad of Pantala hymenaea
(Odonata). The Canadian Entomol. 55: 36-38.
KLOTS, E. B. 1932. Insects of Porto Rico and the Virgin
Islands part 1, Odonata or dragonflies. Scientific

September 2006

Survey of Porto Rico and the Virgin Islands 14: 1-
LAMB, L. 1929. The later larval stages of Pantala (Odo-
nata: Libellulidae). Trans. American Entomol. Soc.
55: 331-334.
MUSSER, R. J. 1962. Dragonfly nymphs of Utah (Odo-
nata: Anisoptera). University of Utah Biological Se-
ries 12(6): 1-71.
NEEDHAM, J. G. 1901. Odonata, pp. 381-612 In Aquatic
Insects in the Adirondacks. Bull. New York State
Mus. 47.
NEEDHAM, J. G., AND H. B. HEYWOOD. 1929. A Hand-
book of the Dragonflies of North America. Spring-
field, C. C. Thomas. 378 pp.
ual of the Dragonflies of North America
(Anisoptera). Berkeley, University of California
Press. 615 pp.
2000. Dragonflies of North America. Gainesville,
Florida, Scientific Publishers. 939 pp.
PAULSON, D. R. 1966. Dragonflies of South Florida
(Odonata: Anisoptera). Ph. D. dissertation, Univer-
sity of Miami, Coral Gables, Florida. 603 pp.
SMITH, R. F., AND A. E. PRITCHARD. 1956. Odonata,
Chapter 4, pp. 106-153 In R. L. Usinger [ed.],
Aquatic Insects of California. University of Califor-
nia Press, Berkeley, California, USA.
WALKER, E. M., AND P. S. CORBET. 1975. The Odonata of
Canada and Alaska. Vol. 3. Toronto, Ont., University
of Toronto Press. 307 pp.
nata, Chapter 12, pp. 164-211 In R. W. Merritt and
K. W. Cummins [eds.], An Introduction to the
Aquatic Insects of North America. Kendall/Hunt,
Dubuque, Iowa.
YOUNG, W. C., AND C. W. BAYER 1979. The dragonfly
nymphs (Odonata: Anisoptera) of the Guadalupe
River Basin, Texas. The Texas J. Sci. 31: 85-97.

Florida Entomologist 89(3)

Lim & Hoy: Overwintering of Citrus Leafminer


'Department of Entomology and Nematology, University of Florida, P.O. Box 110620,
Gainesville, FL 32611-0620 USA

2School of Bioresource Sciences, Andong National University, Andong 760-749, Republic of Korea


The citrus leafminer, Phyllocnistis citrella Stainton (Lepidoptera: Gracillariidae), is an im-
portant pest of citrus that has spread around the world. The citrus leafminer invaded Flor-
ida in 1993 and rapidly colonized all citrus-growing areas. Despite the fact that the citrus
leafminer has been studied extensively, gaps in our knowledge of its biology remain. One is-
sue is how the citrus leafminer overwinters. Citrus leafminers have been observed to over-
winter in many countries, but their capacity to diapause has not been confirmed. To
investigate whether P. citrella overwinters in diapause, we evaluated the ability of eggs to
develop to adulthood on potted citrus trees when reared in both outdoor and growth chamber
conditions. No pupae were found to be in diapause in the outdoor assessment in which three
populations of citrus leafminer were reared in a shade house in Gainesville, Florida begin-
ning on Oct 15, Nov 11, or Nov 22. Larval and pupal mortality significantly increased in the
population set out on Nov 22 compared to the other populations. Adult longevity signifi-
cantly increased in the last two populations compared to the Oct population. However, sim-
ilar degree days were accumulated during the adult lifespan among the three populations.
In the growth chamber assessments, citrus leafminers showed no arrestment in develop-
ment during the pupal stage by short-day treatments. Development (days) from egg to adult
was not different between short- and long-day conditions in both sexes. No signs of repro-
ductive diapause were found from dissections of adult females reared under short- or long-
day conditions. We concluded that P. citrella may overwinter on the few small flushes avail-
able during winter in mid and central Florida, and the relevance of these results for citrus
IPM programs is discussed.

Key Words: photoperiod, adult longevity, immature development, oogenesis

El minador de la hoja de citricos, Phyllocnistis citrella Stainton (Lepidoptera: Gracillarii-
dae), es una plaga important de los citricos que ha sido esparecida alrededor del mundo. El
minador de la hoja de citricos invadio el estado de Florida en 1993 y rapidamente colonize
todas los areas donde se siembran citricos. Al pesar del hecho que el minador de la hoja de
citricos ha sido estudiado extensivamente, todavia hay brechas en nuestro entendimiento de
su biologia. Un asunto es como el minador de la hoja de citricos invera. La inveraci6n del
minador de la hoja de citricos ha sido observada en muchas paises, pero su capacidad para
pasar por el estado de diapausa no ha sido confirmado. Para investigar si P. citrella inverna
en el estado de diapausa, nosotros evaluamos la capacidad de los huevos para desarrollarse
al estado adulto al criados sobre arboles de citricos sembrados en recipients bajo condicio-
nes exteriores y en una camera de crecimiento. Ninguna pupa fue encontrada en el estado
de diapausa en la prueba de condiciones exteriores donde tres poblaciones del minador de la
hoja de citricos fueron criadas en condiciones de sombra en Gainesville, Florida empezando
en el 15 de octubre, el 11 de noviembre, o el 22 de noviembre. La mortalidad de las larvas y
pupas aument6 significativamente en la poblaci6n del 22 de noviembre en comparaci6n con
las otras poblaciones. La longevidad de los adults aument6 significativamente en las ulti-
mas dos poblaciones en comparaci6n con la poblaci6n de octubre. Sin embargo, los dias-grado
acumulados fueron similares durante el period de vida del adulto entire las tres poblaciones.
En las pruebas dentro de la camera de crecimiento, el minador de la hoja de citricos no mos-
tr6 ningun impedimento en su desarrollo durante los estadios de pupa en los tratamientos
de dias cortos. El desarrollo (dias) del huevo hasta el adulto no fue diferente entire las condi-
ciones de dias cortos y dias largos en ambos sexos. Ninguna seal de la diapausa reproduc-
tiva fue encontrada en las disecciones de las hembras adults criadas bajo condiciones de
dias cortos o dias largos. Nosotros concluimos que P. citrella podria estar invernando sobre
unos pocos brotes pequenos de nuevas hojas disponibles durante el invierno en Florida, y se
comenta sobre la importancia de estos resultados en los programs de MIP en la producci6n
de citricos.

Florida Entomologist 89(3)

The citrus leafminer, Phyllocnistis citrella
Stainton (Lepidoptera: Gracillariidae), is an im-
portant citrus pest that is native to Southeast
Asia. Adult females deposit eggs singly upon ten-
der young citrus foliage and larvae immediately
enter the leaf and start feeding on epidermal
cells, producing broad serpentine mines (Hoy &
Nguyen 1997). Mining of immature foliage re-
duces growth rates and yields, and the mines
serve as foci for the establishment of the bacte-
rium causing citrus canker (Gottwald et al. 2001;
Graham et al. 1996; Liu et al. 1999; Ujiye 2000).
Various aspects of the biology of the citrus leaf-
miner have been studied in the laboratory and
field but questions remain, including how this
pest survives the winter in Florida. Typically, this
pest declines to very low densities during the win-
ter and the first spring flush in Mar lacks signifi-
cant populations of this pest (Knapp et al. 1995;
Pena et al. 1996; Villanueva-Jimenez et al. 2000).
The severe reduction in the number of overwin-
tering hosts also reduces populations of the host-
specific parasitoid Ageniaspis citricola Logvi-
novskaya (Hymenoptera: Encyrtidae) in Florida,
which means that populations of this parasitoid
lag behind that of the leafminer until the second
or third major flush cycle during the following
growing season (Villanueva-Jimenez et al. 2000;
Zappala et al. unpublished). Understanding how
the citrus leafminer overwinters could increase
our ability to manipulate pest and natural enemy
populations in Florida's citrus.
Many insects hibernate in diapause, a geneti-
cally determined state of arrested development
that typically occurs prior to the onset of unfavor-
able conditions (Beck 1980; Danilevskii 1965; Den-
linger 2002; Lees 1955; Tauber et al. 1986). Dia-
pause can occur in eggs, larvae, pupae, or adults of
insects, with adult diapause determined by com-
paring their longevity to nondiapausing adults or
by examining the reproductive tracts of males and
females for delays in the development of oocytes or
spermatocytes. In most arthropods, diapause is in-
duced by a combination of cues, including temper-
ature, photoperiod, and host plant condition.
Winter diapause of the citrus leafminer has
not been confirmed. Clausen (1931) and Ayoub
(1960) indicated that the citrus leafminer passes
the winter in the adult stage. Pandey & Pandey
(1964) found no evidence of hibernation or aesti-
vation in India, based on phenology. Barroga
(1968) reported that in the Philippines citrus leaf-
miner populations decreased to low levels during
Jun, Oct, Nov and Apr, which correlated with the
absence of flush. Citrus leafminer populations
also occurred whenever the trees flushed, regard-
less of locality or season of the year in Darwin,
Australia (Wilson 1991) and India (Batra &
Sandhu 1981). Ujiye (2000) reported that because
males of citrus leafminer are sexually active dur-
ing winter, males do not appear to have a winter

diapause, but he proposed reproductive diapause
of females based on the findings by Hamamura
(1980). Ovaries of female citrus leafminer emerg-
ing in autumn are absorbed while exposed to low
temperatures in Japan (Hamamura 1980). Ali
Mafi & Ohbayashi (2004) reported that "Infesta-
tion was recognized even in Dec, Jan, and Feb" in
Japan on Citrus iyo. However, to our knowledge,
no one has attempted to determine whether citrus
leafminer has the genetic capability to enter hi-
bernal diapause by rearing them under controlled
temperatures and daylengths. We report here the
results of rearing the citrus leafminer on potted
citrus trees held in cages outdoors during the win-
ter of 2003-2004 in Gainesville, Florida, and of
rearing them in growth chambers under con-
trolled temperatures and daylengths that could
be expected to induce diapause if this species has
the genetic capacity to do so.


Outdoor Assessment of Pupation and Adult Emergence

Ten potted rough lemon (Citrus jambhiri
Lushington) trees were infested with 100 newly
emerged unsexed citrus leafminers in a PVC pipe-
framed cage (74.5 W x 46.5 Lx 61.5 H cm) covered
with a mesh bag for 3 d in a greenhouse main-
tained at 27C, 80-100% RH, and photoperiod of
16:8 (L:D). Once oviposition was verified, each
tree was covered with a plastic cylinder contain-
ing lateral holes covered with mesh fabric to pre-
vent predation or parasitization and placed in an
outdoor shade house. Once the first adults
emerged, based on daily observations, the leaves
containing citrus leafminers were pruned from
the plants. Leaves containing pupal chambers
were placed in inflated clear plastic bags with pa-
per towels and held in the shade house. Adults
that emerged in the bag were counted daily and,
when adult emergence was discontinued, the
leaves in the bag were examined to find any live
larvae and pupae remaining. Adults that emerged
were placed in diet cups in the shade house indi-
vidually with a honey strip and a water-soaked
cotton ball to evaluate longevity over the winter.
The mortality of these unsexed adults was ob-
served daily, and the diet cup, honey strip, or cot-
ton ball was replaced with new ones as needed.
Longevity was evaluated in two ways: number of
days and number of degree days (DD). Daily 24-h
mean temperatures were obtained from the Flor-
ida Automated Weather Network and used to cal-
culate the number of DD per day, based upon a
threshold of 12.1C (Ujiye 2000; Vargas et al.
2001). These procedures were repeated three
times by setting up lemon trees containing eggs of
citrus leafminer on Oct 15, Nov 11, and Nov 22.
Larval mortality, adult emergence, and longevity
were obtained and analyzed with the Kruskal-

September 2006

Lim & Hoy: Overwintering of Citrus Leafminer

Wallis single factor analysis of variance by rank
(SAS Institute 1995). The Dunn test was used for
post hoc comparison (Zar 1996).

Evaluation of Diapause in Immature P. citrella in
Growth Chambers

To verify the results of the outdoor experi-
ments, the ability of citrus leafminers to diapause
was assessed in a growth chamber under short-
day conditions. Ten lemon trees with new shoots
were exposed to 100 adult citrus leafminers for 1
d under the same conditions as in the outdoor as-
sessment. Half of the trees were placed at 18C
with a photoperiod of 16:8 (L:D) in a growth
chamber, and the other half were placed into a
photoperiod of 8:16 (L:D) at 18C in another
growth chamber. Trees in both growth chambers
were assessed daily for larval hatch. Leaves con-
taining only one larva were marked (115 leaves in
short day condition and 109 in long day condition)
and larval development was assessed daily. De-
velopment (days) from egg to adult was analyzed
with a t-test to find any effect of short daylength.
Pupation rate, emergence rate, and sex ratio in
both treatments and controls were compared by
Fisher's exact test.

Evaluation of Reproductive Diapause inAdult P. citrella
in Growth Chambers

The existence of a reproductive diapause in
adult females was evaluated under short-day con-
ditions. Ten lemon trees with new shoots were ex-
posed to 100 adult citrus leafminers for one day
under the same conditions as in the outdoor as-
sessment. After being individually covered with a
clear plastic cylinder, half of the trees were placed
at 18C and a photoperiod of 16:8 (L:D) in a
growth chamber, and the other half at 18C and a
photoperiod of 8:16 (L:D) in another growth
chamber. Emerging adults were collected daily
and placed with water and honey in diet cups as
described in the outdoor assessment. Six days af-

ter emergence, females were dissected under a
microscope. Females that had mature eggs in the
ovary when dissected were treated as non-dia-
pausing females. This procedure was repeated for
two additional photoperiods, i.e., 14:10 (L:D) and
10:14 (L:D) at 18C.


Larval mortality significantly increased in the
citrus leafminer population set out on Nov 22
compared to populations studied earlier (H =
15.949, df = 2, P < 0.001) (Table 1). This could
have been due to a decline in ambient tempera-
ture (Fig. 1) and leaf hardening. Badawy (1967)
indicated that P citrella larvae develop during
the winter months in Sudan, but that very few
larvae successfully completed their development.
None of the pupae were found to be in diapause
during the outdoor tests (Table 1). Adults did
emerge from the pupae of all the populations
tested in Gainesville. However, fewer adults
emerged in the population set out on Nov 22 than
emerged from those that were set out on the ear-
lier dates (iH = 16.826, df= 2, P < 0.001) probably
due to the increase in larval mortality (Table 1).
Adult longevity was significantly increased to
34 or 35 d in the populations set out on Nov com-
pared to 18 d in the population set out during Oct
(H = 50.811, df= 2, P < 0.001) (Table 1). However,
the similar degree days accumulated by the
adults among the three populations (H = 0.939,
df = 2, P = 0.625) (Table 1) may indicate that cit-
rus leafminers can survive the winter in Florida
as adults without an increase in adult mortality
or arrestment in immature development. Hama-
mura (1980) also found prolonged longevity of
adults during the winter and proposed that citrus
leafminers overwinter as adults in citrus groves
in Hiroshima prefecture in Japan.
Photoperiod is a highly predictable component
of both the tropical and temperate environment
(Denlinger 1986; Pieloor & Seymour 2001) and
has been shown to regulate diapause in several


Mean no. of
Date when Days to first citrus Adult
trees were adult leafminers Larval emergence Proportion of Adult longevity (SD)
placed in emergence per tree at mortality rate pupae in
shade house (degree days) pruning (SD) (SD) (SD diapause days degree days

Oct. 15 18 (142.4) 69.3 (22.4) 0.00 (0.00) a 0.86 (0.06) a 0 18.4 (17.2) a 97.0 (61.6) a
Nov. 11 32 (116.5) 47.0 (18.4) 0.08 (0.06) a 0.65 (0.08) a 0 35.1 (19.8) b 91.0 (52.7) a
Nov. 22 44 (121.1) 63.9(26.8) 0.24 (0.09) b 0.43 (0.11) b 0 34.4 (16.1) b 85.8 (48.9) a

See text for statistics.
Numbers with different letters were different based on Dunn test (P < 0.05).

Florida Entomologist 89(3)




Fig. 1. Mes
during 2003-2
Weather Netv

tropical lepi
L. (Pieloor
osella Dyar
Heliothis ar
house 1982)
ers showed
duced by s
chamber ext
icant differ
emergence r

to short-day or long-day conditions in growth
chambers maintained at 18C. Development time
from egg to adult was also not affected by short-
day conditions in both sexes (female t = 0.967, df
= 74, P = 0.337; male t = 0.141, df = 79, P = 0.888)
(Table 2). Finally, no signs of reproductive dia-
pause were found from the dissections of adult fe-
males reared under short-day or long-day condi-
tions (Table 3). Only one female citrus leafminer
out of 52 that had been placed in 8L: 16D at 18C
lacked mature eggs and had ovarioles reduced in
size. Hamamura (1980) found adult females ab-
sorbed mature eggs when adult females were ex-
posed to 10C and 12L:12D conditions and pro-


These experiments provide no evidence that
the Florida population of the citrus leafminer has
a genetically determined diapause. It is always
frustrating to obtain "negative" data because we
can not prove a negative if we have not tested all
possible combinations of diapause-inducing cues.
Despite this, the weight of the evidence reported
here, and the anecdotal evidence from several
studies of the phenology of the citrus leafminer,
indicate that P citrella lacks the ability to over-
winter in diapause.


Development in days (SD)
Photoperiod Proportion Emergence Proportion
condition n of larvae pupated rate of pupae of females Females Males

Short day 115 0.96 0.74 0.46 33.9 (2.2) 34.4 (2.4)
Long day 109 0.97 0.83 0.51 33.3 (2.1) 34.3 (2.9)
P 0.7221 0.1371 0.6321 0.3372 0.8882

From Fisher's exact test.
From t-test.

vided with honey solution (25%) for a long period
of time (8% absorption from exposure of 10 d and
81% from 40 d). The effects of the parental moth's
experiences also were assessed. Adult citrus leaf-
miners that had been reared at 18C with a pho-
toperiod of 8:16 (L:D) in a growth chamber were
allowed to oviposit on a lemon tree with flushes
and their progeny were allowed to develop under
these same conditions. No signs of developmental
arrestment in eggs, larvae, pupae, or adults were
found from two replications involving 15 and 20
n,15 12s 1s15 2/15 3a/1 parental moths, respectively, indicating that par-
2W3 2W ents reared under potentially diapause-inducing
oat conditions did not influence their progeny's abil-
ity to enter diapause (unpublished data).
an daily temperatures in Alachua County In our studies, the Florida population of the
004 obtained from the Florida Automated citrus leafminer did not show any evidence of ar-
work rested development in immature or adult stages
during the growth chamber or outdoor rearing ex-
periments. Immatures probably overwinter in a
dopterans such as Hypolimnas bolina few tender new shoots or leaves or as adults dur-
& Seymour 2001), Diatraea grandi- ing the winter in mid and central Florida. Be-
Kikukawa & Chippendale 1983), and cause citrus trees produce some flushes, even in
rmigera Hubner (Hackett & Gate- winter, citrus leafminers may not need to dia-
.However, immature citrus leafmin- pause to synchronize with their food resources.
no arrestment in development in- Hence, as Knapp et al. (1995) suggested, suppres-
hort-day conditions in the growth sion of winter flushing by limiting irrigation, in
)eriments at 18C (Table 2). No signif- combination with cool temperatures, may help to
fences were found in pupation and reduce overwintering populations of leafminers in
ates between the populations exposed Florida.

September 2006

Lim & Hoy: Overwintering of Citrus Leafminer


Photoperiod Total number of Number of Proportion of females
condition adults emerged females dissected with mature eggs P

10L:14D 83 47 1.00 1.000'
14L:10D 85 44 1.00

8L:16D 114 52 0.98 1.000'
16L:8D 68 32 1.00

'From Fisher's exact test.


This research was supported by the Davies, Fischer
and Eckes Endowment in Biological Control and a
TSTAR Caribbean Special Research Grant. We thank R.
Wilcox for maintaining the lemon trees.


ALI MAFI, S., AND N. OHBAYASHI. 2004. Seasonal preva-
lence of the citrus leafminer, Phyllocnistis citrella
Stainton (Lepidoptera: Gracillariidae) and its para-
sitoids in controlled and uncontrolled Citrus iyo
groves in Ehime Prefecture, Japan. Appl. Entomol.
Zool. 39: 597-601.
AYOUB, M. A. 1960. Phyllocnistis citrella Stainton, A
main citrus pest in Saudi Arabia. Bull. Entomol. Soc.
Egypt 44: 387-391.
BADAWY, A. 1967. The morphology and biology of Phyl-
locnistis citrella Stainton, a citrus leafminer in
Sudan (Lepidoptera: Tineidae). Bull. Entomol. Soc.
Egypt 51: 95-103.
BARROGA, S. F. 1969. Biological notes and control of cit-
rus leafminer (Phyllocnistis citrella Stainton) affect-
ing citrus seedlings. Philippine J. Plant Industry 33:
BATRA, R. C., AND G. S. SANDHU. 1981. Differential pop-
ulation of citrus leafminer and its parasites on some
commercial citrus cultivars. J. Res. Punjab Agric.
Univ. 18: 170-176.
BECK, S. D. 1980. Insect Photoperiodism, 2nd edition,
Academic Press, New York, 387 pp.
CLAUSEN, C. P. 1931. Two Citrus Leafminers of the Far
East. U. S. Dept. Agric. Tech. Bull. 252, 13 pp.
DANILEVSKII, A. S. 1965. Photoperiodism and Seasonal
Development of Insects. Oliver and Boyd, Edin-
burgh, 283 pp.
DENLINGER, D. L. 1986. Dormancy in tropical insects.
Annu. Rev. Entomol. 31: 239-264.
DENLINGER, D. L. 2002. Regulation of diapause. Annu.
Rev. Entomol. 47: 92-122.
AND T. RILEY. 2001. The citrus canker epidemic in
Florida: the scientific basis of regulatory eradication
policy for an invasive species. Phytopathology 91:
D. S. ACHOR 1996. Citrus leafminer exacerbated the
outbreak of Asiatic citrus canker in South Florida, p.
83 In M. A. Hoy [ed.], Proceedings, International

Meeting: Managing the Citrus Leafminer, 22-25
April 1996, Orlando, Florida. University of Florida,
Gainesville, FL.
HACKETT, D. S., AND A. G. GATEHOUSE. 1982. Diapause
in Heliothis armigera (Hubner) and H. fletcheri
(Hardwick) (Lepidoptera: Noctuidae) in the Sudan
Gezira. Bull. Entomol. Res. 72: 409-422.
HAMAMURA, T. 1980. Studies on the overwintering of
the citrus leaf miner, Phyllocnistis citrella Stainton
(Lepidoptera: Lyonetiidae). Bull. Fruit Tree Res.
Stn., E3: 99-122 (In Japanese with English ab-
HoY, M. A., AND R. NGUYEN. 1997. Classical biological
control of the citrus leafminer Phyllocnistis citrella
Stainton (Lepidoptera: Gracillariidae): Theory, prac-
tice, art and science. Trop. Lepidop. 8 (suppl. 1), 1-19.
adaptations of populations of the southwestern corn
borer, Diatraea grandiosella from tropical and tem-
perate regions. J. Insect Physiol. 29: 561-567.
rus Leafminer, Phyllocnistis citrella Stainton: Cur-
rent Status in Florida-1995. University of Florida,
Institute of Food and Agricultural Sciences, Gaines-
ville, FL.
LEES, A. D. 1955. The Physiology of Diapause in Arthro-
pods, Cambridge University Press, UK, 151 pp.
SON. 1999. Volumes of petroleum spray oil required
for control of Phyllocnistis citrella Stainton (Lepi-
doptera: Gracillariidae) in mature citrus orchards.
Australian J. Entomol. 38: 141-144.
PANDEY, N. D., AND Y. D. PANDEY. 1964. Bionomics of
Phyllocnistis citrella. Indian J. Entomol. 26: 417-
sonal abundance of Phyllocnistis citrella (Lepi-
doptera: Gracillariidae) and its parasitoids in South
Florida citrus. Environ. Entomol. 25: 698-702.
PIELOOR, M. J., AND J. E. SEYMOUR. 2001. Factors affect-
ing adult diapause initiation in the tropical butterfly
Hypolimnas bolina L. (Lepidoptera: Nymphalidae).
Australian J. Entomol. 40: 376-379.
SAS INSTITUTE. 1995. SAS for Windows, Release 6.11
Edition. SAS, Cary, NC.
TAUBER, M. J., C. A. TAUBER, AND S. MASAKI. 1986. Sea-
sonal Adaptations of Insects. Oxford University
Press, New York, 411 pp.

366 Florida Ento

UJIYE, T., 2000. Biology and control of the citrus leaf-
miner, Phyllocnistis citrella (Lepidoptera: Gracillari-
idae) in Japan. Japan Agricultural Research
Quarterly (JARQ) 34: 167-173.
2001. Thermal requirements for ontogenic develop-
ment ofPhyllocnistis citrella Stainton (Lepidoptera:
Gracillariidae). Idesia 19: 35-38.
DAVIES. 2000. Field evaluation of integrated pest man-


ologist 89(3) September 2006

agement-compatible pesticides for the citrus leafminer
Phyllocnistis citrella (Lepidoptera: Gracillariidae) and
its parasitoid Ageniaspis citricola (Hymenoptera: En-
cyrtidae). J. Econ. Entomol. 93: 357-367.
WILSON, C. G. 1991. Notes on Phyllocnistis citrella
Stainton (Lepidoptera: Gracillariidae) attacking
four citrus varieties in Darwin. J. Aust. Entomol.
Soc. 30: 77-78.
ZAR, J. H. 1996. Biostatistical Analysis, 3rd edition,
Prentice Hall, Upper Saddle River, NJ.

Epsky et al.: Lobate Lac Scale Infestation


USDA/ARS, Subtropical Horticulture Research Station, 13601 Old Cutler Rd., Miami, FL 33158

'USDA/APHIS-CPHST, Subtropical Horticulture Research Station, 13601 Old Cutler Rd., Miami, FL 33158

2University of Florida, Tropical Research and Education Center, 18905 SW 280th Street, Homestead, FL 33031


The lobate lac scale, Paratachardina lobata lobata (Chamberlin) was first found in south
Florida in 1999. Reported hosts are present in the germplasm collection located at the
USDA/ARS Subtropical Horticulture Research Station in Miami, and the scale was first
found there in the summer of 2002. A study was initiated to determine the spatio-temporal
dynamics of a lobate lac scale infestation at SHRS from Jul 2003 to Jul 2005. Numbers and
percentages of viable adults, and reproductive success as indicated by ratio of nymphs to vi-
able adults (<2 cm diam and 30 cm long branch sample) were recorded. There were 55 plants
evaluated over the ~80 hectares study site. Infestation increased from 42% of sampled
plants at the start of the study to 75% at the end, and most of the plants had low or moderate
levels of infestation (between 0 and 100 adults per 30 cm branch) over the course of the
study. Percentage of non-viable adults dropped from ~27% at the start of the study to ~7%
by the end of the study, and ratio of nymphs to viable adults dropped from ~9% to ~2%. Spa-
tial analysis showed that initial infestations were along the eastern edge of the sampled
area, with populations declining over the first half of the study but then increasing during
the second half. Over the course of the study, heavy infestations (>100 scales per 30 cm
branch) were found on only seven host plants. Among plants located in areas of high infes-
tation probabilities, individual host susceptibility appeared to be the primary factor regulat-
ing infestation level.

Key Words: Lobate lac scale, infestation level, distribution, spatial analysis


La escama de laca lobulada, Paratachardina lobata lobata (Chamberlin), fue encontrada por
primera vez en el sur del estado de Florida en 1999. Los hospederos reportados estan pre-
sentes en la colecci6n de germplasma localizada en la Estaci6n de Investigaci6n de Horticul-
tura Subtropical (EIHS) de USDA/ARS en Miami, donde se encontr6 la escama por primera
vez en el verano del 2002. Desde julio de 2003 hasta julio de 2005 se inicio un studio para
determinar la dinamica espacial-temporal de la infestaci6n de la escama de laca lobulada en
EIHS. El numero y el porcentaje de los adults viables, asi como su 6xito reproductive indi-
cado por la raz6n de las ninfas con los adults viables en las plants (en muestras de ramas
<2 cm diam y 30 cm de largo) fueron registrados. Habian 55 plants evaluadas por las ~80
hectares en el sitio del studio. La infestaci6n aument6 desde 42% en las plants muestradas
en el inicio del studio hasta 75% al final, y la mayoria de las plants tenian un nivel de in-
festaci6n bajo o moderado entiree 0 y 100 adults por 30 cm de rama) sobre el recorrido del
studio. El porcentaje de los adults no viables baj6 de ~27% en el inicio del studio a ~7%
para el final del studio, y la raz6n de las ninfas con los adults viables baj6 del ~9% al 2%.
El andlisis espacial mostr6 que las infestaciones iniciales estaban localizadas por el borde
que da al este del area muestreado, con las poblaciones declinando en la primera mitad del
studio pero luego aumentando durante la segunda mitad. Sobre el recorrido del studio, in-
festaciones altas (>100 escamas por 30 cm de rama) fueron encontradas en solo siete plants
hospederas. Entre las plants en areas de probabilidades de infestaciones altas, los hospe-
deros en forma individual susceptibles paracen ser el factor principal en regular el nivel de
la infestaci6n.

The lobate lac scale, Paratachardina lobata lo- cus rosasinensis L. in 1999 (Hamon 2001). It was
bata (Chamberlin), an insect native to India and first found in western Miami-Dade County, Flor-
Sri Lanka, was first found in the U.S. when it was ida in two locations in 2000, then in six locations
discovered in Broward County, Florida on Hibis- in 2001 (Howard et al. 2004) and subsequently in

Florida Entomologist 89(3)

over 30 locations throughout Miami-Dade County
by 2002 (FL Dept. Agric. and Cons. Serv., personal
communication). The number of reported host
species increased rapidly from seven in the initial
report to over 120 species in 44 families by Oct
2002 (Howard et al. 2004). The host list has since
increased to include 160 plants in 49 families
(Pemberton 2003a), and includes a number of spe-
cies native to Florida as well as exotic species that
include commercial fruit and ornamental trees.
As part of the National Germplasm Repository,
the USDA/ARS, Subtropical Horticulture Re-
search Station (SHRS) located in Miami, FL,
maintains U.S. clonal collections of tropical and
subtropical plants including mango Mangifera in-
dica L., avocado Persea americana Miller, sugar-
cane Saccharum officinarum L. and related
grasses, ornamentals and other tropical crops. Lo-
cated on the station are 254 genera and 557 spe-
cies planted over an area of ~80 hectares (Anony-
mous 2005). Of the 45 plant families listed as lo-
bate lac scale hosts (Howard et al. 2004), 21 are
represented in SHRS germplasm. SHRS is lo-
cated in eastern Miami-Dade County along Bis-
cayne Bay. An infestation of lobate lac scale was
first found at SHRS on a black olive tree, Bucida
buceras L., during the summer of 2002
(F. Howard, personal communication).
To assess the invasive potential of this insect,
Pemberton (2003b) evaluated an infestation
among plants in a 0.1 hectare yard with a diverse
planting of potential host plants in Broward
County, Florida in Jul 2002. In that study, 37 of 67
plant species had infestations ranging from a few
to many scales and the most severely infested
species was wax myrtle, Myrica cerifera L. Essen-
tially nothing is known about the biology and con-
trol of this pest (Howard et al. 2004), and because
it was not known how long the scale had been
present in the 0.1 hectare yard, it could not be de-
termined if infestation levels observed were re-
lated to time elapsed since initial attack or to sus-
ceptibility of the host plant (Pemberton 2003b).
Laboratory rearing methods and procedures were
not available to address these questions; there-
fore a field study was initiated in the summer of
2003 to study the spatial and temporal aspects of
the relatively recent infestation of the scale at
SHRS. Results of this study will provide informa-
tion on the population dynamics of the lobate lac
scale, an essential prerequisite for implementa-
tion of a pest management program.


The plant database in the USDA/ARS Germ-
plasm Resources Information Network (GRIN)
(Anonymous 2005) was reviewed to determine the
presence of host plants (Howard et al. 2004) in the
collections at SHRS. An initial sampling of repre-
sentative host plants was conducted from Jul to

Aug 2003. Subsequent samplings were performed
at ~ 6 month intervals (during Jan-Feb 2004, Jun-
Jul 2004, Jan-Feb 2005, and Jun-Jul 2005) for a
total of five sampling periods. All plants were as-
signed identification numbers, plant locations
within the ~790 by ~1000 m experimental area at
SHRS were obtained by GPS (GPS III Plus;
Garmin International, Inc., Olathe, KS) and were
recorded in coordinates of longitude and latitude.
Plants were visually inspected by examining
branches that were <2 cm diam (Howard et al.
2004), and presence or absence of scales was re-
corded. If scales were observed, then five subsam-
ples of branches (<2 cm diam and 30 cm long)
were collected. Subsamples were brought to the
laboratory and examined under a stereomicro-
scope (10x), and number of adult scales per sub-
sample was determined. Scales were recorded as
adult if the characteristic lobes were observed
(Chamberlin 1923, 1925; Howard et al. 2004).
Males have not been observed in Florida (Hamon
2001; Howard et al. 2004), and all scales were as-
sumed to be female. Adults were further classified
as viable or non-viable. Viable adult scales had a
shiny red appearance, while non-viable adult
scales had a dull, reddish-purple color and ap-
peared dry. When the viability of the adult was
questionable, the scale was removed from the
branch, and presence or absence of a live insect
was confirmed by dissection. Percentage of non-
viable adults was determined by dividing the
number of non-viable adults by the total number
of adults per subsample. For all but the first sam-
pling period, numbers of nymphs per subsample
were also recorded. The ratio of number of
nymphs to number of viable adults was used as an
indicator of reproductive success of the scales on
the different hosts.
Infestation level on each host was based on the
average number of adults per 30 cm length of
branch. Infestations were rated as heavy (x >
100), moderate (10 < x < 100), low (0 < x < 10), or
not infested (x = 0) (Pemberton 2003b). Two-way
analysis of variance with interaction was used to
determine the effects of infestation level and sam-
pling period on percentage of adults that were not
viable and on the ratio of nymphs to viable adults
using a mixed model in Proc GLM (SAS Institute
1998). The Box-Cox procedure, which is a power
transformation that regresses log-transformed
standard deviations (y + 1) against log-trans-
formed means (x + 1), was used to determine if
transformation was necessary to stabilize the
variance before analysis (Box et al. 1978).
Contour analysis was used to visualize the
spatial distribution of adult scale populations
within the experimental site. This was performed
with Surfer 8 (Golden Software, Inc., Golden,
CO), with GPS coordinates for host locations, a
100 by 65 grid, interpolation by kriging, and a lin-
ear variogram model. For this analysis, the raw

September 2006

Epsky et al.: Lobate Lac Scale Infestation

insect counts (number of adult scales per sample
per plant) were converted to indicator variables
(Brenner 1993; Arbogast et al. 2000) to reflect the
level of infestation. A variable of"1" was assigned
to plants with >10 scales per sample, and a vari-
able of "0" was assigned to plants with <10 scales
per sample. The threshold value of 10 was chosen
to differentiate the moderate and heavy infesta-
tion levels from the low and no infestation levels,
according to the rating system of Pemberton
(2003b). Surfer grids were generated from the in-
dicator variables, and maps were constructed us-
ing probability contours to highlight areas with
moderate to high infestations at each of the five
sampling periods.


The plant species and average number of adult
scales per 30 cm branch per plant for each of the
five sampling periods are given in Table 1. Of the
128 plant species included in the Oct 2002 host
list (Howard et al. 2004), 35 host species (37
plants with two species sampled twice) were eval-
uated in the initial 2003 sampling. In addition, 11
non-listed species from reported host genera were
evaluated, including seven Ficus spp. on which
SHRS personnel noted infestations. Thus, a total
of 48 plants representing 46 species were exam-
ined in Jul-Aug 2003. By fall 2003, heavy scale in-
festations had been discovered on twoAntidesma
species (Euphorbiaceae), specificallyA. dallachy-
anum Baill. andA. bunis (L.). Samples were sent
for confirmation of identification to FDACS, DPI
in Gainesville, FL. These two plants and another
related but uninfested plant, A. venosum E. Mey,
were added to the survey during the Jan-Feb 2004
sampling period, along with two infested Ficus
citrifolia Mill. Some of the host plants could not
be sampled over the full course of the study due to
plant death, construction activities that limited
access to plants, or tree trimming activities by
maintenance personnel that removed most of the
small, low branches suitable for scale infestation
and for sampling.
The total number of plants evaluated over the
5 sampling periods was 55, but the number sam-
pled during a single period ranged from 49 during
Jun-Jul 2004 to 39 during Jun-Jul 2005. There
were four plants sampled in the initial survey
that were not able to be sampled again due to con-
struction activities at SHRS (Terminalia muelleri
Bentham, Peltophorum pterocarpum (DC.), Ficus
benjamin L. and Macadamia integrifolia
Maiden & Betche) and none were infested. A
grapefruit, Citrus x paradisi Macfad., was sam-
pled in the initial survey and a miniature date
palm, Phoenix roebelenii O'Brien, was sampled in
the first two surveys and found to be not infested,
but were not sampled again due to plant death.
Five plants were sampled throughout the study

and were never infested. This included a pond ap-
ple, Annona glabra L., a pitch apple, Clusia rosea
Jacquin, a gumbo limbo, Bursera simaruba (L.)
Sarg., an avocado and a mango.
Among the plants on which scale infestation
was found at some time during the study, there
were few consistent patterns of population in-
crease or decrease (Table 1). Summary statistics
on lobate lac scale population parameters are
given in Table 2. Many of the plants inspected
were not infested during the first sampling pe-
riod, but the number of infested plants increased
over the course of the study. An increase in the
number of plants with low infestations was ob-
served over the first year of the study, but that
number declined by the final sampling period. An
increase also was observed in the number of
plants with moderate infestations, but little
change was noted in the number of plants that
had heavy infestations. Average number of adults
per infestation level remained fairly constant
within the low and moderately infested hosts, but
numbers increased over time in the heavily in-
fested hosts. No interactions between effects due
to infestation level or sampling period were found
on the percentage of scales that were non-viable,
but there was an effect due to sampling period (F
= 7.29; df = 4, 148; P < 0.001; Table 2). The aver-
age ( SD) percentage of non-viable adults ranged
from 27.1 25.89 to 37.9 31.04 during Jul-Aug
2003 to Jun-Jul 2004, but decreased to 6.5 7.46
to 12.0 18.21 during the Jan-Feb 2005 to Jun-
Jul 2005.
Less information was available on the ratio of
nymphs to viable adults (Table 2), but there was
no single factor or interaction effect of infestation
level or sampling period (F = 0.81; df= 11, 117; P
= 0.6266). The reproductive success of adults was
constant, as indicated by the ratio of nymphs to
viable adults, among the three infestation levels
during most of the sampling periods, but a trend
was detected with decreasing ratios as infestation
level increased during the third sampling period.
There was also a trend in decreasing reproductive
success over time, especially for the plants with
low infestation levels.
Spatial aspects of the scale infestation in the
experimental site are shown in Fig. 1, which
shows the location of host plants sampled and the
probability contour maps for moderate-heavy
scale infestations during the five sampling peri-
ods. The shading in the contour maps indicates in-
festation probability level, with the darker shad-
ing representing higher probability that a host
plant within that area will exceed the threshold of
10 adult scales per 30-cm branch. The areas with
the highest probabilities of infestation during the
initial Jul-Aug 2003 sampling were located on the
eastern half of the site (Fig. 1B). This area in-
cluded the black olive tree (#12) where lobate lac
scale was first detected at SHRS in 2002. By Jan-


Common name




Jan-Feb Jun-Jul
2005 2005

Anacardiaceae Mangifera indica L.
Annonaceae Annona cherimola Mi
Annonaceae Annona muricata L.



Annona squamosa L.
Bucida buceras L.
Conocarpus erectus L.
Terminalia catappa L.
Diospyros digyna Jacquin
Elaeocarpus japonicus Turcz.**
Antidesma bunius (L.) Spreng.
Antidesma dallachyanum Baill.*
Antidesma venosum E. Mey.*
Bauhinia sp.**
Inga sp.**
Quercus uirginiana Miller
Lagerstroemia sp.**
Lagerstroemia speciosa (L.) Pers.
Brosimum alicastrum Sw.
Ficus altissima Blume*
Ficus aurea Nutt.
Ficus citrifolia Mill.*
Ficus citrifolia Mill.*
Ficus microcarpa var. rigo (F.M. Bailey)
Ficus racemosa L.*
Ficus religiosa L.*
Ficus superba var. henneana (Miq.)*
Ficus sur Forssk*.
Ficus tinctoria sub. parasitica (Willd.)*
Ficus virens Aiton*
Callistemon viminalis (Sol. ex Gaertn.)
Eugenia uniflora L.
Pimenta dioica (L.)
Pimenta dioica (L.)

sugar apple
black olive
tropical almond
black sapote
Herbert River cherry
tassle berry
orchid tree
live oak
crape myrtle hybrid
queen's crape myrtle
Mayan breadnut
blume council tree
strangler fig
shortleaf fig
shortleaf fig
Indian laurel
cluster fig ficus
bodhi tree peepul tree
port hacking fig superb
cape fig
big leaved fig
weeping bottle brush
Suriname cherry

Infested plants* or genera** not listed previously as hosts for lobate lac scale (Howard et al. 2004).
ns = not sampled due to late addition or loss of plant due to plant death (allspice, rose) or tree trimming activities.



0.0 0.00
0.0 0.00
0.0 0.00
1.4 1.14
129.8 80.35
6.0 4.47
130.0 55.31
4.6 3.05
11.4 18.60
0.0 0.00
0.0 0.00
0.0 0.00
0.0 0.00
0.0 0.00
0.0 0.00
5.8 5.54
8.6 8.44
6.0 5.05
12.6 7.57
0.4 0.89
27.2 19.88
6.6 7.13
24.6 27.52
6.2 13.86
0.0 0.0
0.0 0.00
0.0 0.00
0.0 0.00

0.2 0.45
2.6 4.22
4.2 5.12
21.8 18.05
109.8 84.97
5.8 5.02
9.6 2.61
16.8 15.69
7.0 9.59
197.4 119.77
148.0 104.13
0.0 0.00
0.0 0.00
7.8 3.56
0.0 0.00
0.2 0.45
0.0 0.00
0.0 0.00
4.4 2.41
0.0 0.00
4.2 3.42
5.4 3.65
5.6 9.32
32.0 10.68
2.0 1.87
55.4 32.73
3.0 2.12
1.2 0.84
0.0 0.00
0.0 0.00
0.0 0.00

1.2 1.64
1.6 1.52
4.6 3.29
29.8 10.23
7.2 5.76
5.0 7.91
9.0 8.12
12.6 11.24
6.2 6.65
272.0 159.25
114.0 73.36
0.2 0.45
0.0 0.00
38.6 49.33
0.0 0.00
1.0 1.00
0.4 0.55
0.0 0.00
5.8 5.67
0.6 0.89
16.6 10.83
5.4 5.22
2.2 2.68
17.4 1 7.04
0.2 0.45
31.2 40.20
8.8 19.68
68.8 74.33
2.50 2.55
4.6 4.72
0.0 0.00
0.6 1.34
0.8 1.79

0.4 0.89
0.4 0.55
7.8 3.56
26.6 35.81
27.0 28.20
36.4 22.85
3.8 5.54
64.8 81.20
7.8 9.86
426.6 56.60
238.2 132.23
0.4 0.55
0.0 00
26.6 10.76
0.2 0.45
2.0 1.58
5.2 5.17
0.4 0.55
10.0 17.93
1.8 3.49
10.2 5.76
4.0 8.40
0.2 0.45
6.4 8.38
0.0 0.00
31.4 23.75
2.0 2.00
81.0 96.44
4.0 6.82
0.4 0.55
0.6 0.55
0.0 0.00

0.0 0.00
0.4 0.55
6.4 6.54
38.4 74.19
21.4 9.29
123.0 89.21
13.4 14.57
81.8 47.96
3.4 7.60
0.0 0.00
0.4 0.89
0.0 0.00
1.4 1.14
3.4 1.67
7.0 8.51
13.8 12.46
3.0 5.10
1.8 1.79
0.2 0.45
13.2 14.39
2.8 1.92
55.2 74.52
6.0 3.08
2.0 2.45
0.0 0.00

Epsky et al.: Lobate Lac Scale Infestation












cE .
Sa a
2 5-
c E^

Mc ( 0C

+1 +1 +1

CO 0

+1 +1 + +1
100 0

ccq q

(00 C.]
C<1 0 In
+1 +1 +1

in t-

~,In L-7
, C.] '-I
+1 +1 +1
q cq Dq
00 -1

cc In
t-iC (0
+1 +1 +1
C6 Cq c6
L- C.I "t


C]1 C.] 1-
+1 +1 +1
CMri 1 '
X0 CO '
C. i,
In c Cl

2. ^

'S .P


~S 2

c3 c"~i c"~ C)C) ^ "" r> ^
-S .^ .S I fl a U *, < *

o S ~ C tu t
ca ca ca ca cas u
(D IN~~$ CDC D n
d ^ ^iii^^ I~~~

^^^^^ ii^

Feb 2004 (Fig. 1C), a new focus of high probability
was obtained in the northwest corner of the site.
The three trees sampled in this area included a
custard apple, Annona reticulata L., a sugar ap-
ple, Annona squamosa L., and a black sapote, Di-
ospyros digyna Jacquin (#18). Of those three
trees, infestation increased from none to low in
the custard apple and from low to moderate in the
other two trees. Areas of 0.75 probability of at
least moderate infestation levels within the east-
ern portion were reduced slightly by Jan-Feb
2004, and were greatly reduced by the Jun-Jul
2004 sampling period (Fig. 1D). Probability of at
least moderate infestation levels then increased
by the Jan-Feb 2005 (Fig. 1E) through the Jun-Jul
2005 (Fig. 1F) sampling periods. The populations
detected in Jan-Feb 2004 in the northwest corner,
however, remained stable during the subsequent
sampling periods.
Of the 55 plant species sampled over the five
sampling periods, heavy scale infestations were
found on only seven plants, whose positions are
shown in Figure 1A. These were bignay (#5), Her-
bert River cherry (#6), carambola (#8), black olive
(#12), buttonwood (#17), cattley guava (#47) and
tropical almond (#53). Except for the initial sam-
pling period, most of the remaining plants sam-
pled had either low or moderate infestations. The
three plants with heavy infestations at the start
of the study (black olive, cattley guava, and tropi-
cal almond) dropped to moderate or low infesta-
tions over the time period of the study. The other
four plants either maintained heavy infestations
for several sampling periods or increased to heavy
infestations late in the study. No pesticide appli-
cation or other control measures were applied but
all were subject to pruning and other germplasm
maintenance operations that may have either
limited or increased the amount of small
branches available for additional scale population
growth. Spatial analysis of the adult distribution
showed an increase in infestation in the north-
west corner between Jul-Aug 2003 and Jan-Feb
2004 (due to moderate infestation on black sa-
pote, #18), but level of infestation in that area re-
mained constant over the remaining three sam-
pling periods. In contrast, scale infestations along
the eastern edge of the sampling area decreased
between Jul-Aug 2003 and Jun-Jul 2004 (prima-
rily on black olive and tropical almond) but then
increased over the final two sampling periods (on
other host plants). Between Jun-Jul 2004 and
Jan-Feb 2005, several hurricane-related wind
events impacted the plants in the sampling area.
It is not known why populations rebounded, but
post-hurricane regrowth of small branches may
have stimulated the population increase in sus-
ceptible hosts on which infestation had dropped to
low levels prior to the 2004 hurricane activity.
Over the two years of monitoring changes in lo-
bate lac scale populations on infested plants, indi-

. 0 coc<
0 C! 10
+1 +1 +1 +1
0 C\ I t
+ I L- +I

m0 m

d +I +I d
~, ,
+1 +1


Jul-Aug Jan-Feb Jun-Jul Jan-Feb Jun-Jul
Infestation level n 2003 n 2004 n 2004 n 2005 n 2005

Number of adult scales
none 28 0.0 0.00 19 0.0 0.00 11 0.0 0.00 10 0.0 0.00 10 0.0 0.00
low 9 5.1 2.59 17 4.1 2.76 22 3.1 2.84 22 2.7 2.70 14 3.1 2.30
moderate 8 33.0 22.47 7 34.6 15.23 14 34.6 23.28 12 32.7 20.86 13 31.0 23.98
heavy 3 137.4 12.82 5 167.6 69.47 2 241.3 139.65 3 284.8 125.18 2 319.7 278.18
Percent of adults that were dead
low 9 31.5 16.70 17 32.4 31.91 22 30.3 31.93 22 7.8 12.96 14 5.2 8.22
moderate 8 19.1 14.61 7 47.2 31.43 14 21.6 14.01 12 20.6 25.37 13 8.9 6.62
heavy 3 35.7 9.73 5 43.6 29.53 2 30.8 12.82 3 8.7 3.27 2 1.2 0.73
Ratio of nymphs to live adults
low not determined 17 8.2 + 8.10 20 10.0 31.61 22 0.8 1.29 12 0.9 0.96
moderate not determined 7 10.6 15.17 14 5.8 10.60 12 0.8 0.74 13 2.7 3.55
heavy not determined 5 7.4 7.59 2 1.2 0.55 3 0.8 1.04 2 2.0 1.98

Epsky et al.: Lobate Lac Scale Infestation

A. Host Locations
18 A
^0 A A
A 1 A


C. Jan-Feb2004

C. Jan-Feb 2004

-80.300 -80.296
Longitude (West)

F. Jun-Jul 2005

'--.25) -

-80.292 -80.300


Longitude (West)

Fig. 1. Location of plants sampled for lobate lac scale over ~80 hectares of germplasm maintained at the USDA/
ARS, SHRS in Miami, FL (A); and probability contour maps of moderate to heavy scale infestations detected from
July-August 2003 (B), January-February 2004 (C), June-July 2004 (D), January-February 2005 (E), and June-July
2005 (F). Identification numbers on map A indicate the most heavily infested host plants. Moderate to heavy infes-
tation is specified as more than 10 adult scales per 30 cm length branch (Pemberton 2003b).

vidual host susceptibility appeared to be the pri-
mary factor regulating population level once the
infestation was initiated. Percentage viable
adults and reproductive success was not affected
by infestation level. Scale populations on some
hosts fluctuated rapidly, with populations either
building up or crashing. Other plants had only
low infestation levels or remained uninfested
even though they have been reported as hosts for
the scale and were located near infested plants.
Seven plants evaluated in our study were listed
as highly susceptible to lobate lac scale (Howard

et al. 2004), and we found heavy infestations on
three, i.e., black olive, buttonwood, and caram-
bola. The other four plants (mango, strangler fig,
Indian laurel, and lychee) had either no or low in-
festations. A number of plants included in our
study also were evaluated by Pemberton (2003b)
and, of the highly susceptible plants that were
listed by Howard et al. (2004) and that were in-
cluded in both studies, there was a heavy infesta-
tion on carambola and moderate infestations on
black olive, lychee, and mango. Among the re-
maining plants, sugar apple had a moderate in-





Florida Entomologist 89(3)

festation and atemoya, suriname cherry, and all-
spice had low infestations in both studies; while
soursop had a lower infestation level and sapo-
dilla had a higher infestation level in our study.
Thus, there may be varietal differences among po-
tential hosts that will further influence host sus-
The recent invasion of this insect into the
study site presented a unique opportunity to fol-
low the spatio-temporal dynamics of the invasion
by what was, at that time, a new pest to the area.
The most dynamic changes within the experimen-
tal area throughout the study occurred in the area
of the initial infestation. Comparisons over time
found that the population decreased by the 2nd
and 3rd samplings, but then increased by the 4th
and 5th samplings. Since scales primarily infest
twigs and small branches <2 cm diam (Howard et
al. 2004), the hurricane activity in Aug-Septem-
ber 2004 may have resulted in a flush of new
growth on trees that were susceptible to scale in-
festations resulting in an increase in feeding sites
that promoted scale population growth. Addi-
tional studies are needed to determine growth pa-
rameters for lobate lac scale and to better under-
stand the relationship between host susceptibil-
ity and population growth of this insect. Identifi-
cation of host varieties that are resistant to lobate
lac scale infestation would provide an important
component for IPM approaches for this pest.


The authors are grateful to Pauline Anderson, Tracy
Magellan, Patricia Longas, Willie Hallmon, Alison Walker,
Kimone Anderson (USDA/ARS, Miami, FL), Lizandra
Nieves, Roger Coe (USDA/APHIS, Miami, FL) for techni-
cal help; Wilhemina Wasik, Ray Schnell, Tomas Ayala-
Silva (USDA/ARS, Miami, FL) for assistance with identi-
fication and location of plants monitored in this study;
Yvette Ogle (FLDACS, DPI, Gainesville, FL) for personal
communication on spread of lobate lac scale in Florida;
and Bill Howard, Sibylle Schroer (Univ. ofFL, FLREC, Ft.
Lauderdale, FL), Greg Hodges (FDACS-DPI, Gainesville,
FL), and Terry Arbogast (USDA/ARS, CMAVE, Gaines-
ville, FL) for reviews of this manuscript.


ANONYMOUS. 2005. USDA, ARS, National Genetic Re-
sources Program. Germplasm Resources Informa-
tion Network (GRIN). [Online Database] National
Germplasm Resources Laboratory, Beltsville, MD.
E. MCGOVERN. 2000. Monitoring insect pests in re-
tail stores by trapping and spatial analysis. J. Econ.
Entomol. 93: 1531-1542.
Box, G. E. P., W. G. HUNTER, AND J. S. HUNTER 1978.
Statistics for Experimenters. An Introduction to De-
sign, Data Analysis, and Model Building. J. Wiley &
Sons, New York, NY.
BRENNER, R. J. 1993. Preparing for the 21" century: Re-
search methods in developing management strate-
gies for arthropods and allergens in the structural
environment, In K. B. Wildey and W. H. Robinson,
[eds.], Proc. 1" Intern. Conf. Insect Pests in the Ur-
ban Environment.
CHAMBERLIN, J. C. 1923. A systematic monograph of the
Tachardiinae or lac insects (Coccidae). Bull. Ento-
mol. Res. 14: 147-212.
CHAMBERLIN, J. C. 1925. Supplement to a monograph
of the Lacciferidae (Tachardiinae) or lac insects
(Homoptera, Coccidae). Bull. Entomol. Res. 16: 31-
HAMON, A. 2001. Lobate lac scale, Paratachardina lo-
bata lobata (Chamberlin) (Hemiptera: Kerriidae).
Pest Alert, Florida Department of Agriculture and
Consumer Services, Division of Plant Industry.
FORD. 2004. Lobate lac scale, Paratachardina lobata
lobata (Chamberlin) (Hemiptera: Sternorrhyncha:
Coccoidea: Kerriidae). Univ. of Florida IFAS Exten-
sion EENY-276. http://creatures.ifas.ufl.edu/orn/
scales/lobate lac.htm.
PEMBERTON, R. W. 2003a. Potential for biological con-
trol of the lobate lac scale, Paratachardina lobata lo-
bata (Hemiptera: Kerriidae). Florida Entomol. 86:
PEMBERTON, R. W. 2003b. Invasion of Paratachardina
lobata lobata (Hemiptera: Kerriidae) in south Flor-
ida: a snapshot sample of an infestation in a residen-
tial yard. Florida Entomol. 86: 373-377.
SAS INSTITUTE. 1998. User's manual, version 7.0. SAS
Institute, Cary, NC.

September 2006

Gongalves et al.: Volatile Components from Ceratitis capitata


'Laborat6rio de Quimica Entomol6gica, Departamento de Quimica, Centro de Ci6ncias Exatas e Naturals,
Universidade Federal de Alagoas, Campus A. C. Simoes, 57072-970, Macei6, AL, Brazil

2Departamento de Engenharia Agronomica, Centro de Ci6ncias Biol6gicas e da Saude, Universidade Federal de
Sergipe, Campus Jos6 Aluisio de Campos, Jardim Rosa Elze, 49.100-000, Sao Crist6vao, SE, Brazil


The volatile compounds released by calling males of Ceratitis capitata and those that were
extracted from the salivary glands with n-hexane were analyzed by gas chromatography-
mass spectrometry. Twelve of the 24 compounds identified in the released volatiles, namely,
2-heptanone, 2,5-dimethylpyrazine, 3-octanone, ethyl hexanoate, methyl heptanoate, 2-
ethyl-1-hexanol, limonene, indene, ethyl heptanoate, methyl octanoate, a-trans-bergamo-
tene and (E,E)-a-farnesene, also were detected in the glandular extract. The similarities
found in the chemical profiles of the released volatiles and of the salivary gland suggest that
the latter is the storage site, and also perhaps the site of synthesis, of some of the pheromone
components in this species of fruit fly.

Key Words: Ceratitis capitata, sex attractant, pheromone, aeration extract; salivary gland
extract, gas chromatography-mass spectrometry


Os compostos volateis liberados por machos de Ceratitis capitata em chamamento e os de
suas glandulas salivares foram extraidos e analisados qualitativamente por cromatografia
gasosa acoplada a espectrometria de massas. Comparando a composicio dos extratos de ae-
raCao com a dos extratos de glandulas salivares observou-se que 12 dos 24 compostos iden-
tificados entire os constituintes volateis liberados foram tamb6m encontrados nestas
glandulas. Os compostos coincidentes foram: 2-heptanona, 2,5-dimetilpirazina, 3-octanona,
hexanoato de etila, heptanoato de metila, 2-etil-l-hexanol, limoneno, 1-H-indeno, hepta-
noato de etila, octanoato de metila, a-trans-bergamoteno e (E,E)-a-farneseno. As similarida-
des encontradas na composicio dos dois tipos de extratos sugerem que estas glandulas sao
os locals de armazenamento, e talvez de sintese, dos compostos coincidentes na especie.

Translation by the authors.

The Mediterranean fruit fly, Ceratitis capitata
Wiedemann (1824), is a polyphagous insect (Light
et al. 1988; Liquido et al. 1991; Jang et al. 1989).
It is the most invasive of all members of the Te-
phritidae (Zucchi 2001), and causes extensive
damage to fruit worldwide. It was reported nearly
50 years ago that the volatiles released by C. cap-
itata males attract females (Feron 1959). Since
that time various studies have been conducted in
order to identify the components of the phero-
mone mixture for possible control of field popula-
tions (Jacobson et al. 1973; Ohinata et al. 1977;
Baker et al. 1985; Jang et al. 1989; Coss6 et al.
1995). While the exact composition of the complex
pheromone still remains unknown, numerous vol-
atile components released by males have been
identified (Jacobson et al. 1973; Ohinata et al.

1977; Baker et al. 1985; Jang et al. 1989; Coss6 et
al. 1995).
Although males ofC. capitata release the puta-
tive pheromone components from the proboscis,
the anus and from the cuticular surface, the vola-
tile components may be synthesized in, and/or
stored in, different body structures including the
salivary and pleural abdominal glands, the crop,
the gut and the anal gland (Nation 1981, 1989,
1990; Teal et al. 1999; Lu & Teal 2001). Some of
the components of the pheromone mixture are
produced and released from salivary glands in
some Anastrepha species (Nation 1989, 1990;
Teles 1987; Lima et al. 1996). The salivary glands
may perform the same function in males of C. cap-
itata (Nation 1989, 1990; Teles, 1987; Lima et al.,
1996; Ibaies-L6pez & Cruz-L6pez, 2001), but

Florida Entomologist 89(3)

there are no reports concerning the composition of
the salivary glands. The aim of the present study
was to compare the chemical profiles of the sali-
vary glands with the mixture of volatiles released
by calling males of C. capitata in an attempt to
elucidate the pheromone composition of this com-
mercially important pest.

Insect Population
The wild population of male C. capitata em-
ployed in the present study were obtained from
larvae collected from infested starfruit harvested
from a domestic orchard located in the town of Rio
Largo (0928'42"S / 3551'12"W; altitude 39 m), in
the state of Alagoas, Brazil. Larvae were placed
for pupation in boxes (44 x 35 x 25 cm) con-
structed of expanded polystyrene and containing
vermiculite. After 13 to 15 days, adult male and
female flies emerged and were separated into
glass tanks (30 x 20 x 15 cm) that were main-
tained in the Chemical Entomology Laboratory at
the Federal University of Alagoas (Macei6-AL,
Brazil). Flies were held at a temperature of 25 +
1C with 60% relative humidity under a 14 h pho-
toperiod, and were fed a mixture of sucrose and
brewer's yeast (2:1, w/w).

Extraction of Salivary Glands of Calling Males

Eleven days after the emergence of adult flies,
salivary glands of 10 calling males were removed
under water by using entomological forceps and
the aid of a stereoscopic microscope. The glands
were placed into 2-mL vials, each containing 1 mL
of HPLC grade n-hexane (Aldrich). The vials were
sealed and stored in a freezer until required for
analysis. This procedure was replicated 8 times.

Trapping of Volatiles Released by Calling Males

A group of 20 calling males (11 days post-emer-
gent) was submitted to aeration for a period of 3 h
between 06.00 and 09.00 h when C. capitata is
known to be sexually active (Goncalves 2005). In-
sects were placed in a glass desiccator (180 mm
high, 200 mm diameter) that had been modified by
the addition of an inlet tube containing activated
charcoal to filter the incoming air, and an outlet
tube containing Tenax (100 mg; Chrompack) to
adsorb the released volatiles. Tenax was initially
washed with hexane to remove non-polar contami-
nants and subsequently with methanol to remove
polar contamination. After each solvent, the Tenax
trap was allowed to dry at room temperature in a
modified electric oven (Walita, Microchef Luxo)
with nitrogen flowing through it for a few minutes.
The trap was then heated for 3 h at 280C. The N2
flow rate was 1 L/min, flowing constantly and mea-
sured with an airflow meter (ELE International

Ltd ELE 503-070). An air flow of 0.5 L/min was in-
duced through the desiccator containing the flies
by connecting a water vacuum pump to the outlet
of the tube containing the adsorbent. Water, su-
crose, and brewer's yeast (2:1 w/w), were supplied
throughout the assay period. Following aeration,
the Tenax adsorbent was removed from the outlet
tube of the apparatus and the trapped volatiles
were eluted with 1 mL of HPLC-grade n-hexane.
Eluates were transferred to individual 2-mL glass
ampoules, which were sealed and stored in a
freezer until required for analysis. A blank aera-
tion experiment, in which water and the dietary
materials (but no insects) were placed in the desic-
cator, was carried out in order to determine if any
of the trapped volatile materials derived from the
food. This experiment was replicated 8 times and,
for each replicate, the insects were replaced by new
ones of the same age.

Gas Chromatographic-Mass Spectrometric (GC-MS)

Prior to the analysis of hexane extracts, each
sample was concentrated under a gentle stream of
nitrogen at room temperature to a final volume of
150 pL. Aliquots (1 uL) were injected into a Shi-
madzu model 17A gas chromatograph, equipped
with a Shimadzu non-polar capillary column (30 m
x 0.25 mm i.d.; 0.5 pm polydimethylsiloxane film)
and coupled to a Shimadzu model QP 5050A mass
selective detector. The chromatographic conditions
were as follows: oven temperature initially 30C
and increased to 250C at a rate of 8C/min; injec-
tor (splitless) temperature 200C; detector tem-
perature 270C; carrier gas (helium) flow rate 1
mL/min; MS ionization energy 70 eV.
Components were identified by comparison of
their retention times and MS fragmentation pat-
terns with those of authentic standards (obtained
commercially or laboratory-prepared). Identities
were confirmed by GC analyses of standards and
extracts under identical conditions. A solution
containing an isomeric mixture of a-farnesene, a
gift from Prof E. D. Morgan (Keele University,
Keele, UK), was used to confirm the identity of
(E,E)-a-farnesene. An authentic sample of 2-me-
thyl-4-heptanone was kindly provided by Dr. N.
Oldham (Oxford University, Oxford, UK) for com-
parison purposes.
In cases where standard compounds were not
available, identifications were carried out by com-
parison with reference spectra in the Wiley data-
base 275, the Registry of Mass Spectral Data
(McLafferty & Stauffer 1989), Stokes et al. (1983),
and Rocca et al. (1992).

Preparation of Esters
Authenticate standards of the methyl and the
ethyl esters of hexanoic, heptanoic, and octanoic

September 2006

Gongalves et al.: Volatile Components from Ceratitis capitata

acids were prepared by mixing alcohol (30 pL)
with acetic acid (30 pL) in a Keele micro-reactor
and adding concentrated sulphuric acid (10 pL)
(Attygalle & Morgan 1986). The resulting mixture
was heated for 12 h at 120C, neutralized with the
minimum quantity of sodium bicarbonate and ex-
tracted with n-hexane (500 pL). The extracts were
analyses by GC-MS and the identities of the prod-
ucts confirmed.


Components Extracted from Salivary Glands

The hexane extracts obtained from salivary
glands of C. capitata calling males were charac-
terized by the presence of a mixture of 14 volatile
compounds consisting of an aromatic hydrocar-
bon, alcohols, ketones, esters, nitrogen com-
pounds, and terpenoids (Table 1). 2-Heptanone
was the most abundant compound followed by
(E,E)-a-farnesene, ethyl hexanoate, 2,5-dimeth-
ylpyrazine, and 1-nonanol: together these compo-
nents accounted for 71.9% of the total extract. Li-
monene was the only monoterpene identified in
the salivary gland extract.

Volatiles Released by Calling Males

The volatile mixture released by calling males
of C. capitata was characterized by 24 compo-
nents (Table 2) belonging to chemical classes sim-
ilar to those found in the extracts of salivary
glands. The major component was (E,E)-a-farne-
sene, followed by geranyl acetate and 3-ethy-
loctenoate: together these compounds accounted
for 46.56% of the total extract. Minor components
of the extract included 2-heptanone, ethyl hex-

anoate, 2,5-dimethylpyrazine, indol, and 3-oc-
tanone. Limonene and (E)-p-ocimene were the
only monoterpene hydrocarbons identified in the
volatiles mixture.


Twelve of the 14 compounds identified in the n-
hexane extracts of salivary glands of C. capitata
calling males were present in the more complex
volatile mixture released by the males. While 1-
nonanol and ethyl octanoate were the only com-
pounds exclusive to the salivary gland extracts,
the proportions of 2-heptanone, 2,5-dimeth-
ylpyrazine, ethyl hexanoate, and (E,E)-a-farne-
sene varied significantly between the n-hexane
and the aeration extracts. In contrast, the vola-
tiles released by calling males contained 12 exclu-
sive compounds, namely 1-heptanol, linalool,
(Z,Z)-3,6-nonadien-l-ol, 2-phenylethyl acetate,
ethyl 3-octenoate, 2,6-dimethylpyrazine, indol, 2-
methyl-4-heptanone, (E)-p-ocimene, geranyl ace-
tate, trans-caryophyllene, and a-copaene.
Seventeen compounds, namely, 1-heptanol, 2-
heptanone, 2-phenylethyl acetate, 2,5-dimeth-
ylpyrazine, 2,6-dimethylpyrazine, 2-methyl-4-
heptanone, 3-octanone, 1-nonanol, methyl hep-
tanoate, 2-ethyl-l-hexanol, indene, ethyl hep-
tanoate, (Z,Z)-3,6-nonadien-l-ol, ethyl octanoate,
trans-caryophyllene, a-copaene, and a-trans-ber-
gamotene, were identified in salivary gland ex-
tracts or the aeration mixture of calling males of
C. capitata for the first time. Nine compounds pre-
viously reported present in C. capitata were de-
tected in the aeration or salivary gland extracts
investigated in the present study, namely, ethyl 3-
octenoate, geranyl acetate, (E,E)-a-farnesene,
and linalool (Baker et al. 1985; Jang et al. 1989),


Compound Retention time (min) Percentage compositionb

2-Heptanone 8.14 19.57 ( 3.45)
2,5-Dimethylpyrazine 8.58 12.33 ( 2.06)
3-Octanone 10.26 6.32 ( 3.16)
Ethyl hexanoate 10.60 13.93 ( 1.79)
1- Nonanol 10.71 10.84 ( 2.37)
Methyl heptanoate 11.13 2.93 ( 1.19)
2-Ethyl-l-hexanol 11.29 2.01 ( 0.65)
Limonene 11.39 3.43 ( 1.01)
Indene 11.43 2.99 ( 0.67)
Ethyl heptanoate 12.63 2.93 ( 0.79)
Methyl octanoate 13.16 2.19 ( 0.84)
Ethyl octanoate 14.56 4.49 ( 1.23)
a-trans-Bergamotene 19.79 0.97 ( 0.19)
E,E-a-Farnesene 19.99 15.02 ( 1.25)

Peaks were identified by comparison of retention times and mass spectral data with those of authentic samples.
bValues shown are means ( standard error: n = 8).

Florida Entomologist 89(3)


Compound Retention time (min) Percentage compositionb

2-Heptanol 6.40 2.50 ( 0.86)
2-Heptanone 8.13 9.11 ( 2.14)
2-Phenylethyl acetate 8.17 1.61 (+ 0.53)
2,5-Dimethylpyrazine 8.53 5.68 ( 1.77)
2,6-Dimethylpyrazine 8.73 0.49 (+ 0.13)
2-Methyl-4-heptanone 8.91 0.48 ( 0.20)
3-Octanone 10.25 4.36 ( 1.26)
Ethyl hexanoate 10.59 6.87 ( 1.05)
Indol 10.71 5.17 (+ 2.18)
Methyl heptanoate 11.12 1.21 ( 0.49)
2-Ethyl-l-hexanol 11.29 1.62 ( 0.57)
Limonene 11.39 3.58 ( 0.91)
Indene 11.43 2.04 ( 0.65)
(E)-P-Ocimene 11.71 0.29 ( 0.07)
Ethyl heptanoate 12.63 1.38 ( 0.49)
Linalool 12.78 1.03 ( 0.29)
(Z,Z)-3,6-Nonadien-l-ol 13.16 1.35 ( 0.22)
Methyl octanoate 13.75 0.87 ( 0.36)
Ethyl 3-octenoate 14.56 11.09 ( 3.08)
Geranyl acetate 17.78 13.34 ( 1.57)
trans-Caryophyllene 18.07 2.42 ( 0.54)
a-Copaene 18.45 0.95 ( 0.37)
a-trans-Bergamotene c 19.76 0.42 (+ 0.12)
E,E-a-Farnesene 19.99 22.13 (+ 3.03)

aPeaks were identified (except as indicated otherwise) by comparison of retention times and mass spectral data with those of au-
thentic samples.
bValues shown are means (+ standard error: n = 8).
Peak identified by comparison of mass spectral data with literature values (see Materials and Methods).

ethyl hexanoate, limonene, (E)-p-ocimene, and
methyl octanoate (Jang et al. 1989), and indol
(Cosse et al. 1995).
Considering the 26 compounds presently iden-
tified in the aeration extracts obtained from C.
capitata calling males with the 59 compounds
previously reported by Jang et al. (1989), it is
clear that the volatile mixture obtained in this
study was much less complex than that described
earlier. Such a difference may be due to dissimilar
types of fruit from which the insects were ob-
tained, and/or by different diet given to the in-
sects during the larval phase (Goncalves, 2001;
Silva, 2005). Both factors may exert a direct influ-
ence on the volatiles produced and released by
adult males.
By comparison with previous studies, it is ap-
parent that (Z,Z)-3,6-nonadien-l-ol, trans-caryo-
phyllene, a-trans-bergamotene, (E,E)-a-farne-
sene, limonene, and/or (E)-p-ocimene are com-
mon to the volatile emissions of C. capitata,
Anastrepha suspense and A. ludens (Rocca et al.
1992), and that limonene, (Z)-p-ocimene, and
2,5-dimethylpyrazine are common to C. capitata
and A. fraterculus (Lima et al. 2001). Analogous
to the situation for A. obliqua (Goncalves 2005),

linalool, (Z,Z)-3,6-nonadien-l-ol, (E)-p-ocimene,
geranyl acetate, trans-caryophyllene, and a-co-
paene were identified only in the aeration ex-
tracts obtained from calling males, suggesting
that these compounds are synthesized and
stored in different body structures in insects of
the genera Ceratitis and Anastrepha. The ab-
sence of (Z,Z)-3,6-nonadien-l-ol in solvent ex-
tracts of salivary glands is in agreement with the
proposal of Nation (1989) that the nonenols are
stored in the intestine ofA. suspense. According
to this author, the sesquiterpenes a-trans-ber-
gamotene and (E,E)-a-farnesene are produced
by the salivary glands and released from the pro-
boscis, while the nonenols are released from the
everted anal tissue.
Considering that 12 of the 24 compounds iden-
tified in the aeration extracts from calling males
of C. capitata were also present in extracts of the
salivary glands, it is reasonable to suggest that
such glands are the site of storage, and possibly
also of synthesis, of these compounds in C. capi-
tata. Detailed biosynthetic and histological stud-
ies involving the salivary glands would be neces-
sary in order unambiguously to confirm this hy-

September 2006

Gongalves et al.: Volatile Components from Ceratitis capitata


The authors thank the Coordenacao de Pesquisa e
Ensino Superior (CAPES), Fundacao de Amparo a Pes-
quisa do Estado de Alagoas (FAPEAL, Proc. No.
20031029421-0), and Conselho Nacional de Desenvolvi-
mento Cientifico e Tecnol6gico (CNPq/ Proc. No. 471828/
2004-1) for financial support for this work and for the
provision of MSc and DSc grants.


ATTYGALLE, A. B., AND E. D. MORGAN. 1986. A versatile
microreactor and extractor. Anal. Chem. 58: 3054-
BAKER, R., R. H. HEBERT, AND G. G. GRANT. 1985. Iso-
lation and identification of the sex pheromone of the
Mediterranean fruit fly, Ceratitis capitata (Wied.). J.
Chem. Soc. Chem. Commun. 6: 824-825.
AND T. C. BAKER 1995. Electroantennographic and
coupled gas chromatographic-electroantenno-
graphic response of the Mediterranean fruit fly, Cer-
atitis capitata, to male-produced volatiles and
mango odor. J. Chem. Ecol. 21: 1823-1836.
FERON, M. M. 1959. Attraction chimiche du male de
Ceratitis capitata Wield. (Diptera: Trypetidae) pour
la femelle. Compt. Rend. Acad. Sci. 248: 2403-2404.
GONQALVES, G. B. 2001. Influ6ncia dos Constituintes
Volateis dos Hospedeiros Averrhoa carambola L. e
Mangifera indica L. na Composicio Quimica das
Substancias Liberadas por Machos de Anastrepha
obliqua e Ceratitis capitata (Diptera: Tephritidae).
Dissertacao de Mestrado, Universidade Federal de
Alagoas, 106 p.
GONQALVES, G. B. 2005. Extracgo e identificacao de con-
stituintes volateis das secrecoes de glandulas sali-
vares de machos de Ceratitis capitata e Anastrepha
obliqua (Diptera: Tephritidae) e de carambola e
manga. Ph.D Thesis, Universidade Federal de Ala-
goas, 160 pp.
IBANEZ-LOPEZ, A., AND L. CRUZ-LOPEZ. 2001. Glandulas
salivales de Anastrepha obliqua (Macquart)
(Diptera: Tephritidae) andlisis quimico y mor-
fol6gico, y atividade biol6gica de los components
volatiles. Folia Entomol. Mexicana 40: 221-230.
JONES, AND M. S. FUJIMOTO. 1973. Insect sex attrac-
tants. 13. Isolation, identification, and synthesis of
sex pheromones of the male Mediterranean fruit fly.
J. Med. Chem. 16: 248-251.
AND L. A. CARVALHO. 1989. Electroantennogram re-
sponse of the Mediterranean fruit fly, Ceratitis capi-
tata, to identified constituents from calling males.
Entomol. Exp. Appl. 50: 7-19.
LIGHT, D. M., E. B. JANG, AND J. C. DICKENS. 1988. Elec-
troantennogram response of the Mediterranean
fruit fly, Ceratitis capitata, to a spectrum of plant
volatiles. J. Chem. Ecol. 14: 159-180.
LIMA, I. S., P. E. HOWSE, AND D. R. STEVENS. 1996. Vol-
atile components from the salivary glands of calling
males of the south American fruit fly, Anastrepha
fraterculus: partial identification and behavioral
activity, pp. 107- 113 In B. A. MacPheron and G. J.
Steck [eds.], Fruit Fly Pests. A World Assessment of

their Biology and Management. St. Lucie Press,
London, UK.
LIMA, I. S., P. E. HOWSE, AND R. R. Do Nascimento.
2001. Volatile substances from male Anastrepha
fraterculus (Diptera: Tephritidae) : Identification
and behavioral activity. J. Braz. Chem. Soc. 12:
1991. Host plants of the Mediterranean fruit fly: An
Annotated World Review. Misc. Publ. Entomol. Soc.
Am. 77: 1-52.
Lu, F., AND P. E. A. TEAL. 2001. Sex pheromone compo-
nents in oral secretions and crop of male Caribbean
fruit flies, Anastrepha suspense (Loew). Arch. Insect
Biochem. Physiol. 48: 144-154.
Wiley/NBS Registry of Mass Spectral Data, Vols. 1
& 2. Wiley-Interscience, New York, USA.
NATION, J. L. 1981. Sex-specific glands in tephritid fruit
flies of the genera Anastrepha, Ceratitis, Dacus and
Rhagoletis (Diptera: Tephritidae). Int. J. Insect Mor-
phol. Embryol. 4: 27-30.
NATION, J. L. 1989. The role of pheromones in the mat-
ing system of Anastrepha fruit flies, pp. 189-205 In
A. S. Robinson and G. Hooper [eds.], Fruit Flies:
Their Biology, Natural Enemies and Control.
Elsevier, Amsterdam, Holand.
NATION, J. L. 1990. Biology of pheromone release by
male Caribbean fruit flies, Anastrepha suspense
(Diptera: Tephritidae). J. Chem. Ecol. 16: 553-571.
OTO, AND H. HIGA. 1977. Mediterranean fruit fly:
laboratory and field evaluations of synthetic sex
pheromones. J. Environ. Sci. Heath A 12: 67-78.
BATTISTE. 1992. Comparison of volatiles emitted by
male Caribbean and Mexican fruit flies. J. Chem.
Ecol. 18: 223-244.
SILVA, C. E. 2005. Estudo Comparativo do Ferom6nio de
Agregacao de Ceratitis capitata e Anastrepha obli-
qua (Diptera: Tephritidae) e dos Odores de seus Fru-
tos Hospedeiros Preferencial e Secundario, Averrhoa
carambola L. (Oxalidaceae) e Carica papaya L. (Car-
icaceae). Dissertacao de Mestrado, Universidade
Federal de Alagoas, 76 pp.
SPISHAKOFF, AND K. R. WILZER. 1983. Isolation and
Identification of novel lactones from male Mexican
fruit flies. Agric. Food Chem. 31: 1162-1167.
1999. Isolation and identification of terpenoid sex
pheromone components from extracts of hemolymph
of males of the Caribbean fruit fly. Arch. Insect Bio-
chem. Physiol. 42: 225-232.
TELES, M. C. 1987. Structure and development of spe-
cific sex glands in males of some Brazilian fruit flies
of the genus Anastrepha Schiner, 1868 (Diptera: Te-
phritidae), pp. 179-189 In R. Cavalloro [ed.], Fruit
Flies of Economic Importance. Proc. CEC/IOBC Int.
Symp. A. A. Balkema, Rotterdam, Holland.
ZUCCHI, R. A. 2001. Mosca do Mediterraneo, Ceratitis
capitata (Diptera: Tephritidae), pp. 15-41 In E. F.
Vilela, R. A. Zucchi and F. Cantor [eds.], Hist6rico e
Impacto das Pragas Introduzidas no Brasil. Holos,
Ribeirao Preto, Brazil.

Florida Entomologist 89(3)

September 2006


'USDA-APHIS-PPQ, 41-650 Ahiki Street, Waimanalo, HI 96795

2USDA-APHIS-PPQ, 1600 SW 23rd Drive, Gainesville, FL 32608

3USDA-APHIS-PPQ, 1833 57th Street, Sarasota, FL 34234

In Florida, an ongoing Preventative Release Program utilizes the sterile insect technique to
prevent infestations of the Mediterranean fruit fly (medfly), Ceratitis capitata (Wiedemann).
Unlike other such programs, which use plastic, storage (PARC) boxes, the Florida operation
holds pupae and newly emerged adults in eclosion towers prior to release. Although eclosion
towers save space and labor, few data exist regarding the quality of sterile male medflies
held in towers versus PARC boxes. Here, we present the results of field-cage trials compar-
ing the mating success of sterile males held in towers versus PARC boxes. In addition, pre-
vious research has shown that exposing PARC box-held males to the aroma of ginger root oil
(GRO) increases their mating competitiveness. Consequently, we assessed whether a similar
increase was evident for tower-held males. Finally, we performed a mark-release-recapture
study involving GRO-exposed and non-exposed males and estimated their relative survival
and dispersal in the field using the trap catch data. Data from the mating trials showed that
sterile males held in towers displayed approximately the same mating success as sterile
males held in PARC boxes and that, among tower-held males, GRO significantly increased
mating competitiveness relative to non-exposed males. In the trapping study, significantly
more GRO-exposed males were captured than non-exposed males, and there was no appar-
ent difference in the duration of the post-release interval over which GRO-exposed and non-
exposed males were captured. These findings, along with earlier comparisons of adult
weight, flight ability, and yield suggest no obvious differences in the efficacy of tower and
PARC-box eclosion systems for medfly sterile release programs.
Key Words: Mediterranean fruit fly, sterile insect technique, tower eclosion system

En Florida, un Programa de Liberaci6n Preventativa utiliza la t6cnica del insecto est6ril para
prevenir infestaciones de la mosca mediterranea de la fruta, Ceratitis capitata (Wiedemann). No
como en otros programs que usan cajas plasticas de almac6n (PARC), la operaci6n en Florida
se mantiene las pupas y adults reci6n salidos en torres de eclosi6n antes de ser liberados. Aun-
que las torres de eclosi6n garden espacio y require menos mano de obra, poca informaci6n
existe en cuanto da la cualidad de los machos est6riles de la mosca mediterranea de la fruta
mantenidos en las torres versus los machos en las cajas PARC. Aqui, presentamos los resultados
de las pruebas dejaulas en el campo comparando el 6xito de apareamiento de los machos este-
riles mantenidos en torres versus los mantenidos en cajas PARC. Ademas, ha mostrado en in-
vestigaciones anteriores que al exponer machos mantenidos en las cajas PARC al aroma del
aceite de la raiz de jengibre (ARJ) se aument6 su capacidad para competir en el apareamiento.
Por lo tanto, nosotros evaluamos si habia un aumento similar evidence en los machos manteni-
dos en las torres. Por ultimo, nosotros realizamos un studio de marcar-liberar-recapturar con
machos expuestos y no expuestos al ARJ y estimamos su sobrevivencia y dispersion relative en
el campo usando los datos del numero de machos capturados en trampas. Los datos de las prue-
bas de apareamiento mostraron que los machos est6riles mantenidos en torres tuvieron aproxi-
madamente el mismo 6xito de apareamento que los machos mantenidos en las cajas PARC y que,
entire los machos mantenidos en las torres, el ARJ aument6 significativamente su capacidad
para competir en el apareamiento en relaci6n con los machos no expuestos al ARJ. En el studio
de atrapamiento, significativamente mas machos expuestos al ARJ fueron capturados que ma-
chos no expuestos al ARJ, y no hubo una diferencia aparente en la duraci6n del intervalo pos-li-
beraci6n cuando los machos expuestos y no expuestos al ARJ fueron capturados. Estos hallazgos,
adjunto con las comparisiones anteriores del peso de adulto, habilidad con volar y rendimiento
sugieren que no hay diferencias obvias en la eficiencia de los sistemas de eclosi6n de torres y ca-
jas PARC para los programs de liberaci6n de machos est6riles de la mosca mediterranea de la

Shelly et al.: Mating Success of Male Medflies

The Sterile Insect Technique (SIT) is widely
used to suppress or eradicate infestations of the
Mediterranean fruit fly (medfly), Ceratitis capi-
tata (Wiedemann), a pest that attacks many com-
mercially important fruits and vegetables world-
wide (Hendrichs et al. 2002). Present SIT pro-
grams involve the production and sterilization of
a large number of male pupae (in genetic sexing
strains, a sex-linked, temperature sensitive lethal
[tsl] mutation allows selective elimination of fe-
males in the egg stage, Franz et al. 1996), a pre-
release holding period of 4-5 d during which pu-
pae mature and adults eclose and feed, and aerial
or ground release of the adult males into the envi-
ronment. As even this brief outline suggests, the
SIT is a relatively expensive management strat-
egy both in terms of materials and labor, and con-
sequently there is a persistent need to increase
the efficiency of this protocol and reduce costs.
In 2002, the Florida Preventative Release Pro-
gram against medfly began using a new system,
the Tower Eclosion (TE) system, for emergence
and feeding of adults prior to field release (Sal-
vato et al. 2004). Each tower consists of interlock-
ing, screen-paneled, aluminum frames or trays
(76 x 76 x 2.5 cm, l:w:h) stacked on a wheeled (i.e.,
mobile) base. Pupae are placed in a trough around
the edge of a tray, and food (a sugar-agar gelatin
routinely used in medfly SIT) is placed on each
screen panel. This procedure is repeated for each
tray, and a completed tower consists of 60-80 pu-
pal-holding trays. A small fan (blowing upwards)
is fitted on the top of each tower for ventilation.
Upon emergence, the flies move to the screen and
feed, and the puparia are left behind in the
trough. On the day of field release, towers are
moved into a cold room, where the puparia are
vacuumed from the troughs, and the trays are
manually removed from the tower and turned up-
side down over a container to collect the chilled
In Florida, the TE system replaced the Plastic
Adult Rearing Container (PARC) system, which is
still used in the ongoing medfly SIT programs in
California and Guatemala. In the PARC system,
pupae are placed in paper bags, which, in turn,
are placed in plastic (PARC) boxes (48 x 60 x 33
cm, l:w:h) with screened panels on the lid and
sides for ventilation. The sugar-agar gelatin is
placed on the lid screen, and the boxes are stacked
for storage. On the day of field release, boxes are
moved into a cold room, bags are removed, and
the boxes are turned upside down over a con-
tainer to collect the chilled flies.
The TE system offers several advantages over
the PARC system as follows: (1) the TE system re-
quires much less space than the PARC system to
hold a given number of flies, (2) the TE system re-
duces labor costs considerably, owing to auto-
mated pupal loading, puparia separation and dis-
posal, and tray washing, and (3) by eliminating

the use of paper bags, the TE system reduces sup-
ply costs, generates less waste, and reduces 'fly
loss' (flies remaining inside the discarded paper
bags) better than the PARC system. In addition to
economic issues, however, a comprehensive com-
parison requires data on the performance of
males held in the TE versus PARC systems. Po-
tential economic benefits may become less com-
pelling if, for some reason, males from the TE sys-
tem are of poor quality relative to males from the
PARC system. To date, only one study by Salvato
et al. (2004), who found no significant differences
in yield (emergence), adult weight, or flight abil-
ity between them, has compared the quality of
medfly males from the two systems.
To provide further comparisons between the 2
holding systems, the present study, which was
conducted in Hawaii and Florida, had 3 main ob-
jectives. First, in Hawaii we compared the mating
success of sterile tsl males from eclosion towers or
PARC boxes in competition with males from a re-
cently established wild colony for females from
that same colony. Second, in both Hawaii and
Florida, we assessed the effectiveness of ginger
root oil (GRO) in enhancing the mating perfor-
mance of sterile tsl males held in eclosion towers.
Prior work (Shelly 2001) has shown that exposure
to the aroma of GRO significantly increases the
mating performance of male medflies (a protocol
termed 'aromatherapy'). Application of GRO to in-
dividual PARC boxes has already been shown to
enhance male mating success (Shelly et al. 2004),
and here we test for a similar effect with eclosion
towers. Third, in Florida we performed a release-
recapture study comparing trap captures of GRO-
exposed versus non-exposed sterile tsl males to
identify potential differences in post-release dis-
persal and longevity. Previous studies (Shelly et
al. 2004; Levy et al. 2005) have not detected any
negative effect of GRO on survival of male med-
flies (except for a specific diet-related effect, Levy
et al. 2005), but these were conducted in field or
laboratory cages, and data are lacking from the
open field. Similarly, GRO exposure had no appar-
ent effect of male performance on a laboratory
flight-mill (S. Opp, personal communication), but
no data are available examining potential effects
of GRO exposure on male dispersal in the field.


Study Insects and Mating Trials-Hawaii

In Hawaii, tsl males were from the Vienna-7/
Tol-99 strain produced by the California Depart-
ment of Food and Agriculture Hawaii Fruit Fly
Rearing Facility (Waimanalo, Oahu, HI). This
strain has been mass-reared at the USDA-Mos-
camed facility in El Pino, Guatemala, since 1999,
and =1.25 million eggs from this facility were
used to start the colony in Hawaii in 2001. Males

Florida Entomologist 89(3)

used in the current study were dyed fluorescent
pink (DayGlo Color Corporation, Cleveland, OH)
and irradiated as pupae 2 d before eclosion in hy-
poxia at 150 Gy of gamma irradiation from a 13Cs
Owing to the low availability of wild flies, we
used flies from a recently established colony
(REC, 4 generations removed from the wild) in
the mating trials. This colony was derived from
300-500 adults reared from coffee berries col-
lected on the island of Kauai. Adults were held in
screen cages and provided with a sugar-protein
(yeast hydrolysate) mixture (3:1 by weight), wa-
ter, and an oviposition substrate (perforated plas-
tic vials containing small sponges soaked in
lemon juice). Eggs were placed on standard larval
diet (Tanaka et al. 1969) in plastic containers over
vermiculite for pupation. Adults used in the mat-
ing trials were separated by sex within 24 h of
eclosion, well before reaching sexual maturity at
5-8 d of age (T.E.S., unpublished data) and kept in
screen-covered buckets (5-liter volume; 100-125
flies per bucket) with ample food (sugar-protein
mixture) and water.
Five experiments were conducted in Hawaii.
In experiment 1, tsl males from eclosion towers or
PARC boxes competed against REC males for
REC females. On a given day, tsl pupae from the
same batch were placed in 2 towers and 2 PARC
boxes on the day of irradiation. The towers and
boxes were kept in the same room under the same
environmental conditions (25-27C, 60-85% RH,
12:12 L:D). Because the aforementioned rearing
facility supports the ongoing SIT program in Cal-
ifornia, our allotment of pupae was insufficient to
use fully loaded eclosion towers (1 tower holds =
1.25 million pupae). Consequently, we placed pu-
pae (350 mL, where 1 mL =60 pupae) in a single
tray per tower and left all other trays empty. For
all trials, pupae were placed in the 30th tray from
the bottom in towers built 60 trays high. A single
slab of sugar agar gel of the same size used in the
Florida program (15 x 9 x 3 cm, l:w:h) was placed
on the pupae-containing trays. For each PARC
box, we followed the protocol used in the Califor-
nia program and put 100 mL of pupae in each of 6
paper bags, which were then placed on the box
floor, and placed a 20 x 15 x 3 cm (l:w:h) slab of the
sugar agar gel on the screen panel on the box lid.
Four d after pupal placement (i.e., 2 d post-peak
emergence), flies were moved into a cold room
(4C for 10-15 min), and 100-200 flies from each
tower tray or PARC box were transferred to a
screen-covered, plastic bucket. These flies were
provided sugar-agar gel and held at 25-28 C until
testing 2 d later (i.e., when most tsl males were 4
d old).
In experiments 2-5, tsl males from towers with
or without GRO treatment competed against REC
males for REC females. In each of these experi-
ments, tsl pupae from the same batch were placed

in 4 towers on the day of irradiation (as above, 350
mL of pupae were placed in the 30th tray of a given
tower, and sugar agar was provided for food). Two
towers were placed in each of 2 separate rooms
under the same environmental conditions as
above. In one room, the towers received no GRO
exposure, and the flies (control males) were han-
dled in the same manner described above. In the
other room, we exposed each of the towers to GRO
by applying 1 mL of GRO to a cotton wick (2.5 cm
length, 1 cm diameter), placing the wick in an alu-
minum foil-lined Petri dish, and placing the Petri
dish on the floor beneath the tower.
In experiments 2-4, GRO exposure commenced
between 0800-0900 h on the day after peak emer-
gence and continued for 24 h, at which time the
flies (treated males) were chilled and collected. In
experiment 2, the flies were tested 2 d after collec-
tion (i.e., as above, the majority of tsl males were
4 d old when tested), and in experiment 3, we as-
sessed the long-term effectiveness of GRO expo-
sure by holding the tsl males for 5 d before testing
(i.e., control and treated tsl males were 7 d old
when tested). In experiment 4, we followed the
same protocol as experiment 2, except that we at-
tempted to simulate fully-loaded towers by plac-
ing approximately 22,500 grains of rice ("flies";
=45 grains per mL; 500 mL used per tray) and a
15 x 9 cm piece of cardboard ("sugar agar") on all
trays in the towers. This simulation was intended
only to mimic the physical environment affecting
the upper movement of air within towers, and it is
recognized that odor absorption (possibly affect-
ing GRO dispersion) may have differed between
simulated and fully operational towers. Finally, in
experiment 5, GRO was applied at the time of pu-
pal loading and left until adults were chilled and
collected (i.e., 4 d later). Trials were then con-
ducted 2 d later (when tsl males were 4 d old).
Mating trials were conducted at the USDA-
ARS-PBARC facility in Honolulu, Oahu, HI, dur-
ing Apr-May, 2005. For the experiment comparing
the mating success oftsl males held in towers ver-
sus PARC boxes, we released 75 REC males, 75
REC females, and either 75 tower-held tsl males
or 75 PARC box-held tsl males in nylon-screen,
field cages (diameter 3.0 m, height 2.5 m). For the
experiments comparing the mating success of tsl
males held in GRO-exposed towers versus non-ex-
posed towers, we released 75 REC males, 75 REC
females, and either 75 GRO-exposed tsl males or
75 non-exposed tsl males in field cages. When
tested, REC males were 7-13 d old and REC fe-
males were 8-14 d old. REC males were not ex-
posed to GRO in any trial. In both experiments,
the tsl males released within a given field cage de-
rived from the same tower or PARC box, and the
tsl males from a given tower or PARC box were
used in a single cage only (i.e., 4 tents were run
per test day, with each containing tsl males from
a single tower or PARC box). The field cages con-

September 2006

Shelly et al.: Mating Success of Male Medflies

trained 2 artificial trees (each 2 m tall with =450
leaves resembling those of Ficus benjamin L.)
Artificial trees were used because they provided a
chemically neutral substrate on which the flies
display the entire complement of natural activi-
ties. Flies were released between 0800-0830
hours, mating pairs were collected over the fol-
lowing 4 h, and males were identified with a black
light. Over 20% of females mated in all trials (the
minimum proportion defining an acceptable trial,
FAO/IAEA/USDA 2003), consequently no data
were excluded. Air temperature ranged between
25-30 oC during the trials.

Study Insects and Mating Trials-Florida

In Florida, tsl males were from the same strain
used in Hawaii, with pupae shipped directly by
air from the Guatemalan rearing facility to the
eclosion facility in Sarasota, FL. Prior to ship-
ping, tsl pupae were dyed (fluorescent pink) and
irradiated 2 d before eclosion in hypoxia at 145 Gy
of gamma irradiation from a 60Co source. Because
wild (or REC) flies were unavailable, we used
males and females from a standard, bisexual
strain (Petapa) reared in Guatemala. Pupae from
the Petapa strain also were dyed and irradiated
before shipping, but a different dye color (green)
was used for identification. Adults of the Petapa
strain were separated by sex within 24 h of eclo-
sion and held in transparent, plexiglass cages (40
x 30 x 30 cm, l:w:h; 300-500 flies per cage) with
screen panels on the top and were provided sugar-
agar gelatin as a source of food and water. Flies
from the Petapa strain were 5-7 d old when
In Florida, we conducted 3 mating experi-
ments, all of which involved comparisons between
non-exposed and GRO-treated males from eclo-
sion towers, with GRO always applied at the time
of pupal loading. In experiments 6 and 7, we ap-
plied 0.50 and 0.25 mL of GRO, respectively, to
each of 10 filter paper squares (5 by 5 cm) and
placed 1 square on trays 1 (bottom), 5, 10, 15, 20,
25, 30, 40, 50, and 60 (top). Thus, a total volume of
5.0 and 2.5 mL of GRO was used per tower in ex-
periments 6 and 7, respectively. In experiment 8,
we applied 1 mL of GRO to a cotton wick, placed
the wick in a Petri dish, and placed the Petri dish
on the bottom tray of the tower. In all 3 experi-
ments, pupae and the sugar-agar gelatin were
placed on all trays of a given tower (in the same
amount as in Hawaii), except that trays holding
GRO were left empty. GRO-exposed towers were
kept in a separate room from non-exposed towers
under the same environmental conditions (19-
240C, 60-80% RH). Peak emergence of adult males
occurred 2 d and chilling occurred 4-5 d after pu-
pal loading, and samples of males were taken
from 4-6 trays from the middle of the tower (trays
20-40) during chilling. These males were trans-

ferred to cages, provided sugar agar gel, and held
until testing 1-2 d later (i.e., tsl males were 3-5 d
old when tested).
Mating trials were conducted at the eclosion
facility in Sarasota, FL, during Feb-Apr, 2003, in
the same manner as those in Hawaii except that
(i) Petapa flies were used instead of REC (or wild)
flies, (ii) tsl males from a given tower were used in
2 cages on each of 2 successive d (as in Hawaii, 4
cages 2 with non-exposed and 2 with GRO-ex-
posed tsl males were run per day), (iii) the cages
each contained a single potted ruby red grapefruit
tree (Citrus paradisi M.), and (iv) tests were run
between 1000-1400 h, owing to the relatively cool,
winter temperatures (21-280C).

Release-Recapture Study-Florida

The release-recapture study was conducted as
part of the routine operation of the Florida Pre-
ventative Release Program. Test flies were re-
leased by air on 7 dates (at roughly 1-month inter-
vals) between March-August, 2004, in 2 areas
(Hillsborough and Sarasota Counties) included
within the established trapping grid. On a given
date, approximately 6 million (4 fully loaded tow-
ers) GRO-exposed or non-exposed tsl males were
released in each area (on 1 date, GRO-exposed
males were mistakenly released away from the
target sites, resulting in a total of 7 releases for
non-exposed males and 6 releases for GRO-ex-
posed males; Table 3). Test flies were dyed green
or blue to distinguish them from the pink-dyed
flies released in the ongoing control program. To
expose males, we applied 1 mL of GRO to a cotton
wick placed in a Petri dish and placed the dish on
the bottom tray of the tower at the time of pupal
placement. GRO-exposed and non-exposed towers
were kept in separate rooms under the same en-
vironmental conditions as above. Treatments as
well as dye colors were alternated between the 2
test areas on successive releases; alternating col-
ors allowed us to distinguish males from succes-
sive releases in the same area (because males
were unlikely to survive >60 d, males from succes-
sive releases of the same color in a given area
were unlikely to occur contemporaneously).
Releases were made from a small aircraft
(Beechcraft BE90 King Air) flying at an approxi-
mate ground speed of 160 mph (250 km/h) at an
altitude of 600-800 m in Hillsborough Co. and
800-1,100 m in Sarasota Co. Releases were made
in 4 sites in Hillsborough Co. and 5 sites in Sara-
sota Co., where each site represented an east-
west, oriented line along which 6 Jackson traps
(baited with trimedlure) were placed at 264 m in-
tervals for a total transect length of 1584 m (1
mile). Thus, the Hillsborough and Sarasota study
areas contained 24 (4 sites, 6 traps per site) and
30 traps (5 sites, 6 traps per site), respectively. All
sites were in residential areas, and traps were

Florida Entomologist 89(3)

typically placed 4-5 m above ground in citrus
trees. During a release, the aircraft made 3 passes
perpendicular to, and evenly spaced along, a site
(trap line). Approximately equal numbers of flies
were released per pass in each area. Traps were
serviced daily, excluding weekends, and captured
flies were examined under a black light.

Statistical Analyses

Comparisons of mating success were based on
raw data with the t-test as the assumptions of
normality and equal variance were met in all
cases (excepting one instance, where data were
log10 transformed to normalize the data). Propor-
tions were likewise compared with the t-test but
using arsine transformed values. Pairwise com-
parisons involving trapping data from the re-
lease-recapture study were made by using the
Mann-Whitney test as data were not normally
distributed. Preceding analyses were made with
SigmaStat software (version 2.0). Slopes of re-
gression lines showing temporal decline in male
captures were compared following Zar (1996).


Mating Trials-Hawaii

Table 1 presents the results of mating trials
conducted in Hawaii. There was no significant dif-
ference found in the mating success of tsl males
held in PARC boxes versus eclosion towers (exper-
iment 1). On average, males from PARC boxes ob-
tained 19% of all matings per replicate compared

to 22% for males from towers (P > 0.05). The addi-
tion of GRO to eclosion towers after adult emer-
gence (experiment 2) significantly increased the
number of matings obtained by tsl males, with
GRO-exposed males, on average, achieving over
twice as many matings (20.1/8.8 = 2.3) per repli-
cate as non-exposed tsl males. In this experiment,
REC males had a mating advantage over both
GRO-exposed and non-exposed, tsl males. How-
ever, the magnitude of this advantage varied with
GRO treatment: on average, GRO-exposed males
obtained 44% of all matings per replicate com-
pared to only 20% for non-exposed males (P <
0.001). The effect of GRO was evident even 5 d af-
ter the exposure period (experiment 3). As before,
GRO-exposed tsl males obtained, on average,
twice as many matings (23.5/11.7 = 2.0) per repli-
cate as non-exposed tsl males. In this experiment,
REC males accounted for significantly more mat-
ings than non-exposed males, but no difference in
mating frequency was found between REC males
and the GRO-exposed males. Results from the full
tower simulation (experiment 4) were similar to
those obtained for the other experiments. GRO-
exposed, tsl males obtained, on average, about
60% more matings (19.8/12.2 = 1.62) per replicate
than non-exposed, tsl males, and there was no sig-
nificant difference in mating frequency between
REC and GRO-exposed, tsl males. GRO conferred
a mating advantage even when applied at the
time of pupal placement (experiment 5). In this
case, GRO-exposed tsl males obtained signifi-
cantly more matings than control tsl males and a
similar number of matings as REC males. The
timing of GRO application (post- or pre-adult


Matings per replicate

Experiment n Holding Unit GRO4 tsl males REC males

1 8 PARC box 8.0 1.2 a, D 33.4 3.2 b, E
Tower -7.7 1.4 c, D 27.4 3.0 d, E
2 12 Tower -8.8 1.2 a, D 35.6 1.7 b, F
Tower + (A/day 4) 20.1 1.9 c, E 25.8 1.7 d, G
3 8 Tower -11.7 + 1.4 a, D 29.1 3.2 b, F
Tower + (A/day 7) 23.5 2.3 c, E 23.1 1.4 c, F
4 8 Tower -12.2 + 2.1 a, D 29.3 3.6 b, F
Tower + (A/day 4, rice) 19.8 2.4 c, E 24.0 2.2 c, F
5 10 Tower -9.1 1.3 a, D 28.1 3.5 b, F
Tower + (P/day 4) 19.2 + 2.1 c, E 20.8 + 2.8 c, F

'Control = No GRO (-), Treated = 1 mL GRO (+). A = GRO applied after adult emergence, P = GRO applied at time of pupal place-
ment, day number = age of adult sterile males when tested, rice = rice placed on trays to simulate full tower.

September 2006

Shelly et al.: Mating Success of Male Medflies

emergence) did not affect the relative mating suc-
cess of GRO-exposed males; the average propor-
tion of total matings achieved by treated males
was similar in experiments 1 (43%) and 5 (48%),
respectively (P > 0.05).

Mating Trials-Florida

In Florida, application of GRO to eclosion tow-
ers at the time of pupal placement boosted the
mating success of tsl males at all 3 doses tested
(Table 2). Each of these experiments yielded the
same basic results: (i) GRO-exposed, tsl males
achieved significantly more matings per replicate
than non-exposed, tsl males, (ii) Petapa males ac-
counted for significantly more matings per repli-
cate than non-exposed, tsl males, and (iii) Petapa
and GRO-exposed, tsl males obtained similar
numbers of matings per replicate.


Based on data pooled over both areas, GRO-ex-
posed tsl males were captured in significantly
greater numbers than non-exposed males (t =
59.0, P < 0.05, n,= 7, n2 = 6, Mann-Whitney test;
Table 3). On average, 298 GRO-exposed males
were captured per release compared to only 76
non-exposed males. The small number of repli-
cates precluded rigorous comparisons of the treat-
ment groups within each of the release areas, but
higher trap catches were noted for GRO-exposed
than non-exposed males in both Hillsborough
(1,018 versus 199, respectively) and Sarasota
(772 versus 331, respectively).
In addition, we compared GRO-exposed and
non-exposed tsl males with respect to the post-re-
lease interval over which males were trapped.

Combining data from both release areas, we
found no significant difference in the length of the
'capture interval' between GRO-exposed (mean =
16.7 d, range = 8 27 d) and non-exposed (mean =
11.3 d, range = 4 26 d) males (t = 52.0, P > 0.05,
n1= 7, n2 = 6, Mann-Whitney test).
The above results show that, although GRO-
exposed males were trapped in higher numbers
than non-exposed males, they were not trapped
over substantially longer post-release intervals.
Collectively, these findings suggest that the cap-
ture rate of GRO-exposed males declined more
rapidly than for non-exposed males. However,
closer inspection of daily trap captures reveals
that numbers of GRO-exposed and non-exposed
males declined at a similar rate in the day just af-
ter a release and that a steeper decline was evi-
dent for GRO-exposed males only later (Fig. 1). A
logo transform of the raw data yielded linear de-
clines for both types of males, with nearly identi-
cal slopes between 1-10 d after release (t = 0.2, P
> 0.05, df = 16). In contrast, the rate of decline in
male captures from 11-30 d post-release was sig-
nificantly greater for GRO-exposed males than
non-exposed males (t = 2.4, P < 0.05, df = 36). This
finding appeared to derive from the fact that
nearly all (512/530 = 97%) of non-exposed males
were captured in d 1-10 following release, result-
ing in a nearly horizontal line for capture rate
over the later days. For GRO-exposed males, a
larger proportion of trapped males were captured
after d 10 (139/1,790 = 8%), with the temporal de-
cline in their capture rate being more evident.
Because tsl males were released aerially over
relatively large areas, the trapping data do not al-
low for rigorous comparison of the dispersal abil-
ity of GRO-exposed versus non-exposed tsl males.
However, analysis of presence/absence data on a


Matings per replicate

Experiment n Holding Unit GRO4 tsl males REC males

6 8 Tower 4.0 + 0.9 a, D 23.7 + 2.7 b, F
Tower + (5 mL) 12.6 + 2.3 c, E 17.0 + 2.6 c, F
7 10 Tower 10.3 + 1.8 a, D 21.0 + 2.5 b, F
Tower + (2.5 mL) 20.0 + 3.3 c, E 15.9 + 2.4 d, G
8 10 Tower 8.0 + 1.5 a, D 28.4 + 7.1 b, F
Tower + (1 mL) 17.7 + 2.3 c, E 19.5 + 4.4 c, G

'Control = No GRO (-), Treated = GRO applied at time of pupal placement (+). In all cases, sterile
males were tested when 3-4 d old.

Florida Entomologist 89(3)


Number of trapped males
Release Release
number date Non-exposed GRO-exposed

1 3/2 105 (S) 391 (H)
2 3/30 74 (H) 573 (S)




September 2006

= 59.0, P < 0.05, n, = 7, n2 = 6, Mann-Whitney
test). The greater value observed for GRO-ex-
posed was not unexpected given the higher num-
ber of GRO-exposed males captured overall. How-
ever, similar dispersal ability between the 2 male
types is indicated by the fact that the relative dif-
ference in trap-day occurrence (47/19 = 2.5) be-
tween GRO-exposed and non-exposed males was
similar to the relative difference in the total num-
ber of GRO-exposed versus non-exposed males re-
captured in the week following releases (1506/495
= 3.0). In other words, relative abundance alone
was a crude indicator of distribution independent
of the GRO status of the males.


4/28 45 (S) 416 (H) The present study shows that mating competi-
5/25 54 (H) 96 (S) tiveness, as measured in field-cage trials, is simi-
6/29 134 (S) [3] (H)# lar for sterile tsl males held in eclosion towers or
7/27 71 (H) 103 (S) PARC boxes. Along with the results of Salvato et
8/31 47 (S) 211 (H) al. (2004), which showed no difference in yield,
l 530 1,790 weight, or flight ability, this finding indicates that
there are no major differences in the overall qual-
Data were excluded from analysis, because flies were mis- ity of tsl males held in these 2 eclosion systems.
nly released away from target sites. Although longevity or dispersal ability have not
yet been compared, the observed similarity in
overall male vigor suggests that economic consid-
)-day basis suggests, preliminarily at least, rations, rather than concern over fly quality, will
GRO-exposed and non-exposed males show be the key determinant in programmatic deci-
ilar dispersal. For the 1-week interval follow- sions to switch from PARC to TE systems.
a release, we scored the presence/absence of The present study also demonstrated a posi-
flies over all traps and computed the propor- tive effect of GRO exposure on the mating perfor-
of total trap-days that were 'positive' (i.e., >1 mance of sterile, tsl males held in eclosion towers
male present). Over all releases in both areas, similar to that previously observed for males held
3-exposed males were present in 47% of the in PARC boxes. For trials involving GRO applica-
)-d compared to 19% for non-exposed males (t tion after adult emergence, the mating frequency
of treated males (in competition with REC males
for REC females) was approximately twice that
0mo0 observed for non-exposed males in both eclosion
SNo RO (solid lines) systems. In the PARC box system, GRO-exposed
o ROexposoed (dashe lines) males (1 mL for 3 h) achieved 52% of the total
o matings compared to only 24% for non-exposed
"o males (Shelly et al. 2004), while in the TE system,

0 matings compared to only 20% for the non-ex-
".g -o-- oo2_oo o posed males (experiment 2).
--- -- pB8 -In contrast to PARC boxes, placement of GRO
at the time pupal loading invariably boosted the
mating success of tsl males held in eclosion tow-
01 s 15 i 20 2 ers. For PARC boxes, application of GRO at the
Days after release time of pupal loading conferred a mating advan-
tage to the subsequently emerged males for a dose
'ig. 1. Number of GRO-exposed and non-exposed tsl of 1.0 mL but not for a dose of 0.25 mL (Shelly et
es captured on individual days following releases. al. 2004). Thus, for PARC boxes, a volumetric
es represent average numbers (+1) computed over dose-to-container ratio of 0.0025 (0.00025 m3
leases; note ordinate is log,0 scale. Simple linear re- GRO/0.1 m3 PARC box) did not result in increased
sions for log,,transformed male numbers were as
ws: Non-exposed males, d 1-10: Y = 1.37 0.14X, r2 male mating success. For tower-held males, how-
80; Non-exposed males, d 11-30: Y = 0.21 0.01X, r2 ever, a mating advantage was consistently ob-
43; GRO males, d 1-10: Y = 1.94 0.13X, r2 = 0.91; served following application of 1 mL GRO at pu-
d, 11-30: Y = 0.75 0.02X, r = 0.53. pal loading or at a dose-to-container ratio of only

all r
= 0.8
= 0.

Shelly et al.: Mating Success of Male Medflies

0.0004 (0.001 m3 GRO/2.5 m3 eclosion tower) or
16% of the ineffective ratio noted above for PARC
boxes. Why GRO, when applied at pupal place-
ment, was more effective in towers than PARC
boxes is not known. GRO exposure appears effec-
tive only when applied to adults: exposing pupae
(but not adults) did not influence the mating per-
formance of the subsequently emerged adults
(Shelly 2001). Thus, the difference observed be-
tween PARC boxes and towers likely reflects a dif-
ference in the amount of GRO remaining in cotton
wicks (placed under the towers; 3-dimensional
dispenser with surface area of approximately 10
cm2) as opposed to blotter paper (placed on PARC
boxes; two dimensional dispenser with surface
area of 25 cm2). The difference could have also de-
rived from differences in air circulation between
the 2 holding systems. In PARC boxes, GRO was
placed on a screened panel on the lid, and the odor
was not directed downward into the box. In tow-
ers, on the other hand, GRO was placed beneath
the towers, and its odor was drawn directly and
continuously over the pupae and the emerged
adults in the towers.
Consistent with earlier results (Shelly et al.
2004), the release-recapture study suggested
there was no negative effect of GRO exposure on
the survival of sterile males. GRO-exposed males
were recaptured in significantly greater numbers
and over similar post-release capture intervals as
non-exposed males. As noted, the present study
does not allow for a rigorous comparison of dis-
persal ability, but the analysis of trap-day occur-
rence suggests similar movement by GRO-ex-
posed and non-exposed males. A planned field ex-
periment in Hawaii will provide additional data
on vagility by monitoring male captures at differ-
ent distances from a central release point.
In conclusion, based on results from an ongo-
ing study, it seems unlikely that the higher cap-
ture of GRO-exposed males in the Florida re-
leases was an artifact of the GRO exposure itself.
Preliminary field data from Hawaii indicate that
males exposed to GRO in the laboratory are, in
fact, less likely to be captured in trimedlure-
baited traps in the field than naive, non-exposed
males. Thus, it does not appear that exposure in
the laboratory to one attractant (GRO) increases
male responsiveness to another lure (trimedlure).

This finding, if validated, suggests that the Flor-
ida trapping data actually provide a conservative
estimate of the relative abundance of GRO-ex-
posed males in the environment.


We thank Mark Peebles, Amy Young, James Edu,
Elaine Pahio, David Dean, Pedro Rendon, Sue McCombs,
Don McInnis, Neil Wright, and Roger Corrales for assis-
tance. Comments by Don McInnis improved the paper.


FAO/IAEA/USDA. 2003. Manual for Product Quality
Control and Shipping Procedures for Sterile Mass-
Reared Tephritid Fruit Flies, Version 5.0. Interna-
tional Atomic Energy Agency, Vienna, Austria.
DRICHS. 1996. Development and application of ge-
netic sexing systems for the Mediterranean fruit fly
based on a temperature sensitive lethal, pp. 185-191
In B. A. McPheron and G. J. Steck [eds.], Fruit Fly
Pests: A World Assessment of Their Biology and
Management. St. Lucie Press, Delray Beach, FL.
ENKERLIN. 2002. Medfly areawide sterile insect
technique programmes for prevention, suppression
or eradication: the importance of mating behavior
studies. Florida Entomol. 85: 1-13.
LEVY, K., T. E. SHELLY, AND B. YUVAL. 2005. Effects of
the olfactory environment and nutrition on the abil-
ity of male Mediterranean fruit flies to endure star-
vation. J. Econ. Entomol. 98: 61-65.
2004. Efficacy of tower medfly eclosion systems. Bio-
control Sci. Tech. 14: 77-80.
SHELLY, T. E. 2001. Exposure to a-copaene and a-co-
paene-containing oils enhances mating success of
male Mediterranean fruit flies (Diptera: Tephriti-
dae). Ann. Entomol. Soc. Am. 94: 497-502.
2004. Aromatherapy in the Mediterranean fruit fly
(Diptera: Tephritidae): sterile males exposed to gin-
ger root oil in prerelease storage boxes display in-
creased mating competitiveness in field-cage trials.
J. Econ. Entomol. 97: 846-853.
MOTO. 1969. Low-cost larval rearing medium for
mass production of oriental and Mediterranean fruit
flies. J. Econ. Entomol. 62: 967-968.
ZAR, J. H. 1996. Biostatistical Analysis. 3rd ed. Prentice
Hall, Upper Saddle River, NJ.

Florida Entomologist 89(3)


'Unidad de Zoologia, Facultad de Biologia, Universidad de Salamanca. 37071- Salamanca, Spain, e-mail:

2Dipartimento di Biologia, Sezione di Zoologia e Citologia, Universita degli Studi di Milano-Via Celoria, 26, 20133,
Milan, Italy


The prepupa of Chalybion femoratum (Fabricius) is described and illustrated. The most sa-
lient character state shown by the mature larva of this species lies in the presence of scat-
tered setae on the integument. Morphological characters with diagnostic value for the final
instar of the tribe Sceliphrini are discussed.

Key Words: Hymenoptera, Sphecidae, Chalybion femoratum, larval morphology


Se describe, y compare con la previamente descrita del g6nero, la prepupa de Chalybion fem-
oratum (Fabricius). El estado de caracter que permit distinguir la larva madura de esta es-
pecie radica en la presencia de setas esparcidas en el tegumento. Adicionalmente, se discuten
caracteres morfol6gicos con valor diagn6stico de las larvas maduras de la tribu Sceliphrini.

Translation by the authors.

Sphecid wasps of the genus Ci..', i-i .... Dahlbom
normally nest in preexisting cavities and provision
their nests with spiders (Bohart & Menke 1976).
Of the 45 species of the genus (Pulawski 2005),
only the mature larva ofCh. californicum (de Sau-
ssure, 1867) is known (Evans and Lin 1956; Evans
1959). The aim of this paper is to describe the pre-
pupa of Ch. femoratum (Fabricius, 1781).


The description is based on 1 postdefecated
mature larva (prepupa) obtained by Carlo Poli-
dori in Castelleone (Cremona Province, Italy) in
July 2004, from a nest of Sceliphron caementar-
ium (Drury 1770). Following the method used to
conserve and prepare larval specimens (Asis, un-
published), the specimen was fixed in alcohol
(70%) for later treatment with KOH (10%, 70 'C)
for study under the microscope. The following ab-
breviations are used: d = diameter, h = height, 1 =
length, w = width. The voucher specimen is depos-
ited at the "Torres-Sala" Entomological Founda-
tion (Valencia, Spain).

Description of the Mature Larva
Chalybion femoratum (Fabricius) (Figs. 1-6)

Body (Fig. 1) (1 = 11 mm, maximum w = 5 mm)
yellow, fusiform, strongly curved at level of third

thoracic segment. Anus transverse, in terminal
position; supra- and infra-anal lobes very similar
in size. Pleural lobes well developed. Integument
with tiny spinules (1 = 5 pm) and few dispersed se-
tae (1 = 10 pm). Thoracic and abdominal spiracles
(Fig. 2) of the same size, all in a single line; atria
(d = 120 pm) with walls lined with ridges forming
irregular hexagons, subatria short and subglo-
bose, opening into subatria armed with spines.
Thorax with a pair of oval callosities, one on each
side of the mid-dorsal line, just anterior to the
first pair of spiracles.
Cranium (Fig. 3) [w = 0.7 mm, h (measured to
apex of clypeus) = 1.1 mm] higher than wide, with
sparse setae (1 = 20 pm), more abundant on the
lower sides of the cephalic capsule, and puncta-
tions (5-7 pm). Coronal suture present, parietal
bands absent; antennal orbits (d = 48 pm) subcir-
cular, with a small circular central area with
three sensilla. Clypeus with scattered setae (10
pm) and punctations. Labrum (Fig. 4a) (maxi-
mum w = 420 pm; maximum h = 300 pm) bilobate;
anterior margin with six barrel-shaped sensilla
on each side of the median line; each lobe surface
with several setae (1 = 10 pm) and sensory pores
(d = 8-10 pm); epipharynx (Fig. 4b) spinulose cen-
trally and laterally, with several sensory pores (d
= 8-10 pm).
Mouthparts. Mandible (1 = 435 pm, maximum w
= 335 pm) robust, sclerotized, brown, with four api-

September 2006

Tormos et al.: Mature Larva of Chalybion femoratum

0. 1 mm

0.2 mm

3 -
0.4 mm

0.1 mm

60. 2 mm



Figs. 1-6. Mature larva of Chalybion femoratum: (1) Body, ventro-lateral view; (2) Anterior thoracic spiracle
(atrium (a), subatrium (b), tracheal trunk (c)); (3) Cranium (frontal view); (4) Labrum; (5) Epipharynx; (6) Maxilla;
(7) Labium.

cal teeth, with several setae close to the basal ex-
ternal margin. Maxilla (Fig. 5) with the lacinial
area spinulose and external margin with setae (1 =
5 pm) and punctations; maxillary palpus (1 = 82
pm) with 3 apical sensilla; galea (1 = 120 pm) simi-
larly with 2 apical sensilla. Labium (Fig. 6) setose
(1 of the setae = 12 pm); labial palpus longer than
wide (46 x 36), with two pointed sensilla; salivary
orifice (1 = 328 pm) transverse with lips raised.

The mature larva of Ch. femoratum is similar
to that of Ch. californicum, the only other of this
genus described to date. It differs from the latter
in having scattered setae on the integument.
Although in both mature larvae it is possible to
consider that the epipharyngeal sensory pores are
well removed from the anterior margin, this char-
acter state is not very differentiated in Ch. femo-
ratum. Additionally, in the latter larva the con-
vexities on the top and sides of the cephalic cap-
sule are almost indiscernible. These two charac-
ter states should not be used in classification keys
until more mature larvae of this genus become

The variability observed in these two charac-
ter states means that currently there is no clear
autapomorphy that allows the differentiation of
the last larval stage of the two genera encom-
passed within the subtribe Sceliphrina: Chaly-
bion and Sceliphron. To date, these two genera,
respectively, are defined by exhibiting (Evans &
Lin 1956; Evans 1959), the following: (a) vertex
and sides of head capsule strongly roughened by
small convexities/vertex and sides of head capsule
weakly roughened, and (b) epipharyngeal sensory
pores removed from the anterior margin of the la-
brum /epipharyngeal sensory pores scattered,
some being close to the anterior margin of la-
The last larval stage of Sceliphrina can be dif-
ferentiated from that of Podiina, the other sub-
tribe that with Sceliphrina forms the tribe Sce-
liphrini, by the presence, in the former, of abun-
dant setae on the head capsule (Evans 1959). In
the subtribe Podiina, only the mature larvae of
species of Penopodium and Podium are known
(Evans & Lin 1956; Evans 1964; Buys 2001; Buys
et al. 2004).
Currently, the last larval stage of Sceliphrini
can be defined by the combination of the following

0.5 cm



0.2 mm

390 Florida Ento

character states: (a) Thoracic and abdominal spi-
racles of the same size, all in a single line, (b) la-
brum with large setae and prominent sensory
cones, and (c) thorax with two or three pairs of
prominent subdorsal callosities.


The laboratories of the Fundacion Entomologica
"Torres-Sala" were used to carry out this study. Finan-
cial support for this research was provided by the Junta
de Castilla y Le6n, project SA012A05.


BOHART, R. M., AND A. S. MENKE. 1976. Sphecid Wasps
of the World. A Generic Revision. University of Cali-
fornia Press, Berkeley, Los Angeles, London. 1 color
plate, IX + 695 pp.


ologist 89(3) September 2006

BUYS, S. C. 2001. Last instar larva of Penepodium du-
bium (Hymenoptera: Sphecidae). Rev. Biol. Trop. 49
(1): 327-330.
BUYs, S. C., E. F. MORATO, AND C. A. GAROFALO. 2004.
Description of the immature instars of three species
of Podium Fabricius (Hymenoptera, Sphecidae) from
Brazil. Rev. Brasil Biol. 21(1): 73-77.
EVANS, H. E. 1959. Studies on the larvae of digger
wasps. Part V: Conclusion. Trans. American Ento-
mol. Soc. 85: 137-191.
EVANS, H. E. 1964. Further studies on the larvae of dig-
ger wasps. Trans. American Entomol. Soc. 90: 235-
EVANS, H. E., AND C. S. LIN. 1956. Studies on the larvae
of digger wasps (Hymenoptera, Sphecidae). Part I:
Sphecinae. Trans. Amererican Enomol. Soc. 81: 131-
PULAWSKI, W. J. 2005. Catalog of Sphecidae. http://

Diagne et al: Phyllophaga ephilida Feeding Preferences


Department of Entomology, Louisiana Agricultural Experiment Station, Louisiana State University
Agricultural Center, Baton Rouge, LA 70803 USA


Limited biological information about Phyllophaga ephilida, a major sweet potato pest in
Louisiana, is available. In 2001 and 2002, a study was conducted in the laboratory to inves-
tigate the feeding preference of adult Phyllophaga ephilida (Say) (Coleoptera: Scarabaeidae)
for the foliage of eight woody plant species: water oak (Quercus nigra L.), live oak (Quercus
virginiana Mill.), red maple (Acer rubrum L.), slash pine (Pinus caribaea Morelet), pecan
(Carya illinoensis (Wangenh) K. Koch), sweetgum (Liquidambar styraciflua L.), southern
magnolia (Magnolia grandiflora L.), and American elm (Ulmus americana L.). Beetles were
placed in an arena with the eight host plants and allowed to feed for 24 h (choice test). Leaf
area consumed and change in leaf weight were recorded. In 2001 and 2002, host plant had
a significant effect on both leaf area and weight consumed. In 2001, mean leaf area (mm2)
consumed was pecan (504), followed by elm (314), water oak (237), maple (176), live oak
(38.0), and sweetgum (4.00). Southern magnolia and slash pine were not consumed. In 2002,
mean leaf area (mm2) consumed was pecan (628), followed by elm (390), water oak (204), ma-
ple (75.0), and live oak (30.0). Southern magnolia, sweetgum, and slash pine were not con-
sumed. In 2001, mean leaf consumption (mg) was pecan (8.400), water oak (3.700), maple
(3.500), live oak (1.300), elm (0.300), and sweetgum (0.060). Southern magnolia and slash
pine were not consumed. In 2002, mean leaf consumption (mg) was pecan (10.00), elm
(4.200), water oak (3.200), maple (1.500), and live oak (1.000). Southern magnolia, sweet-
gum, and slash pine were not consumed. Phyllophaga ephilida exhibited a preference for pe-
can, oak, and elm. They avoided slash pine and southern magnolia.

Key Words: adult feeding, phytophagous, Phyllophaga ephilida, sweet potato


Informaci6n biol6gica sobre Phyllophaga ephilida, una plaga principal de batata en el estado
de Louisiana es limitada y poca disponible. En 2001 y 2002, un studio fue realizado en el
laboratorio para investigar la preferencia alimenticia de adults de Phyllophaga ephilida
(Say) (Coleoptera: Scarabaeidae) para el follaje de ocho species de plants lenosas: roble
americano (Quercus nigra L.), roble de Virginia (Quercus virginiana Mill.), arce rojo (Acer
rubrum L.), pino (Pinus caribaea Morelet), pacana (Carya illinoensis (Wangenh) K. Koch),
ocozol (Liquidambar styraciflua L.), magnolio (Magnolia grandiflora L.) y olmo americano
(Ulmus americana L.). Los escarabajos fueron puestos en una arena de ocho plants hospe-
deras y permitieron alimentarse por 24 h (prueba de escoger). El area de la hoja consumida
y el cambio en el peso de la hoja fueron notados. En 2001 y 2002, el hospedero de plant tuvo
un efecto significativo en el area y el peso de la hoja consumida. En 2001, el promedio del
area de hoja (mm2) consumida de pacana fue 504, siguido por olmo (314), roble americano
(237), arce (176), roble de Virginia (38.0) y ocozol (4.00). El magnolio y el pino no fueron con-
sumidos. En 2002, el promedio del area de hoja (mm2) consumido fue pacana (628), seguido
por olmo (390), roble americano (204), arce (75.0) y roble de Virginia (30.0). El magnolia su-
refio, olmo y el pino no fueron consumidos. En 2001, el promedio de hoja consumida (mg) fue
pacana (8.400), roble americano (3.700), arce (3.500), roble de Virginia (1.300), olmo (0.300)
y ocozol (0.060). El magnolio y el pino no fueron consumidos. En 2002, el promedio de hoja
consumida (mg) fue pacana (10.00), olmo (4.200), roble americano (3.200), arce (1.500) y ro-
ble de Virginia (1.000). El magnolio, ocozol y pino no fueron consumidos. Phyllophaga ephil-
ida mostr6 una preferencia para pacana, roble y olmo. Ellos evitaron el pino y el magnolia.

Adult June beetles are nocturnal defoliators of may be classified as monophagous (host range in-
trees, shrubs, and grasses in diverse ecological re- cludes plants of one or a few closely related spe-
gions (Richter 1958; Vallejo et al. 1998). Several cies within a genus), oligophagous (host range in-
reports describe adult Phyllophaga spp. feeding cludes several genera within a family), or polyph-
on the foliage of woody plants (Davis 1916; agous (host range includes several families in one
McLeod 1986; Potter 1998). However, the host or more orders of plants) (Metcalf & Luckman
range and feeding preference of adult Phylloph- 1975). Often the common traits of the preferred
aga ephilida have not been investigated. Insects hosts are phytochemical.

Florida Entomologist 89(3)

The larvae ofP. ephilida damage sweet potato
roots in Louisiana (Rolston & Barlow 1980). Most
commercial sweet potato fields in Louisiana are
relatively small (about thirty hectares) and bor-
dered by tree lines or woody areas. These trees
may provide food for adults. Knowledge of the
host range and feeding preference of adults for
the common tree species in south central Louisi-
ana would help in understanding the ecology of
the pest and perhaps explain the specific distribu-
tion of the pest among grower fields.
The purpose of this study was to investigate
the feeding preference of P ephilida adults for the
foliage of eight trees common in the sweet potato
growing areas of Louisiana.


Host preference tests were conducted in an
arena constructed from an 11.3-L Rubbermaid
plastic container (40.6 x 28.5 x 27.5 cm) with a
snap cover. An opening (10 by 40 cm) was cut in the
center of the plastic cover and flexible screen (1.02
mm) was glued to the edges of the opening with a
hot glue gun to allow for airflow and to prevent
beetles from escaping. A layer of sand (3 cm deep)
was placed in the bottom of the container and a sty-
rofoam board (2 cm thick) was placed on top. Cut
into the board were circular holes to accommodate
20-ml glass scintillation vials. These vials stood
erect in the board and held the host plant stems.
Vials were filled with distilled water, the host plant
stems inserted into the top, and parafilm was
wrapped over the top of the vial to hold the host
plant in place. Young leaves from eight species of
trees were harvested 1 h before the experiment.
The evaluated leaves were: water oak (Quercus ni-
gra L., Fagaceae), live oak (Quercus virginiana
Mill., Fagaceae), red maple (Acer rubrum L., Acer-
aceae), slash pine (Pinus caribaea Morelet, Pi-
naceae), pecan (Carya illinoensis (Wangenh) K.
Koch, Juglandaceae), sweetgum (Liquidambar
styraciflua L., Hamamelidaceae), southern magno-
lia (Magnolia grandiflora L., Magnoliaceae), and
American elm (Ulmus americana L., Ulmacaeae).
All eight test plants were randomly assigned a po-
sition in a circular arrangement within the arena.
In the field, adult beetles were collected in Japa-
nese beetle pheromone traps (Trece Incorpo-
rated). Pheromones traps consisted of 4 mL of me-
thyl ester of L-isoleucine impregnated in a rubber
septum. The traps were placed in commercial
sweet potato fields in St Landry Parish, LA, on 22
May 2001 and 4 Jun 2002 and inspected weekly.
Beetles were kept in the laboratory 24 h before
their use in the experiment at 24C. The photope-
riod was maintained at 16:8 (L: D). Twenty adult P
ephilida males were placed in the center of the
arena and allowed to feed for 24 h. After 24 h, the
consumption was determined by measuring the
leaves before and after insect feeding with a leaf

area reader (Li-Cor). Leaf area consumption was
calculated by subtracting the final leaf area from
the initial leaf area. Leaf weight was measured
with a Mettler Toledo@ scale. Control leaves with-
out beetles were used to adjust for weight loss due
to desiccation. The experiment was conducted in
2001 and 2002, with a randomized complete block
design with six replicates for each trial. After each
bioassay, the genitalia of all beetles were dissected
and a species determination made by A. Diagne.
Identified adults were sent to Dr. E. R. Woodruff for
confirmation (Gainesville, Florida). The voucher
specimens were deposited at the Louisiana State
University Entomology Museum. A total of six tri-
als were conducted each year. Data were analyzed
by SAS General Linear Model (GLM), and Least
Significant Difference (LSD) was used for mean
separation (SAS Institute 1990).


Leaf area consumed by male P ephilida was
different among the eight host plants in 2001 in
trial 1 (F = 4.09; df= 7, 5;P < 0.0001), in trial 2 (F
= 2.24; df= 7, 5; P < 0.0001), in trial 3 (F = 8.62;
df= 7, 5; P < 0.0001), in trial 4 (F = 6.39; df= 7, 5;
P < 0.0001), in trial 5 (F = 9.85; df = 7, 5; P <
0.0001), in trial 6 (F = 9.85; df= 7, 5; P < 0.0001)
(Table 1). Leaf area consumed by male P ephilida
was significant among the host plants in 2002 in
trial 1 (F = 3.08; df= 7, 5; P < 0.0001), in trial 2 (F
= 3.32; df= 7, 5; P < 0.0001), in trial 3 (F = 15.65;
df= 7, 5; P < 0.0001), in trial 4 (F = 3.56; df= 7, 5;
P < 0.0001), in trial 5 (F = 9.26; df = 7, 5; P <
0.0001), in trial 6 (F = 8.06; df= 7, 5; P < 0.0001)
(Table 2). No discernible losses in leaf area were
detected in the controls, hence adjustment for wa-
ter loss was not made to leaf area measurements
of the treatments. In 2001, mean leaf area con-
sumed by male P ephilida was pecan, 504 mm2,
followed by elm (314 mm2), water oak (237 mm2),
maple (176 mm2), live oak (38.0 mm2), and sweet-
gum (4.00 mm2). Southern magnolia and slash
pine were not consumed. In 2002, leaf area con-
sumed was pecan, with a mean of 628 mm2, fol-
lowed by elm (390 mm2), water oak (204 mm2),
maple (7.50 mm2), and live oak (3.00 mm2). South-
ern magnolia, sweetgum, and slash pine were not
consumed. Leaf area measurement can be a less
reliable measurement of beetle feeding than
weight due to the variation in leaf density and
leaf thickness between plant species. Therefore,
consumption (mg) based on fresh weight of foliage
consumed also was measured.
The consumption by male P ephilida was dif-
ferent among the eight host plants in 2001 in trial
1 (F = 4.27; df = 7, 5; P = 0.0001), in trial 2 (F =
2.18; df= 7, 5; P < 0.0001), in trial 3 (F = 7.52; df
= 7, 5; P < 0.0001), in trial 4 (F = 3.08; df = 7, 5; P
< 0.0001), in trial 5 (F = 12.38; df = 7, 5; P <
0.0001), in trial 6 (F = 5.24; df= 7, 5; P < 0.0001)

September 2006

Diagne et al: Phyllophaga ephilida Feeding Preferences


Mean area (mm2) consumed/trial1

Host Plant 1 2 3 4 5 6

Pecan 241 a 836 a 458 a 632 a 723 a 139 b
Elm 137 a 78 b 170 be 473 ab 236 b 785 a
Maple 140 a 69 b 102 cd 203 be 0 c 543 a
Live oak 0 b 0 b 8 cd Oc Oc 225 b
Water oak 0 b 150 b 329 ab 345 b 53 be 550 a
Southern magnolia 0 b 0 b d Oc Oc Oc
Sweetgum 0 b 28 b 0 d Oc Oc Oc
Slash pine 0 b 0 b d Oc Oc Oc

1Means with the same letter are not significantly different based on LSD test, P = 0.05.


Mean area (mm2) consumed/trial'

Host Plant 1 2 3 4 5 6

Pecan 220 a 546 a 851 a 564 a 1152 a 439 a
Elm 240 a 364 ab 312 b 440 ab 667 b 324 a
Maple 0 b Oc 80 c 151 be 217 c 7b
Live oak Ob 36 b 0 c 54 c 63 c 30 b
Water oak 7 b 272 abc 143 be 448 ab 270 c 85 b
Southern magnolia 0 b Oc Oc Oc Oc 0 b
Sweetgum 0 b 0 c Oc Oc Oc 0 b
Slash pine 0 b Oc Oc Oc Oc 0 b

'Means with the same letter are not significantly different based on LSD test, P = 0.05.

(Table 3). The consumption by male P ephilida
also was significant among the host plants in
2002 in trial 1 (F = 3.09; df = 7, 5; P < 0.0001), in
trial 2 (F = 3.56; df= 7, 5; P < 0.0001), in trial 3 (F
= 20.68; df = 7, 5; P < 0.0001), in trial 4 (F = 2.87;
df= 7, 5; P < 0.0001), in trial 5 (F = 7.84; df = 7, 5;
P < 0.0001), in trial 6 (F = 7.31; df = 7, 5; P <
0.0001) (Table 4). No discernible losses in weight
were detected in the controls, hence adjustments
were not made to the leaf weight measurements.
In 2001, mean leaf consumed (mg) was pecan
(8.400 mg), water oak (3.700 mg), maple (3.500
mg), live oak (1.300 mg), elm (0.300 mg), and
sweetgum (0.06 mg). Southern magnolia and
slash pine were not consumed at all. In 2002,
mean weight consumed was pecan (10.00 g), elm
(4.200 mg), water oak (3.200 mg), maple (1.500
mg), and live oak (1.000 mg). Southern magnolia,
sweetgum, and slash pine were not consumed.
Within the range of host plants evaluated, P.
ephilida showed a preference for the foliage of pe-
can, elm, water oak, and maple. Southern magno-
lia and slash pine were avoided completely, while
sweetgum was barely eaten.

Measurements of leaf area and leaf weight
consumed revealed a similar ranking of host
plant preference. These results suggest that adult
P ephilida will feed on plant species in at least
four plant families (Juglandaceae, Ulmaceae, Fa-
gaceae, and Aceraceae) indicating a polyphagous
feeding habit. Several reports were made based
on observations of adult Phyllophaga feeding on
leaves of trees (Travis 1939; Sweetman 1931).
Phyllophaga hirticula, P. tristis, and P fraterna
were reported feeding on hickory and oak (Davis
1916). Sweetman (1931) reported that P impli-
cata, P. fusca, and P drakei selectively feed on
woody plants. Phyllophaga implicata adults were
reported to feed on the foliage of elm, ash, poplar,
and willow (Lago et al. 1979; McLeod 1986). Adult
Phyllophaga, not identified to species, have been
observed feeding on walnut, persimmon, birch,
elm, poplar, hickory, and oak foliage (Potter 1998).
Travis (1939) reported that P lanceolata adults
were recorded in Iowa feeding on close to thirty
host plants, among them corn, soybean, and po-
tato. Sweetman (1931) reported that P anxia
feeds on a broad range of herbaceous plants.

Florida Entomologist 89(3)


Mean weight (mg) consumed/trial'

Host Plant 1 2 3 4 5 6

Pecan 4.0 a 14.0 a 7.6 a 10.5 a 12.0 a 2.3 bc
Elm 1.4 be 0.8 b 1.8 b 5.1 b 2.5 b 8.5 a
Maple 2.8 ab 1.4 b 2.1 b 4.1 b 0 b 11.1 a
Live oak 0 c 0 b 0.2 b 0 c 0 b 7.5 ab
Water oak 0 c 2.3 b 5.2 a 5.4 b 0.8 b 9.7 a
Southern magnolia Oc 0 b 0 b c 0 b Oc
Sweetgum Oc 0.4 b 0 b Oc 0 b Oc
Slash pine Oc 0 b 0 b c 0 b Oc

1Means with the same letter are not significantly different based on LSD test, P = 0.05.


Mean weight (mg) consumed/trial'

Host Plant 1 2 3 4 5 6

Pecan 3.5 a 8.7 a 13.6 a 9.0 a 18.4 a 7.0 a
Elm 2.6 a 3.9 b 3.4 b 4.7 abc 7.2 b 3.5 b
Maple 0 b 0 b 1.6 be 3.1 be 4.4 be 0.1 c
Live oak 0 b 1.2 b Oc 1.8 be 2.2 be 0.9 be
Water oak 0.1 b 4.3 ab 2.2 be 7.1 ab 4.2 be 1.3 be
Southern magnolia 0 b 0 b c Oc Oc Oc
Sweetgum 0 b 0 b c Oc Oc Oc
Slash pine 0 b 0 b c Oc Oc Oc

Means with the same letter are not significantly different based on LSD test, P = 0.05.

There is no observation that P ephilida feeds on
the foliage of host plants other than woody plants.
The underlying determinant of the host range
of P ephilida is likely to be the presence of phy-
tochemicals and/or volatile compounds acting as
an attractant or phagostimulant to the beetle.
Sweetman (1931) suggested that the odor of
plants can influence the direction of flight in some
species of Phyllophaga. Phyllophaga implicita is
attracted to willow twigs, particularly when its fo-
liage has been bruised (Sweetman 1931). The
chemical ecology of P. ephilida has not been stud-
ied. Insects feeding on plants may induce the pro-
duction of toxins and digestability reducers
(Dicke 1999). Constitutive secondary plant chem-
icals have similar effects on several herbivore spe-
cies, with only specialist species able to overcome
the negative effects of the chemicals. It is proba-
ble that secondary compounds are present in the
host plants evaluated in this study, and their
presence could contribute to the preferences
shown by P. ephilida. The phenological stage of
the host plant is accompanied by changes in the
quality or quantity of secondary compounds, wa-
ter, and oil content, and can affect the feeding be-

havior of insects. In our study, only the new
growth leaves were used so as to keep their age
relatively constant.
Phyllophaga ephilida exhibited preference for
pecan, oak, and elm. These species commonly are
found in tree lines along the edges of commercial
sweet potato fields in Louisiana. Their proximity
to sweet potato production areas helps explain the
abundance of both adult P ephilida that are cap-
tured in blacklight traps and pheromone traps,
and the presence of damaging populations of lar-
vae in the fields. The feeding sites also are mating
sites for the beetles. Therefore, a feeding site can
be a good choice for the placement of pheromones
for mating disruption of beetle around sweet po-
tato fields. The determination of the nutritional
components of the preferred host plants can pro-
vide a foundation for a specific diet that will allow
further biological studies on P ephilida.


We thank E. R. Woodruff for the confirmation of the
identification of this species and E. Webster, M. Baur,
and B. Castro for reviewing this manuscript.

September 2006

Diagne et al: Phyllophaga ephilida Feeding Preferences


DAVIS, J. J. 1916. A progress report on white grub inves-
tigations. J. Econ. Entomol. 9: 261-283.
DICKE, M. 1999. Specificity of herbivore-induced plant
defences, pp. 43-59 In D. J. Chadwick and J. A. Goode
[eds], Insect-plant Interactions and Induced Plant
Defence. John Wiley and Sons Ltd, West Sussex, En-
gland. 281 pp.
MCLEOD, J. M. 1986. White Grub Management. North
Dakota State University. Cooperative Extension
Service Circular E-901. 4 pp.
METCALF, R. L., AND W. H. LUCKMAN. 1975. Introduc-
tion to Pest Management. John Wiley and Sons, New
York City, NY. 587 pp.
LAGO, P. K., R. I. POST, AND C. Y. OSETO. 1979. The phy-
tophagous Scarabaeidae and Troginae of North Da-
kota. North Dakota Insects. 12: 1-131.
POTTER, D. 1998. Destructive Turfgrass Insects: Biol-
ogy, Diagnosis and Control. Ann Arbor Press,
Chelsea, Michigan. 344 pp.

RICHTER, P. 0. 1958. Biology of Scarabaeidae. Annu.
Rev. Entomol. 3: 311-334.
ROLSTON, L. H., AND T. BARLOW. 1980. Insecticide con-
trol of a white grub Phyllophaga ephilida Say (Co-
leoptera: Scarabaeidae) on sweet potato. J. Georgia
Entomol. Soc. 15: 445-449.
SAS INSTITUTE. 1990. SAS Users' Manual, Version 8.
SAS Institute, Cary, NC. 890 pp.
SWEETMAN, H. L. 1931. Preliminary report on the phys-
ical ecology of certain Phyllophaga (Scarabaeidae,
Coleoptera). Ecology. 12: 401-422.
TRAVIS, B. V. 1939. Habits of the June beetle, Phylloph-
aga lanceolata (Say), in Iowa. J. Econ. Entomol. 32:
VALLEJO, F., M. A MORON, AND S. ORDUZ. 1998. First re-
port and description of immature stages of Phylloph-
aga obsoleta (Blanchard) (Coleoptera: Melolonthidae)
in Colombia. Coleopts. Bull. 52: 109-117.

Florida Entomologist 89(3)

September 2006


PSIS/Division of Entomology, University of Massachusetts, Amherst, MA, 01003, USA


Releases of predacious mites are recommended for use in greenhouse flower crops for sup-
pression of western flower thrips, Frankliniella occidentalis (Pergande). Control from preda-
cious mites alone, however, is not adequate and must be supplemented with the use of
insecticides. The principal material currently used by growers in the northeastern United
States for western flower thrips control is spinosad (Conserve). In laboratory tests on di-
rect toxicity, we found that fresh residues (2 h) of this material were not toxic to motile
stages of Neoseiulus (=Amblyseius) cucumeris (Oudemans) (74 vs 78% survival for the
treated group and the untreated water controls, respectively), the principal species of pre-
dacious mites used for control of western flower thrips, but did lower survival of Iphiseius de-
generans (Berlese) (56 vs. 73% survival for the treated group and the untreated water
controls, respectively). There were no differences for either species from exposure to older
(24 h) residues. In contrast, using the same assay we observed 10 and 3% survival of first in-
star and adult western flower thrips. We found no indication of that either mite species was
repelled by freshly dried (2 h post application) residues of this compound. Spinosad did, how-
ever, reduce oviposition of mites when confined in glass vials with pollen, a water source, and
pesticide-treated foliage. Oviposition in the first 24 h period after confinement was not af-
fected but in the second and third days, it was reduced by 48 and 76% for N. cucumeris and
41 and 70% for I. degenerans, compared with oviposition in the same periods by mites in un-
treated vials. These data indicate that the use of spinosad may not be compatible with re-
leases of these predacious mites in a western flower thrips suppression program.

Key Words: Frankliniella occidentalis, Neoseiulus (= Amblyseius) cucumeris, Iphiseius de-
generans (Berlese), pesticide compatibility


Liberaciones de acaros depredadores son recomendadas para el uso en los cultivos de flores
en los invernaderos para suprimir el trips occidental de flores, Frankliniella occidentalis
(Pergande). Sin embargo, usando solamente acaros depredadores para controlarlos, no es
adecuado y debe ser suplementado con el uso de insecticides. El material principal usado ac-
tualmente por los agricultores en el noreste de los Estados Unidos para el control del trips
occidental de flores es spinosad (Conserve). En pruebas del laboratorio sobre la toxicidad
direct, nosotros encontramos que los residues frescos (2 h) de este material no fueron t6xi-
cos a los estadios m6viles deNeoseiulus (=Amblyseius) cucumeris (Oudemans) (74 vs 78% so-
brevivencia en el grupo tratado y en el grupo no tratado de control que recibi6 solo aqua,
respectivamente), la especie principal de acaros depredadores usado para controlar el trips
occidental de flores, pero bajo la sobrevivencia de Iphiseius degenerans (Berlese) (56 vs. 73%
sobrevivencia en el grupo tratado y en el grupo no tratado de control que recibi6 solo aqua,
respectivamente). No hubo una diferencia en ambas species expuestas a los residues mas
viejos (24 h). En contrast, usando el mismo bioensayo nosotros observamos 10 y 3% sobre-
vivencia del primer estadio y adulto del trips occidental de flores. No encontramos ninguna
indicaci6n que las dos species de acaros fueron repelidos por los residues reci6n secados (2
h despu6s de la aplicaci6n) de este compuesto. Sin embargo, spinosad reduj6 la oviposici6n
de acaros confinados en ampolletas de vidrio con polen, una fuente de agua, y follaje tratado
con pesticide. La oviposici6n no fue afectada en el primer period de 24 h despu6s del sobre-
parto pero en el segundo y tercer dia, fue reducida a 48 y 76% en N. cucumeris y 41 y 70%
en I. degenerans, comparada con la oviposici6n en los mismos periods para los acaros en las
ampolletas no tratadas. Estos datos indican que el uso de spinosad puede ser incompatible
con las liberaciones de estos acaros depredadores en un program de supresi6n del trips oc-
cidental de flores.

Most greenhouse crops in the northeastern cidentalis (Pergande) (Thysanoptera: Thripidae).
United States and many other regions are af- This pest must be managed to prevent distortion
fected by western flower thrips, Frankliniella oc- of flowers and leaves by thrips feeding, and trans-

Van Driesche et al.: Compatibility of Spinosad

mission of tospoviruses (Yudin et al. 1986; Robb
1989; Daughtrey et al. 1997; Jacobson 1997;
Lewis 1997). Producers of spring flower crops in
the northeastern United States most often rely on
application of pesticides for control of this pest,
and currently spinosad, a "reduced-risk" material
derived from a soil microorganism, is the most
frequently used pesticide. This product currently
provides excellent control, but western flower
thrips has frequently developed resistance to pes-
ticides, including abamectin, bifenthrin, chlorpy-
rifos, cyfluthrin, dimethoate, methomyl, and per-
methrin (Robb 1989; Immaraju et al. 1992; Robb
et al. 1995; Jensen 2000).
To reduce the risk of resistance developing to
spinosad, the potential for managing western
flower thrips in greenhouse flower crops with
predators has been investigated. The predators
with most potential for use in short term flower
crops are predatory mites, including Neoseiulus
(=Amblyseius) cucumeris (Oudemans) and Ip-
hiseius degenerans (Berlese) (Acari: Phytoseiidae)
(Van Driesche et al. 1998). Efficacy ofN. cucum-
eris in greenhouses varies among crops, being
most effective in crops such as peppers, which
have long production periods and provide pollen
(a resource used by this predator) (de Klerk & Ra-
makers 1986). Efficacy of this species in floral
crops has not been widely substantiated, despite
its being recommended by natural enemy produc-
ers. In California, a single release ofN. cucumeris
on caged chrysanthemums at 2.5 mites per leaf
reduced western flower thrips to about 2-7 per
leaf over a three week test period (Hessein & Par-
rella 1990), but this density was too high for com-
mercial crops. Gill (1994) reported that releases of
N. cucumeris in open rearing units ("sachets")
lowered pesticides needed for western flower
thrips control in bedding plants from 3.6 to 0.4 ap-
plications per crop, but thrips densities in control
and treated greenhouses were not reported. In
the United Kingdom, Bennison et al. (2001), ob-
tained effective western flower thrips control on
impatiens with weekly applications of higher re-
leases ofN. cucumeris (180 mites per m2, 3-4x the
commercially recommended rate). Similarly, De
Courcy Williams (2001) in the U.K. found that
weekly releases of 200 N. cucumeris mites per m2
reduced western flower thrips on cyclamen. Trials
in Massachusetts with this higher rate found that
it provided better control than the commercially
recommended rate in short term spring flower
crops, but even at this higher rate N. cucumeris
alone did not strongly suppress western flower
thrips (Van Driesche et al. 2005). Even low densi-
ties of western flower thrips may pose a risk to the
crop by spreading tospoviruses such as impatiens
necrotic spot virus (INSV). No information is
available on the efficacy ofL. degenerans in flower
crops. While more effective per mite as a thrips
control agent than N. cucumeris, this species is

much more expensive to rear and so has not been
commercially popular, although it remains avail-
Based on the partial effectiveness of biological
control agents for western flower thrips in spring
bedding plants and the frequency with which this
thrips has evolved pesticide resistance, it may be
useful to combine spinosad (to achieve high thrips
suppression) and predatory mite releases (to re-
duce survival of potentially pesticide-resistant
thrips that might survive pesticide applications).
Spinosad has been reported to be generally com-
patible with predatory mites (Miles et al. 2003;
Ahn et al. 2004; Kim et al. 2005). The objective of
our study was to further assess spinosad compat-
ibility with N. cucumeris and I. degenerans. For
both mite species, we assessed (1) contact toxicity
to mites vs thrips (adults and larvae), (2) repel-
lency of residues of spinosad to mites, and (3) ef-
fects of residues on mite oviposition.


Sources of Mites and Thrips

Western flower thrips used in all tests were
from a colony reared in our laboratory (University
of Massachusetts), started about 1997 with mate-
rial from a laboratory colony from Texas A & M
University (K. Heinz laboratory). Thrips were
reared on excised bean leaves by the method of
Doane et al. (1995) modified by Lim et al. (2001).
Neoseiulus cucumeris mites were purchased from
Koppert Biological Systems (Koppert B. V., The
Netherlands), shipped in bran with grain mites
iT ,.,1. ..,,y... putrescentiae [Schrank], Acarina:
Acaridae) and quality of these mites was consis-
tently high, with large numbers of live mites in
containers. Neoseiulus cucumeris mites were used
in experiments immediately upon receipt. Iphi-
sieus degenerans, packed in sawdust, were pur-
chased from Biobest N. V. (Westerlo, Belgium).
Shipments of L degenerans repeatedly arrived in
poor condition with many mites being dead or
moribund. Before being used in experiments, live
mites were placed on castor bean (Ricinus com-
munis L., Euphorbiaceae) leaves dusted with red
apple pollen and allowed to feed for at least 16 h
to ensure that only mites in good health were
used in experiments.

Experiment #1: Toxicity to Adult Thrips

Toxicity of spinosad at the full labeled rate
(0.1198 g a.i./ml solution) to adult and larval
thrips was tested as a positive control in view of
the low toxicity that we expected to find for the
beneficial mites. For thrips, we examined only the
effects of 2 h-old residues.
For tests with adult thrips, excised leaves of
young red kidney bean (Phaseolus vulgaris L., Fa-

Florida Entomologist 89(3)

baceae) plants produced in growth chambers were
sprayed to the point of run-off with a hand held
sprayer. Control leaves were sprayed with tap wa-
ter. The nozzle of a hand powered spray bottle was
held approximately 10 cm from leaf surfaces and
a total of five sprays were applied to each leaf sur-
face. Leaves were then allowed to dry at room
temperature on the lab bench for 2 h at which
time 2 x 1.5 cm leaf rectangles were cut and
placed individually in 1-dram glass shell vials
(Kimble Glass, Inc, Vineland, NJ). Vial caps had
ventilation holes cut in them, and a small piece of
Nitex screening (03-95/33) (Sefar America, Inc.,
Briarcliff Manor, NY) was held in place by the
perforated cap. A disk (6 mm dia) of filter paper
(#3 Qualitative) (Whatman Limited, Kent, U.K.)
cut with a standard office hole punch was moist-
ened with tap water and added to each vial to pro-
vide humidity.
Five adult female thrips from our laboratory
colony were placed in each of 46 pesticide-treat-
ment and 46 water-treatment vials. Vials were
placed on trays and held for 24 h in a growth
chamber at 22C on a light cycle of 16:8 (L:D), and
about 30% RH. After 24 h, thrips were checked for

Experiment #2: Toxicity to Larval Thrips

Spinosad-treated red kidney bean leaves were
prepared as described above for tests with adult
thrips up to the point of completed air drying. At
that point, whole leaves were fixed to a disc of
white paper cut to fit inside a plastic 15 x 1.5 cm
petri dish (Falcon, Becton-Dickinson, Franklin
Lakes, NJ). The edges of the leaves were taped
with scotch tape to the disc and a thin barrier of
tangletrap sticky material (Tanglefoot Co., Grand
Rapids, MI) was applied around the whole edge of
the leaf to produce a thrips-proof border. This en-
sured that test animals were continuously con-
fined on the treated surface, which was not possi-
ble for winged adults.
With a fine paint brush, 10 immature (first or
second instar larvae) thrips from the UMASS col-
ony were placed near the center of the treated
bean leaf in each of 24 pesticide-treated and 25
water-treated petri dishes. Petri dishes then were
placed on trays and incubated for 24 h in a growth
chamber at 22C on a light cycle of 16:8 (L:D), and
about 30% RH. After 24 h, petri dishes were
closely examined and the number of dead thrips
larvae counted.

Experiment #3: Toxicity to Both Mite Species

Bean leaves were treated with spinosad follow-
ing the same procedures as described for adults
thrips and allowed to dry in a greenhouse for 2,
24, or 48 h to simulate the natural degradation of
the material expected under conditions of green-

house use. Spinosad was applied at full label rate
(0.1198 g a.i./ml solution) for all tests. Leaves
with aged residues were taped into petri dishes
and a tangletrap barrier was created around the
leaf as in the experiment with larval thrips. Un-
like conditions in the thrips tests, in the mite
tests a dusting of red apple pollen and a 6-mm
disk of moistened filter paper were placed on the
bean leaf to supply mites with food and water. In
tests with N. cucumeris and I degenerans, 10
mites (any motile stage) were added per petri
dish. Petri dishes were stacked on trays and
placed in a growth chamber and held at 22C,
with a photoperiod of 16:8 (L:D) and about 30%
R.H. Numbers of petri dishes per treatment var-
ied from 6 to 27, as indicated in Table 2. Water
disks were remoistened twice during the 24 h in-
cubation period. After 24 h, the arenas were in-
spected and the numbers of live, dead, or missing
mites recorded.

Experiment #4: Repellency to Both Mite Species

This test was run with 2-h-aged residues of
spinosad, applied at the full label rate (0.1198 g
a.i./ml solution) to the leaves of red kidney bean,
as described above. Freshly sprayed leaves were
allowed to dry indoors for 2 h. A test arena was
constructed by cutting a spinosad-treated leaf in
half along one side of the mid-vein and then tap-
ing it to half of a water-treated control leaf cut
down the opposite side. The two leaf halves were
overlapped slightly and white glue used to join
them seamlessly. On the surface formed by these
joined leaf halves, tanglefoot was applied as a
barrier to create a square arena (10 cm on a side)
centered on the dividing line so that exactly half
of the enclosed surface was treated and half un-
treated. To initiate the test, 10 mites (the same
procedure for both mite species) were placed on
the line between the treated and untreated leaf
halves. The arena was then continually observed
for 15 min and mites on the treated and un-
treated areas were counted every 3 min. The five
resulting counts were then summed for an indi-
vidual replicate, such that a complete lack of re-
pellency would be indicated by a 25:25 division of
mite-observations between the two parts of the
arena. For N. cucumeris, the responses of 300
mites (30 replicates) were examined and for Ide-
generans, 240 mites (24 replicates) were ob-

Experiment #5: Effect of Spinosad on Mite Oviposition

Spinosad-treated red kidney bean leaves were
prepared as in the toxicity experiments described
above, and then leaf squares (2 x 1.5 cm) were cut
from the treated leaves after 2 h drying time and
placed individually in 1-dram glass shell vials, as
above. Vial caps were cut for ventilation and

September 2006

University of Florida Home Page
© 2004 - 2010 University of Florida George A. Smathers Libraries.
All rights reserved.

Acceptable Use, Copyright, and Disclaimer Statement
Last updated October 10, 2010 - - mvs