Title: Florida Entomologist
Full Citation
Permanent Link: http://ufdc.ufl.edu/UF00098813/00050
 Material Information
Title: Florida Entomologist
Physical Description: Serial
Creator: Florida Entomological Society
Publisher: Florida Entomological Society
Place of Publication: Winter Haven, Fla.
Publication Date: 1997
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: VID00050
Source Institution: University of Florida
Holding Location: University of Florida
Rights Management: Open Access

Full Text


The 80th Annual Meeting of the Florida Entomological Society will be held August
4-7, 1997, at the Adam's Mark Hotel, Daytona Beach Resort, 100 North Atlantic Ave.,
Daytona Beach, FL 32118. Phone: (800) 444-2326; FAX: (904) 253-8841.


The deadline for submission of papers for the 80th Annual Meeting of the Florida
Entomological Society will be Friday, May 9, 1997. Time allotted for submitted oral
papers will be eight minutes for presentation and two minutes for discussion. Confir
nation of receipt of papers will be sent to the first author. There will be oral student
paper sessions with awards as in previous years. A description of the format for judg
ing the student papers is printed in the Newsletter. Students participating in the
judged sessions must be members of the Florida Entomological Society and registered
for the meeting.

Joe Funderburk, Chairman
Program Committee
NFREC, University of Florida
1700 SW 23rd Drive
Route 3, Box 4370
Quincy, FL 32351 USA
Phone: (904) 875-7146
FAX: (904) 875-7148


Oral Presentation
Project Exhibit (Poster) Session
Student Paper (Judged)
Student Poster (Judged)

Return to: Joe Funderburk
NFREC, University of Florida
Route 3, Box 4370
Quincy, FL 32351

DEADLINE: May 9, 1996
Author's Name
Title of Paper

Affiliation and Address
of the First
'-I. 1. -.,: Author

Abstract: Mustbe Provided. Do notuse more than 75 words.



Hall et al.: Citrus Rust Mite & the Negative Binomial 1


'Research Department, United States Sugar Corporation
P.O. Drawer 1207, Clewiston, FL 33440

University of Florida, Citrus Research & Education Center
Lake Alfred, FL 33850

3DowElanco, Tampa, FL 33607

4University of Florida, Department of Entomology and Nematology
Gainesville, FL 32611 0620


Count data for the number of citrus rust mites per cm2 on fruit across a 4-ha (10
acre) area of orange trees followed the negative binomial probability distribution 79%
of the time based on chi-square tests. A correlation of r = 0.993 was found between ob
served counts and counts projected based on the distribution. A common k of 0.149
was computed but generally appeared more suitable for mean densities of 3.0 to 55.0
than 0.5 to 3.0 citrus rust mites per cm2. For mean densities of from 0.5 to 55 citrus
rust mites per cm2, the parameter k of the negative binomial was related to the mean
density (x): k = 0.081 + 0.1139*(log,x). Estimated k values were used to draw ex
pected count data profiles for several mean densities ranging from 1 to 40 citrus rust
mites per cm2. Due to the skewness of the count data, the number of mites per cm2 ex
pected in most individual samples was always considerably smaller than the average
density. Based on the negative binomial, mean rust mite densities could be estimated
from the percentage of samples with at least one mite. Results of the study provide a
means to predict the relative frequency histogram of densities associated with a mean
density of citrus rust mites per cm2 across an area of trees.

Key Words: citrus rust mite, Phyllocoptruta oleivora, sampling, negative binomial dis


Los datos del nfmero de Phyllocoptruta oleivora por cm2 en frutos de naranja en
un area de 4 ha (10 acres) siguieron una distribuci6n de probabilidad binomial nega
tiva en el 79% de los casos, basada en pruebas de chi-cuadrada. Fue encontrada una
correlaci6n de r = 0.993 entire los conteos observados y los proyectados sobre la base
de la distribuci6n. Una k comfn de 0.149 fue computada, aunque en general pareci6
ser mas adecuada para densidades medias de 3.0 a 55.0 que para densidades de 0.5 a
3.0 acaros por cm2. Para densidades promedio de 0.5 a 55 por cm2, el parametro k de
la binomial negative estuvo relacionado con la densidad promedio (x): k=0.081 +
0.1139*(log,0x). Los valores estimados de kfueron usados para calcular los perfiles de
los datos de conteo esperados para varias densidades medias en el rango de 1 a 40 aca
ros por cm2. Debido a la desviaci6n de los datos de los conteos, el nimero de acaros por
cm2 esperado en la mayoria de las muestras individuals fue siempre considerable
mente menor que la densidad promedio. Tomando como base la binomial negative, las
densidades medias de acaros podrian ser estimadas a partir del porcentaje de mues
tras con al menos un acaro. Los resultados del studio proven medios para predecir

Florida Entomologist 80(1)

el histograma de frecuencia relative de densidades asociadas con una densidad pro
medio de P oleivora por cm2 en un area con arboles.

Average densities of the citrus rust mite (CRM) [Phyllocoptruta oleivora (Ash
mead)] per cm2 on fruit across an area of orange trees can be estimated from counts of
the number of mites present within a one-cm2 surface area per fruit (Hall et al. 1994).
The number of fruit and trees that must be sampled depends upon both the desired
precision of estimates and the density of mites at which this precision is required.
If the probability distribution (e.g., see Gomez & Gomez 1984) associated with
CRM count data is known, the frequency histogram of individual counts associated
with a particular mean density can be projected. This would be useful for projecting
damage by a CRM population and for establishing control levels. Histograms of CRM
counts taken within individual trees usually followed the negative binomial probabil
ity distribution (Hall et al. 1991). No information was available on probability distri
butions describing CRM count data from fruit over an area of trees.
We had a considerable amount of CRM count data from fruit samples taken across
4-ha (10-acre) areas of 'Hamlin and 'Valencia' orange trees in Florida (Hall et al.
1994). Previous analyses of the data indicated that the counts usually followed an ag
gregated distribution (Hall et al. 1994). Because aggregated dispersions often follow
the negative binomial probability distribution (Southwood 1978), and because CRM
count data from individual trees usually followed the negative binomial, we evaluated
this distribution for projecting the frequency histograms of our count data.


Count data were obtained on the number of CRM per cm2 on fruit across 32 4-ha
blocks of 'Hamlin' and 'Valencia' orange trees using a transect sampling plan (Hall et
al. 1994). This plan consisted of 192 1 cm2 samples per block -two samples per fruit,
four fruit per tree (1 from each compass quadrant), 12 trees along one transect be
tween the northeast and southwest corners of the block, and 12 trees along a second
transect between the northwest and southeast corners of the block. All CRM except
eggs within a 1 cm2 sample were counted using a 10X magnifier fitted with a 1 cm2
grid of 25 equal-sized subdivisions. In cases where >35 CRM per cm2 were present, the
number of mites was sometimes estimated by counting the number of mites in a diag
onal row of five grid subdivisions and multiplying by 5. The block samples were taken
during May through December within several different citrus growing areas in Flor
ida. The only treatment applied to the blocks during the study was a summer spray
of copper and oil. No samples were taken until at least 6 wk after this treatment.
The negative binomial probability distribution is characterized by two parame
ters, the mean (x) and a coefficient k (Johnson & Kotz 1969). The value of the k pa
rameter defines the shape of the negative binomial distribution and serves as a
general indicator of aggregation, with smaller values of k indicating increased aggre
gation (Southwood 1978). An iterative solution was used to manually estimate k for
each block averaging at least 0.5 CRM per cm2 (25 blocks):

N og 1 ] X-]

March, 1997

Hall et al.: Citrus Rust Mite & the Negative Binomial

with N= total number of samples, log = natural logs, and Ax= the sum of all frequent
cies of sampling units containing more than individuals (Bliss & Fisher 1953, South
wood 1978).
The 25 estimates were then evaluated using regression procedures presented by
Bliss & Owen (1958) and Bliss (1958) to determine if a single, common k (k) existed.
This involved regressing two statistics for each block, y'[= s2 x] on x'[= x2 (s2/N)],
where x was the mean, s2 the variance and N the number of individual counts per
block. The regression was forced through the origin, and k was estimated from the in
verse of the slope of the regression. The adequacy of this k estimate was evaluated us
ing a regression analysis of 1/k on logo(x): a trend between these variables discredits
the suitability of a single kc (Bliss & Owen 1958, Southwood 1978).
One way to write negative binomial probabilities is:

(x k 1)! r k Ix ] = 0,1,2... (2)

where p, is the probability of a sample having x mites (Williamson & Bretherton
1963). To determine the histogram of CRM counts expected in each block according to
the negative binomial, we used observed mean densities and estimates of k in the fol
lowing iterative probability formula:

Px+l x=+1 ~- Px x = 0,1,2... (3)

with the probability of no mites (x=0) being

p= k k-i

where p, = the probability of a sample containing xmites. Equation (3) was obtained
by writing successive terms for p,, p, P2... from equation (2) and noting the common
multiplier. Using this iterative method avoids brute force calculation of the combine
trials which often cause computer overflow for large values of x. We programmed
SAS (SAS Institute Inc. 1990) software to compute the successive probabilities. Chi
square tests (a = 0.05) according to guidelines presented by Gomez & Gomez (1984)
and correlation analyses were used to test the fit of the observed CRM counts to those
expected under the negative binomial based on estimated k values.


The mean density of CRM observed in the 25 blocks ranged from 0.5 to 112.5 per
cm2. A regression analysis indicated that the maximum density of CRM (y) observed
in each block could be estimated from the mean density (x): y= 32.5 + 17.3*x; r2 = 0.85,
F- 123.1, PR >F 0.0001, d.f. 23.
Individual estimates of the negative binomial k ranged from 0.0199 to 1.58 ( =
0.2147, s = 0.3025) (Fig. 1). With respect to investigations into kI, an initial plot of y'
on x' indicated that one data point clearly deviated from the main trend of the regres
sion (Fig. 2). This data point, which was associated with a mean density of 112.5 CRM
per cm2 and a k value of 1.58, was excluded from further investigations into k but in
dicated that CRM aggregation may substantially decrease as population densities in
crease to as high as 100 or more CRM per cm2.

Florida Entomologist 80(1)

Among the 24 sets of count data retained for k, determination, mean densities
ranged from 0.5 to 54.9 CRM per cm2 (x = 12.96 per cm2, s = 16.9). The individual kes
timates for the blocks varied from 0.0199 to 0.3580 (k= 0.158, s = 0.1052). A kof 0.149
was calculated (F = 141.7, Pr > F = 0.0001, r2 = 0.86, d.f. 24) (r2 corrected for the mean
= 0.83, d.f. = 23) (Fig. 2). A statistically insignificant relationship (a = 0.05) was found
between 1/k and x number CRM per cm2, but a weak relationship (r2 = 0.395) was
found between 1/k and logo(x) number CRM per cm2 (Fig. 3), which indicated the k of
0.149 may not have been a suitable substitute for all of the individual k values. A sim
ilar problem was reported with respect to determining a k associated with a set of
wireworm counts (Bliss & Owen 1958). Variability in individual k values associated
with small mean CRM densities (e.g., 0.5 to 3.0 CRM per cm2) was responsible for this
trend; no significant trend was found between 1/k and log,,(x) number CRM per cm2
among mean densities of from 3 to 55 CRM per cm2 (N = 15), and the same k. (0.149)
was calculated across these densities.
Because k tended to be a poor substitute for individual values at mean densities be
low around 3.0, as an alternative to k we conducted a regression analysis and deter
mined an equation for estimating k across different mean densities (x): k = 0.081 +

1.75 I I -






0.25 *


0 20 40 60 80 100 120

Mean no. crm/cm2

Fig. 1. Negative binomial k values associated with observed mean densities of cit
rus rust mites (crm) per cm2 on fruit across 4-ha areas of orange trees.

March, 1997

Hall et al.: Citrus Rust Mite & the Negative Binomial

25000 I I

20000 Y =6.699XI

/ 1/kc = 6.699

15000 kc =0.149


U5000 data point excluded'
from regression

0 1 rI l ,

0 5000 10000

Fig. 2. Regression analysis used to calculate a single, common k (k) for a negative
binomial distribution describing the number of citrus rust mites per cm2 on fruit
across 4-ha areas of orange trees. The open circle represents a data point excluded
from the regression; this data point was associated with a mean of 112.5 mites/cm2
and a kof 1.58.

0.1139*(logx); F = 27.11; Pr > F = 0.0001; r = .55; d.f. 23. A comparison of some histo
grams generated from individual, common and regressed k values is presented in Fig. 4.
Chi-square tests indicated that CRM counts across a 4-ha area of trees followed
the negative binomial distribution in 19 of 24 (79%) areas based on individual k val
ues and in 16 of 24 (67%) areas based on either kc or regressed k values. Among the
observed count histograms that did not follow the negative binomial based on chi
square tests, these histograms visually resembled the distribution (e.g., Fig. 4). Over
all 24 sets of CRM count data, the correlation between observed counts and counts
projected using the negative binomial was 0.993 based on individual k values, 0.976
based on regressed k values, and 0.965 based on the kI estimate. Among the 24 count
sets, the lowest correlation between observed counts and counts expected under the
negative binomial was 0.929, 0.921 and 0.872 based on individual k values, regressed
k values and kI, respectively.
Overall, counts of the number of CRM per cm2 on fruit across a 4-ha area of trees
appeared to be at least reasonably described by the negative binomial distribution
when mean densities were in the range of 0.5 to 55 per cm2. The distribution in con

6 Florida Entomologist 80(1) March, 1997

55 .
50 -* *
4 Y = 17.61- 0.32X
40 _- *-
SF=3.62 (Pr>F=0.07)
35 =0.14 Y= 22.4-13.1X
4K 30 df=23 -
0 23 F=14.4, Pr>F=0.001
r" 25 r. r = 0.395
20 d.f.= 23-

10 7 i
5 ' .* .- Y .-
0- I- Iiii I I
0 10 20 30 40 50 0 1 2

Mean no. crmlcm2 Log,0 mean no. crm/cm2

Fig. 3. Relationship between 1/k and x compared to 1/k and log,,x number of citrus
rust mites (crm) per cm2 on fruit across a 4-ha area of orange trees.

junction with the k parameter could therefore be used to project the frequency histo
gram of CRM densities at any mean density within this range, which in turn could be
used in combination with models projecting how much surface damage to fruit a given
density of CRM will cause (e.g., see Allen 1976, Yang et al. 1995) for an overall esti
mate of damage a CRM population will cause. The distribution could also be used to
develop a sequential sampling plan (Southwood 1978), which might reduce the cost of
sampling CRM. While histograms projected based on k were similar to those based on
regressed k values, overall our analyses favored histograms based on the regressed
estimates. As a word of caution, values and the goodness-of fit of the negative bino
mial distribution could be negatively influenced by extraneous factors that affect mite
dispersion in a grove (e.g., chemical applications).
Expected profiles of CRM counts for a number of mean densities ranging from 1 to
40 mites per cm2 were projected based on the negative binomial using regression esti
mates of k in formula #3 (Fig. 5). Differences were relatively small between means of
1 to 40 mites per cm2 with respect to the projected probability of any individual count
in the range of 5 to 15 mites per cm2. Due to the skewness of CRM count data, the num
ber of mites per cm2 expected in most individual samples was always considerably
smaller than the average density. For example, at an average density of 5 CRM per
cm2, fewer than 5 mites per cm2 were expected to be present in around 80% of the in
dividual samples across a 4-ha area. This information would be important to a citrus
grower who might mistakenly assume that an average density based on scouting data
reflects the midpoint of densities present. As the mean density increased, the probabil
ity increased that any particular large count would be observed. For example, the ex
pected percentage of counts above 30 mites per cm2 increased from about 4% at a mean
of 5 mites per cm2 up to about 18% at a mean of 20 mites per cm2. The skewness of CRM
count data supported contentions made by McCoy et al. (1976), namely that a control
threshold should take into consideration the frequency histogram of mite counts.
Given that count data follow the negative binomial and kis known, expected mean
densities can be estimated from the percentage of samples containing at least one an

Hall et al.: Citrus Rust Mite & the Negative Binomial

0.08 served
.0 , ^^^^ NB. individual k I
S0.04 mmon
Q" 0 1^ <^ NB k re ssed*
0 20 40 60

Mean = 4.7 mites

S 0.12

S0.08 L
J, NB. individual k'
2 0.04 NB omnk
0. 0 lllllI iamN k regr d
0 25 50 75 100 125

Mean = 26.5 mites

S 0.08
2 0.04 B1commonkI
0 75 150 225 300
Number of rust miteslcm2
Fig. 4. Probability (proportion) of observed counts of citrus rust mites per cm2 com
pared to counts projected from negative binomial (NB) distributions derived from in
dividual, common and regressed k values (largest observed counts shown,
probabilities below 0.0001 or above 0.15 not shown). An asterisk (*) indicates the ob
served histogram followed the projected histogram based on a chi-square test (a

Florida Entomologist 80(1)


. 0.6

0.4 E=40/cm

0.3 x=20/c- 2
-- :L :7,=10/cm
M i=5/cm
C) 0.1 I cm

0 50 100 150 200 250 300
Number of citrus rust mites/cm2
in individual samples

Fig. 5. Projected frequency distributions for citrus rust mite counts per cm2 on fruit
across a 4-ha area of orange trees based on the negative binomial and regressed k val
ues, means of 1, 5, 10, 20 and 40 mites per cm2 (class probabilities below0.00025 and
above 0.7 not shown). The zero-class probability extended up to 0.81 at a mean of 1
mite per cm2.

imal (see Southwood 1978). The relationship between mean CRM density and the per
centage of infested samples based on the negative binomial is presented in Fig. 6.
Similar relationships between percentages of infested samples and mean CRM den
sities have been observed without the use of a probability distribution (Knapp & Fa
sulo 1983, McCoy et al. 1976). The percentage of samples infested became
increasingly poorer as an indicator of mean density as CRM densities increased. Ben
efits and precautions associated with using the percentage of infested samples as an
indicator of mean CRM densities have been discussed (McCoy et al. 1976).


The authors acknowledge and thank Clay McCoy, Jorge Pena and Phil Stansly for
their constructive reviews during the preparation of this manuscript. This article is
Florida Agricultural Experiment Stations Journal Series No. R-05232.


ALLEN, J. C. 1976. A model for predicting citrus rust mite damage on Valencia orange
fruit. Environ. Entomol. 5: 1083-1088.

March, 1997

Hall et al.: Citrus Rust Mite & the Negative Binomial

C4 T IJ.VV-A) I \OU.4 A)
E 50 r=0.93)
r= 0.93

40 -


1030 40 50 60 70 80 90100

0 10 20 30 40 50 60 70 80 90100

X = Percent samples infested
Fig. 6. Relationship between the mean density of citrus rust mites per cm2 on fruit
across a 4-ha area of trees and the percentage of 1-cm2 samples with at least one rust
mite. Parameter estimates (standard error) associated with the hyperbola were 80.94
(1.166) and 3.994 (0.1905). A 95% confidence interval is given by the dotted lines. Per-
cent samples infested (x) can be estimated from the mean number of mites/cm2 (y) by:
x= (80.94 )/(3.99 + ).

BLISS, C. I. 1958. The analysis of insect counts as negative binomial distributions.
Proc. XInt. Congr. Entomol. 2: 1015-1032.
BLISS, C. I., AND R. A. FISHER 1953. Fitting the negative binomial distribution to bi
logical data and note on the efficient fitting of the negative binomial. Biomet-
rics. 9: 176-200.
BLISS, C. I., AND A. R. G. OWEN. 1958. Negative binomial distributions with a common
k. Biometricka. 45: 3758.
GOMEZ, K. A., AND A. A. GOMEZ. 1984. Statistical Procedures for Agricultural Re
search, 2nd ed. Wiley & Sons, New York. 680 pp.
HALL, D. G., C. C. CHILDERS, AND J. E. EGER 1991. Estimating citrus rust mite (Acari:
Eriophyidae) levels on fruit in individual citrus trees. Environ. Entomol. 20:
HALL, D. G., C. C. CHILDERS, AND J. E. EGER 1994. Spatial dispersion and sampling
of citrus rust mite (Acari: Eriophyidae) on fruit in 'Hamlin and 'Valencia' or
ange groves in Florida. J. Econ. Entomol. 87: 687-698.

10 Florida Entomologist 80(1) March, 1997

JOHNSON, N. L., AND S. KOTZ. 1969. Discrete Distributions. Houghton Mifflin Co.,
Boston. 328 pp.
KNAPP, J. L., AND T. R. FASULO. 1983. Citrus rust mite. In Florida citrus integrated
pest and crop management handbook. SP-14, Florida Coop. Ext. Serv., IFAS,
University of Florida, Gainesville.
McCoY, C. W., R. F. BROOKS, J. C. ALLEN, AND A. G. SELHIME. 1976. Management of
arthropod pests and plant diseases in citrus agroecosystems. Proc. Tall Timbers
Conf. Ecolog. Animal Control Habitat Management. 6: 117.
SAS INSTITUTE INC. 1990. SAS Language: Reference, Version 6, First Edition. Cary,
NC. 1042 pp.
SOUTHWOOD, T. R. E. 1978. Ecological Methods, 2nd ed. Wiley/Halsted, New York.
WILLIAMSON, E., AND M. H. BRETHERTON. 1963. Tables of the Negative Binomial Dis
tribution. John Wiley & Sons, N.Y. 275 pp.
YANG, Y., J. C. ALLEN, J. L. KNAPP, AND P. A. STANSLY. 1995. Relationship between
population density of citrus rust mite (Acari: Eriophyidae) and damage to
'Hamlin' orange fruit. Environ. Entomol. 24: 1024-1031.

Florida Entomologist 80(1)


Department of Entomology, P.O. Box 231, Cook College
Rutgers University, New Brunswick, NJ 08903


The effect of insecticides currently used in commercial eggplant fields to control
the Colorado potato beetle, Leptinotarsa decemlineata (Say) on two egg predators, Co
leomegilla maculata DeGeer and Chysoperla carnea (Stephens) was evaluated. Mor
tality from contact exposure to leaf residues, topical applications, and ingestion of
contaminated eggmasses was compared for the following insecticides: esfenvalerate
alone and in combination with piperonyl butoxide (PBO); oxamyl; PBO; and rotenone
alone and in combination with PBO. Topical exposure and feeding studies were con
ducted using concentrations 1.00, 0.90, 0.80, 0.70, 0.60, 0.50, 0.40, 0.30, 0.20, and
0.1OX the maximum labeled dose; leaf exposure studies were conducted using concern
trations 1.00, 0.75, 0.50, and 0.25X the maximum labeled dose. Mortality of C. macu
lata adults and larvae from topical exposure was high after 48 h of exposure for all
chemicals and doses. Mortality from topical exposure was low for C. carnea larvae in
all cases when compared to PBO alone. Mortality from exposure to leaf residues was
low in all cases for C. maculata adults but varied, depending on dose and chemical, for
both C. maculata and C. carnea larvae. For all treatments, ingestion of treated eggs
negatively affected the feeding and survival of C. maculata adults and larvae and C.
carnea larvae. Esfenvalerate combined with PBO had the greatest effect on C. macu
lata adults; rotenone combined with PBO had the greatest effect on C. maculata lar
vae; esfenvalerate combined with PBO affected C. carnea larvae the most.

Key Words: Coleomegilla maculata, Chrysoperla carnea, insecticides, eggplant, IPM,
Leptinotarsa decemlineata

March, 1997

Hamilton & Lashomb: Insecticides and CBP Predators 11


Fue evaluado el efecto de los insecticides comunmente usados en campos comer
ciales de berenjena para el control del escarabajo de Colorado, Leptinotarsa decemli
neata (Say), sobre dos depredadores de huevos, Coleomegilla maculata DeGeer y
Chysoperla carnea (Stephens). Fue comparada la mortalidad por exposici6n a resi
duos en las hojas, aplicaciones t6picas, e ingestion de masas de huevos contaminadas
con los siguientes insecticides: esfenvalerate solo y en combinaci6n con piperonyl bu
toxide (PBO); oxamyl; PBO; y rotenone solo y en combinaci6n con PBO. Los ensayos
de exposici6n t6pica fueron efectuados utilizando concentraciones de 1.00, 0.90, 0.80X,
0.70, 0.60, 0.50, 0.40, 0.30, 0.20, y 0.10X de la dosis maxima recomendada para los
products. Los studios de exposici6n de las hojas fueron conducidos usando concern
traciones de 1.00, 0.75, 0.50, y 0.25X de la dosis maxima recomendada. La mortalidad
de los adults y larvas de C. maculata mediante exposici6n t6pica fue alta luego de 48
horas de exposici6n a todos los products y dosis. La mortalidad de las larvas de C.
carnea mediante exposici6n t6pica fue baja en todos los casos cuando se compare con
con el PBO solo. La mortalidad mediante exposici6n a residues en las hojas fue baja
en todos los casos para los adults de C. maculata pero vari6 con la dosis y el product
en las larvas de C. maculata y C. carnea. En todos los tratamientos, la ingestion de
huevos tratados afect6 negativamente la alimentacidn y sobrevivencia de los adults
y larvas de C. maculata y de las larvas de C. carnea. El esfenvalerate combinado con
PBO tuvo el mayor efecto en los adults C. maculata, el rotenone combinado con PBO
tuvo el mayor efecto en las larvas de C. maculata, el esfenvalerate combinado con PBO
fue el product que mas afect6 a las larvas de C. carnea.

In New Jersey, the Colorado potato beetle, Leptinotarsa decemlineata (Say), is a
key pest in eggplant and, if left uncontrolled, subsequent defoliation can completely
destroy the crop (Cotty & Lashomb 1982). Colorado potato beetle populations have be
come resistant to most chemical insecticides resulting in difficulty in protecting the
crop (Forgash 1985). During the late 1980s, a biological control intensive pest man
agement program (BCIPM), using the egg parasitoid Edovum puttleri Grissell (Hy
menoptera: Eulophidae), was implemented to reduce grower reliance on insecticides
(Lashomb 1989). This program uses field scouting, weekly parasitoid releases, and in
secticide applications pre and post-release if either eggmass or larva/adult economic
thresholds are reached.
A side benefit to this successful program has been increased populations of several
indigenous natural enemies of Colorado potato beetle including Coleomegilla macu
lata DeGeer and Chrysoperla carnea (Stephens) (Lashomb unpublished) not found in
non-BCIPM fields. Fields not utilizing biocontrol are repeatedly treated with various
insecticides including esfenvalerate, oxamyl and rotenone (Hamilton 1995). The lack
of natural enemies suggests that insecticide usage may be adversely affecting non-tar
get species such as C. maculata and C. carnea. Several studies have documented the
effect of various insecticides on C. maculata and C. carnea; however, these materials
are not used in eggplant. This study evaluated the effect of insecticides commonly ap
plied to eggplants for Colorado potato beetle control on C. maculata and C. carnea.


C. maculata DeGeer (larvae and adults) and C carnea (Stephens) (larvae) were
used in all studies. C. maculata adults and larvae were obtained from the USDA Ben

Florida Entomologist 80(1)

eficial Insect Laboratory (Mission, TX); C. carnea from Rincon-Vitova Insectaries, Inc.
(Ventura, CA). Before each study, all individuals were held in a Precision" growth
chamber maintained at 26 1 C, 45 5% RH and a photoperiod of 15:9 [L:D].
Three commonly used insecticides, esfenvalerate (Asana XL", E. I. Dupont, Wilm
ington, DE), oxamyl (Vydate L", E.I. Dupont, Wilmington, DE), and rotenone (Ro
tenox", Fairfield American, Frenchtown, NJ) and 1 synergist, piperonyl butoxide
(PBO) (Butoxide', Fairfield American, Frenchtown, NJ) were tested. Combinations of
esfenvalerate and PBO, and rotenone and PBO were also tested. The topical exposure
and egg feeding studies were conducted using concentrations [g (AI) per liter] of ap
proximately 1.0, 0.90, 0.80, 0.70, 0.60, 0.50 0.40, 0.30 and 0.20X of the maximum la
beled dose recommended for controlling Colorado potato beetle (esfenvalerate 0.12,
0.11, 0.096, 0.08, 0.07, 0.06, 0.05, 0.036, and 0.02X respectively; oxamyl 1.20, 1.08,
0.96, 0.84, 0.72, 0.60, 0.48, 0.36, and 0.24X, respectively; PBO 2.39, 2.15, 1.91, 1.67,
1.43, 1.20, 0.96, 0.72, and 0.48X, respectively; rotenone 7.66, 6.89, 6.13, 5.34, 4.60,
3.83, 3.06, 2.30, and 1.53X, respectively) and a water control. Leaf exposure studies
were conducted using concentrations [g (AI) per liter] of approximately 1.0, 0.75, 0.50
and 0.25X of the maximum labeled dose recommended for controlling Colorado potato
beetle (esfenvalerate 0.12, 0.09, 0.06, and 0.03X, respectively; oxamyl 1.20, 0.90,
0.60, and 0.30X, respectively; PBO 2.39, 1.79, 1.20, and 0.60X, respectively; rotenone
- 7.66, 5.75, 3.83, and 1.92X, respectively) and a water control.

Topical Exposure Tests

The effect of topically applied insecticides on C. maculata adults and larvae, and
on C. carnea larvae was evaluated. For each material and dose, 100 C. maculata
adults were treated with 10 pl of insecticide using a Burkhardt" metered micro-sy
ringe applicator. Treated individuals were then placed into vented 9-cm petri dishes
(10 per dish) containing moistened filter paper and held for 48 h. Mortality was re
corded at 24 and 48 h post exposure. All individuals were removed after 48 hours, re
counted, and the mean percent mortality was determined. This procedure was
repeated for both C. maculata and C. carnea larvae with the exception that treat
ments were made using 1 pl of insecticide.

Leaf Exposure Tests

The effect of insecticide leaf residues on C. maculata adults and larvae, and C. car
nea larvae was evaluated using treated leaf disks. For each material, dose and species,
10 eggplant leaves were excised and the petioles inserted into an Oasis' rootcube
moistened with water and trimmed to 63.5 cm2 leaf disks using a 9-cm plastic petri
dish placed over the midrib. Each leaf disk was dipped into 100 ml of the respective
concentration for each material, air dried, and placed into vented 9-cm petri dishes.
Ten individuals were introduced into each petri dish and held for 48 hours. Mortality
was recorded at 24 and 48 h post exposure. All individuals were removed after 48
hours, recounted, and the mean percent mortality was determined.

Feeding and Survival Tests

The effect of insecticides topically applied to L. decemlineata eggmasses on feeding
and survival of C. maculata adults and larvae, and C. carnea larvae was determined.
Eggmasses were obtained by rearing Colorado potato beetle larvae to adults and al
lowing them to lay eggs on caged potato plants maintained in the greenhouse under

March, 1997

Hamilton & Lashomb: Insecticides and CBP Predators 13

25.0 + 2.0'C temperature and a photoperiod of 12:12 (L:D). Eggmasses were collected
from plants and trimmed to 10 eggs per mass. For each species, one hundred trimmed
eggmasses per concentration per chemical were treated with 10 Pl of material per egg
mass using a Burkhardt" metered micro-syringe applicator, allowed to air dry, and
transferred to sealed Solo" plastic condiment cups (59.1 ml) (1 eggmass per cup). A
single individual was then placed into each cup. Each day following the initiation of
the test, all eggmasses were removed from cups, examined for evidence of feeding, and
replaced with freshly treated eggs. Daily monitoring continued until either pupation
or mortality occurred. The mean number of eggs fed upon and the survival rate of in
dividuals was determined.

Statistical Analysis

Topical and leaf exposure percent mortality data were corrected using Abbott's
(1925) and analyzed using probit analysis (Robertson & Preisler 1992, SAS 1987);
feeding and survival data were transformed to SQRT(X + 1) (Snedecor and Cochran
1978) and analyzed using linear regression (SAS 1987).


Topical Exposure Tests

Topical exposure of C. maculata adults and larvae and C. carnea larvae to all ma
trials tested resulted in mortality in all cases (Table 1). Mortality levels were consis
tently higher at 48 h post-exposure than at 24 h post-exposure. C. maculata adults
were most sensitive to topical applications of esfenvalerate in combination with PBO
after both 24 h and 48 h when compared to PBO alone (Table 2). Exposure to rotenone
alone resulted in the lowest toxicity ratios observed. Overall, toxicity ratios for esfen
valerate alone and in combination with PBO for C. maculata larvae were higher than
those observed for adults but lower for all other materials. Similar levels of mortality
have been reported when C. maculata was topically exposed to cypermethrin, car
baryl, fenvalerate, malathion and permethrin (Coats et al. 1979, Lecrone & Smilowitz
1980, Roger et al. 1994). C. carnea larvae were least affected by topical applications
when compared to PBO, thus supporting evidence that C. carnea is tolerant of certain
pyrethroid and carbamate insecticides (Shour & Crowder 1980, Ihaaya & Casida
1981, Grafton-Cardell & Hoy 1985, 1986, Pree et al. 1989). Toxicity ratios at 48 h post
exposure were highest for esfenvalerate in combination with PBO, followed by ox

Leaf Exposure Tests

Exposure to foliar residues of the insecticides also resulted in high levels of mor
tality (Table 3). For each insecticide, mortality was highest at 48 h post-exposure. C.
maculata adults, however, showed no response to low levels (0.25) of esfenvalerate
alone, rotenone alone or in combination with PBO, or oxamyl. Exposure to leaves
treated with esfenvalerate in combination with PBO resulted in the highest toxicity
ratios observed for each species, followed by esfenvalerate alone and oxamyl (Table 2).
C. maculata adults were least affected by exposure to the insecticides tested when
compared to larvae. These findings contradict the work by Plapp & Bull (1978) that
showed C. carnea to be less susceptible to contact residues of pyrethroid insecticides

Florida Entomologist 80(1)


, c'J




c LO~ O

r '-7 co

o 0 t CO

0 00 0
0 0 CD 0


I c




O ^
m m
LO ^

0 t-

CD '?

0 CD
0 C

R a

,-- c


LO 0

+1 +1 +1
, 00 0
CM C 00

a e'
0 "

o ,
o cs-

C 'J



o i
- ND


^-1 f
i- L

v v L
m- M
CO 0




cc -
c'J o,

Ni M i M CD M -I O N
00 000000
+1 +1 +1 a) +1 +1 +1 +1 +1 +1

C0 LO 2 M O Q 0 C O

o o
m m

G 0

March, 1997

Hamilton & Lashomb: Insecticides and CBP Predators 15


+1 +1 +1 +1
Cf t C 0 t0

OO cjoC

SCl o

+ u cis o -
cc 0M
S00 N t C'J

KC o6 C

0 00 CM CM

H M oo o

t i O li

Ct 1l 0 C'J

z o) +1 +1 +1 +1
a +1 00 t

o c.j QO
-c c0 C M

I 00D 0

^ u uo O

Florida Entomologist 80(1)

LOC N- 0 --

CD LflC'J -
CD' N- 0c'

N- N-d C Ct
C' -- -m

- c'J 0

ict C Cl d C
CD-00 f

- c'J~r~

- ~bN- D D
C'J CD i'C' C

CD 0 0
LOl o 0


Cti 0
c'j 0


a C

m C C
3 C
m ai ai

w c c

CD 0 0 0



e za

~CCLfO Lfl
000CC N-
0 0 0- 0

~m 0 0 N-
6dovi CD

Cfl c 0
m Lf 0 0N-
cC 0 0

0 nm

Lfl 0 -
C'J Lflo

C'Jt N- cf
0~CD0 CD,

0-ato Lfl
000- b'

00 -DLOf0 0

0 0OCD-0 0

0 0C-



0 00

> aa

eai me w w ^
C o C C E $
2t2a, a oo ,

V V 0 0 Z f

March, 1997

Hamilton & Lashomb: Insecticides and CBP Predators

oC ~ LfLf
cc c'j cc cj
- cfl 'J -

cco C C
c'J 1,
- cc cc

0 '- C Lfl C
o at C'J N- C
0 t--00

o 0 CD C
o L C 0 0
-i- CM LO i
LO i-


C 00 0 0
0-C CM C -


c'J -

-ff CM '- C^

-00 CC 00 -

*^f at N- C C

Florida Entomologist 80(1)


L _


2 ccE-



CD "
t c 'J0

- cc -

cz cc
Scc M

+1 +1
0 t-



cc -
00 -







c'j Ln

oo M
c'J 'J
- c'J

CO _^


0^ c c) LO
S0 O

- c'


O 00

+1 +1

O oo

- Cc

cs J

c J

S= 0LO

0 0 O C

D D -D

+1 +1 +1 +1 +1 +1


m C
d 8 E

a aco C

C 2OL 2
ho ~~ o
^ T3 T c c
w >> oo >

- C'



- CO

+1 +1
0 i

i 0

od D


o D


March, 1997

Hamilton & Lashomb: Insecticides and CBP Predators 19

M 00
-^ COJ Cf 0 CM

ECD c0 tN-

3r 00 I "

g M n
j c l b LO
r CO i- 00 CM CO
go LL o

L, "

So) +1 +1 +1 +1
+1 CtO CD

L Ll 0 C 00 0 C

E -C
C tL t- at C
= CDCD C'J t

0 0 0 0 LO 0

S CD" t- C M O

H o

Florida Entomologist 80(1)

March, 1997

Ct fl fl Cf
o o C C C C0

t- 00 L C


o o CC C
o o CC C


o Q


N *' *- Ll -
t-~ f ^1 O t- t-
bt' 00 0C 00 00

O CD b' N Ct

o o CC C
o o CC C




B -D

00 Lf Lf
St 00

CT Cn -
M C'j -
+1 +1 +1

+1 +1 +1

CD tN-

0 0C
+1 +1 +1


000 C

+1 +1 +1

Cn CD N-


+1 +1 +1

, c'i cn

o o

Hamilton & Lashomb: Insecticides and CBP Predators 21


SI^ tN- 00 tN-
po oo

Q -L +1 +1 +1
M C CD oo LO
W e 0m

E c o o o
M 3 Q 000
Lf +1 +1 +1


t +1 +1 +1

o co

O C-x

at -

5 ES +1 +1 +1

S +1 CD C C

7 0

0 00


Florida Entomologist 80(1)

when compared to various carbamate and organophosphate materials. The data also
suggest that field applications of rotenone could have a detrimental effect on the lar
val populations of both predators.

Feeding and Survival Tests

All insecticide treatments negatively affected the feeding and survival of C. mac
ulata and C. carnea (Table 4). Significant linear relationships (P < 0.05) between dose
and feeding were found for all insecticides and species tested with the exception of C.
maculata larvae exposed to eggs treated with rotenone (r = 0.34) and esfenvalerate
combined with PBO (r = 0.26). Esfenvalerate combined with PBO and PBO alone had
the greatest effect on feeding by C. maculata adults (b = 2.62 and 2.84, respectively),
whereas rotenone combined with PBO had the greatest effect on C. maculata larvae
(b = 2.15), followed by oxamyl alone (b = 1.93). Feeding by C. carnea larvae was most
affected by esfenvalerate in combination with PBO and rotenone in combination with
PBO (b = 1.69 and 1.63, respectively). Overall, survival of larvae for both species was
most affected by esfenvalerate combined with PBO. C. maculata adult survival was
most affected by oxamyl (b -4.90). PBO alone, however, had little impact on the sur
vival of C. maculata and C. carnea larvae but greatly reduced the survival of C. mac
ulata adults (b -4.79). Egg mortality, as the result of insecticide treatments, may in
part explain the decreased feeding levels and subsequent reduced survival observed.
Insecticide repellency might also account for the drop in egg consumption. Finally,
mortality of individuals from ingestion of surface residues could be responsible for re
duced feeding. Adverse effects from the ingestion of treated food items have been re
ported for other insecticides. Singh & Varma (1986) found reductions in C. carnea
survival from ingestion of Corcyra cephalonica Stainton eggs treated with several in
secticides including, endosulfan, carbaryl, and cypermethrin. Giroux et al. (1994)
demonstrated that ingestion of Colorado potato beetle eggs by C. maculata was re
duced when eggs were treated with Bacillus thuringiensis var. san diego. Reduced
survival due to ingestion of treated food material has also been reported for other coc
cinellids such as Hippodamia convergens Gu6rin-Meneville (Hurej & Dutcher 1994).
Our results show that C. maculata adults and larvae, and C. carnea larvae are sus
ceptible to chemical insecticides commonly used to control Colorado potato beetle.
This finding is important in terms of the design of a pest management program. It
suggests that changing the types of insecticides applied may allow predator survival
in fields thereby helping to reduce pest populations.


We thank M. E. Balzer and the undergraduate students who assisted with this
project. This study was supported by USDA agreement TPSU-RU-3361-527, CSRS
NE-NAPIAP Special Projects. New Jersey Agricultural Experiment Station Publica
tion Number D-08245 17-96, supported by State funds and U.S. Hatch Act.


ABBOTT, W. S. 1925. A method of computing the effectiveness of an insecticide. J.
Econ. Entomol. 18: 265-267.
COATES, S. A., J. R. COATES, AND C. R. ELLIS. 1979. Selective toxicity of three syn
thetic pyrethroids to eight Coccinellids, a Eulophid parasitoid, and two pest
Chrysomelids. Environ. Entomol. 8: 720-722.

March, 1997

Hamilton & Lashomb: Insecticides and CBP Predators 23

COTTY, S., AND J. LASHOMB. 1982. Vegetative growth and yield response of eggplant
to varying first generation Colorado potato beetle densities. J. New York Ento
mol. Soc. 90: 220-228.
FORGASH, A. J. 1985. Insecticide resistance in the Colorado potato beetle, p. 33-53. in
D. N. Ferro and R. H. Voss [eds.], Proceedings of the symposium on the Colorado
potato beetle. XVIIth International Congress of Entomol. Res. Bull. #704,
Mass. Agric. Expt. Sta., Amherst.
GIROUX, S., J-C. COTE, C. VINCENT, P. MARTEL, AND D. CODERRE. 1994. Bacteriologi
cal insecticide M-One effects on predation efficiency and mortality of adult Co
leomegilla maculata lengi (Coleoptera: Coccinellidae). J. Econ. Entomol. 87: 39
GRAFTON-CARDWELL, E. E, AND M. A. HOY. 1985. Short-term Effects of permethrin
and fenvalerate on oviposition by Chrysoperla carnea (Neuroptera: Chrysopi
dae). J. Econ. Entomol. 78: 955-959.
GRAFTON-CARDWELL, E. E., AND M. A. HOY. 1986. Genetic improvement of common
green lacewing, Chrysoperla carnea (Neuroptera: Chrysopidae): Selection for
carbaryl resistance. Environ. Entomol. 15: 1130-1136.
HAMILTON, G. C. 1995. A comparison of eggplant grown under conventional and bio
logical control intensive pest management conditions in New Jersey. Rutgers
Cooperative Extension Bull. E196. 12 pp.
HUREJ, M., AND J. D. DUTCHER 1994. Indirect effect of insecticides on convergent lady
beetle (Coleoptera: Coccinellidae) in pecan orchards. J. Econ. Entomol. 87:
IHAAYA, I., AND J. E. CASIDA. 1981. Pyrethroid esterases may contribute to natural
pyrethroid tolerance of larvae of the common green lacewing. Environ. Ento
mol. 10: 681-684.
LASHOMB, J. 1989. Use of biological control measures in the intensive management of
insect pests in New Jersey. Amer. J. Alternative Agric. 3: 7783.
LECRONE, S., AND Z. SMILOWITZ. 1980. Selective toxicity of pirimicarb, carbaryl, and
methamidophos to green peach aphid, Myzus persicae (Sultzer), Coleomegilla
maculata lengi (Timberlake), and Chrysopa oculata Say. Environ. Entomol. 9:
PLAPP, F. W. JR., AND D. L. BULL. 1978. Toxicity and selectivity of some insecticides
to Chrysopa carnea, a predator of the tobacco budworm. Environ. Entomol. 7:
PREE, D. J., D. E. ARCHIBALD, AND R. K. MORRISON. 1989. Resistance to insecticides
in the common green lacewing Chrysoperla carnea (Neuroptera: Chrysopidae)
in southern Ontario. J. Econ. Entomol. 82:29-34.
ROBERTSON, J. L., AND H. K. PREISLER 1992. Pesticide bioassays with arthropods.
CRC Press. Boca Raton, FL. 127 pp.
ROGER, C., D. CODERRE, AND C. VINCENT. 1994. Mortality and predation efficiency of
Coleomegilla maculata lengi (Coleoptera: Coccinellidae) following pesticide ap
plications. J. Econ. Entomol. 87: 583-588.
SAS INSTITUTE. 1985. SAS/STAT Guide for Personal Computers -Version 6. Cary,
N.C. 378 pp.
SINGH, P. P., AND G. C. VARMA. 1986. Comparative toxicities of some insecticides to
Chrysoperla carnea (Chrysopidae: Neuroptera) and Trichogramma brasiliensis
(Trichogrammatidae: Hymenoptera), two arthropod natural enemies of cotton
pests. Agric., Ecosys. and Environ. 15: 2330.
SHOUR, M. H., AND L. A. CROWDER 1980. Effects of pyrethroid insecticides on the com
mon green lacewing. J. Econ. Entomol. 73: 306-309.
SNEDECOR, G. W., AND W. G. COCHRAN. 1978. Statistical Methods. Iowa University
Press. Ames, IA. 593 pp.

Florida Entomologist 80(1)


Entomology and Nematology Department
University of Florida
Gainesville, FL 32611


Encarsia polaszekiEvans n. sp., reared from the Bemisia tabaci complex from Bra
zil, is described and illustrated.

Key Words: Sweetpotato whitefly, cotton whitefly, silverleaf whitefly, Bemisia, Encar
sia, biological control


Se describe e ilustra Encarsia polaszeki Evans, criada del complejo de Bemisia ta
baci de Brasil.

The sweetpotato whitefly (SPWF), Bemisia tabaci (Gennadius), was described
from tobacco in Greece in 1889 and was first reported in the New World (in Florida)
by Quaintance in 1900. B. tabaci is widely distributed throughout most tropical and
subtropical areas of the world. Increasing evidence suggests that there may be several
closely related species and/or biotypes (or strains or races) of B. tabaci occurring in
various parts of the world. The B. tabaci species complex is composed of three closely
related or sibling species, namely, B. tabaci, B. argentifoliiPerring and Bellows and B.
poinsettia Hempel, and their biotypes. Perring et al. (1993) estimated the damage
caused by the silverleaf whitefly (SLWF), B. argentifolia, to U.S. agriculture at over a
half a billion dollars.
Encarsia (Hymenoptera: Aphelinidae) species are among the most common and ef-
fective parasitoids of whiteflies and have been used successfully in biological control
programs aimed at several different pests species. The greenhouse whitefly, (Trialeu
rodes vaporariorum (Westwood)), citrus blackfly (Aleurocanthus woglumi Ashby) and
citrus whitefly (Dialeurodes citri (Ashmead)), have been brought under biological con
trol in most areas of the world primarily by Encarsia formosa Gahan, E. opulenta (Sil
vestri) and E. lahorensis (Howard), respectively.
The search for natural enemies of the SPWF has been focused in the Orient and
Middle East region, which was believed to be the native home of this pest (Mound
1963); (Lopez-Avila 1986). Gill (1992) provided evidence indicating that the SPWF (or
SLWF) may have originated in the New World and suggested that the search for nat
ural enemies of the SPWF species be focused on the Neotropics. Encarsia polaszeki
was reared from the B. tabaci complex in abundant numbers at the two sites where it
was collected and may be an effective natural enemy of the SPWF or its relatives in
other areas of the world.

March, 1997

Evans: Encarsia polaszeki n. sp.

Encarsia polaszeki Evans, sp. nov. (Figs. 1 7)


Length: Range = 0.45-0.55 mm, mean = 0.53 mm (based on 10 specimens)
Coloration: (Fig. 1) Body yellowish with head, pronotum, central portion of mesos
cutum, axillae apices, metanotum, base of metasomal tergite I, dorsolateral margins
of tergites I-V, and transverse band on tergite VI, dark brown; tergite VII dusky; eyes
red; legs and antennae pale with F6 slightly darker than other segments; wings hya
Structure: Head -postocellar bars prominent; mandibles tridentate; antenna (Fig.
4) comprised of radicle (R), scape (S), pedicel (P), 3 funicular segments (F l3) and 3 club
segments (F4-6) each having the following length/width ratios: 2.5, 4.6, 1.3, 1.6, 1.8,
2.0, 2.1, 2.0 and 2.5; relative lengths of segments R-F6 to length of Fl: 1.1, 3.3, 1.4, 1.0,
1.3, 1.4, 1.5, 1.4 and 1.8; flagellum with the following number of linear sensilla: F1:0,
F2:1, F3:2, F4:2, F5:3, F6:3, basiconic setae present on F2F6. Mesosoma -mesoscu
tum 1.3 times as wide as long with broad hexagonal sculpturing and 2 pairs of slender
setae; each parapsis with 2 setae; each axilla with 1 short seta, scutellum with 2 pairs
of setae, Scl not reaching base of Sc2, distance between placoid sensillae 2.5-3.0 times
the diameter of 1 sensillum; endophragma reaching base of metasomal tergite II; tibial
spur of middle leg (Fig. 6) 0.7 times as long as corresponding basitarsus; tarsal formula
5-5-5; fore wing (Fig. 3) almond-shaped, with 5-6 costal setae, 2 basal group setae, 2
submarginal setae, marginal vein with 5 long and stout setae along the anterior mar
gin, 2 large setae at its base and 6-8 smaller setae along its interior, alary fringe about
0.6 times as long as greatest width of disk. Metasoma -dorsum with imbricate lateral
margins on tergites I-V, tergite VI and VII with weak striations; lateral margin of terg
ites II-V with 1 pair of long, slender setae, tergites V and VII with 1 pair of medial se
tae, lateral margin of tergite V with an additional pair of short setae, tergite VII with
2 pairs of long, slender setae; venter with 2 pairs of slender setae between the base of
the metasoma and the ovipositor, ovipositor arising near the center of tergite III, 0.9
times as long as tibia of middle leg, valvulae III broad, 0.4 times as long as ovipositor.

Male (Fig. 7)

Length: Range = 0.560.70 mm, mean = 0.64 mm (based on 10 specimens)
Coloration: Head and mesosoma similar in coloration as that of female except only
the basal quarter of each axilla is pale; metasoma dark brown; wings hyaline.
Structure: Similar to that of female except flagellar segments Fl-F4 subequal in
length and F5 and F6 fused.
Distribution: Brazil.
Host: Bemisia tabaci complex.
Holotype: Female, Brazil: Pernambuco, Olinda, 18 v 1991, F D. Bennett, reared
from Bemisia tabaci complex on Chamaesyce sp. deposited in the USNM. Paratypes:
39 females and 19 males with the same data as holotype. Additional specimens: 5 fe
males and 2 males, Brazil, Pernambuco, Salvador, 22 V 1991, F D. Bennett, reared
from Bemisia tabaci complex on Chamaesyce sp. deposited as follows: U.S. National
Museum, Washington, D.C. (USNM), Natural History Museum, London, UK
(BMNH), Florida State Collection of Arthropods (FSCA), Aligarh Muslim University
(AMU) and G.A. Evans personal collection.
Comments: Encarsia polaszeki Evans is most similar in structure and coloration
to Encarsia brevivalvula Hayat and Encarsia septentrionalis Hayat, which were both

Florida Entomologist 80(1)

Figs. 1-7. Encarsia polaszeki (1 6 female, 7 male): 1) Habitus, with metasoma di
vided left side dorsum, right side venter; 2) head, frontal view; 3) fore wing; 4) an
tenna; 5) hind tibia; 6) middle tibia; 7) antenna.

described from India. E. polaszeki may be distinguished from E. brevivalula by its
more elongate third valvular segment and having only 2 pairs of setae on the mesos
cutum (the third valvular segment of E. brevivalvula is very short and the mesoscu
tum has 4 pairs of setae). E. polaszeki differs from E. septentrionalis by its short F 1
antenna segment which is only slightly longer than wide, and by having the distal
submarginal vein seta as long as the proximal submarginal vein seta (the Fl segment
of E. septentrionalis is approximately 2 times as long as wide, and the distal submar
ginal vein seta is much longer than the proximal submarginal vein seta).
Etymology: Encarsia polaszeki is named in honor and recognition of Dr. Andrew
Polaszek for his contribution to the systematics of the genus Encarsia.


I thank F D. Bennett who collected this species and A. B. Hamon for the identify
cation of the whitefly species and review of this manuscript. Financial support for this
investigation was provided under CSRS Special Grant 89-34135-4581 'Biological Fac

March, 1997

Evans: Encarsia polaszeki n. sp. 27

tors Affecting the Abundance of the Sweetpotato Whitefly in the Caribbean including
Florida'. Florida Agricultural Experiment Station Journal Series No. R-04817.


GILL, R. 1992. A review of the sweetpotato whitefly in southern California. Pan-Pacific
Entomol. 68: 144-152.
LOPEZ-AVILA, A. 1986. Taxonomy and biology, pp. 3-11 in M. J. W. Cock, [ed.] Bemisia
tabaci A literature survey on the cotton whitefly with an annotated bibliogra
phy CAB International Institute of Biological Control, Silwood Park, Ascot,
Berks., UK.
MOUND, L. A. 1963. Host-correlated variation in Bemisia tabaci (Gennadius). Pro
ceedings of the Royal Entomological Society, London, (A) 38: 171-180.
1993. Identification of a whitefly species by genomic and behavioral studies.
Science 259: 7477.

Abou-Setta et al.: Effect of Diet on Proprioseiopsis rotendus


'University of Florida, Institute of Food and Agricultural Sciences
Citrus Research and Education Center
700 Experiment Station Road
Lake Alfred, FL 33850

2Faculty of Agriculture, Mansoura University, Mansoura, Egypt


Proprioseiopsis rotendus (Muma) (Acari: Phytoseiidae) developed and oviposited
when provided with all life stages of Tetranychus urticae Koch (Acari: Tetranychidae),
and pollen of ice plant, Malephora crocea (Jaquin), live oak, Quercus virginiana
Miller, or cattail, Typha latifolia (L.), as food sources under laboratory conditions of 26
+ 1C and 75-85% RH. Developmental times on the different foods were 6.58 0.36,
8.17 + 0.92, 7.29 + 0.51, and 7.41 + 0.89 d (mean SD) for females, and 6.12 0.49,
7.96 + 0.94, 6.68 + 0.72, and 6.75 + 0.60 d for males, respectively When T urticae was
provided as the food source, the highest net reproductive rate (R = 23.69), female lon
gevity (45.7 6.26 d), mean generation time (T = 19.54), intrinsic rate of increase (rm
0.162), and finite rate of increase (e" 1.176) were obtained. Pollen of M. crocea was
the superior food source with Ro 21.73, female longevity 44.1 13.3 d, T 22.57,
rm 0.136, and e" 1.46, followed by Q. virginiana. Cattail pollen was the least fa
vorable food source tested with Ro 15.08, female longevity 56.1 4.83 d, T = 23.96,
r 0.113, and e" = 1.120. The sex ratio was 57 1:43 1 (female:male) for all diets
tested. Male longevity was 47.3 + 6.08 d when fed T urticae compared with 26.9-35.2
d when fed pollen. P rotendus adult females cannibalized newly hatched larvae. The
mean daily ovipositional rate was 1 per d (max. 2) when fed on T urticae or 0.5 per d

Florida Entomologist 80(1)

(max. 1) when fed on cattail pollen. Duration of the oviposition period was 5 times
longer than the generation time (egg to egg) of P rotendus.

Key Words: Biology, food range, developmental time, two-spotted spider mites, ovipo


Proprioseiopsis rotendus (Muma) (Acari: Phytoseiidae) se desarroll6 y ovoposit6
cuando fue alimentado con todos los estadios de Tetranychus urticae Koch (Acari: Te
tranychidae), y polen de Malephora crocera (Jaquin), Quercus virginiana Miller, o
Typha latifolia (L.) en condiciones de laboratorio de 26 1C y 75 85% RH. Los tiem
pos de desarrollo en los diferentes alimentos fueron 6.58 0.36, 8.17 + 0.92, 7.29
0.51, y 7.41 + 0.89 d (media DT) para las hembras, y 6.12 0.49, 7.96 0.94, 6.68
0.72, y 6.75 + 0.60 d para los machos, respectivamente. Cuando T urticae fue sumi
nistrado como alimento, fueron obtenidos los mas altos valores de tasa neta de repro
ducci6n (Ro 23.69), longevidad de la hembra (45.7 6.26), tiempo promedio de
generaci6n (T = 19.54), tasa intrinseca de incremento (rm 0.162) y tasa finita de in
cremento (e" = 1.176). El polen de M. crocera fue la fuente de alimento superior, con
Ro. 21.73, longevidad de la hembra = 44.1 13.3 d, T = 22.57, rm = 0.136, y e" 1.46,
seguido por Q. virginiana. El polen de T latifolia fue la fuente de alimento menos fa
vorable, con Ro 15.08, longevidad de la hembra 56.1 + 4.83 d, T 23.96, rm = 0.113,
y e" 1.120. La relaci6n sexual fue 57 1:43 1 para todas las dietas probadas. La
longevidad del macho fue 47.3 + 6.08 d cuando fue alimentado con T urticae compa
rado con 26.9-35.2 d cuando fue alimentado con polen. Las hembras adults de Pro
prioseiopsis rotendus canibalizaron las larvas reci6n eclosionadas. La tasa media
ovoposicional diaria fue de uno por dia (max. 2) cuando se alimentaron con Tetran
ychus urticae, o 0.5 (max. 1) cuando se alimentaron con polen de Typha latifolia. La
duraci6n del period de ovoposici6n fue 5 veces mas larga que el tiempo de generaci6n
(de huevo a huevo) de P rotendus.

Thirty-eight species of phytoseiid mites have been reported on Florida citrus; how
ever, the biology of only a few have been studied (Muma & Denmark 1970, Abou-Setta
& Childers 1987, 1989, Caceres & Childers 1991, Yue et al. 1994, Fouly et al. 1995).
Proprioseiopsis rotendus (Muma) (Acari: Phytoseiidae) was recorded from Florida cit
rus litter and bark (Muma & Denmark 1970) as well as from a wide range of plants
in Arizona and Pennsylvania (Moraes et al. 1986).
Some phytoseiid mites (especially in the genus Euseius) can use alternative food
sources such as pollen. This phenomenon increases species survival when animal prey
are scarce. Such species are considered low density regulators, density independent,
and have the ability of population increase in advance of their prey in late spring and
early summer (McMurtry & Johnson 1965, Kennett et al. 1979, Abou-Setta &
Childers 1987, Flechtmann & McMurtry 1992). To our knowledge, the biology of this
genus has not been previously studied.
This study was conducted to determine the impact of different diets on the biology
and life table parameters of P rotendus. Two spotted spider mites, Tetranychus urti
cae Koch (Acari: Tetranychidae), and pollen of ice plant, Malephora crocea (Jaquin),
live oak, Quercus virginiana Miller, or cattail, Typha latifolia (L.) were selected for
this purpose.
T urticae is a common phytophagous mite that feeds on more than 150 species of
host plants (Jeppson et al. 1975). It is only an occasional pest on Florida citrus under

March, 1997

Abou-Setta et al.: Effect of Diet on Proprioseiopsis rotendus 29

greenhouse conditions (Childers 1994). M. crocea pollen has been used as a diet for
other phytoseiid mites in studies which were conducted in California and Florida (Mc
Murtry & Johnson 1965, Abou-Setta & Childers 1987, Fouly et al. 1995). T latifolia
and Q. virginiana are annual and perennial plants, respectively, that occur in and
around citrus groves in Florida. Their pollens may be natural food sources for P ro
tendus during flowering.
This study was part of an ongoing effort to understand the biology of natural ene
mies associated with citrus in Florida.


The P rotendus culture was established from individuals collected on lower canopy
leaves in a 'Pineapple' orange grove at Fort Ogden, DeSoto County, Florida in March
1993. The main culture was maintained on plastic arenas similar to ones used by
Swirski et al. (1970). The rearing arena consisted of a water wick and a black painted
plastic surface surrounded with a sticky barrier to prevent mites from escaping. Are
nas were covered with plastic Petri dish bottoms to provide an enclosed environment
about 2 cm high. A small piece of black construction paper was provided as arresting
place and a few non-absorbent cotton fibers were attached to the water wick as egg
deposition sites. The culture was provided with ice plant pollen obtained from the
University of California-Riverside.
The food sources evaluated included all stages of T urticae (on small pieces of Lima
bean leaves), or pollen of M crocea, Q. virginiana, or T latifolia. Pollens were collected
when available in the field and stored in the refrigerator at 5C; small amounts were
added to the arena as needed using a fine 5-0 sable hair brush.
The newly deposited eggs were transferred individually to 2.5 cm diam arenas.
Newly hatched larvae were provided with one of the food sources listed. As new fe
males matured, they were exposed to males from the same food group or from the
main culture. A male was present with each female until her death.
The arenas were held in an environmental chamber at 26 + 1'C with a photoperiod
of 12:12 (L:D). Relative humidity (75-85%) was measured at 1 cm height above the
arena surface using a thermocouple for temperature and relative humidity (Abou
Setta & Childers 1987).
Individual development, survival, and egg deposition were observed daily. Life ta
ble parameters were calculated using a BASIC computer program (Abou-Setta et al.


Behavioral Observations

Rearing of P rotendus was successful using the plastic rearing arena with either
one of the pollens or motile stages of T urticae as the food source. Eggs, larvae,
nymphs and adults of T urticae were consumed by adult P rotendus.
P rotendus eggs were whitish, crystalline, elongate-oval, and with a sticky surface
when newly deposited. Their color changed to light reddish-brown before hatching.
Larvae were whitish and slightly larger than the egg.
The protonymph and deutonymph stages were progressively larger than the larval
stage with developing body color becoming increasingly brownish. Males were smaller
than females and with the same brown coloration. Morphology of this species was con
sidered by Fouly et al. (1994).

Florida Entomologist 80(1)




C- h

a1 C'j -t 0


C !?
La Lb


iq; b ^

March, 1997

Florida Entomologist 80(1)

Mating took place just as a female reached maturity and continued for about 4 h.
Mated females became more spherical as the egg developed compared to unmated fe
males. No multiple matings were observed.
Eggs were deposited on the arena's surface close to the water wick, on the con
struction paper, or on the cotton fibers attached to the water wick. Egg deposition sites
were the same regardless of food source.
P rotendus was easily disturbed in the arena when exposed to bright light after
darkness. Avoidance of light may contribute to their inhabiting the inner or lower can
opy and surface debris on the ground.
Cannibalism was observed in crowded cultures, especially by adults feeding on
newly emerged larvae. Feeding was not observed on any other stage.

Developmental Time and Adult Longevity

Type of diet tested and gender significantly affected individual developmental
time and adult longevity. Mean male developmental time (6.95 1.0 d) and longevity
(34.20 + 9.46 d) were significantly shorter than that for female (7.40 + 0.9d and 48.25
+ 9.5 d respectively).
Both sexes developed significantly slower on ice plant pollen than other diets. Fe
male mean developmental time, 6.58 0.36 d, was the shortest on T urticae (Table 1).
Mean female longevity was significantly longer on cattail pollen than other diets
(56.1 4.8 d) while oviposition was the lowest (Tables 1 & 2).
Mean male longevity was significantly longer on T urticae (47.3 6.1 d) followed
by live oak, cattail, and ice plant pollens (Table 1).

Life Table Parameters

Sex ratio was not affected by food source (Table 2). Survival curves of P rotendus
under laboratory conditions followed a type I pattern in which most eggs developed to
maturity and death occurred gradually after an extended ovipositional period (Fig. 1,
Maximum daily oviposition per female did not exceed 2 eggs per female per day (m.
value of 1.19 expected female progeny per female per day) when fed upon T urticae
(Fig. 1). Average daily oviposition ranged from a maximum of 1 egg per day when fed
T urticae to a low of 0.5 egg per day on cattail pollen.


Food Source (%) F' Ro' T (d)' rm' e""1

Tetranychus urticae 57 41.46 23.69 19.54 0.162 1.176
Malephora crocea 56 39.11 21.73 22.57 0.136 1.146
Quercus virginiana 56 29.16 16.33 21.41 0.130 1.139
Typha latifolia 58 26.05 15.08 23.96 0.113 1.120

'F mean total fecundity (eggs per female); R., net reproductive rate; T, mean generation time; r_ intrinsic rate
of increase; e", finite rate of increase.

March, 1997

Florida Entomologist 80(1)

1. 1. uricae 1.0

1.0 M LX- 0.5


0.0 i 0.0
S1. B 5M.C roce 1.0
2 Lx
S1.0 Mx 0.5
0i .5
0.0 I I I I I 0.0 I
E 00 0.0 0
1 C Q.virginiana 1.0
C 1.5

0 1.0 Mx 0.5 0)
0. 0V5

E 0.0 I 0.0
LL D T. latifolia 1.0

1.0 -
Mx 0.5

0.0 0.0
0 10 20 30 40 50 60 70 80
Mite age (days)

Fig. 1. Age-specific fecundity (M) and survivorship (L) of Proprioseiopsis rotendus
feeding on different sources of food at 26'C. (A) Tetranychus urticae (mite); (B) Male
phora crocea (pollen); (C) Quercus virginiana (pollen); (D) Typha latifolia (pollen).

Feeding on T urticae resulted in the highest mean total fecundity of 41.46 eggs per
female, net reproductive rate (R) value of 23.69 expected females per female, and
shortest mean generation time (T) of 19.54 d. Feeding on pollen of different plant spe
cies resulted in lower rm values with ice plant providing the maximum rm (Table 2). A
similar ranking for the diets was observed for Typhlodromalus peregrinus (Muma)
(Fouly et al. 1995).

P rotendus developed more slowly than most species of Phytoseiidae previously
studied. This slow development may be a requirement to compensate for lower levels

March, 1997

Abou-Setta et al.: Effect of Diet on Proprioseiopsis rotendus 33

of available food in both litter and the lower tree canopy than more arboreal phy
toseiid species. Other phytoseiid species completed their development from egg to
adult female at a constant temperature of 25-26'C within a range of 4-6 d (Sheriff
1982, Abou-Setta & Childers 1987, 1989, Bonde 1989, Caceres & Childers 1991,Fouly
et al. 1995) compared with 6.58-8.17 d for P rotendus.
The ovipositional period for P rotendus was about 5 times greater than the gener
ation time with a maximum ovipositional rate of 2 eggs per female per day This ratio
did not exceed 3-4 times for other phytoseiid species studied from Florida citrus (Eu
seius mesembrinus (Dean), Galendromus helveolus (Chant) and T peregrinus) with
maximum oviposition rates of 3-4 eggs per female per day at the same temperature
(Abou-Setta & Childers 1987, Caceres & Childers 1991, Yue et al. 1994, Fouly et al.
1995). The number of eggs produced by a female over a period equal to one or two gen
eration times was responsible for most of the calculated rm value for mites and insects
(Abou-Setta & Childers 1991). The longest survival of P rotendus adult females oc
curred on cattail pollen, the least suitable food source.


We would like to thank Mr. H. A. Denmark, Florida Department of Agriculture and
Consumer Services, Division of Plant Industry, for identifying P rotendus. This study
was partially supported by an Egyptian Peace Fellowship grant. Florida Agricultural
Experiment Station Journal Series No. R-05466.


ABOU-SETTA, M. M., AND C. C. CHILDERS. 1987. Biology of Euseius mesembrinus (Ac
ari: Phytoseiidae): life tables on ice plant pollen at different temperatures with
notes on behavior and food range. Exp. Appl. Acarol. 3: 123-130.
ABOU-SETTA, M. M., AND C. C. CHILDERS. 1989. Biology of Euseius mesembrinus (Ac
ari: Phytoseiidae): life tables and feeding behavior on tetranychid mites on cit
rus. Environ. Entomol. 18: 665-669.
ABOU-SETTA, M. M., AND C. C. CHILDERS. 1991. Intrinsic rate of increase over differ
ent generation time intervals of insects and mite species with overlapping gen
erations. Ann. Entomol. Soc. America 84: 517-521.
ABOU-SETTA, M. M., R. W. SORRELL, AND C. C. CHILDERS. 1986. Life 48: a BASIC com
puter program to calculate life table parameters for an insect or mite species.
Florida Entomol. 69: 690-697.
BONDE, J. 1989. Biological studies including population growth parameters of the
predatory mite Amblyseius barker (Acari: Phytoseiidae) at 25'C in the labor
tory. Entomophaga 34: 275 287.
CACERES, S., AND C. C. CHILDERS. 1991. Biology and life tables of Galendromus hel
veolus (Acari: Phytoseiidae) on Florida citrus. Environ. Entomol. 20: 224-229.
CHILDERS, C. C. 1994. Biological control of phytophagous mites on Florida citrus uti
lizing predatory arthropods, pp. 255-288 in D. Rosen, F D. Bennett and J. L.
Capinera [eds.]. Pest management in the subtropics-Biological control-A
Florida perspective. Intercept Ltd., Andover, UK.
FLECHTMANN, C. H. W., AND J. A. McMURTRY. 1992. Studies on how phytoseiid mites
feed on spider mites and pollen. Intern. J. Acarol. 18: 157-162.
FOULY, A. H., M. M. ABOU-SETTA, AND C. C. CHILDERS. 1995. Effects of diet on the bi
ology and life tables of Typhlodromalus peregrinus (Acari: Phytoseiidae). Envi
ron. Entomol. 24: 870-874.
FOULY, A. H., H. A. DENMARK, AND C. C. CHILDERS. 1994. Description of the immature
and adult stages of Proprioseiopsis rotendus (Muma) and Properioseiopsis ase
tus (Chant) from Florida (Acari: Phytoseiidae). Intern. J. Acarol. 20(3): 199-207.

34 Florida Entomologist 80(1) March, 1997

JEPPSON, L. R., H. H. KEIFER, AND E. W. BAKER 1975. Mites injurious to economic
plants. Univ. California Press, Berkeley
KENNETT, C. E., D. L. FLAHERTY, AND R. W. HOFFMANN. 1979. Effect of wind-borne
pollens in the population dynamics of Amblyseius hibisci (Acarina: Phytosei
idae). Entomophaga 24: 8398.
MCMURTRY, J. A., AND H. G. JOHNSON. 1965. Some factors influencing the abundance
of the predaceous mite Amblyseius hibisci in southern California (Acarina:
Phytoseiidae). Ann. Entomol. Soc. America 58: 49-56.
MORAES, G. J. DE, J. A. McMURTRY, AND H. A. DENMARK. 1986. A catalog of the mite
family Phytoseiidae: references to taxonomy, synonymy, distribution and habi
tat. Brasilia: EMBRAPAOOT. 353 pp.
MUMA, M. H., AND H. A. DENMARK. 1970. Arthropods of Florida and neighboring land
areas, Vol. 6. Phytoseiidae of Florida. Florida Dept. Agric. and Consumer Ser
vices, Div. Plant Ind., Gainesville.
SHERIFF, A. A. 1982. Comparative studies on the influence of temperature on phy
toseiid mites (Acarina: Phytoseiidae). Ph.D. Thesis, Rutgers, New Jersey
SWIRSKI, E., S. AMITAI, AND N. DORZIA. 1970. Laboratory studies on the feeding hab
its, post-embryonic survival and oviposition of the predaceous mites Ambly
seius chilenensis Dosse and Amblyseius hibisci Chant (Acarina: Phytoseiidae)
on various kinds of food substances. Entomophaga 15: 93-106.
YUE, B., C. C. CHILDERS, AND A. H. FOULY. 1994. A comparison of selected plant pol
lens for rearing Euseius mesembrinus (Acari: Phytoseiidae). Intern. J. Acarol.
20: 103-108.

Florida Entomologist 80(1)


University of Florida
Fort Lauderdale Research & Education Center
3205 College Avenue
Fort Lauderdale, FL 33314, USA


Larvae of Hypsipyla grandella attacked the seed capsules of West Indies mahoga
nies, Swietenia mahagoni Jacquin, in spring (March -April) after the capsules de
hisced and the seeds were exposed, which occurred prior to flushing. One to 5 larvae
occurred per capsule. The seeds apparently were a preferred food source and 50-96%
of the seeds in capsules examined in June were damaged by larvae. Seed capsules
during their period of expansion from early summer to winter were virtually free of
borer attack, and during this period neither hardened-off shoots nor persistent cap
sule cores from previous seasons served as food sources for more than a few larvae.
The hardness of the capsule valves is apparently a factor in preventing penetration by
the larvae. Although the persistence of seeds in the capsules is transitory, and the
availability of capsules more limited and more variable than that of shoots, the seed
capsule contents appeared to be preferred as a food source, as higher percentages of
dehisced seed capsules than new shoots were attacked when both were simulta

March, 1997

Howard & Giblin Davis: Hypsipyla in seed capsules

neously available. The damage by H. grandella to mahogany seeds impacts regener
ation of this tree species.


Las larvas de Hypsipylla grandella atacan a las capsulas de las semillas de la
caoba de las Indias Occidentales, Swietenia mahagoni Jacquin, en la primavera
(marzo-abril) cuando 6stas se abren y las semillas estau expuestas, lo cual ocurre an
tes del brote de nuevas hojas. De una a cinco larvas se encontraron por capsula. Las
semillas aparentemente fueron la fuente de alimento preferida. El 50-96% de las se
milas en las capsulas exminadas enjunio estuvo danado por las larvas. Las capsulas
durante su period de expansion, a comienzos del verano, y hasta el invierno estaban
virtualmente libres del ataque de los barrenadores. Durante este period tanto los
brotes endurecidos como los corazones persistentes de las capsulas sirvieron como ali
mento a de unas pocas larvas. La dureza de las valvas de la capsula es aparentemente
un factor de impedimento a la penetraci6n por las larvas. A pesar de que la persisten
cia de las semillas en las capsulas es transitoria, y la disponibilidad de las capsulas
es mas limitada y mas variable que la de los brotes, el contenido de la capsula de las
semillas parece ser preferido como alimento, debido a que mas porcentaje de capsulas
abiertas que de nuevos brotes fueron atacados cuando ambos estaban simultanea
mente disponibles. El dano causado por H. grandella a las semillas de caoba afecta la
regeneraci6n de este arbol.

Two species of Hypsipyla are important pests of timber trees of the family Meli
aceae. One species, H. grandella (Zeller), known as the mahogany shoot borer, is con
sidered the major pest of mahoganies (Swietenia spp.) and cedros (Cedrelaspp.) at the
nursery and young plantation stage in the American tropics. The larvae kill the apical
shoot, inducing a secondary shoot that results in a crooked stem and excessive side
branching (Dourojeanni Ricordi 1963, Grijpma 1974, Howard & Meerow 1994, Lamb
1966, Yamazaki 1992). Hypsipyla robusta (Moore) plays a similar role on meliaceous
trees in the Eastern Hemisphere tropics (Gray 1972). Research on the biology and
control of Hypsipyla spp. was recently reviewed by Newton et al. (1993).
Most studies of the feeding habits of these insects have focused on injury to shoots,
but both species also have been reported to infest seed capsules of meliaceous hosts
(Beeson 1961, Betancourt 1987, Bruner 1936, Monte 1933, Roberts 1966, Solomon
1995, Tillmanns 1964,Wagner et al. 1991). Monte s (1933) observations indicated that
H. grandella was highly adapted to utilizing seed capsules: larvae that hatched from
eggs laid on green seed capsules of cedros penetrated into the interior of the fruit and
fed on seeds. Before pupating, the mature larva made a hole in the capsule valve by
which it later exited as an adult. Adults of H. grandella reared from larvae infesting
seed capsules were larger than those obtained from shoots (Becker 1976), indicating
the importance of fruits in the development of this insect. Hypsipyla robusta is listed
as a pest of fruits of meliaceous timber trees in West Africa (Wagner et al. 1991). In
northern India and Burma, where H robusta attacks toon (Toona ciliata H. Roemer),
it is known as the toonn fruit and shoot borer (Beeson 1961).
Some observations have been made on the seasonal occurrence of feeding on seed
capsules by Hypsipyla spp. Hochmut & Manso (1975) reported that in Cuba H. gran
della attacks seed capsules of meliaceous trees during the dry period when young
shoots of these hosts are not available. In northern India, the first generation of the
growing season of H. robusta feeds on flowers of T ciliata, the second on seeds in the

Florida Entomologist 80(1)

green fruit capsule, and the remaining generations (third through fifth) in shoots of
the current year (Beeson 1961). In Nigeria, Hypsipyla sp. feeds on flowers of African
mahogany, Khaya ivorensis A. Chev., from September to November, on seed capsules
from November to February, and on shoots during the remainder of the growing sea
son (Roberts 1966).
In southern Florida, the host of H. grandella is the West Indies mahogany, (Swi
etenia mahagoni [Linnaeus] Jacquin), which is native to Florida and is the only rep
resentative of Meliaceae native to the continental United States (Harlow & Harrar
1968,Pennington 1981). Although the insect attacks exotic species of Meliaceae, few of
these are planted in southern Florida. Attacks on shoots of West Indies mahogany by
H. grandella peak in May of each year, coinciding with the spring flush (production of
new leaves and shoots) at the beginning or just before the advent of the wet season
(June -September). During the remainder of the wet season, shoot production is spo
radic and attack by H. grandella diminishes greatly, and is virtually nil during the dry
season (October -May) (Howard 1991). This paper elucidates the seasonal abundance
and feeding damage of Hypsipyla grandella (Lepidoptera: Pyralidae) in seed capsules
of Swietenia mahagoni in southern Florida.


The study was conducted during the period March 1995-January 1996. We identi
fled the immature stages of H. grandella by diagnostic characters illustrated by
Ramirez Sanchez (1964). We collected 20 larvae from different seed capsules and
reared them to the adult stage on excised mahogany shoots with their bases in water
or on mahogany seeds. We compared the specimens of adults with illustrations in
Becker 1976, Grijpma 1974, Ramirez Sanchez 1964, and Roovers 1971. For confirm
tion of the identification, 4 of the specimens of adults were examined and identified as
H. grandella by J. B. Heppner (Florida State Collection of Arthropods, Gainesville).
All trees in this study were 5-20 year old West Indies mahoganies planted on the
Fort Lauderdale Research & Education Center. A total of 338 trees of this species are
planted on this site.
Physical characteristics of the seed capsules and seeds that might influence larval
feeding were observed. The length of time that the winged seeds persisted on the cap
sule and thus remained available as food for larvae was determined. Capsules were
marked with the date that they opened and observed on April 17 and 24 for the num
bers of seeds persisting and the numbers that had been shed as indicated by placental
The thickness of capsule valves were measured at their midpoints with calipers in
April and in October. The diameters and lengths of capsules were measured in July,
October and January. In April and October, the resistance to penetration of seed cap
sules and their contents was determined with a Model RP-3T Missouri Type Rind
Penetrometer (Alan Machine Co., Ames, Iowa 50010). Points of penetration included
seeds, cores, and valves at midpoint and seams between valves. A common bottle cork
(processed bark of Quercus suberL.) was used as a standard.
The first field observations in this study were made on March 29, 1995, prior to the
spring flush. Twenty dehisced capsules, all of which showed larval damage (gallery
plus frass typical of this species), were removed from trees and dissected and exam
ined to confirm the presence of larvae of H. grandella and to determine where the lar
vae had fed.
Between April 4 -May 8, three trees, all about 6 m in height, were selected arbi
trarily among those bearing unopened seed capsules. These trees were observed every
1 3 days to determine the dates of flushing, dehiscence of the capsule, and initial dam

March, 1997

Howard & Giblin Davis: Hypsipyla in seed capsules

age to shoots and/or seed capsules. The dates that the capsules dehisced and that
borer damage was first evident were written on 2 cm x 9 cm aluminum tags fixed to
the peduncles. On May 3 and May 8 the percentage of marked seed capsules and of 50
randomly selected new shoots with borer damage was determined and all marked
seed capsules (n = 26) were dissected and examined for larvae.
From May 1995 -January 1996, West Indies mahogany trees on the Research Cen
ter continued to be examined frequently for evidence of damage to seed capsules or
shoots. In January 1996, a total of 400 capsules of the current season were clipped
from trees and examined in the laboratory for evidence of borer damage. One hundred
of these were examined with a 10 x hand lens for eggs of H. grandella. We had previ
ously observed that cores of a large portion of the seed capsules of West Indies mahog
any persist after valves and seeds have been shed and may remain on the trees for up
to 2 seasons. To determine whether these served as food sources for H. grandella lar
vae in winter, 78 cores of the past summer's and 22 of the previous year's seed capsule
which remained on trees were examined. In addition, the stems and branches of 75
trees < 2 m in height were examined closely for larvae or damage by them.


Seed capsules of West Indies mahogany (Fig. 1) develop from small (5 mm diam)
flowers that bloom in June-July in Florida (Howard et al. 1995), expanding rapidly in
summer. By the end of July the capsules were about half their mature size of about 65
mm x 75 mm, which they reached in January when they began to dihisc. Hypsipyla
grandella rarely penetrated the capsules until they dehisced, after which they readily
attacked them. The relatively thick, hard capsule valves are no doubt a factor in pre
venting most H. grandella larvae from penetrating them to reach the food-rich seeds.
Newly dehisced valves were a mean of 8.9 mm thick (range 7-15 mm, n = 20) mea
sured at midpoint and 5 10 x harder than the cores, as measured by penetrometer
readings. The resistance to penetration of the cores was similar to that of common bot
tle cork, while that of the seeds was lower (Table 1).
In both October and April, sites along the seams at the juncture of valves were gen
erally 1/3 to half as resistant as the valves at midpoint. However, in April about 30%
of the sites along the seams had about the same resistance as seeds, i. e., about 10%
the resistance of tissue at valve midpoints, indicating that the seams were weakening
as the capsules dehisced. However, with the exception noted below, larvae did not uti
lize this potentially easy entry point.
On March 29, 1995, the first day of field observations, about 30 trees had seed cap
sules. Only a small portion of the total capsules had dihisced and all these had H.
grandella damage. Of the 20 of these that were sampled, 11 were infested with larvae
and 2 contained cocoons of this species. There was never more than one late-instar
larva per capsule, but earlier instars occurred up to 5 per capsule. The 9 capsules that
did not contain H. grandella immature stages showed evidence of their damage, but
the larvae had left.
One predehisced mature capsule observed in April had a hole along a seam. The
capsule was dissected and found to contain a larva of H. grandella. In August, a young
capsule 42 x 44 mm had a hole interior to the margin of the valve about 3.6 mm in
diam with frass identifiable as that of H. grandella. The hole was filled with the
gummy exudate characteristic of wounds in mahoganies. Dissection of the capsule re
vealed a late instar of H. grandella.
West Indies mahogany seeds persisting in capsules apparently were a preferred
food source for H. grandella larvae. They usually bored through the layers of seed
wings and, upon reaching the core, attached themselves between the core and the

Florida Entomologist 80(1)

Fig. 1. Parts of seed capsules of Swietenia mahagoni; A) Dehisced capsule with
damage and frass of H. grandella larvae; B) Dehisced insect-free capsule (s, seeds,
which are born on a central core; v, valves); C) Capsule in initial stage of dehiscence;
D) Core after seeds have been shed, with galleries made by H. grandella larvae.

seeds, hollowing out the seeds from their proximal sides so that from the outside of the
capsule the larvae were not visible and the seeds appeared sound. Early and late in
stars were observed feeding on seeds, but only late instars were observed in cores,
suggesting that they fed on seeds first and then bored into cores.
The seeds are susceptible to attack by H. grandella only as long as they persist in
capsules, and this parameter is highly variable. Five of the marked capsules shed
most of their 50-80 seeds in a few days. Others shed seeds more gradually. In 3 cap
sules observed, all of the seeds persisted 18 days after they had dehisced, in spite of
winds on one day estimated at up to 25 kph that buffeted the tree branches and
caused the seeds to flutter. Once larvae began to attack seeds, they enveloped them in
webbing, which prevented them from falling from the capsules.
Most seed capsules on the 3 sample trees dehisced on different days between
March 31 -April 13 and were attacked by H. grandella larvae prior to flushing of these
trees. Flush occurred during the period April 14-19. Hypsipyla grandella larvae ap
parently entered 2 of the capsules the day that these split open as evidenced by frass
issuing from between the seeds. In the other 24 capsules, frass was first seen 7 to 30
days after the capsules opened.
Higher percentages of dehisced seed capsules than new shoots were attacked
when both were simultaneously available. When counts were made on May 3, 70
100% of the dehisced capsules on the 3 sample trees (n = 26), compared to 14 22% of
the new shoots (n = 150), had been attacked. A week later, there was a slight increase
in the percentage of dehisced capsules attacked but no discernible increase in shoots
Capsules (n = 22) dissected on June 5 revealed that larvae of H. grandella had fed
on seeds and excavated galleries in the cores of almost all capsules and were still
present in 81.8% of them. Six to 37 seeds persisted per capsule, but 50-96% of the

March, 1997

Howard & Giblin Davis: Hypsipyla in seed capsules


Plant Part N Mean + SD

Common cork 10 3.7 +0.1

April, dehisced capsules:
Valves, midpoints 25 16.4 + 3.7
Valves, seams 15 9.3 + 4.5
Cores 20 3.9 + 0.6
Seeds 10 1.6 0.3

October, predehisced capsules:
Valves, midpoints 11 13.5 + 2.9
Valves, seams 11 9.7 + 4.3

seeds in different capsules were damaged by larvae, either with large holes or com
pletely hollowed out from the inside. A maximum of 5 larvae were found in one cap
sule, these being of early instars. Most capsules had 1 or 2 larvae. A single mature
(fifth instar) larva was found in each of nine capsules. Two capsules had pupae. Cap
sules (n = 4) dissected on July 31 were similarly damaged, and one contained a single
pupa of H. grandella.
Of the 400 capsules examined in January, a total of 5 capsules had dehisced. Lar
vae of H. grandella were in 4 of the dehisced capsules and in one predehisced capsule,
distributed as follows: Two capsules had apparently been dehisced at least since De
cember and each harbored a fifth instar H. grandella larva. Two dehisced capsules
each harbored 1 early instar of H. grandella. There were no entrance holes in the
valves, indicating that the larvae had entered after dehiscence of the capsule. An
early instar larva was on the outside of a predehisced capsule. In the laboratory, this
larva bored into the valve during the next 4 days, but eventually died. There were 2
capsules with superficial feeding scars that may have been caused by H. grandella
larvae, but larvae were not present. Two capsules had initial entrance holes with
pitch tubes, which may have expelled attacking larvae.
With the exception noted, neither eggs nor borings by larvae were observed in the
predehisced capsules in January. At this time, the previous summer's leaves persisted
on the West Indies mahogany trees and no new leaves or shoots were produced. There
was no evidence of shoot or stem attack by H. grandella.
Also in January 1996, 69.2% of the past summer's and 90.9% of the previous year's
cores had borer damage. Many of the cores of the capsules of the current year con
trained deteriorated silk and frass that had persisted since the spring and one had an
empty pupal case of H. grandella. However, no live immature H. grandella were found
in the cores. The galleries of many of them were occupied by other arthropods, mostly
predators, including Pseudomyrma sp. (Hymenoptera: Formicinae) workers with
brood, other ant species, wasps, and spiders. The persistent cores were dry and dete
riorated and of doubtful value as a food source for H. grandella, which is adapted to
feed on living plant tissue.
None of <2 m trees (n=75) closely inspected in January had H. grandella larvae or
damage to stems.

Florida Entomologist 80(1)


Our observations indicate that H. grandella larvae attack seed capsules in spring
with dehiscence of the capsules and exposure of the seeds, which occurs prior to flush
ing. About 70-100% of the open capsules on different trees are attacked. The larvae
hollow out seeds and penetrate the core. This insect's apparent preference for seed
capsules is consistent with Becker's (1976) observation that larvae reared from cap
sules are larger than those reared from shoots. When the trees flush, < 25% of the
shoots on the same trees are attacked. Larvae rarely penetrate the capsule valves,
probably because of the thickness and hardness relative to that of the seeds and cap
sule cores. Seed capsules of the current year are virtually free of borer attack during
their period of expansion from spring to early winter, as are persistent capsule cores
from previous seasons. We have occasionally found larvae in shoots in late summer,
but they apparently are very scarce to absent from this host plant from midsummer
to the next spring flush. The question of where and under what conditions H gran
della passes this period in Florida remains a gap in our knowledge of their life history.
Gravid females are presumably present and either abundant or efficient enough to
find the few dehisced seed capsules present during January, more than 3 months be
fore an abundant supply of dehisced capsules and new shoots are available in April.
Meliaceous trees other than S. mahagoni are ruled out as alternate hosts, because
they are extremely rare in southern Florida.
Both excised shoots and seeds of West Indies mahogany support the development
of larvae to maturity in the laboratory. Although in the field a higher percentage of
capsules than shoots were attacked and apparently are a richer food source, their
availability is less certain than shoots. Young trees less than 5 years old do not pro
duce seed capsules, and mature trees produce many more shoots than capsules. An
nual production of seed capsules is highly variable. We have routinely observed that
mature West Indies mahoganies produce from 0 to about 50 seed capsules compared
to many hundreds of shoots. A maximum of about 300 capsules was observed on a tree
in 1995.
The results of this study indicate that H. grandella may severely restrict the re
generation of West Indies mahogany in Florida. Where seed production is important,
West Indies mahogany capsules should be protected against H. grandella attack, but
methods have not been investigated.


We thank J. V DeFilippis and Martha Howard for field assistance, Omelio Sosa,
Jr, for lending us the penetrometer, and Kimberly Klock and Thomas Weissling for re
viewing the manuscript. This is Florida Agricultural Experiment Station Journal Se
ries No. R-05083.


BEESON, C. F. C. 1961. The Ecology and Control of the Forest Insects of India and
Neighbouring Countries. Govt. of India, 2nd Edit.
BECKER, V. O. 1976. Microlepidopteros asociados con Carapa, Cedrela y Swietenia en
Costa Rica, pp. 75-101, in, J. L. Whitmore, (ed.) Studies on the shootborer, Hyp
sipyla grandella (Zeller), Lep.: Pyralidae. Vol. II. CATIE Misc. Publ. No. 1., CA
TIE, Turrialba, Costa Rica.
BETANCOURT BARROSO, F. 1987. Silvicultura Especial de Arboles Maderables Tropi
cales. Editorial Cientifica-Tecnica, Ministerio de Cultura, La Habana, Cuba.
427 pp.

March, 1997

Howard & Giblin Davis: Hypsipyla in seed capsules

BRUNER, S. C. 1936. El taladrador (Hypsipyla) y otras plagas del cedro en Cuba. Rev.
Agric. (Havana) 19(2): 73-80.
DOUROJEANNI RICORDI, M. 1963. El barreno de los brotes (Hypsipyla grandella) en ce
dro y caoba. Agronomia (Peru) 30(1): 35-43.
GRAY, B. 1972. Economic tropical forest entomology Ann. Rev. Entomol. 17: 313-354.
GRIJPMA, P. 1974. Contributions to an integrated control programme of Hypsipyla
grandella (Zeller) in Costa Rica. Doctoral Thesis, Wageningen, Netherlands,
Landbouwhogeschool te Wageningen.
GRIJPMA, P., AND G. I. GARA. 1970. Studies of the shootborer, Hypsipyla grandella
Zeller. II. Host preference of the larva. Turrialba 20: 241-247.
HARLOW, W. M., AND E. S. HARRAR 1968. Textbook of Dendrology McGraw-Hill Book
Co., New York, St. Louis, San Francisco, Toronto, London, Sydney. 512 pp.
HOCHMUT, R., AND D. M. MANSO. 1975. Protecci6n contra las plagas forestales in
Cuba. Editorial Cientifica-T6cnica, Ministerio de Cultura, La Habana, Cuba.
290 pp.
HOWARD, F. W. 1991. Seasonal incidence of shoot infestation by mahogany shoot
borer (Lepidoptera: Phycitidae) in Florida. Florida Entomol. 74: 150-151.
HOWARD, F. W., AND A. W. MEEROW 1994. Effect of mahogany shoot borer on growth
of West Indies mahogany in Florida. J. Trop. Forest Sci. 6: 201 203
HOWARD, F. W., S. NAKAHARA, AND D. S. WILLIAMS. 1995. Thysanoptera as apparent
pollinators of West Indies mahogany, Swietenia mahagoni (Meliaceae). Ann.
Sci. For. 52: 283-286.
LAMB, F. B. 1966. Mahogany of Tropical America. Its Ecology and Management. The
University of Michigan Press, Ann Arbor. 220 pp.
MENENDEZ, J. M., AND M DEL C. BERRIOS. 1992. Apuntes sobre modificaciones obser
vadas en la forma de ataque de Hypsipyla grandella (Lepidoptera: Phycitidae).
Revista Forestal Baracoa 22: 4146.
MONTE, 0. 1933. Hypsipyla grandella Zeller, uma praga da silvicultura (Lep. Phyciti
dae). Rev. Entomol. (Rio de Janeiro) 3: 281-285.
The mahogany shoot borer -prospects for control. Forest Ecology and Manage
ment 57: 301-328.
PENNINGTON, T. D. 1981. A monograph of neotropical Meliaceae. The New York Bo
tanical Garden, New York.
RAMIREZ SANCHEZ, J. 1964. Investigaci6n preliminary sobre biologia, ecologia y control
de Hypsipyla grandella Zeller. Bol. Inst. Forest. Latino Americano: 54-77.
ROBERTS, H. 1966. A survey of the important shoot, stem, wood and flower and fruit
boring insects of the Meliaceae in Nigeria. Nigerian Forest Information Bulle
tin (New Series) No. 15: 138.
ROBERTS, H. 1968. An outline of the biology of Hypsipyla robusta Moore, the shoot
borer of Meliaceae of Nigeria, together with brief comments on two stem borers
and on other lepidopteran fruit bore also found in Nigerian Meliaceae. Com-
monwealth Forestry Rev. 47: 225-232.
ROOVERS, M. 1971. Observaciones sobre el ciclo de vida de Hypsipyla grandella
(Zeller) en Barinitas, Venezuela. Bol. Inst. Forest. Latino-Americano de Invest.
yCapac. 38: 146.
SOLOMON, J. D. 1995. Guide to insect borers of North American broadleaf trees and
shrubs. Agric. Handbk. 706, U. S. Department of Agriculture, Forest Service.
735 pp.
TILLMANNS, H. J. 1964. Apuntos bibliograficos sobre Hypsipyla grandella Zeller. Bul.
Inst. Forest. Latino-Americano de Invest. y Capac. 16: 82-92.
WAGNER, M. R., S. K. N. ATUAHENE, AND J. R. COBBINAH. 1991. Forest Entomology in
West Tropical Africa: Forest Insects of Ghana. Series Entomologica 47. Kluwer
Academic Publishers, Dordrecht, London, Boston, 210 pp.
mahogany shoot borer, Hypsipyla grandella Zeller (Lepidoptera: Pyralidae), on
the meliaceous trees in the Peruvian Amazon. Appl. Entomol. Zool. 27: 31 38.

Florida Entomologist 80(1)


Department of Entomology & Nematology
P.O. Box 110620, University of Florida
Gainesville, FL 32611 0620


Metaseiulus occidentalis (Nesbitt) is a phytoseiid mite which is commercially
available as a biological control agent of spider mites. Genetic manipulation of this
phytoseiid species has yielded transgenic strains, but none have been released into
the environment. Previous data suggested that M. occidentalis could not survive the
wet, humid summers in Florida. A non-transgenic strain of M. occidentalis was re
leased into field plots in Gainesville on soybean plants infested with the two-spotted
spider mite, Tetranychus urticae Koch. Populations were monitored from April-Octo
ber 1994, and weather data were gathered at the release site. Permethrin-treated bar
rier rows were monitored to determine if the mites dispersed outside the plots, and
aerial dispersal was monitored with sticky traps. Predator and spider-mite popular
tions repeatedly crashed during the summer months, and population growth was neg
atively correlated with rainfall. CLIMEX, a population growth model which uses
climatic factors to determine whether a given poikilothermic species can colonize and
persist in new geographic areas, also indicated that M. occidentalis cannot persist
through the wet season in Florida, although it may be able to establish and persist
through the fall, winter and spring months.

Key Words: Spider mites, phytoseiid mites, genetic improvement, climate models, bi
logical control, risk assessment


Metaseiulus occidentalis (Nesbitt) es un acaro fitos6ido disponible comercialmente
como agent de control biol6gico de acaros fit6fagos. La manipulaci6n gen6tica de esta
especie de fitos6ido ha producido colonies transg6nicas, pero ninguna ha sido liberada al
ambiente. La informaci6n previa sugiere que M. occidentalis no podria sobrevivir el ve
rano lluvioso y humedo de la Florida. Una colonia no transg6nica de M occidentalis fue
liberada en parcelas de campo en Gainesville en plants de soya infestadas con el acaro
de dos manchas, Tetranychus urticae Koch. Las poblaciones fueron muestreadas desde
abril hasta octubre de 1994; los datos climaticos fueron colectados en el sitio de liberal
ci6n. Se muestrearon las hileras de barrera tratadas con permetrina para determinar si
los acaros se dispersaron fuera de las parcelas, y la dispersion area fue monitoreada con
trampas pegagosas. Las poblaciones del depredador y del fitdfago colapsaron repetida
mente durante los meses de verano; ademas, el creciemiento de la poblaci6n estuvo co
rrelacionado negativamente con la lluvia. CLIMEX, un modelo de crecimiento
poblacional que usa factors climaticos para determinar cuando una especie poikilot6r-
mica puede colonizar y persistir en nuevas areas geograficas, indic6 tambi6n queM. oc
cidentalis no puede persistir durante la estaci6n de lluvias en la Florida, aunque podria
establecerse y persistir durante los meses de otono, invierno y primavera.

March, 1997

McDermott & Hoy: M. occidentalis Persistence

The western predatory mite, Metaseiulus (= Typhlodromus or Galendromus) occi
dentalis (Nesbitt) (Acari: Phytoseiidae), is an obligatory predator and successful bio
logical control agent of spider mites (Tetranychidae) in vineyards and apple and
almond orchards in the western United States (Hoyt 1969, Flaherty & Huffaker 1970,
Hoy 1985a). M. occidentalis is marketed commercially as a biological control agent
and is recognized as having a potentially world-wide role in integrated spider mite
control programs (Hoy 1985a,b).
The biology and bionomics of M occidentalis are well known. Numerous life-table
studies have examined the effects of different temperatures and prey availability on
M. occidentalis (Laing 1969, Tanigoshi et al. 1975, Bruce-Oliver & Hoy 1993). Hoy et
al. (1985a) demonstrated that hungry adult females display an explicit aerial dis
persal behavior in low to moderate wind speeds. Well-fed mites do not show aerial dis
persal behavior, indicating that food availability may be a component in stimulating
aerial dispersal.
Muma & Denmark (1970) do not list M. occidentalis among the species of phy
toseiid mites occurring in Florida, and Denmark (Personal communication) indicated
no subsequent records of M. occidentalis are available to indicate this species has
since established in Florida, despite numerous commercial importations. Hoying &
Croft (1977) examined literature and museum specimens and, aside from one report
from eastern Wisconsin and one specimen from southern Alberta, Canada, found no
reports of M. occidentalis occurring east of the Rocky Mountains.
A number of phytoseiid species, including M. occidentalis, have been genetically
manipulated via artificial selection to produce pesticide-resistant or non-diapausing
strains (Hoy 1992). Genetic manipulation using recombinant DNA techniques could
improve the efficiency of genetic manipulation of biological control agents by reducing
the time required to identify variability upon which to select, and by providing genes
which do not occur naturally in the species. Presnail & Hoy (1992) used a maternal
microinjection technique to transform M. occidentalis with a plasmid containing the
B1galactosidase gene (lac Z construct) from Escherichia coli under the control of the
Drosophila melanogaster Meigen heatshock 70 promoter. Because so much is known
about the biology of M. occidentalis, it is an ideal arthropod to use for evaluating the
risks of releasing transgenic arthropods into the environment.
The U.S. Department of Agriculture, Agricultural Biotechnology Research Advi
sory Committee (ABRAC) (1991) provides guidelines for risk assessment in trans
genic releases. Our tests aim to answer the following questions that are raised in the
ABRAC guidelines: 1) What is the organism's potential to establish itself in the access
sible environment? 2) What is the potential for monitoring and control in the access
ble environment?
If M. occidentalis cannot permanently establish in Florida due to its inability to
survive the unfavorable summer climate, then Florida could be an ideal site for exper
mental transgenic releases. Experimental plots could be maintained throughout the
favorable fall, winter, and spring months, with the summer climate serving as an ad
ditional safe-guard against accidental establishment. This study examines the ability
of non-transgenic M. occidentalis to persist in Florida through the unfavorable wet
summer months in experimental field plots. The plots were also designed to deter
mine whether M. occidentalis can be contained within the experimental plots and
kept from dispersing aerially. In addition, a climatic model is used to determine the
likelihood of M occidentalis establishing and persisting in Florida.

Florida Entomologist 80(1)



Sixty to seventy-five pinto bean seeds were planted (3:2 potting soil to vermiculite
mixture) in eight liter pots. A total of 75 pots were arranged into five plots, with each
plot containing three rows of five pots. Plots were laid out on an east-west axis at a
University of Florida field station in Gainesville, FL. The rows were spaced 60 cm
apart, and the plots were placed 152 cm apart (Fig. 1). Single pots of beans were
placed in line with each of the rows between each of the plots to act as "trap" plants
between the plots. On March 27, 1994 (Julian day 86; all subsequent dates refer di
rectly to days of the Julian calender), when the bean plants had reached the 3 to 5-leaf
stage, the center row of each plot was infested with Tetranychus urticae (Koch) by lay
ing cut foliage containing T urticae atop the uninfested potted plants. As the cut foli
age dried, the T urticae adults transferred to the green foliage. On day 92, a 10-leaf
subsample was taken from each plot to determine the approximate density of T urti-
cae. Three paraffin-coated paper disks containing adult M. occidentalis females were
spaced equally along the center row of each plot to approximate a 20:1 spider mite to
predator ratio.
On day 113, two pots were removed from each row and replaced with two new pots
of bean plants. The foliage from the pots that were removed was cut and laid over the
new plants to allow the predators and T urticae present on the cut foliage to transfer
to the new foliage. From that point, two new pots were cycled into each row every two
weeks, and the pots were rearranged so that the oldest pot was in the center of each
row, flanked by the two newest pots. All new plants were sprayed with carbaryl (1.1
kg a.i./ha) one day before placement into the field to eliminate other predators and
herbivores. The strain of M. occidentalis used (COS) is resistant to carbaryl, sulfur,
and organophosphorus insecticides (Hoy 1984).
Plots were sampled once weekly starting on day 99. Five leaves were sampled from
each plant for a total of 25 leaves per row per plot. Each 25-leaf sample was placed into
a paper bag, chilled and taken to the laboratory, where a mite brushing machine was
used to brush the mites from each 25-leaf sample onto a glass plate. Numbers of all
stages of M. occidentalis and T urticae were counted under a dissecting microscope,
and the mean number of mites per leaf was determined for each row in each plot.
Following a population crash, plants were re-infested with spider mites on day
142, and M. occidentalis was added again on day 148. Sampling resumed on day 155.
Additional T urticae and M. occidentalis were added on day 197, and T urticae only
were added on day 225. Weekly sampling continued through day 281 (October 8,
1994). To determine if mite populations would rebound on their own, a subset of two
pots per plot were removed on day 197 and replaced with new bean plants. The re
moved pots were placed in a new location with one new pot for each two pot subset.
These five new plots were not re-infested, but fresh pots were cycled in each week.


The two outside rows of each plot, as well as the trap plants between the plots were
sprayed with permethrin (0.06 kg a.i./ha) every two weeks starting on day 92. T urti
caeis unaffected by permethrin at this rate, but the M. occidentalis strain used in this
study is highly susceptible to this insecticide. Thus the outside rows of each plot were
designed to act as barrier rows to dispersal by M. occidentalis.
Twenty-two 1.8 m cedar stakes were spaced at 1.5-m intervals around the plot
(Fig. 1). Each stake held three 185 mm x 78 mm plexiglass plates coated with a thin

March, 1997

McDermott & Hoy: M. occidentalis Persistence 45

E Legend
SPlotas infested wvtfh
X X M. occidenta

0 06m o o Unrnfsted trap rows
1 Stakes holding peiglass
X X scqkytrops
0 Trap plants


o 0


o 0

o o





Fig. 1. Schematic diagram of field plot layout showing the center row infested with
M. occidentalis and T urticae, uninfested barrier rows and trap plants, and plexiglass
sticky traps.

layer of gear oil. Plates were suspended on hooks set approximately 165, 110, and 54
cm above the ground. Plates were removed once a week, labeled as to the height above
ground, and geographical axis to the plot, and taken to the lab. Plates from like
heights and axes were placed into trays and soaked in tap water and automatic dish

Florida Entomologist 80(1)

washing detergent to loosen material stuck to the grease. The slurry from the trays
was then filtered through a fine mesh screen, and the contents were examined under
a dissecting microscope to determine if any M. occidentalis were stuck to the plates,
which would indicate that M. occidentalis was dispersing aerially.

Weather Data

Meteorological data was gathered from the site by a Campbell CM-10 datalogger
and weather station (Campbell Scientific, Logan, UT). Temperature, precipitation,
relative humidity, and wind speed and direction were recorded every 10 min. Maxi
mums, minimums and totals for each 24-h period were compiled. Data were down
loaded roughly once a week.

The CLIMEX Model

A computer climate modeling system was used to determine the likelihood of M
occidentalis surviving and establishing in Florida. CLIMEX is a computerized climate
matching system which uses biological data to predict the potential relative abun
dance and distribution of poikilothermic animals in a given geographic area (Sutherst
& Maywald 1985). The CLIMEX model utilizes climatic data from around the world,
along with what is known of the biology and distribution of a given species to deter
mine that species' potential to survive and proliferate in a given environment.
The CLIMEX model calculates an Ecoclimatic Index (EI) which utilizes weekly
temperature, moisture, and daylength indices, and yearly cold, dry, heat, and wet
stress indices. The El, scaled between 0 and 100, is determined by the following equa
El 100(GI)/52x [(1-CS) x (1-DS) x (1 HS) x (1-WS)
where CS, DS, HS, and WS are yearly cold stress, dry stress, heat stress, and wet
stress indices scaled between 0 and 1. GI is the weekly population growth index,
which is the product of the weekly temperature, moisture, and daylength indices.
The derivations of these indices are described in more detail in Sutherst & May
wald (1985).
Optimal and upper and lower threshold temperatures for M. occidentalis popular
tion growth were obtained from the literature (Tanigoshi et al. 1975) and used in the
model. Unknown moisture parameters and threshold indices were then systemati
cally altered until a distribution map approximating the known distribution of M oc
cidentalis in western North America was achieved (Hoying & Croft 1977). The model
then graphed the predicted population growth curves for M. occidentalis populations
in Jacksonville and Tampa, FL. These cities were chosen as they are the two cities
closest to Gainesville that are included in the model's meteorological database.



Both species remained in the plots at relatively stable levels (at a roughly 28:1
prey:predator ratio) through the month of April (Fig. 2). The mean densities of M oc
cidentalis and T urticae for each sampling date crashed at the beginning of May (day
127), corresponding to a storm on day 124 that dumped 80.7 mm of rain in 5 h. Spider
mite populations were reduced four fold, to just over 1.5 T urticae per leaf. Predator

March, 1997

McDermott & Hoy: M. occidentalis Persistence 47

mite densities were cut in half, lowering the prey:predator ratio. The food supply for
M. occidentalis continued to decline over the subsequent 7-d period, and only two
male M. occidentalis found in the entire 125 leaf sample from all five plots on day 141.
Reinfestation of the plots in late May was done with higher densities of both mite
species, although the prey:predator ratio was maintained roughly the same as in the
first infestation. Between days 155 and 162, T urticae populations declined 55%,
while M. occidentalis populations declined 65% (Fig. 2). Sampling on day 162 occurred
in a light rain, and 26 mm of rain fell that afternoon, immediately before the sample
was taken. More T urticae were added to each plot on day 164, resulting in the popu
nation upswings seen for both species in the samples taken on day 169.
Population trends of both species were negatively correlated with increasing rain
fall (df 16, p = 0.0036 [Fig. 3A], df 16, p = 0.0015 [Fig. 3B]). Percent population
change was not calculated for those weeks in which more mites were added to the
plots. The subplots that were started on day 170 showed that neither mite population
would rebound without reinfestation (Table 1). No M occidentalis were found in this
subplot after day 211, although some T urticae persisted at very low densities
through the rest of the sampling dates.


A total of five M. occidentaliswere found on the 66 plexiglass plates designed to de
tect aerial dispersal. The mites were found only on two of the 27 sampling dates (Table
2). Three were found on plates on the north side of the plot at the 43 cm level on day
211, and two were discovered on plates on the day 260 sampling date; one on the north

-0- T. urticae
-*- M. occidentalis (x 10)
Precipitation previouss 7 days)

18 100

16 90

' 14J I 7S

s12 1 70
S- 60

I: I \ 0
| 5 0A

4 20

92 106 120 134 148 162 176 190 204 218 232 246 260 274 288
Sampling Date (Julian)

Fig. 2. Mean densities of M. occidentalis and T urticae and cumulative total rain
fall for each sampling date.

Florida Entomologist 80(1)

100 *
& 80
S so *
S 60 r= 0.421
S 20

- -20

-60 -
S- *

0 10 20 30 40 50 60 70 80 90 100

Total Rainfall (mm.)

g 60
c 20
H 0
- -40
. -60

0 10 20 30 40 50 60 70 80 90 100

Total Rainfall (mm.)
Fig. 3. A. Percent population change of M. occidentalis as a function of total rainfall
in Gainesville, FL from Julian day 99, 1994 to Julian day 281, 1994. B. Percent popu
lation change of T urticae as a function of total rainfall. Population change is calcu
lated weekly, excluding those weeks that mites were released into the plots.

March, 1997

McDermott & Hoy: M. occidentalis Persistence


Sampling Date Mean M. occidentalis Mean T urticae
(Julian) per Leaf per Leaf

204 0 1.25
211 0.008 0.86
218 0 0.27
225 0 0.54
232 0 1.01
239 0 0.66
246 0 0.42
253 0 0.20
260 0 0.49
267 0 0.25
274 0 0.12
281 0 0.44

side of the plots at the 54 cm high level, and one on the east side at the 110 cm high
level. Spider mites were found on the plates on all sample dates. The prevailing wind
direction at the site is from the south-southwest, although easterly winds prevailed
during some periods of stormy weather. Calculating the area along each side and each
end of the plot to a height of 1.8 m (the height of the stakes holding the plexiglass
plates) yields a total area of 61.32 m2 around the periphery of the plot. The 66 plexi
glass plates cover a total area of 0.952 m2, or 1.55% of the peripheral area to a height
of 1.8 m. Extrapolating that the five M. occidentalis collected on the plexiglass plates
represent 1.55% of the total number aerially dispersing within that area, we conclude
that aerial dispersal of M. occidentalis involved several hundred females. Although
this extrapolation may statistically seem of little value, Hoy et al. (1984) used the same
type of extrapolation in a California almond orchard to estimate that the numbers of
dispersing mites could be in the millions over the same time interval. Thus, the plot
management scheme adopted appears to be useful in reducing rates of aerial dispersal.


Date Number of Height Wind Direction
(Julian) M. occidentalis (cm) Axis ()

211 3 110 N 187
260 1 54 N 97
260 1 110 E 97

Florida Entomologist 80(1)

Four living and two dead adult female M. occidentalis were found on the per
methrin-treated trap crops and barrier rows over the course of the experiment. All
represented single individuals found on different dates (days 176, 190, 197, 225, 232,
267). No M. occidentalis eggs were found on the barrier rows, suggesting that a pop
ulation had failed to develop there despite the presence of prey. Spider mites had no
difficulty dispersing from the infested center row to the outer barrier rows, and mean
spider mites per leaf ranged from 0.2 to 19.4 per leaf, so sufficient prey was available
to sustain M. occidentalis if they had dispersed there.

The CLIMEX Model

The upper portion of each graph shows 30 year average monthly temperatures and
precipitation from the CLIMEX meteorological database (Figs. 4A, 4B). The bottom
portion of each graph shows the population growth index for M. occidentalis. The line
labeled "GI" indicates predicted population growth during the year. Population
growth is maximized where GI = TI. Figures 4A and 4B indicate that M. occidentalis
may not enter a winter diapause in much of Florida, and that it can indeed survive the
drier and cooler spring, fall, and winter months. Both graphs indicate that M. occiden
talis populations should decrease to zero in the summer months (July, August, Sep
tember). The model predicts that populations start to crash earlier in Tampa (mid
June), than in Jacksonville (early July). This could be expected since Tampa has a
higher average rainfall for the month of June. The model also predicts that M. occi
dentalis populations could establish earlier in the fall in Tampa (late September) than
in Jacksonville (mid-October). This is because September is the wettest month of the
year for Jacksonville, while July and August are the wettest months in Tampa. Since
M. occidentalis does not have a summer estivation to carry it through the stressful
months of July, August, and September, the Ecoclimatic Index for both cities is zero,
indicating that M. occidentalis will not permanently establish in Florida.


Both the CLIMEX model and the field plot data suggest that M. occidentalis pop
ulations will not survive the summer months in Florida without reintroductions. The
CLIMEX model indicates that M. occidentalis could establish in Tampa and Jackson
ville during the drier fall, winter and spring months, but persistence in Gainesville is
only suggested from the experimental data during April. Populations of both M. occi
dentalis and T urticaewere negatively impacted by high rainfall (Fig. 2). According to
Sutherst & Maywald (1985), the three most important aspects of the climate in deter
mining distribution and abundance of animals are temperature, moisture, and for
some species, daylength. Field & Hoy (1986) showed that M occidentalis larvae do not
mature well at high relative humidities and that egg hatch is inhibited. Herne (1968)
found that immersion of the European red mite, Panonychus ulmi (Koch), arrested
feeding, oviposition, and molting activities. Klubertanz et al. (1990) suggested that
wetted leaf canopy may temporarily retard spider mite population growth. Akinlosotu
(1982) and Yaninek et al. (1987, 1996) found that in the absence of significant preda
tors, weather was the greatest limiting factor in cassava green mite (CGM), Monony
chellus tanajoa (Bondar), populations in Africa. CGM populations were highest
during the dry season and were lowest during the wet season, when precipitation ex
ceeds evaporation (Yaninek et al. 1987). What we observed in this experiment may be
a combination of both direct mortality from the rainfall and population decline from
the wetted canopy Most of Florida's summer rains occur in the late afternoon and

March, 1997

McDermott & Hoy: M. occidentalis Persistence

IA M J I J AS 0 N 1 I

GI = 56 El = 0


G.- TI
-- rIi TI

J Ir M A M I JI_

A S o N D

GI = 48 El = 0

Fig. 4. A. Predicted population growth curves for M. occidentalis based on Tampa,
Florida meteorological data from the CLIMEX model. B. Predicted population growth
curves for M. occidentalis based on Jacksonville, Florida meteorological data from the
CLIMEX model. Upper portion of graphs shows average monthly temperatures (C,
line) and average monthly rainfall (mm, bars). Bottom portion of graphs depicts Tem
perature Index (TI) and Growth Index (GI) of M. occidentalis. Population growth is
maximized where GI TI. An Ecoclimatic Index (EI) of 0 indicates climatic conditions
are not favorable for permanent survival of M. occidentalis.

200 --


0.0 T -- J--

Florida Entomologist 80(1)

early evening hours and are followed by high nighttime relative humidities peaking
at >95% between 3:00 and 6:00 am. A combination of rain and dew can keep the can
opy wet for almost all of the evening and nighttime hours.
Results from this study indicate that the permethrin-treated barrier rows did pro
vide an effective barrier to ambulatory dispersal of M. occidentalis, although some
aerial dispersal did occur. While we recognize that this experimental design may not
be optimal from some standpoints, it is very pragmatic for risk assessment studies
which require small easily sampled and easily mitigated treatment plots. Although
we detected low rates of aerial dispersal, our sampling method undoubtedly underes
timated the incidence of aerial dispersal. However, all other evidence suggests that
any transgenic M. occidentalis that do disperse will be unlikely to permanently estab
lish and persist in Florida.


We wish to thank Dr. Jon Allen for his assistance with the CLIMEX model. We also
thank Juan Villanueva for his assistance in translating the Resumen. This is Florida
Agricultural Experiment Station Journal Series No. R-04736.


AKINLOSOTU, T. A. 1982. Seasonal trend of green spider mite, Mononychellus tanajoa
population on cassava, Manihot esculenta and its relationship with weather
factors at Moor Plantation. Insect Sci. Appl. 3: 251-254.
BRUCE-OLIVER, S. J., AND M. A. HOY. 1990. Effect of prey stage on life table attributes
of a genetically-manipulated strain of Metaseiulus occidentalis (Nesbitt) (Ac
ari: Phytoseiidae). Exp. Appl. Acarol. 9: 201-217.
FIELD, R. P., AND M. A. HOY. 1986. Evaluation of genetically improved strains of
Metaseiulus occidentalis (Nesbitt) (Acarina: Phytoseiidae) for integrated con
trol of spider mites on roses in greenhouses. Hilgardia 54: 132.
FLAHERTY, D. L., AND C. B. HUFFAKER 1970. Biological control of Pacific mites and
Willamette mites in San Joaquin Valley Vineyards. I. Role of Metaseiulus occi
dentalis. II. Influence of dispersion patterns of Metaseiulus occidentalis. Hil
gardia 40: 267-330.
HEADLEY, J. C., AND M. A. HOY. 1987. Benefit/cost analysis of an integrated mite man
agement program for almonds. J. Econ. Entomol. 80: 555-559.
HERNE, D. H. C. 1968. Some responses of European red mite to immersion in water.
Canadian Entomol. 100: 540-541.
HoY, M. A. 1984. Genetic improvement of a biological control agent: multiple pesticide
resistances and nondiapause in Metaseiulus occidentalis (Nesbitt), pp. 673-679
in D. A. Griffiths, C. A. Bowman [eds.], Acarology VI, Vol. 2. Ellis Horwood,
HoY, M. A. 1985a. Almonds (California: Integrated mite management for California
almond orchards, pp. 299-310 in W. Helle, M. W. Sabelis [eds.], Spider mites,
their biology, natural enemies and control. Vol. II. Elsevier Press, Amsterdam.
HoY, M. A. 1985b. Recent advances in genetics and genetic improvement of the Phy
toseiidae. Annu. Rev. Entomol. 30: 345-370.
HoY, M. A. 1992. Criteria for release of genetically improved phytoseiids: An exami
nation of the risks associated with the release of biological control agents. Exp.
Appl. Acarol. 14: 393-416.
HoY, M. A., AND D. L. FLAHERTY 1975. Diapause induction and duration in vineyard
collected Metaseiulus occidentalis. Environ. Entomol. 4: 262-264.
HoY, M. A., H. E. VAN DE BAAN, J. J. R. GROOT, AND R. P. FIELD. 1984. Aerial move
ments of mites in almonds: Implications for pest management. California Ag
ric. 38(9): 21-23.

March, 1997

McDermott & Hoy: M. occidentalis Persistence

HOYING, S. A., AND B. A. CROFT. 1977. Comparisons between populations of Typhlo
dromus longipilus Nesbitt and T occidentalis Nesbitt: Taxonomy, distribution,
and hybridization. Ann. Entomol. Soc. America 70: 150-159.
HOYT, S. C. 1969. Integrated chemical control of insects and biological control of mites
on apple in Washington. J. Econ. Entomol. 62: 7486.
KLUBERTANZ, T. H., L. P. PEDIGO, AND R. E. CARLSON. 1990. Effects of plant moisture
stress and rainfall on population dynamics of the two-spotted spider mite (Ac
ari: Tetranychidae). Environ. Entomol. 19: 1773-1779.
LAING, J. E. 1969. Life history and life table of Metaseiulus occidentalis. Ann. Ento
mol. Soc. America 62: 978-982.
MUMA, M. H., AND H. A. DENMARK. 1970. Phytoseiidae of Florida. Arthropods of Flor
ida and neighboring land areas. Vol. 6. Florida Department of Agriculture and
Consumer Services. Gainesville, FL. 150 pp.
PRESNAIL, J. P., AND M. A. HOY. 1992. Stable genetic transformation of a beneficial ar
thropod, Metaseiulus occidentalis (Acari: Phytoseiidae) by a microinjection
technique. Proc. Natl. Acad. Sci. USA 89: 7732-7736.
SUTHERST, R. W., AND G. F. MAYWALD. 1985. A computerized system for matching cli
mates in ecology. Agric. Ecosys. Environ. 13: 281-299.
TANIGOSHI, L. K., S. C. HOYT, R. W. BROWN, AND J. A. LOGAN. 1975. Influence of tem
perature on population increase of Metaseiulus occidentalis (Acarina: Phytosei
idae). Ann. Entomol. Soc. America. 68: 979-86.
U.S. DEPARTMENT OF AGRICULTURE. 1991. Part III. Proposed guidelines for research
involving the planned introduction into the environment of organisms with de
liberately modified hereditary traits; Notice. Federal Register Vol. 56 (22), Feb
ruary 1, 1991: 4134-4155.
YANINEK, J. S., H. R. HERRIN, AND A. P. GUTIERREZ. 1987. The biological basis for the
seasonal outbreak of cassava green mites in Africa. Insect Sci. Applic. 8: 861
YANINEK, J. S., H. R. HERRIN, AND A. P. GUTIERREZ. 1996. Dynamics of the cassava
green mite, Mononychellus tanajoa (Bondar)(Acari: Tetranychidae), in Africa:
Seasonal factors affecting phenology and abundance. Environ. Entomol. (in

Florida Entomologist 80(1)


Center for Medical, Agricultural, and Veterinary Entomology
U.S. Department of Agriculture, Agricultural Research Service
P. O. Box 14565, Gainesville, FL 32604, USA

Department of Entomology and Nematology
University of Florida
Gainesville, Florida 32611


Two rows of collard greens (Brassica oleracea var. acephala L.) were planted be
tween two cabbage fields in Bunnell, Flagler County, Florida in spring 1995. More lar
vae of the diamondback moth (DBM), Plutella xylostella (L.), were found on collard
plants than on cabbage plants in the adjacent fields. The parasitism rate of DBM lar
vae collected from the collard plants reached 72% in early May and was higher than
for larvae collected from the cabbage plants in adjacent fields. Parasitoids recovered
from DBM larvae were mainly Diadegma insulare (Cresson). The damage to collard
plants caused by DBM larvae was greater than on cabbage plants. At harvest, there
was no significant difference in damage ratings of cabbage heads sampled near the
middle of the field and damage to heads on rows nearest the collards. The results sug
gest that collard may have potential as a trap crop of DBM in cabbage fields, and that
collard can play an important role in maintenance of the natural enemy, D. insulare.

Key Words: Plutella xylostella, Conura side, Spilochalcis, population regulation, Bras
sica oleracea


Fueron plantadas dos hileras de acelga, Brassica oleracea var. acephala, entire dos
campos de col en Bunnell, condado de Flagler, Florida, en la primavera de 1995. Fue
ron encontradas mas larvas de Plutella xyllostella (L.) en las acelgas que en las coles
de los campos adyacentes. Las tasas de parasitismo de las larvas de P xylostella en la
acelga alcanzaron el 72% a principios de mayo y fueron mas altas que en la col. Los
parasitoides recuperados fueron principalmente Diadegma insulare (Cresson). El
dano causado por P. xylostella fue mayor en las plants de acelga. En el moment de
la cosecha, no hubo diferencia significativa en el dano de las coles muestreadas junto
a las acelgas o en el centro del campo. Los resutados sugieren que la acelga podria te
ner potential como cultivo trampa para P. xylostella en campos de col y puedejugar un
papel important en el mantenimiento de D. insularis.

The diamondback moth (DBM), Plutella xylostella (L.), is the most destructive pest
of cabbage and other crucifers throughout the world. The annual cost for control of
this pest is estimated to be U.S. $1 billion (Talekar & Shelton 1993). This pest typi
cally has been controlled with pesticides (Shelton et al. 1993a). The diamondback

March, 1997

Mitchell et al.: Diamondback Moth in Cabbage and Collard 55

moth, however, has become resistant to synthetic insecticides used against it in many
countries (Shelton et al. 1993a, Talekar & Shelton 1993). In the USA, control failures
have occurred in several states including Florida, Georgia, North Carolina, Texas,
Wisconsin, and New York (Shelton et al. 1993b). Therefore, other control tactics, in
cluding biological control, cultural control and the use of pheromones (McLaughlin et
al. 1994), should be integrated in the management strategy for this pest.
Cultural practices can be efficient and ecologically sound methods for control of
DBM. Successful use of Indian mustard [Brassicajuncea (L.) Czern] as a trap crop for
management of DBM on cabbage has been recorded from India (Srinivasan & Krishna
Moorthy 1992). Intercropping cabbage with garlic or tomato has been reported in Cen
tral America (Andrews et al. 1992), but substantial reduction of DBM infestation in
cabbage has not been reported. In Hawaii, however, interplanting cabbage with tomato
has shown significant reduction of larval density of DBM in cabbage (Bach & Tabash
nik 1990). The objective of this study was to compare DBM densities, damage to cab
bages, and larval parasitism on collard plants and cabbage plants in adjacent fields.


Experimental Location

The cabbage fields used in the study were located in Bunnell, Flagler County, Flor
ida. Cabbage seedlings (Brassica oleracea var. capitataL.) were planted into two adja
cent fields 4 January (field B) and 20 January (field A), 1995, respectively. Two rows of
collard seedlings (Brassica oleracea var. acephala L.) were planted between these two
fields 27 January. Field B (5.26 ha) was on the north side of the collard rows, and field
A (5.06 ha) was on the south side (Fig. 1). Cabbage and collard plants were planted in
rows 0.76 m apart with 0.23 m plant spacing. The length of each row was 275 m.

Insect Sampling

DBM larvae (1st to 4th instars) on cabbage were sampled weekly beginning 17
January for field B and 9 February for field A; and the larvae and cocoons on collard
plants were sampled beginning 13 March. The collards were sampled at 10 different
sites, each 20 m apart along the rows. Six sites in a grid pattern were sampled in each
cabbage field (Fig. 1). Three sites in a row were in each outer third (35 m from the
edge) of the cabbage field, with 3 sites next to the collards and 3 sites at the opposite
side of the field (away from the collards). The number of plants sampled at each site
decreased as their size increased, from 65 cabbage plants the first wk to 13 the last wk
of sampling and from 47 collard plants the first wk to 5 the last wk of sampling.
DBM larvae were brought into the laboratory and held for emergence of parasi
toids and diamondback moths. If nothing emerged, the hosts were dissected (Day
1994) to examine parasitoids. The larvae were reared in 0.26-liter food cups (10 cm
high x 5 cm diam) under laboratory conditions of 21 C, 50 60% RH, and 12:12 [L:D]
photoperiod. A 5 cm diam hole was cut through the center of the cup's lid, and the hole
was coated with PorexTM porous plastics (Porex Technologies, Fairburn, Georgia) for
ventilation. Fresh collard leaves were supplied for food daily.

Cabbage Damage Rating

At harvest, 13 consecutive mature cabbage heads >15.2 cm diam at each site were
individually rated using the rating scale developed by Greene et al. (1969) and modi

Florida Entomologist 80(1)


11 14 11 14

21 Field B 15 21 Field A 15

31 16 31 16

Fig. 1. Schematic of collard plantings and adjacent cabbage fields. Two vertical dot
ted lines between field A and B are the collard planting. Vertical bars in the cabbage
fields indicate sampling sites for DBM immatures and cabbage head damage ratings.
Horizontal bars in field A indicate cross-ratings of the first 12 rows of cabbage heads
from edges of the field.

fled by Leibee et al. (1995). The ratings were: 1) no damage on head or 4 wrapper
leaves; 2) no head damage but minor feeding damage on wrapper leaves; 3) no damage
on head but obvious feeding damage on wrapper leaves; 4) very minor feeding damage
on head, but no feeding through outer head leaves; 5) feeding damage through outer
head leaves; and 6) severe damage to head and wrapper leaves. Leibee et al. (1995)
categorized cabbage heads rated < 3 marketable under normal market conditions.
However, the growers with whom we were working considered only cabbage heads
rated in categories 1 and 2 as acceptable to the market in spring 1995.
Besides rating cabbage at the permanent sample sites in each field, damage rat
ings also were made on cabbage at selected sites along the field edges. Five sampling
sites, 50 m apart, were chosen along each edge of field A (Fig. 1); five cabbage heads
were rated in each of the first 12 rows from the edge at each site. Unfortunately, this
was not done for field B because the cabbage heads were harvested before we could
collect data.

Statistical Analysis

The variation of DBM larval counts, the percentage of parasitism and cabbage rat
ings at the permanent sites between collard and cabbage plants in both fields were an
alyzed using general linear models procedure (GLM), and differences between the
means were tested with least significant difference multiple range test (LSD; SAS In
stitute, 1990). The raw numbers were transformed by log (n + 1) to meet the assump
tions of GLM (Marks 1990) before performance of the analysis. Average damage

March, 1997

Mitchell et al.: Diamondback Moth in Cabbage and Collard 57

ratings of cabbage along edges of field A were analyzed with an independent student
t test for each of the 12 rows.


DBM Larval Abundance
Numbers of DBM larvae per collard plant were inconsequential until mid-March
and then increased rapidly to a peak in late April (Fig. 2). Numbers of DBM cocoons
per collard plant showed a trend similar to that observed for DBM larvae (Fig. 2).
Densities of DBM larvae on collard plants on each collection date were greater
than densities on cabbage plants in fields A and B (Fig. 3). Mean numbers of DBM lar
vae per plant from 13 March to 10 April were significantly higher on collard plants
than on cabbage plants in fields A and B (13 March, F= 4.12; df = 2,19; P< 0.05; 20
March, F= 7.98; df= 2,19; P< 0.01; 28 March, F= 13.75; df = 2,19; P< 0.01; April 3,
F= 38.43;df 2,19;P< 0.01; 10 April, F= 16.68;df 2,19; P< 0.01), but no significant
differences were shown between the two cabbage fields (P> 0.05). These results sug
gest that the collards were more attractive than cabbage to gravid DBM females.


The parasitism rates of DBM larvae from collard showed an increase from 3.2% on
13 March to 72% on 1 May (Fig. 2). By contrast, parasitism rates of DBM larvae on
cabbage remained very low throughout the season, even at harvest (Fig. 4). The per


M 150

C- 120
o 90

M 60
d 30

13 20 28 3 10 17
Mar. Apr.
Sampling Dates 1995

24 1

Fig. 2. Average numbers of diamondback moth larvae and pupae per collard plant
and the larval parasitism (% + SD) per site by D. insulare.

Florida Entomologist 80(1)

n- --- collards
CO 0- cabbage field A
S6 --- cabbage field B



17 26 3 9 14 21 28 7 13 20 28 3 10

Jan. Feb. Mar. Apr.
Collection Dates 1995
Fig. 3. Average numbers of diamondback moth larvae per collard plant and per
cabbage plant from adjacent fields; shows that means are significantly different on
the same sampling dates.

cent parasitism of DBM larvae on the collards from 4 and 10 April was significantly
higher than on cabbage (F= 13.35 and 7.54, respectively; df= 2,19; P< 0.01). The dif
ference in DBM larval parasitism in fields A and B was not significant (P > 0.05).
Of 1,812 parasitoids found, 1,683 were reared to adults and 129 were dissected at
the larval, pupal or pharate adult stage. D. insulare was the most abundant parasi
toid (99.5%), and the sex ratio was 1:1.1 : d. No obviously biased sex ratio was found
from each collection throughout the season. Eight Conura (Spilochalcis) side (Walker)
(Hymenoptera: Chalcididae) (0.5%) were reared from DBM cocoons collected in April.
The numbers of parasitoids collected and the densities of DBM larvae were not
correlated (r= 0.3403, df = 14, P > 0.05), but percent parasitism and the DBM larval
densities were correlated (r= 0.7183, df = 14, P < 0.05). There were significant corre
nations (r= 0.8876 and 0.9723, respectively; df = 14; P< 0.01) between the numbers
of the DBM larvae collected at any particular week and the percent parasitism and
the numbers of parasites collected 2 weeks later (i.e., 2-week-lag, Fig. 5).
After the field collections were finished, the collard plants (heavily damaged by
feeding of DBM larvae) were brought into the laboratory, and the cocoons of DBM and
D. insulare were collected and checked for parasitoids. From 607 cocoons collected,
284 parasitoids emerged (46.8%). The parasitoids included 225 D. insulare (79.3%), 39
C. side (13.7%) and 20 unidentified hymenopterous parasitoids (7%). The sex ratio of
D. insularewas 1:1.5 (:d6).

March, 1997

Mitchell et al.: Diamondback Moth in Cabbage and Collard 59


S80 -A- cabbage field B
U) .. o. cabbage field A
+1 60-

E 40-

CO 20


13 20 28 3 10
Mar Apr.
Collection Dates 1995
Fig. 4. Percent parasitism of diamondback moth larvae by D. insulare per site on
collard and cabbage plants in adjacent fields; shows that means are significantly dif
ferent on the same sampling dates.

Cabbage Damage Rating
Percentage of marketable cabbage heads in rows 1-12 did not show significant dif
ferences between the north and south sides (P > 0.05). The cabbage damage ratings
for the plots from inner 1/3 of fields A (1.36 0.55) and B (1.31 0.60) were not sig
nificantly different (F= 0.3088; df 3, 152; P> 0.05) from the outer 1/3 of these fields
(farthest away from the collards, Fig. 1) of field A (1.41 0.62) and B (1.26 0.40).
There also was no significant difference in the percentage of marketable cabbage
heads among those sampling locations (F= 1.2381; df = 3, 8; P> 0.05). This suggests
that DBM populations did not spread from the collards to the adjacent cabbage fields
even though the DBM population reached very high levels in the collards. Leaves of
the collard plants were observed to have much heavier damage by DBM larvae than
cabbages throughout the season.


Compared with the cabbage plants in the adjacent fields, collard plants had
greater DBM larval infestation and suffered greater damage. Therefore, collard may
be evaluated as a trap crop in cabbage fields for control of DBM. Detailed interplant
ing plans may be needed for large area trial as the use of Indian mustard in cabbage
fields (Srinivasan & Krishna Moorthy 1992). Planting collards earlier than cabbage
may help to attract early arrivals of DBM. Planting collards over a larger area in and

Florida Entomologist 80(1)

100 35
parasitoids 0 9
r =0.9723, P < 0.01 30
25 m
E w
60 20

20 4
40 0

20 10

parasitism *

0 -r = 0.8876, P<0.01
0 10 20 30 40 50 60

DBM Larvae per Plant

Fig. 5. Correlation showing the relationship between % parasitism and the num
bers of parasitoids per collard plant and the numbers of DBM larvae per plant. The
parasites were collected 2 weeks later than were the DBM larvae (i.e., 2-week-lag).

around cabbage fields may offer growers a significant level of protection of their cab
bage crop from attack by DBM.
In the study of Harcourt (1957), collards were shown to have greater numbers of
DBM larvae than six other cultivated crucifers (including cabbage), which agrees
with our results.
It is not clear why DBM is more attracted to collards than cabbage. It is reported
that DBM is attracted to crucifers that contain chemical stimulants (Talekar & Shel
ton 1993) for feeding (e.g. glucosides) and oviposition (e.g., sulfur-containing glucosi
nolates). Collards may contain higher levels of those volatile chemicals than does
D. insularehas been recorded from Southern Canada to Venezuela and west to Ha
waii (Fitton & Walker 1992) and is a major parasitoid of DBM in north America (Har
court 1960, Idris & Grafius 1993, Lasota & Kok 1986). High parasitism in collards and
the significant correlations shown in our study between the 2-week-lagged DBM lar
val abundance and parasitism by D. insulare suggest that this species was response
ble for regulating DBM populations in the collards. When the parasitism reached a
certain level (64%, April 24), the population of DBM in collards started to decrease
(Fig. 2). However, D. insulare did not show the same relationship with DBM in the
cabbage fields, possibly because of the low densities of the host.


We appreciate the help ofW. Copeland, N. Doran, R. Furlong, J. Gillett, J. Leach,
E. Lanehart, and J. Rye (CMAVE, USDA-ARS, Gainesville, FL), of G. S. Evans (De

March, 1997

Mitchell et al.: Diamondback Moth in Cabbage and Collard 61

apartment of Entomology and Nematology, University of Florida, Gainesville) for iden
tifying parasitoids, V. Chew and M. Mayer (CMAVE, USDA-ARS, Gainesville, FL) for
analyzing data, and of R. Hawkins, T Turner, R. Mitchell, and Q. Emery (Flagler
County, FL.) for the use of their cabbage crop and land.
This article reports the results of research only. Mention of a proprietary product
does not constitute an endorsement or the recommendation for its use by USDA.

ANDREWS, K. L., R. J. SANCHEZ, AND R. D. CAVE. 1992. Management of diamondback
moth in Central America, pp. 487-497 in Talekar, N. S. [ed.], Management of di
amondback moth and other crucifer pests: Proceedings of the Second Interna
tional Workshop. Shanhua, Taiwan. Asian Vegetable Research and
Development Center.
BACH, C. E., AND B. E. TABASHNIK. 1990. Effects of nonhost plant neighbors on popu
nation densities and parasitism rates of the diamondback moth (Lepidoptera:
Plutellidae). Environ. Entomol. 19: 987-994.
DAY, W. H. 1994. Estimating mortality caused by parasites and diseases of insects:
Comparisons of the dissection and rearing methods. Environ. Entomol. 23: 543
FITTON, M., AND A. WALKER 1992. Hymenopterous parasitoids associated with dia
mondback moth: the taxonomic dilemma, pp. 225-232 in Talekar, N. S. [ed.],
Management of diamondback moth and other crucifer pests: Proceedings of the
Second International Workshop. Shanhua, Taiwan: Asian Vegetable Research
and Development Center.
looper control in Florida: a cooperative program. J. Econ. Entomol. 62: 798-800.
HARCOURT, D. G. 1957. Biology of the diamondback moth, Plutella maculipennis
(Curt.) (Lepidoptera: Plutellidae), in eastern Ontario, II. life-history, behavior,
and host relationships. Canadian Entomol. 89: 554-564.
HARCOURT, D. G. 1960. Biology of the diamondback moth, Plutella maculipennis
(Curt.) (Lepidoptera: Plutellidae), in eastern Ontario, III. natural enemies. Ca
nadian Entomol. 92: 554-564.
IDRIS, A. B., AND E. GRAFIUS. 1993. Field studies of the effect of pesticides on the dia
mondback moth (Lepidoptera: Plutellidae) and parasitism by Diadegma insu
lare (Hymenoptera: Ichneumonidae). J. Econ. Entomol. 86: 1196-1202.
LASOTA, J. A., AND L. T. KOK. 1986. Diadegma insularis (Hymenoptera: Ichneu
monidae) parasitism of the diamondback moth (Lepidoptera: Plutellidae) in
Southwest Virginia. J. Entomol. Sci. 21: 237-242.
LEIBEE, G. L., R. K. JANSSON, G. NUESSLY, AND J. L. TAYLOR 1995. Efficacy of ema
mectin benzoate and Bacillus thuringiensis at controlling diamondback moth
(Lepidoptera: Plutellidae) populations on cabbage in Florida. Florida Entomol.
78: 8296.
MARKS, R. G. 1990. Analyzing Research Data. Robert E. Krieger Publ. Co., Inc. Mal
abar, FL.
MCLAUGHLIN, J. R., E. R. MITCHELL, AND P. KIRSCH. 1994. Mating disruption of dia
mondback moth (Lepidoptera: Plutellidae) in cabbage: Reduction of mating
and suppression of larval populations. J. Econ. Entomol. 87: 1198-1204.
SAS INSTITUTE. 1990. User's Guide: Statistics. Ver. 6, 4th ed. SAS Institute Inc., Cary,
North Carolina.
MAHR, AND S. D. EIGENBRODE. 1993a. Insecticide resistance of diamondback
moth (Lepidoptera: Plutellae) in North America. J. Econ. Entomol. 86: 1119.
PREISLER, W. T. WILSEY, AND R. J. COOLEY. 1993b. Resistance of diamondback
moth (Lepidoptera: Plutellidae) to Bacillus thuringiensis in the field. J. Econ.
Entomol. 86: 697-705.

62 Florida Entomologist 80(1) March, 1997

SRINIVASAN, K., AND P. N. KRISHNA MOORTHY. 1992. The development and adoption
of integrated pest management for major pests of cabbage using Indian Mus
tard as a trap crop, pp. 10-14 in Talekar, N. S. [ed.], Diamondback moth and
other cruciferous pests: Proceedings of the second International Workshop.
Shunhua, Taiwan. Asian Vegetable Research and Development Center.
TALEKAR, N. S., AND A. M. SHELTON. 1993. Biology, ecology, and management of the
diamondback moth. Annu. Rev. Entomol. 38: 275-301.


Florida Entomologist 80(1)


1University of Florida Institute of Food and Agricultural Sciences
Tropical Research and Education Center
Homestead, FL 33031

Dept. of Ecology & Evolutionary Biology
University of Connecticut
Storrs, CT 06269


Cymophyes nesocoris New Species and Xyonysius acticola New Species are de
scribed from the Turks & Caicos Islands, British West Indies. Species of Cymophyes
have previously been known to occur only in the Eastern Hemisphere. The immature
stages are described and the hosts and habitats discussed.


Cymophyes nesocoris Nueva Especie y Xyonysius acticola Nueva Especie son
descritas de las islas Turks y Caicos, Antillas Britamicas. Anteriormente, las species
de Cymophyes eran solo conocidas del Hemispherio Oriental. Son descritos los estados
inmaduros y discutidos los hospedantes y habitats.

During the course of our work on the lygaeid fauna of the West Indies two unusual
new species have been collected on the Turks and Caicos Islands. The most striking
of these is an undescribed species of Cymophyes Fieber, a genus which has not been
known previously to occur in the Western Hemisphere.
We also recognize a new species of the orsilline genus Xyonysius Ashlock & Lattin
whose closest relative appears to be an endemic species from the Galapagos Islands.
All measurements are in millimeters.

March, 1997

Baranowski & Slater: New Species ofLygaeidae

Cymophyes nesocoris Baranowski and Slater, New Species
(Fig. 1)

DESCRIPTION. General coloration stramineous, heavily punctate. Conspicuously
differentiated brown to black punctures as follows: on midline of head, antenniferous
tubercles, first, second, basal half of third antennal segments, midline and lateral
margins of pronotum, a single row on either side of midline of scutellum, a few irreg
ularly spaced on corium, forming a longitudinal vitta through the middle of pro,
meso-, and metapleuron, mesally on meso and metasternum, all of femora and tibiae
dark brown; punctures on rest of body stramineous, concolorous with body surface.
Abdominal terga with a pair of dark brown vittae, composed of dark brown punctures,
merging mesally on last two segments. Wings not reaching end of abdomen.
Head elongate, strongly tapering anteriorly, apex of tylus slightly exceeding distal
end of first antennal segment. Length head 0.70, width 0.60, interocular space 0.38.
Pronotum slightly narrowing from posterior to anterior margin, lateral margins very
slightly sinuate, anterior margin concave, posterior margin straight. Length prono
tum 0.82, width across anterior margin 0.52, width across posterior margin 0.90.
Scutellum slightly elevated along midline, but lacking a definite carina. Length
scutellum 0.40, width 0.42. Length claval commissure 0.36. Distance along midline
from apex clavus to apex corium 0.76. Distance along midline from apex corium to
apex wing membrane 1.06. Forefemora strongly incrassate, armed below with a series
of major and minor spines with the apices darkened. Labium extending between fore
coxae. Length labial segments I 0.29, II 0.29, III 0.14, IV 0.23. Antennae thick, all seg
ments with short hairs, segment IV impunctate, segment I enlarged distally. Length
antenna segments I 0.23, II 0.40, III 0.33, IV 0.38. Total body length 4.70.
TYPES. Holotype. 6 Turks and Caicos Isl., B.W.I., North Caicos Island: Horse
Stable Beach Rd., 22 X-1993 (on Sporobolus domingensis), R. M. & H. V Baranowski. In
U.S. National Museum of Natural History. Paratypes. 1336, 1892, same data as holo
type; 166, 242, same except 26-XI-1994; 3d, 52, same except Mudhole Pond; 3d, 22,
same except 0.5 mi S. Horse Stable Beach, 26 VI-1993; 1086, 1882, same except 27-VI
1993. Middle Caicos Island: 21 192, Julia Williams Point, 24-X-1993, (on Sporobo
lus domingensis), R. M. & H. V Baranowski; Providenciales Island: 4 d, 7 0.5 mi N.
Downtown, 19-X-1993, (on Sporobolis domingensis), R. M. & H. V Baranowski; 124d,
136 Leeward, 19-X-1993 (on Sporobolis domingensis), R. M. & H. V Baranowski;
206, 202, Leeward Hwy, Pole #A139, 20-X-1995, (on Sporobolis domingensis), R. M.
& H. V Baranowski; 1326, 1862, 0.5 mi S. Downtown, 20-X-1993, (on Sporobolis
domingensis), R. M. & H. V Baranowski; 4d, 5 same except 23-XI-1994. In U.S. Na
tional Museum of Natural History, American Museum of Natural History, Natural
History Museum (England), Florida State Collection of Arthropods, R. M. Baranowski
and J. A. Slater collections.
ETYMOLOGY. Referring to the island distribution.
This species will key to C. ochroleuca Fieber in Seidenstucker (1953) (see English
translation in Slater 1955), but it is not actually closely related. Cymophyes ochro
leuca is the most elongate of the species in the nominal subgenus but is considerably
less elongate than is C. nesocoris. The latter has a body at least five and one-fourth
times the maximum width. C. ochroleuca has a maximum length/width ratio not
greater than four and one-fourth times as long as wide. None of the described species
of Cymophyes have conspicuous black punctures on the antennal segments. The genus
Stenophyella does have punctate antenna segments but also has a conspicuously bi
fid apex on the abdomen which is not the case with the new species described here.
Stenophyella has been thought to be confined to Australia but we have examined spec
imens from Thailand, Macao, Vietnam, Papua New Guinea and New Caledonia.

Florida Entomologist 80(1)


Fig. 1. Cymophyes nesocoris Baranowski and Slater, New Species, dorsal view.

March, 1997


Baranowski & Slater: New Species ofLygaeidae

Lindberg (1958) recognized the similarity of Cymophyes and Stenophyella when he
described an elongate species from the Cape Verde Islands as Stenophyella africana.
Slater (1966) noted that Lindberg's species lacked a bifid apex on the abdomen and
transferred S. africana to Cymophyes. Linnavuori (1978) erected the subgenus Afro
phyella in the genus Cymophyes for C. africana because of its extremely elongate body.
Cymophyes (Afrophyella) africana is a very elongate species. It is, in fact, much
more elongate than is C. nesocoris, the length/width ratio being at least 7.5 and some
times over 8, whereas in C. nesocoris the ratio is less than 5.5. It also lacks the black
antenna punctures and is an overall very pale species throughout with at most a
faint trace of darkened punctures as a faint line along the pleural surfaces. In addi
tion to the type locality Linnavuori (1978) reported C. africana from the Sudan, Ethi
opia and Pakistan. We have examined Ethiopian and Pakistan specimens and agree
that they appear to be conspecific.
We treat C. nesocoris in the nominal subgenus Cymophyes despite its more elon
gate body and black antennal puncture.
DISTRIBUTION. Cymophyes nesocoris was collected on the islands of Providen
ciales, Middle Caicos and North Caicos. It was not found on Grand Turk nor on the is
lands of Andros and Long Island in the Bahamas even though the host plant was
present. Subsequent to the completion of this manuscript, Dr. Horatio Grillo of the
Universidad Central de Las Villas, Santa Clara, Cuba, brought to the attention of the
junior author a question he had first raised as early as 1978 concerning the possibility
of an insect similar to Cymophyes occurring on Cuba. Dr. Grillo has been kind enough
to send pictures of specimens from Cuba and also to allow us to include the informa
tion in this paper. His photographs clearly indicate that Cuban material is conspecific
with Cymophyes nesocoris. Thus, not only is the species also present in the Greater
Antilles, but if it is an introduced species, it was established sometime before 1978.
The discovery of a species of this otherwise Eastern Hemisphere genus in the West
Indies, more specifically the Bahama Archipelago, raises questions as to whether we
are dealing with an introduction or an endemic but previously overlooked taxon. The
most probable scenario seems to us to consider C. nesocoris to be an introduced species
from an as yet unknown place in the Eastern Hemisphere. Much of the West African
lygaeid fauna is still poorly known and seems a likely area for investigation. Given the
prevailing east to west trade winds at the latitude of the Turks and Caicos Islands, the
possibility of the species having reached the islands by aerial transport seems higher
than by introduction in commercial or recreational ships or planes. On the other hand
it must be recognized that C. nesocoris is not really extremely closely related to any of
the known species of Cymophyes and does have similarities to species of Stenophyella.
One must thus take into account the former presence of a member of this complex in
the past in the Western Hemisphere. Sailer & Carvalho (1957) described a Miocene
fossil species from the Mojave desert in California as Procymophyes lithax
Thus we face the fascinating question of whether we are dealing with a previously
unknown species from somewhere in the Eastern Hemisphere or a hitherto uncol
elected vicariant species, native to the Western Hemisphere.


Cymophyes nesocoris was collected only on Sporobolus domingensis (Trin.) Kunth.
(Poaceae), a common roadside grass in the Greater Antilles, the Bahama Archipelago
and south Florida. All stages were found in the seedheads during the periods col
elected. Eggs are deposited, typically singly, between the seed and sheath. Nymphs and
adults were observed feeding on the seeds. Other species of grasses were swept at sev
eral sites where S. domingensis was present without collecting C. nesocoris.

Florida Entomologist 80(1)


Fifth instar (in alcohol)

Elongate, slender, stramineous in color. Head and body impunctate. Antennae,
femora and tibiae with brown punctation. Pro meso and metapleuron with a light
brown longitudinal vitta. Wing pads, lateral margins of pronotum and midline of
scutellum pronotum and head brown. Legs light brown, eyes red. Each abdominal
tergite with a pair of small brown spots. Length head 0.68, width 0.63, interocular
space 0.43. Length pronotum 0.63, width 0.90. Length wing pads 1.33. Length abdo
men 3.0. Length labial segments I 0.28, II 0.24, III 0.16, IV 0.22. Length antennal seg
ments I 0.20, II 0.38, III 0.32, IV 0.38. Total body length 4.50.

Fourth instar (in alcohol)

Shape and color similar to preceding instar. Length head 0.60, width 0.48, interoc
ular space 0.34. Length pronotum 0.40, width 0.34. Length wing pads 0.62. Length
abdomen 1.80. Length labial segments I 0.16, II 0.20, III 0.16, IV 0.16. Length anten
nal segments I 0.14, II 0.26, III 0.22, IV 0.32. Total body length 3.10.

Third instar (in alcohol)

Shape and color similar to preceding instar. Length head 0.40, width 0.42, interoc
ular space 0.30. Length pronotum 0.26, width 0.52. Length wing pads 0.22. Length
abdomen 1.40. Length labial segments I 0.16, II 0.24, III 0.12, IV 0.16. Length anten
nal segments I 0.10, II 0.18, III 0.18, IV 0.28. Total body length 2.36.

Second instar (in alcohol)

Shape and color similar to preceding instar. Length head 0.38, width 0.32, interoc
ular space 0.22. Length pronotum 0.18, width 0.40. Length abdomen 1.0. Length la
bial segments I 0.12, II 0.16, III 0.10, IV 0.14. Length antennal segments I 0.06, II
0.12, III 0.12, IV 0.24. Total body length 1.70.

First instar (in alcohol)

Shape and color similar to preceding instar. Length head 0.28, width 0.28, interoc
ular space 0.24. Length pronotum 0.12, width 0.30. Length abdomen 0.56. Length la
bial segments I 0.10, II 0.14, III 0.06, IV 0.14. Length antennal segments I 0.06, II
0.10, III 0.10, IV 0.22. Total Body length 1.10.

Egg (in alcohol)

Elongate, tapering to both ends, operculum flat with 6-10 micropylar projections,
opposite end rounded. Length 0.74, width at middle 0.26, operculum 0.10.

Xyonysius acticola Baranowski and Slater New Species
(Fig. 2)

DESCRIPTION. General coloration brown to griseus; head brown with a pale mid
line vitta extending from base anteriorly to approximately middle of eyes, a black

March, 1997

Baranowski & Slater: New Species ofLygaeidae

vitta on either side of midline extending anteriorly around ocelli to anterior eye mar
gin; antennal tubercles marked with black laterally; ventral surface of head pale with
a short black vitta on either side of labium; pronotum brown, median longitudinal ca
rina of posterior pronotal lobe and humeri pale. Scutellum with a pale vitta extending
from apex to midpoint. Hemelytra mottled brown; membrane with faint brownish
markings. Upper half of pleuron brown, lower half yellowish, acetabula white. Fem
ora brownish with proximal one-third yellow, tibiae and tarsi yellowish. Distal half of
first antennal segment, all of fourth segment brown; proximal half of first, all of sec
ond and third yellowish.
Head nondeclivent, impunctate, tylus almost reaching distal end of first antenna
segment. Length head 1.0, width 1.0, interocular space 0.60. Pronotum uniformly
punctate with a faint medial longitudinal carina; anterior pronotal lobe with a trans
verse impression interrupted by the median longitudinal carina; lateral margins
slightly sinuate, posterior margin slightly concave. Length pronotum 1.0, width 1.6.
Scutellum punctate with three raised ridges, one extending from the base to the mid
point, the other two extending from the lateral margins of the base to the midpoint.
Length scutellum 0.70, width 0.94. Length claval commissure 0.64. Midline distance
apex clavus-apex corium 1.04. Midline distance apex corium-apex membrane 0.96.
Length labial segments I 0.80, II 0.80, III 0.76, IV 0.40. First and fourth antennal seg
ments enlarged, second and third slender. Length antennal segments I 0.36, II 0.78,
III 0.66, IV 0.50. Total body length 5.25.
TYPES. Holotype. 6 North Caicos Isl., Horse Stable Beach, 23-X-93, on Iva im
bricata, (R. M. & H. V Baranowski). In U.S. National Museum of Natural History
(NMNH). Paratypes. 546 282, same data as holotype; d1, SAME except 26-XI-1994;
76 142,Whitby 27-XI-94, on Iva imbricata, (R. M. & H. V Baranowski). Providen-
ciales Isl. 16, Grace Bay, 18-X-1993 (R. M. & H. V Baranowski). In U.S. National
Museum of Natural History, American Museum of Natural History, Natural History
Museum, (England), Florida State Collection of Arthropods, R. M. Baranowski and J.
A. Slater collections.
ETYMOLOGY. Referring to the beach habitat of the host plant.
Xyonysius acticola is readily distinguishable from the other West Indian species of
Xyonysius by virtue of the elongate tylus and the relatively short fourth antennal seg
ment. In X acticola the length of the head measured along the midline from the level
of the anterior margin of the eyes to the apex of the tylus is subequal to, or greater
than, the length of antennal segment four. In both X californicus and X basalis, not
only is the tylus less acuminate, but the fourth antennal segment is relatively much
longer, more than one and one-half times the distance from the front margin of the eye
to the apex of the tylus (1.66 is the lowest ratio in a series measured). The fourth an
tennal segment is also slightly shorter than segment three in X. acticola, but much
longer than segment three in X. californicus and X. basalis.
Xyonysius acticola typically has a complete dark brown annulus on the distal half
of the first antennal segment. Some specimens of X californicus also have this com
plete annulus, but most specimens have only irregular dark markings rather than a
complete distal annulus.
Actually X acticola more closely resembles X naso (Van Duzee) which is endemic
(but widespread) on the Galapagos Islands. Like X acticola, X. naso has the distance
from the anterior margin of the eye subequal to the length of the fourth antennal seg
ment. Both thus have noticeably more elongate acuminate heads than do other spe
cies of Xyonysius.
Xyonysius naso is readily separated from X. acticola by its much longer labium
which extends posteriorly onto abdominal sternum three. In X acticola the labium

Florida Entomologist 80(1)


Fig. 2. Xyonysius acticola Baranowski and Slater, New Species, dorsal view.

March, 1997


Baranowski & Slater: New Species ofLygaeidae

reaches between the metacoxae but not onto the abdominal sternum. This is reflected
in the relatively much longer third labial segment in X naso where labial segment
three is slightly longer than segment two whereas in X. acticola it is shorter.
Xyonysius acticola is a relatively dark, often griseous appearing species with at
least indications of four pale calloused patches across the middle of the pronotum and
the calli cicatrices are never completely black. In the specimens of X naso that we
have examined the color is pale yellow (from Fernandina and Santa Cruz Islands)
without any indication of pale calloused pronotal areas and with completely black
pronotal cicatrices. Ashlock (1972) notes however that X naso is quite variable in
color (not geographically correlated incidentally) so that these color differences, al
though striking may not be definitive.
One of the most striking differences between these two long headed species is the
shape of the bucculae. In X. naso the bucculae are very broad anteriorly, but narrow
quickly and reach posteriorly only to the level of the anterior end of the antenniferous
tubercles. In X acticola, by contrast the bucculae are relatively low anteriorly but
slope gradually and extend much further posteriorly to terminate at about the level
of the anterior margin of the compound eyes.
It is interesting that both of these elongate headed species appear to be restricted
in their use of host plants. Ashlock (1972) reported that X. naso was found breeding
only on species of the endemic composite Scalesia and, as noted below, X. acticola is also
restricted to a species of composite. This is in contrast to those mainland species of the
genus for which biological data is available and which feed on a wide variety of plants.


Adults and nymphs of X acticola were found in the seed heads of Iva imbricata
Walt. (Asteraceae). According to Correll and Correll (1982) this plant is found in the
southeastern United States, the Bahamas and Cuba. Eggs, frequently more than one,
are deposited between the seed and sheath. Adults and nymphs appear to feed only on
the seeds. This plant was found only in the sand dune areas of the islands, a habitat
similar to that of sea oats, Uniola paniculata L.


Fifth instar (in alcohol)

Elongate, oval. Head, thorax, including wing pads a reticulated cream and brown
with a few longitudinal brown irregular vittae. Abdomen a reticulated pink and
cream, scent gland sclerites dark brown. Antennal segment I dark brown with distal
tip cream, segments II and III yellow to tan, segment IV brown, segments I and IV
slightly enlarged, segments II and III uniformly slender. Femora dark brown with
proximal one third and distal tip cream; tibiae cream with distal one third brown;
tarsi brown. Length head 1.10, width 1.0, interocular space 0.62. Length pronotum
0.62, width 1.60. Length wing pads 1.48. Length abdomen 2.55. Length labial seg
ments I 0.70, II 0.70, III 0.72, IV 0.52. Length antennal segments I 0.28, II 0.52, III
0.46, IV 0.52. Total body length 5.22.

Fourth instar (in alcohol)

Similar in shape and color to fifth instar. Length head 0.70, width 0.80, interocular
space 0.54. Length pronotum 0.38, width 1.10. Length wing pads 0.50. Length abdo

Florida Entomologist 80(1)

men 1.84. Length labial segments I 0.44, II 0.54, III 0.44, IV 0.38. Length antenna
segments I 0.20, II 0.32, III 0.30, IV 0.38. Total body length 3.25.

Third instar (in alcohol)

Similar in shape and color to preceding instars. Length head 0.60, width 0.60,in
terocular space 0.40. Length pronotum 0.26, width 0.76. Length wing pads 0.20.
Length abdomen 1.40. Length labial segments I 0.38, II 0.46, III 0.44, IV 0.36. Length
antenna segments I 0.12, II 0.20, III 0.18, IV 0.26. Total body length 2.55.

Second instar (in alcohol)

Similar in shape and color to preceding instars except head straw-colored with two
irregular, longitudinal, tan, vittae on each side of midline. Thorax and abdomen retic
ulated cream and pink; thorax with one irregular, longitudinal, tan vitta on each side
of midline. Length head 0.40, width 0.40, interocular space 0.30. Length pronotum
0.12, width 0.48. Length abdomen 0.78. Length labial segments I 0.30, II 0.36, III
0.30, IV 0.28. Length antennal segments I 0.10, II 0.12, III 0.14, IV 0,22, Total body
length 1.46.

First instar (in alcohol)

Shape more elongate than preceding instars. Head and thorax brown. Abdomen
similar to second instar. Legs colored as in preceding instar, but paler. Antennae light
brown. Length head 0.38, width 0.30, interocular space 0.20. Length pronotum 0.14,
width 0.30. Length abdomen 0.60. Length labial segments I 0.20, II 0.24, III 0.24, IV
0.24. Length antennal segments I 0.08, II 0.10, III 0.10, IV 0.20. Total body length

Egg (in alcohol)

Elongate, straw-colored; length 1.1, width 0.3. Operculum 0.09 in diameter with 8
12 stalked, knobbed micropyles. Opercular end flattened, opposite end rounded.
The genus Xyonysius is confined to the Western Hemisphere. Previously nine spe
cies were recognized ranging from Chile and the Galapagos north to southern Canada
(Slater and Baranowski 1990). Three species, including X acticola are now recognized
from the West Indies.


We thank Dr. Edwin L. Bridges (Curator, Fairchild Tropical Garden Herbarium,
Miami, FL) for the identification of plant material, Dr. Robert W Brooks (Snow Ento
mological Museum, U. Kansas, Lawrence, Kansas) for loan of specimens of Xyonysius
naso, Mrs. Holly Glenn and Mrs. Julieta Brambila, Tropical Research and Education
Center, for reviewing the manuscript and for assistance in labeling, pinning and
curating material.We especially thank Mrs. Helen Baranowski who has accompanied
the senior author and assisted in field collecting on many of the West Indian islands
for many years. Florida Agricultural Experiment Station Journal Series No. R-04878.


ASHLOCK, P. D. 1972. The Lygaeidae of the Galapagos Islands (Hemiptera: Het
eroptera). Proc. California Acad. Sci. 4th Series: 39: 87103.

March, 1997

Baranowski & Slater: New Species ofLygaeidae 71

CORRELL, D. S., AND H. B. CORRELL. 1982. Flora of the Bahama Archipelago. 1692 pp.
J. Cramer.
LINDBERG, H. 1958. Hemiptera Insularum Caboverdensium Systematik, Okologie &
Verbrietung der Heteropteren & Cicadinen der Kapverdischen Inseln. Com-
ment. Biol. Helsingf. 19: 246 pp.
LINNAVOURI, R. 1978. Hemiptera of the Sudan, with remarks on some species of the
adjacent countries 6. Aradidae, Meziridae, Aneuridae, Pyrrhocoridae, Steno
cephalidae, Coreidae, Alydidae, Rhopalidae, Lygaeidae. Act. Zool. Fenn. 153: 1
SAILER, R. I., AND J. C. M. CARVALHO. 1957. Miocene arthropods from the Mojave
Desert California. Order Hemiptera -Suborder Heteroptera, in A. R. Palmer,
Geol. Survey Professional Paper 294-G. 237-280.
SEIDENSTUCKER, G. 1953. Neue Cymophyes-arten aus Syrien und Kasakstan (Hem.
Het., Lygaeidae). Ann. Entomol. Fenn. 19: 168-174.
SLATER, J. A. 1955. A revision of the subfamily Pachygronthinae of the world. (Hemi
ptera: Lygaeidae). Philippine J. Sci. 84: 1160.
SLATER, J. A. 1966. A contribution to our knowledge of the Pachygronthinae
(Hem.:Lyg.). J. Entomol. Soc. Queensland. 5: 51 65.
SLATER, J. A., AND R. M. BARANOWSKI. 1990. Lygaeidae of Florida (Hemiptera:Het
eroptera). Arthropods of Florida and Neighboring Land Areas. Vol. 14, 211 pp.
Florida Dept. Agr. and Consumer Serv. Div. of Plant Ind. Gainesville.


Thomas: Degree-Days for Mexican Fruit Fly


U.S. Department of Agriculture, Agricultural Research Service
Subtropical Agriculture Research Laboratory, 2301 So. International Blvd.
Weslaco, TX 78596


Degree-day accumulations and puparial duration of the Mexican fruit fly, Anas
trepha ludens (Loew), in the field was found to fit closely with a degree-day accumu
lation model developed by Leyva-Vazquez (1988) with laboratory data. Larval
development time was more variable, however, and did not agree well with the labo
ratory based degree-day model. This may have been caused by a tendency of the lar
vae to remain in the fruit beyond the necessary development time and for subsequent
egression to be spread over a period of weeks. Duration of the pre-imaginal stages is
strongly a function of season. The puparial stage may be prolonged up to three months
in the winter or be as brief as three weeks in the summer. There was no evidence of a
winter diapause.

Key Words: Degree-days, population model, citrus pest, diapause, Anastrepha ludens


Se encontr6 que las acumulaciones de grados-dias y la duraci6n del estado pupal
en el campo de la mosca mexicana de las frutas, Anastrepha ludens (Loew), se ajusta

Florida Entomologist 80(1)

ban a un modelo de acumulaci6n de grados-dias desarrollado por Leyva-Vazquez
(1988) a partir de datos de laboratorio. El tiempo de desarrollo larval fue mas variable
y no se ajust6 bien al modelo. Esto pareci6 deberse a la tendencia de las larvas a per
manecer en el fruto mas alla del tiempo de desarrollo necesario, y a que su salida del
fruto ocurre durante un period de semanas. La duraci6n de los estados preimagina
les esta fuertemente en funci6n de la estaci6n. El estado pupal podria prolongarse
hasta tres meses en el invierno, o ser tan breve como de tres semanas en el verano. No
hubo evidencia de diapausa de invierno.

Pre-imaginal development in the Mexican fruit fly, Anastrepha ludens (Loew), has
been studied extensively under laboratory conditions (Darby & Kapp 1933, Baker
1944, Baker et al. 1944, Flitters & Messenger 1965, Celedonio-Hurtado et al. 1988).
Flitter & Messenger (1965) state that development time for this species, egg to adult,
ranges from 40-90 days under "normal" conditions.
Specific knowledge of pest phenology is an essential ingredient of effective pest
management. Models based on population dynamics and the environmental parame
ters which drive them, almost invariably include the effect of temperature on devel
opment time. For example, the appearance of temperate tephritid pests, notably the
apple maggot, Rhagoletis pomonella (Walsh), and the cherry fruit fly, Rhagoletis in-
differens Curran, vary greatly from year to year but can be predicted by the standard
degree-day method. Suppression operations are planned accordingly (Aliniazee 1976,
Laing & Heraty 1984). The potential of a degree-day model for predicting outbreaks
of the Mexican fruit fly, an intermittent but serious pest of citrus along the southern
border of the United States, has long been recognized. The efforts to develop data for
such a model are detailed in the present article and the efficiency of a degree-day
based model of development time is assessed.
Leyva-Vazquez (1988) was the first to determine the degree-day accumulations for
development time for each pre-imaginal stage under laboratory conditions. He re
ported 316 10 degree-days for the puparial stage and 291 + 57 combined degree-days
for the egg and three larval instars. Using linear regression applied to the same data,
the lower threshold of development was estimated to be at 9.4'C.
The purpose of the present investigation was to validate the degree-day calcula
tions of Leyva-Vazquez (1988) obtained from laboratory experiments by determine
tion of the duration and degree-day accumulations of the pre-imaginal stages under
field conditions.


All insects used in these experiments were from laboratory cultures maintained at
the USDA facility in Weslaco, Texas, using the rearing methods described by Spisha
koff & Hernandez-Davila (1968).
Duration of the larval stage was determined by placing infested grapefruit in an
outdoor screened enclosure located in the center of a grove of citrus at the Weslaco
site. For the purposes of this experiment, the duration of the larval stage was mea
sured from the day of oviposition until the day the larvae egressed the fruit. At 2 week
intervals, three fresh grapefruits were placed in a laboratory cage containing 15-d-old
adult Mexican fruit flies and exposed to oviposition for a period of 4 hours. At the end
of this exposure period each grapefruit was placed separately in 5-liter plastic tubs
containing 10 cm of clean sand. The sides and bottoms of the tubs were punctured to

March, 1997

Thomas: Degree-Days for Mexican Fruit Fly

allow drainage. Two of the tubs were then placed in the outdoor enclosure, partly bur
ied in the soil so that the level of the sand was at ground level. The third grapefruit
was held in the laboratory at 25 C as a control. Beginning after 2 weeks, the minimum
time for egg hatch and larval development, the sand in the tubs was sifted daily, ex
cept on weekends, to detect the presence of egressed larvae (larvae found on Monday
were pooled and scored as if found on Saturday). The fruit was left in the tub until 2
weeks after the last larva had egressed. A recording hygrothermograph was main
stained in the enclosure to provide ambient temperature data. This experiment began
in July 1994 and continued through December 1995.
Duration of the puparial stage was determined by sprinkling 100 10-d-old larvae
into each of five 5-liter capacity plastic tubs containing clean sand to a depth of 10 cm.
Before placing the larvae, a small amount of water was sprinkled on the sand and
holes poked in the surface with a narrow rod. The larvae were allowed to inter them
selves in the sand, a process which normally required less than 10 minutes. Four of
the tubs were then transported to the field and placed outdoors. The experiment was
conducted from May 1992 to April 1993 at two sites in the state of Nuevo Leon, Mex
ico, an area in which the Mexican fruit fly is indigenous. One of the sites was a citrus
orchard located near the town of General Teran. The other was a grove of wild citrus,
Sargentia greggiWats., in a mountain canyon 15 km west of the town of Linares. The
experiment was also replicated at the Weslaco site between June 1993 and June 1994,
again in the screened enclosure. Details of the environment at these locations and
data on seasonal survival rates of the immature stages of the Mexican fruit fly have
been described in a separate study (Thomas 1995). Briefly, freezing temperatures are
rare in this region and during this study occurred on only one winter night when the
temperatures reached -1C. The winters and springs are typically dry with most rain
fall in the summer months.
After two weeks of exposure a pyramid-shaped emergence cone, 60 cm2 at the base,
was placed over each of the individual tubs. Emerged adults accumulated in an in
verted glassjar on the top of the cone. These cones were checked daily throughout the
study, except on weekends. A recording hygrothermograph was maintained at each
site to provide ambient temperature data.
Degree-days (D) were calculated using the standard weather bureau formula,
also known as the Means Method (Pruess 1983, Fry 1983),
[Max + Min/2] -base.
This formula was used by Leyva-Vazquez (1988) for the Mexican fruit fly and has
been used successfully for other tephritid pests as well (Aliniazee 1976, Reissig et al,
1979). All temperature values were rounded to the nearest C including the base, for
which a value of 9 C was used. For statistical analysis, the correlations between de
gree-days and development time and day length and development time, were calcu
lated using least squares regression (Sokal & Rohlf 1973). In the regression equation,
day length was represented as the difference, in days, between the oviposition date
and the summer solstice.


Laboratory studies have shown that temperature is a dominant factor determine
ing larval development time. Flitters & Messenger (1965) reported 11-12 days larval
development time at constant 27 C but were able to extend the larval stage to 125
days using a 12 + 10 C temperature regime. In the present field study over the course
of the year, the duration of the larval stage in grapefruit was found to range from as

Florida Entomologist 80(1)

few as 19 days in May to as many as 69 days for an oviposition in mid-November (Ta
ble 1). Although this result would seem to be in accord with the expected effect of sea
sonally prevailing temperatures, temperature alone may not have been the most
dominant factor. There was typically a 1-2 week lag between the first and last larval
egress in each test although presumably these larva were exposed to at least similar
temperatures. In one test, with an oviposition date in mid-February, there was a 19
day spread between the first and last larval egress. Even in the controls, infested fruit
maintained at constant temperature (24'C) in the laboratory, mean larval duration
was 23.1 to 37.3 days, with an overall range from 19 to 54 days. It is noteworthy that
some larvae in May, a warm month, did not egress until 29 days post-oviposition,
while in November, a cool month, the first larva egressed also in 29 days. Accordingly,
the time spent inside the fruit by any particular larva is not necessarily reflective of
development time per se. Under optimal conditions the egg and larva can complete de
velopment in 16 days. It would appear that conditions inside the fruit were not uni
form, or at least, did not induce uniformity in the behavior of the larvae with respect
to egression. Some reports suggest that larvae egress the fruit in response to environ
mental cues, rather than as a conclusion to the completion of development or deple
tion of the food source. McPhail & Bliss (1933) likewise found a range of 18 to 35 days
for the larval stage in mangoes held in the laboratory. In field-collected mangoes (Cu
ernavaca, Mexico) left outdoors (exact date of oviposition unknown) the maximum
crawl off date was 44 days. They noted that egression from mangoes was stimulated
by rainfall, and that this effect could be induced by drumming or vibrating the fruit.


Oviposition No. Range Mode First Modal
Date Pupae (Days) (Days) ('D) (D)

Jan 03 65 42-59 58 392 565
Jan 26 244 39-56 41 339 360
Feb 16 153 33-52 44 338 470
Mar 09 189 26-36 28 312 338
Mar 23 70 25-32 26 316 334
Apr 13 200 21-39 25 327 402
May 04 41 19-29 27 332 473
Jun 14 7 33-33 33 659 659
Jul31 32 23-33 33 381 552
Aug 15 117 22-28 22 420 420
Sep 02 177 20-28 26 346 445
Sep 07 192 22-33 27 388 477
Oct01 231 20-32 28 323 446
Oct 19 9 39-56 41 447 463
Nov 10 66 29-41 32 395 410
Nov 14 25 63-69 63 479 479

Mean D 387 + 87: 456 + 84.

March, 1997

Thomas: Degree-Days for Mexican Fruit Fly

They reported that in the absence of rainfall the larvae typically egressed in the early
morning hours, which would be the time of highest humidity Thus, one might con
clude that in the absence of a specific entrainment, the larvae trickle out of the fruit
over a period of weeks rather than making a synchronized mass exodus. In accordance
with the findings of Baker et al. (1944), there was no evidence of a gender related dif
ference in development time as has been reported for Anastrepha suspense (Loew) by
Sivinski & Calkins (1990). Of the 616 flies which closed on the earliest emergence
date of each replicate in the Mexican field studies, 317 were females and 299 males.
The actual temporal spread in egression was mainly a function of the number of
larvae produced by the fruit. Among the June replicates only one fruit produced lar
vae and in this fruit only seven larvae completed development. In this case, all larvae
egressed on the same day. Evidently the high summer temperatures inhibit survival,
possibly because of dessication of the fruit. Of the twelve replicates between May 24
and August 18, only three produced larvae that egressed and pupariated, while the
control fruit in the laboratory each produced in excess of 100 larvae per fruit. By con
trast, during the preceding springtime replicates, eleven out of twelve fruit produced
larvae. The numbers ranged from 1 to 201 larvae per fruit with a mean of 75 larvae
egressing to pupariate per fruit. Not surprisingly, larger numbers of surviving larvae
produce a wider egression pattern. This was especially obvious in the control fruit
where optimal temperature conditions and a lack of environmental cues triggering
egression resulted in as many as 353 larvae developing in one fruit and a spread of as
much as 32 days between the first and last larval egress.
Under these circumstances, degree-days was not a good predictor of development
time as defined here, time between oviposition and larval egression. Leyva-Vazquez
(1988) determined the mean accumulation of degree-days for egg + larval develop
ment to be 291 + 57'd. In this study the mean accumulation for the first egressing lar
vae in each replicate was 387 + 87'd and for the modal egression date, 456 + 84'd.
Thus, the larval stage was prolonged relative to that which would be determined from
ambient temperature alone and accumulated degree-days was not a good predictor of
larval egress. The coefficient of determination (r2) between degree-days and modal
egress was only 0.565, and for first larval egress only 0.490. Those values were not
much higher than the predictive value of calendar date relative to day length. The co
efficient of determination (r2) for day length vs modal egress was 0.448. Reissing et al.
(1979) also found the degree-day method to be a better predictor of emergence than
mean historical calendar date for the apple maggot.
One well known cause of disparity between degree-day predictions and actual de
velopment time is the Kaufmann or Rate Summation effect (Worner 1992). Often in
the field, development at low temperatures is faster, and development at high temper
atures slower than predicted from laboratory studies. This is especially true of motile,
herbivorous insects. Various explanations have been offered to account for this effect
(Wagner et al. 1984; Hagstrum & Milliken 1991), but it is unlikely that the Kaufmann
effect can be evoked as the cause of the disparity. Firstly, the prolongation of the larval
stage occurred even in those replicates wherein the ambient temperatures were far
from the extremes at which development was retarded in the laboratory (less than
9C or in excess of 31 C). Secondly, it is doubtful that the temperatures inside a grape
fruit resting on the ground would reach the extremes that were experienced in ambi
ent temperatures. Moreover, since the duration of the larval stage in the controls was
also extended beyond the minimum development time with egression spread out over
a period of weeks, it is doubtful that any temperature based effect was the dominant
factor determining the delay in the date of egress.

76 Florida Entomologist 80(1) March, 1997

In contrast with the results from larval duration, the puparial development time
closely paralleled the degree-day predictions from the laboratory studies. Leyva
Vazquez (1988) reported 316 + 10'd mean accumulation between pupariation and
adult eclosion. In the field studies conducted at Weslaco the mean puparial stage du
ration was 304 + 25'd for the first adult eclosion and 310 + 24'd for the modal adult
eclosion (Table 2). These values were very close to predicted and thus temperature
was the dominant factor determining intra-puparial development time. Furthermore,
the correlation between eclosion and degree-day accumulation was very high. For ear
liest eclosion the coefficient (r2) was 0.966 and for modal eclosion 0.979.
The results obtained from the Mexican portion of the study were more ambivalent.
The mean accumulation for the modal egression date at the citrus grove near General
Teran was 329 42'd, in reasonably close agreement with the results from the labo
ratory and the Texas field study. However, the mean accumulation for the modal
egression date at the mountain canyon site was substantially higher, 410 + 46'd. The


Pupariation No. Range Mode First Modal
Date Flies (Days) (Days) ('D) ('D)

Jan 10 112 43-50 43 345 345
Jan 24 121 36-42 38 277 297
Feb 07 165 29-36 32 284 299
Feb 22 157 29-34 30 289 305
Mar 07 173 28-31 28 313 313
Mar 21 136 22-24 22 281 281
Apr04 106 22-24 22 302 302
Apr 18 4 21-21 21 313 313
May 02 65 17-18 17 288 288
May 16 1 16-16 16 277 277
May 23 21 14-15 15 283 305
Jun28 7 15-15 15 279 279
Sep 09 4 18-18 18 329 329
Sep 20 3 16-17 16 271 271
Oct04 3 16-21 18 276 303
Oct 18 4 31-35 35 316 350
Nov01 1 35-35 35 336 336
Nov 15 8 36-37 36 326 326
Nov29 3 43-43 43 323 323
Dec 13 115 45-56 50 322 328
Dec 27 133 49-52 49 349 349

Mean D 304 + 25:310+ 24.

Thomas: Degree-Days for Mexican Fruit Fly







Fig. 1. Length of the puparial stage of the Mexican fruit fly at two localities in
northern Mexico: a citrus grove and a mountain canyon.

effect of the slightly cooler mean temperatures at the mountain location is reflected in
the graphic representation of these results (Fig. 1). The puparial stage is uniformly
prolonged at the higher elevation site relative to the commercial citrus grove location
in accordance with expectations. The duration of the puparial stage was shortest
(about 3 weeks) during the warmest summer months, and longest during the winter
season (prolonged as much as 3 months). Since degree-day accumulations predict this
prolongation of the puparial stage, and it can be duplicated in the laboratory, the data
suggest that the Mexican fruit fly naturally overwinters in the puparial stage, but not
in a true diapause. The greater apparent accumulation of degree-days at the moun
tain site suggests that actual thermal unit accumulation was less than that calcu
lated by the Means Method. Yellow chapote grows on the east facing slope of the
Sierra Madre Oriental, and there may be a montane shadow effect that causes cooler
temperatures in the late afternoon compared to the open lowland sites where com
mercial citrus is cultivated. If so, then an hourly rather than daily heat unit accumu
nation model may be necessary to predict adult eclosion date in this habitat.
In summary, the results of these studies indicate a significant seasonal effect on
generation time. Ultimately the prediction of demographic events such as adult eclo
sion, seasonality of infestation and number of annual generations will have to incor
porate data from the adult reproductive cycle. Temperature is more strongly
influential of puparial stage duration than of larval stage duration. Puparial develop
ment is so prolonged by low temperatures that overwintering in this stage naturally
results without induction of diapause.

Florida Entomologist 80(1)


The author expresses his gratitude to Jorge Leyva-Vazquez and Robert V. Dowell
for helpful comments on the manuscript. Celestino Cervantes, Ronay Riley, Francisco
Daniel, Jose Galvan, and Reyes Garcia provided essential technical assistance. The
author is indebted to Sr. Abel J. Martinez of the Huerta El Bosque and Sr. Ruben
Bravo of Rancho Los Pinos for permission to conduct experiments on their properties.
The work in Mexico was conducted under the auspices of the Instituto Nacional de In
vestigaciones Forestales y Agropecuarias (INIFAP).


ALINIAZEE, M. T. 1976. Thermal unit requirements for determining adult emergence
of the western cherry fruit fly (Diptera: Tephritidae) in the Willamette Valley of
Oregon. Environ. Entomol. 5: 397-402.
BAKER, A. C., W. E. STONE, C. C. PLUMMER, AND M. MCPHAIL. 1944. A review of stud
ies on the Mexican fruit fly and related Mexican species. USDA Misc. Publ. 531:
BAKER, E. W. 1944. Studies on the response of fruit flies to temperature. J. Econ. En
tomol. 37: 280-284.
CAREY. 1988. Demography of Anastrepha ludens, A. obliqua and A. serpentina
(Diptera: Tephritidae) in Mexico. Florida Entomol. 71: 111-119.
DARBY, H. H., AND E. M. KAPP. 1933. Observations on the thermal death points of
Anastrepha ludens (Loew). USDA Tech. Bull. No. 400. Washington D.C. 20 pp.
FLITTERS, N. E., AND P. S. MESSENGER 1965. Effect of temperature and humidity on
development and potential distribution of the Mexican fruit fly in the United
States. USDA Tech. Bull. No. 1330. Washington D.C. 36 pp.
FRY, K. E. 1983. Heat-unit calculations in cotton crop and insect models. USDA-ARS
Advances in Agricultural Technology, Western Series No. 23, Oakland CA, 23
HAGSTRUM, D. W., AND G. A. MILLIKEN. 1991. Modeling differences in insect develop
mental time between constant and fluctuating temperatures. Ann. Entomol.
Soc. America 84: 369-379.
LAING, J. E., AND J. M. HERATY. 1984. The use of degree-days to predict emergence of
the apple maggot, Rhagoletis pomonella (Diptera: Tephritidae), in Ontario. Ca
nadian Entomol. 116: 1123-1129.
LEYVA-VAZQUEZ, J. L. 1988. Temperature umbral y unidades calor requeridas por los
estados inmaduros de Anastrepha ludens (Loew) (Diptera: Tephritidae). Folia
Entomol. Mexicana 74: 189-196.
MCPHAIL, M., AND C. I. BLISS. 1933. Observations on the Mexican fruit fly and some
related species in Cuernavaca, Mexico, in 1928 and 1929. USDA Circular No.
235, Washington D.C. 24 pp.
PRUESS, K. P. 1983. Day-Degree methods for pest management. Environ. Entomol.
12: 613-619.
SPISHAKOFF, L. M., AND J. G. HERNANDEZ-DAVILA. 1968. Dried torula yeast as a sub
stitute for Brewer's yeast in the larval rearing medium for the Mexican fruit fly.
J. Econ. Entomol. 61: 859-860.
diction of apple maggot fly emergence from thermal unit accumulation. Envi
ron. Entomol. 8: 5154.
SIVINSKI, J. M., AND C. O. CALKINS. 1990. Sexually dimorphic developmental rates in
the Caribbean fruit fly (Diptera: Tephritidae). Environ. Entomol. 19: 1491
SOKAL, R. R., AND F. J. ROHLF. 1973. Introduction to Biostatistics. W. H. Freeman,
San Francisco, CA.

March, 1997

Thomas: Degree-Days for Mexican Fruit Fly 79

THOMAS, D. B. 1995. Predation on the soil inhabiting stages of the Mexican fruit fly
Southwestern Entomol. 20: 61-71.
Modeling insect developmental rates: a literature review and application of a
biophysical model. Ann. Entomol. Soc. America 77: 208-225.
WORNER, S. P. 1992. Performance of phenological models under variable temperature
regimes: consequences of the Kaufmann or Rate Summation effect. Environ.
Entomol. 21: 689-699.

Gould and Hennessey: Carambola Cold Water Treatment 79


USDA-ARS, Subtropical Horticulture Research Station
Miami, FL 33158


The Caribbean fruit fly, Anastrepha suspense (Loew), is a pest of quarantine sig
nificance of carambolas. The fruits are subjected to cold storage quarantine treatment
when shipped to areas outside of the known range and where the fly could survive. In
this study, rapid cooling in cold water increased mortality of Caribbean fruit fly larvae
in carambolas over passive air cooling. Air-cooled carambolas required more than 24
h to cool to the treatment temperature of 1.1 C, while water-cooled fruits required
only about 45 min. After 1 day, Anastrepha suspense larvae had greater than 65%
mortality in water-cooled carambolas, while mortality of larvae in air-cooled fruits
was only 20%. Mortality of larvae in water-cooled fruits was 98% at 4 days, and 100%
(1,900 larvae treated) after 9 days. Twenty six larvae were recovered from air-cooled
fruits after 4 days (1,900 larvae treated), and one larva after 11 days of treatment.
Larval mortality from cold-water-treated fruit reached probit 9 in 8 days, about 2/3
the time (13 days) required for the same level of mortality of larvae in air-cooled fruits.
This difference in mortality is probably due to the rapidity of the cooling. It may be
possible to use this modification to shorten the current cold treatment of 12 days at
1.1 C for Florida carambolas.

Key Words: Caribbean fruit fly, Anastrepha suspense, commodity treatment, cold stor


La mosca del Caribe, Anastrepha suspensa(Loew), es una plaga de la carambola,
Averrhoa carambola L., con importancia cuarentenaria. Las carambolas de la Florida
son sometidas a tratamiento cuarentenario de almacenaje en frio antes de ser envia
das a lugares donde esa mosca no existe pero tendria posibilidades de sobrevivir. En
este studio, el enfriamiento rapido en agua increments la mortalidad de las larvas de
la mosca del Caribe en comparaci6n con el enfriamiento pasivo mediante aire. La ca
rambolas tratadas con aire frio requirieron un minimo de 24 h para alcanzar la tem-
peratura de tratamiento de 1.1 C, mientras que las enfriadas con agua requirieron

Florida Entomologist 80(1)

solamente cerca de 45 minutes. Despues de un dia, las larvas de Anastrepha suspense
Loew tuvieron mas del 65% de mortalidad en las carambolas tratadas con agua fria,
mientras que las larvas de frutas tratadas con aire frio tuvieron solamente 20% de
mortalidad. Las larvas de las frutas enfriadas con agua tuvieron un 98% de mortali
dad a los 4 dias y ninguna larva viva (de 1900 tratadas) fue recuperada 9 dias despues
del tratamiento. Veintiseis larvas fueron recuperadas de frutas tratadas con aire frio
despues de 11 dias de tratamiento. La mortalidad de las larvas en las frutas tratadas
con agua fria lleg6 al Probit 9 en 8 dias, cerca de 2/3 del tiempo (13 dias) requerido por
el mismo nivel de mortalidad de las larvas en las frutas tratadas con aire frio. Esta di
ferencia de mortalidad es probablemente debida a la rapidez del enfriamiento. El tra
tamiento de agua fria pudiera ser usado para acortar el tiempo de 12 dias a 1.1C
actualmente usado para las carambolas de la Florida.

The carambola, Averrhoa carambola L. (Oxalidaceae), is grown throughout the
tropics for its oblong, finned fruits. In south Florida the carambola is grown on 215
hectares and the acreage is increasing (Florida Agricultural Statistics Service 1996).
The carambola in Florida is a poor host for the Caribbean fruit fly, Anastrepha sus
pensa (Loew) (Swanson & Baranowski 1972). A cold quarantine treatment was devel
oped to market expanded production of carambolas (Gould & Sharp 1990, Gould
1996). Other producers of carambolas have used similar treatments (Armstrong et al.
1995). The present treatment takes 12 d at 1.1 C to produce probit 9 mortality
(99.9968% mortality), a standard often used by regulatory agencies in assessing risk
of introduction of exotic pests (Shannon 1994). This requires considerable lead time
for marketing and shipping fruits, and the cost of keeping the fruits refrigerated at
the proper temperature for 12 d is significant.
Many commodities are cooled as soon as they reach the packing house (precooling)
to preserve the initial market quality of the product (Hardenburg et al. 1986). This in
volves rapidly cooling the produce with water, ice, or cold air. In this study we exam
ined the effect of rapidly cooling carambolas with water at 1.0 + .5C on the mortality
of Caribbean fruit fly larvae infesting carambolas.


Carambolas (mean weight 175 23.9 g, n = 100, 16 count) were obtained from
Brooks Tropicals, Homestead, Florida. The carambolas did not have a detectable in
festation in the field so they were exposed to approximately 100,000 Caribbean fruit
flies in a cage for two days. The infested fruits were held at ambient conditions (26
3C) until large 3rd instar larvae developed (6-7 days). This was verified by cutting
samples of five fruits.
The carambolas were then divided randomly into two treatment groups of 120
fruits and a control group of 80 fruits. The control was subdivided into 8 groups of 10
fruits. The control was held without any cold treatment and the larvae emerging were
used to estimate the population present in the treated fruits. One treatment was not
precooled and consisted of carambolas in commercial 16-count cardboard boxes placed
directly into a walk-in cooler at 1.1 + .75'C.
For the second treatment, fruits were precooled with cold water. Carambolas were
placed in nylon mesh bags (15 x 28 cm, 35 liters) and immersed in 0.3 + 0.1 C ice wa
ter until the fruit core temperatures approximated the water temperature (37-42
min.). The ice water was held in a 209 liter tank with a water circulating pump. Water

March, 1997

Gould and Hennessey: Carambola Cold Water Treatment 81

temperature was monitored at 5 min intervals with a national standards traceable
thermometer, and ice was added as necessary to maintain the temperature of the wa
ter. Thermocouples (30-gauge type t copper-constantin) were placed at the centers of
three fruits in each treatment. Temperatures were monitored at 5-min intervals.
Fruits were removed from the water when all of the monitored center temperatures
reached 1.1 C and were placed in 16-count commercial boxes in a walk-in cooler with
the other fruit from treatment No. 1 (1.1 .75'C). At one-day intervals, 10 carambolas
were removed from each treatment up to 12 days. Both treated and control fruits were
held over sand for larval emergence. This experiment was replicated five times at one
week intervals.
Statistical analysis was performed using probit analysis (SAS Institute 1988) and
also by fitting linear and non-linear models using TableCurve 2D (Jandel Scientific


The passively air-cooled carambolas in this experiment required at least 24 h to
cool down to the treatment temperature (at 24 h the temperature was 2.15 .26'C).
Water-cooled fruits cooled down in 40 (0.97 .35'C) to 45 min. (0.93 .27'C). Gould
& Sharp (1990) found similar cooling curves, but did not investigate mortality from
ice-water cooling.
Infestations in control fruits ranged from 134 to 753 per 80 carambolas, with a
mean of 387 and std. dev. of 192. This is an average of 4.8 larvae per fruit with std. dev.
of 2.4 larvae per fruit.
The mortality of insects from the two treatments was dramatically different. After
one day, larvae in hydro-cooled fruits had greater than 65% mortality, while larvae in
air-cooled fruits had about 20% mortality (Fig. 1A). Larvae in water-cooled fruits had
98% mortality at four days, and no larvae (from approximately 1900 larvae treated)
were recovered more than 9 days after treatment. Twenty six larvae were recovered
from air-cooled fruits (from approximately 1900 treated) after 8 days of treatment,
and one larva was recovered after 11 days of treatment. These differences are less
ened when the data are shifted one day to take into account the time it takes air
cooled fruits to cool down to treatment temperature (Fig. 1B).
The probit transformation gave the best 'fit' for the data of equations tested (using
the f statistic and Tablecurve which fits 3000+ equations). Predictions for 75, 90, 99
and 99.9968% mortality are given in Table 1. Both the predictions for probit 9 for lar
vae in air-cooled and the shifted air-cooled fruits are close to that found by Gould &
Sharp (1990). The prediction for larvae in water-cooled fruits differs greatly from pre
dictions of mortality for larvae of either of the air-cooled data sets and the 95% fiducial
limits do not overlap indicating the differences are significant. Probit 9 mortality in
cold-water treated fruits was reached in just over half the time (59.8%) in air-cooled
fruits. This difference in mortality is probably due to the rapidity of the cooling rather
than water immersion. Taschenberg et al. (1974) found that Caribbean fruit fly larvae
could survive 24 h with low mortality in water, so the water itself probably did not
have a major effect on the mortality of the larvae.
Hallman (1994) found that rearing temperature did not significantly effect the
mortality of Caribbean fruit fly larvae treated at 1'C. The coldest rearing tempera
tures used in that study, 20'C, may not have been cold enough to bring about a phys
biological acclimation response in the insect. Other studies have shown that in some
species of flies, adults can acclimate to cold temperatures (Meats 1976, Czajka & Lee
1990, Chen & Walker 1994).

Florida Entomologist 80(1)

March, 1997

100 I-



0 20

> 0
> 1001
C) 80



20 l
0 .

0 2 4 6 8 10 12
Fig. 1. Survivorship of (A) Caribbean fruit fly larvae in carambolas treated with
cold water (solid square) and cold air (open square). Survivorship of (B) Caribbean
fruit fly larvae in carambolas treated with cold water (solid square) and cold air (open
square) adjusted by subtracting one day from cold-air treatment time to account for
delayed cooling time.

The mortality of larvae over the longer period of time is presumed to be due to fac
tors other than cold shock. Lee (1991) referred to this type of mortality as 'indirect
chilling injury' with the mechanism of death presumably due to depletion of cellular
resources. Lee (1991) termed the short term type of mortality as direct chilling injury
or cold shock. Postulated mechanisms for cold shock include breakdown of the lipid
portions of the cell wall in the temperature range of 1-2 C which allows the cellular
contents to leak (Lee & Chapman 1987).

Gould and Hennessey: Carambola Cold Water Treatment 83


95% Fiducial Limits
Treatment Mortality (Days) Lower Upper

Water-Cooled 75% 1.65 1.11 2.06
90% 2.80 2.38 3.39
99% 4.78 4.02 6.21
99.9968% 7.94 6.44 10.95

Air-Cooled 75% 4.48 4.17 4.80
90% 6.08 5.69 6.56
99% 8.84 8.18 9.71
99.9968% 13.27 12.09 14.82

Air-Cooled 75% 3.35 3.13 3.56
Shifted 1 d 90% 5.14 4.88 5.43
99% 8.21 7.73 8.80
99.9968% 13.14 12.22 14.28

The more rapid mortality of larvae in water-cooled fruits was probably from cold
shock including the lack of time for the larvae to acclimate to the cold temperatures.
Whatever the mechanism of death, this study has shown that rapid cooling brings
about rapid mortality It may be possible to incorporate this into the current cold
treatment of 12 days at 1.1 C for Florida carambolas and reduce treatment time sub
stantially. Reductions in the treatment may make the treatment less costly and also
less damaging to the fruits.


We thank W. Montgomery of USDA-ARS for his assistance, E. Schnell of USDA
ARS for translation of the abstract to Spanish, and G. Hallman, USDA-ARS, Weslaco,
TX, L. Neven, USDA-ARS, Yakima, WA, and M. Trunk, Brooks Tropicals, Homestead,
FL, for critically reviewing and improving this manuscript.


ARMSTRONG, J. W., S. T. SILVA, AND V. M. SHISHIDO. 1995. Quarantine cold treatment
for Hawaiian carambola fruit infested with Mediterranean fruit fly, melon fly,
or oriental fruit fly (Diptera: Tephritidae) eggs and larvae. J. Econ. Entomol.
88: 683-687.
CHEN, C. P., AND V. K. WALKER 1994. Cold-shock and chilling tolerance in Droso
phila. J. Insect Physiol. 40: 661 669.
CZAJKA, M. C., AND R. E. LEE, JR 1990. A rapid cold-hardening response protecting
against cold shock injury in Drosophila melanogaster. J. exp. Biol. 148: 245-254.
FLORIDA AGRICULTURAL STATISTICS SERVICE. 1996. Tropical fruit production and
value. 2 pp.
COULD, W. P. 1996. Cold treatment, the Caribbean fruit fly, and carambolas, pp. 489
493 in B. A. McPheron and G. J. Steck [eds.], Fruit fly pests: A world assessment

Florida Entomologist 80(1)

of their biology and management. St. Lucie Press, Delray Beach, Florida. 586
GOULD, W. P., AND J. L. SHARP. 1990. Cold-storage quarantine treatment for caram
bolas infested with the Caribbean fruit fly (Diptera: Tephritidae). J. Econ. En
tomol. 83: 458-460.
HALLMAN, G. J. 1994. Mortality of third-instar Caribbean fruit fly (Diptera: Tephriti
dae) reared at three temperatures and exposed to hot water immersion or cold
storage. J. Econ. Entomol. 87: 405-408.
HARDENBURG, R. E., A. E. WATADA, AND C. Y. WANG. 1986. The commercial storage of
fruits, vegetables, and florist and nursery stocks. U.S. Dept. of Agriculture, Ag
riculture Handbook No. 66 (revised), 130 pp.
JANDEL SCIENTIFIC. 1994. TableCurve 2D Windows v 2.0 User's Manual. Jandel Sci
entific, San Rafael, California. 404 pp.
LEE, R. E., JR. 1991. Principles of insect low temperature tolerance, pp. 17-46 in R. E.
Lee, Jr. and D. L. Denlinger [eds.], Insects at low temperature. Chapman and
Hall, New York. 513 pp.
LEE, D. C., AND D. CHAPMAN. 1987. The effects of temperature on biological mem
branes and their models. Symp. Soc. Experimental Biol. 41: 3552.
MEATS, A. 1976. Developmental and long-term acclimation to cold by the Queenslang
fruit-fly (Dacus tryoni) at constant and fluctuating temperatures. J. Insect
Physiol. 22: 1013-1019.
SAS INSTITUTE. 1988. SAS/STAT User's Guide. Release 6.03 Ed. SAS Institute, Cary,
SHANNON, M. J. 1994. APHIS, pp. 1-10 in J. L. Sharp and G. J. Hallman [eds.], Quar
antine treatments for pests of food plants. Westview Press, Boulder, CO. 290 pp.
SWANSON, R. W., AND R. M. BARANOWSKI. 1972. Host range and infestation by the
Caribbean fruit fly, Anastrepha suspense (Diptera: Tephritidae), in South Flor
ida. Proc. Florida State Hortic. Soc. 85: 271-274.
TASCHENBERG, E. F., F. LOPEZ, AND L. F. STEINER 1974. Response of maturing larvae
of Anastrepha suspense to light and to immersion in water. J. Econ. Entomol.
67: 731-734.

March, 1997

Florida Entomologist 80(1)


Center for Medical, Agricultural and Veterinary
Entomology Agricultural Research Service
U.S. Department of Agriculture
Gainesville, Florida 32604


Teflubenzuron baits were active against laboratory colonies of the red imported
fire ant, Solenopsis invicta Buren. Worker brood production ceased soon after treat
ment and by four weeks posttreatment, most colonies were devoid of brood. Worker
ants did not exhibit any direct effects from treatment with teflubenzuron. As is typical
with most insect growth regulators, colony mortality was slow and dependent on old
age attrition of the worker ants. A few (<25) female alates were produced in one of the
laboratory colonies at 12 weeks posttreatment.

March, 1997

Williams et al.: IFA Control with Teflubenzuron 85

The teflubenzuron baits reduced field colonies of S. invicta by 75-79% within 6
weeks after treatment, 83-86% within 13 weeks, and 77-91% within 17 weeks. At 17
weeks posttreatment, the presence of worker brood in the plots treated with the lower
rates, 0.1125% and 0.0225%, gave evidence of recovery of some colonies. However, the
results of the field tests indicate that teflubenzuron has excellent potential for control
of field populations of S. invicta.

Los cebos de teflubenzuron fueron activos contra colonies de laboratorio de la hor
miga roja importada de fuego, Solenopsis invicta Buren. La producci6n de obreras
ces6 en breve tiempo despu6s del tratamiento, y a las 4 semanas la mayoria de las co
lonias qued6 desprovista de ellas. Las obreras no mostraron efectos director del tra
tamiento con teflubenzuron. Como es tipico en la mayoria de los regluadores del
crecimiento de insects, la mortalidad de la colonia fue lenta y dependiente del des
gaste por edad de las obreras. Unas pocas (<25) hembras aladas fueron producidas en
las colonies de laboratorio a las 12 semanas del tratamiento. Los cebos de tefluben
zuron redujeron las colonies de campo de S. invicta en un 75-79% en 6 semanas, en un
83-86% en 13 semanas, y en un 77-91% en 17 semanas despu6s del tratamiento. Alas
17 semanas del tratamiento, la presencia de obreras inmaduras en las parcelas trata
das con las dosis mas bajas, 0.1125% y 0.0225%, fue una evidencia de la recuperacidn
de algunas colonies. Sin embargo, los resultados de pruebas de campo indicaron que
teflubenzuron tiene un excelente potential para el control de poblaciones de campo de
S. invicta.

Insect growth regulators (IGRs) are highly active against the red imported fire
ant, Solenopsis invicta Buren, (Banks 1986, Banks et al. 1978, 1983, 1988, Phillips et
al. 1985, 1989, Vinson & Robeau 1974, Vinson et al. 1974). The most effective IGRs
prevent the replacement of the worker caste in colonies through mortality of develop
ing immatures, degeneration of the reproductive organs of the queen, and/or a shift in
caste differentiation from worker to sexual forms. This lack of worker replacement
usually results in colony death because the existing worker ants die and dependent
castes and immatures succumb from neglect. Juvenile hormone mimics have been the
most effective IGRs, and two of these materials, fenoxycarb and 1 (8-methoxy-4,8
dimethylnonyl)-4-(1 methylethyl) benzene, have been used in commercial baits for
fire ant control.
Another group of IGRs, i.e. Benzoylphenyl urea (BPU) compounds, of which di
flubenzuron (dimilin) is the best known, has been successfully developed as control
agents for a number of other insects. These chemicals are commonly known as chitin
inhibitors because they interfere with normal endocuticular deposition and molting in
insects. They also are ovicidal in some cases. Because of the low solubility of the BPUs
in soybean oil or other food attractants, this group of IGRs has not been used success
fully against fire ants. A newer BPU, teflubenzuron [1 (3,5-dichloro-2,4-difluorophe
nyl)-3-(2,6-difluorobenzoyl)-urea] (American Cyanamid Co., Wayne, NJ 07470 USA),
is much more soluble in food attractants and may offer promise in fire ant manage
ment systems. Teflubenzuron is considerably more physiologically active than di
flubenzuron against a number of agricultural pests and has effectively controlled
some insects, such as the diamondback moth, Plutella xylostella (Linnaeus), and the
red flour beetle, Tribolium castaneum (Herbst), that are highly resistant to other
types of insecticides (Ishaaya & Klein 1990). Herein, we report the results of labor
tory and field studies with teflubenzuron against S. invicta.

Florida Entomologist 80(1)


Laboratory Tests

Three laboratory tests were conducted with laboratory-reared queenright S. in-
victa colonies (Banks et al. 1981). For each test, teflubenzuron (10% emulsifiable con
centrate) was combined with once-refined soybean oil to produce baits containing
0.1% and 0.5% active ingredient (wt/wt). In each test, three colonies were exposed to
0.5 ml of each bait concentration. The 0.1% solution (0.5 mg per colony AI) and the
0.5% solution (2.5 mg per colony AI) were tested against colonies with 20-25 ml brood
and 20,000-40,000 workers, and 30-35 ml brood and 50,000-70,000 workers, respect
tively. Three colonies were exposed to 0.5 ml of neat once-refined soybean oil and
served as non-treated controls. The test colonies were allowed ad libitum feeding on
the oil solutions which were offered in micropipets. The colonies were returned to nor
mal diet (Banks et al. 1981) 24 h after treatment and maintained in the laboratory at
27 + 2'C. Monthly observations (including numbers of workers, reproductive, and
amount of brood) were made until the colonies died, returned to their normal pre
treatment index level, or for one year, whichever occurred first.
Effectiveness of the treatments was based on comparison of the before and after
treatment size index of each colony. This index was derived by multiplying the as
signed values for worker numbers, i.e. 1-6, by the quantity of worker brood, i.e. 1-25,
(Table 1); e.g. a colony with a rating of 5F would have a colony index of 125 (5 x 25)
(Banks & Lofgren, 1991). For each of the three tests, data were combined for the three
colonies. Mean percent reduction in colony indices were analyzed using an analysis of
variance and Tukey's Studentized Range (HSD) test (SAS Institute 1988).

Field Tests

Pregel defatted corn grit baits containing 0.01125, 0.0225, or 0.045% teflubenzu
ron were prepared in our laboratory for the field tests as follows. Technical tefluben
zuron (97.5%) was dissolved in dimethyl formamide (0.5-1.5% by weight of oil in the
formulation) and the solution was incorporated into warm (20-25'C) once-refined soy
bean oil. The oil solution was slowly poured over the corn grits as they were stirred in
a large food mixer. Stirring continued for about 10 minutes to insure thorough mixing
of the oil and grits.


Estimated Quantity of Worker Brood
Estimated Number of Worker Ants (gms)

Rating Value Rating Value

<100 1 1 0 A 1
101-5000 2 2 1-5 B 5
5001-20000 3 3 5-10 C 10
20001-35000 4 4 10-20 D 15
35001-50000 5 5 20-30 E 20
>50000 6 6 >30 F 25

March, 1997

Williams et al.: IFA Control with Teflubenzuron 87

Each bait was broadcast with a tractor-mounted granular applicator (Williams et
al. 1983) at a rate of 1.12 kg/ha on 0.2-ha plots with an average of 12 mounds per hect
are in nongrazed permanent pasture in Union County, Florida. Three plots were
treated with each teflubenzuron concentration; three plots were treated with Logic
(fenoxycarb, Ciba-Geigy, Greensboro, NC) at a rate of 1.12 kg/ha as a standard, and
three plots were left untreated as a control. Efficacy of the treatments was evaluated
by comparison of the before and after (6, 13 and 17 weeks) treatment population in
dices using standard methods established for determination of population indices of
S. invicta (Banks et al. 1988). Mean reductions in population indices were analyzed
using an analysis of variance and Tukey's Studentized Range (HSD) test (SAS Insti
tute 1988).


Laboratory tests

Teflubenzuron was very active against laboratory colonies of S. invicta (Table 2). In
test one, worker brood production ceased soon after treatment and by four weeks post
treatment two colonies at the 0.5 mg dosage were devoid of brood and only a few pu
pae remained in the third colony. All colonies at the 2.5 mg rate were devoid of brood
at four weeks. The only worker brood production thereafter through the one-year test
occurred in one colony at the 0.5 mg rate; about 0.5 ml was present at the one-year
posttreatment evaluation. In test one, all three colonies subjected to the 2.5 mg rate
died by 36 weeks; however, two colonies at the 0.5 mg rate were alive at one year, al
though neither contained more than 500 workers and only one colony had a queen
present. It is doubtful that these colonies would have survived under field conditions.
Colony reduction did not occur as quickly in the second test. All three colonies
treated at 0.5 mg and two treated at 2.5 mg still contained some worker brood at four
weeks; however, all treated colonies were devoid of worker brood by eight weeks and
remained so until the test was discontinued after 32 weeks. At the conclusion of test
two, fewer than 500 workers remained alive in any treated colony
In test three, with the exception of one 0.5 mg treatment, all colonies at both dos
ages (0.5 mg and 2.5 mg) were devoid of worker brood by sixteen weeks and remained
so until the test was discontinued after 32 weeks. The one colony at 0.5 mg still had
worker brood present until the end of the test. At the end of test three, fewer than 100
workers remained alive in any treated colony with all colonies having fewer than 25
workers, including one colony at 0.5 mg that contained a small amount of brood. All
of the queens in the treated colonies were dead or were not producing eggs except one
colony treated with 0.5 mg.
Worker ants did not exhibit any direct effect of treatment with teflubenzuron in
any test. Thus, as is typical with most insect growth regulators, colony mortality was
slow and dependent on old-age attrition of the worker ants. No alate production oc
curred in any of the treated colonies in test one and test three; however, in test two, a
few (<25) female alates were produced at 12 weeks posttreatment in two replicates at
the 2.5 mg rate.
The untreated colonies in test one showed no change or increase in size through 24
weeks posttreatment, but began a decline thereafter that left all three devoid of
worker brood and reduced in size by one year. The controls in test two began a decline
at 12 weeks that resulted in termination of the test after 32 weeks. The control colo
nies in test three were significantly different than the treatments until week 16. After
this time, although they were noticeably different containing large physogastric

Florida Entomologist 80(1)

Co o6 Lf
0^ ^ M


C -
S00 L
at LLO 0

- 00 L

t- N- Lfl

0C N- Cfl

at at


C0 0 01

m M
0 00c C


0c 00 cc


co o c'

at at
cO cO N

CC L~fl


00 00


- LOt-

at CM CV
N- Nt- c

CM at CD

00 t


March, 1997








Williams et al.: IFA Control with Teflubenzuron

at -

cc o










OO t- 0 Cf LOf
00 t- O COD t
t~ 00 00 00 CD

t- t- CD 0t 0t
Ll C^ C^ Cl CM

LO o o


o ia a

NLn N N o
cC' CLn C -

30 30 30 .Co

= =2 m 5

Florida Entomologist 80(1)

queens, large amounts of brood, and numerous workers, the colony indices were not
statistically different.
Significant (P < 0.05) reductions in colony indices of treated colonies compared
with untreated controls were observed from 4-24 weeks in test one, up to 8 weeks in
test two and from 4-12 weeks in test three. These results indicated that the reductions
were caused by teflubenzuron and not attributable to a decline observed in the un
treated controls late in the tests. The decline in the control colonies is not readily ex
plainable; however, they might have been inadvertently exposed to teflubenzuron
over the long test period. Food and water tubes for all of the colonies, both treated and
controls, were handled using the same gloves and forceps. This technique has not ap
peared to affect control colonies in other tests with chemicals. However, teflubenzuron
is biologically effective at extremely low dosages and they may have been exposed to
minute residues in these tests.

Field tests

The teflubenzuron baits reduced the population indices of field colonies of S. in-
victa by 75-79% within 6 weeks after treatment (Table 3). No significant difference
was noted in the effectiveness of any of the teflubenzuron baits or the Logic standard
at either the 6 or 13 week evaluation. By 17 weeks, however, the presence of worker
brood in 6 and 5 colonies, respectively, in plots treated with the 0.01125% and
0.0225% baits gave some evidence of recovery. No surviving colonies in plots treated
with the 0.045% teflubenzuron or Logic baits contained any worker brood after 17
weeks. The number of workers in the surviving colonies after 17 weeks had been sub
stantially reduced at the highest dosage of teflubenzuron with 82.1% of the colonies
having <10,000 workers, 48.1% having <1,000 which was not statistically different
than Logic standard (100% of the colonies had <10,000 workers, 91.7% had <1,000
The results of the field tests indicate that teflubenzuron has excellent potential for
control of field populations of S. invicta. The levels of teflubenzuron tested in the field
were extremely low when compared with Logic; therefore, higher levels of teflubenzu
ron may produce even better control.


We thank Karen Vail for her laboratory and statistical assistance and J. Hogsette,
B. Forschler, and D. Oi for their comments on the manuscript. The second and third
authors are retired from the USDA.


BANKS, W. A. 1986. Insect growth regulators for control of imported fire ants, pp. 387
398 in C. S. Lofgren and R. K. Vander Meer [eds.], Fire ants and leaf cutting
ants: Biology and management. Westview, Boulder, CO.
BANKS, W. A., AND C. S. LOFGREN. 1991. Effectiveness of the insect growth regulator
pyriproxyfen against the red imported fire ant (Hymenoptera: Formicidae). J.
Entomol. Sci. 26: 331 338.
BANKS, W. A., C. S. LOFGREN, AND J. K. PLUMLEY. 1978. Red imported fire ants: effects
of insect growth regulators on caste formation, and colony growth and survival.
J. Econ. Entomol. 71: 75-78.
WILLIAMS, D. P. WOJCIK, AND B. M. GLANCEY. 1981. Techniques for collecting,
rearing, and handling imported fire ants. USDA-SEA AATS-S-21. 9 pp.

March, 1997

Williams et al.: IFA Control with Teflubenzuron

BANKS, W. A., L. R. MILES, AND D. P. HARLAN. 1983. The effects of insect growth reg
ulators and their potential as control agents for imported fire ants. Florida En
tomol. 66: 172-81.
BANKS, W. A., D. F. WILLIAMS, AND C. S. LOFGREN. 1988. Effectiveness of fenoxycarb
for control of red imported fire ants (Hymenoptera: Formicidae). J. Econ. Ento
mol. 81: 83-87.
ISHAAYA, I., AND M. KLEIN. 1990. Response of susceptible laboratory and resistant
field strains of Spodoptera littoralis (Lepidoptera: Noctuidae) to teflubenzuron.
J. Econ. Entomol. 83: 5962.
PHILLIPS, S. A., JR., D. M. CLABORN, AND H. G. THORVILSON. 1985. An insect growth
regulator (ProDrone) for effective management of the red imported fire ant
(Hymenoptera: Formicidae). J. Entomol. Science 20: 194-98.
PHILLIPS, S. A., JR., AND H. G. THORVILSON. 1989. Use of fenoxycarb for area-wide
management of red imported fire ants (Hymenoptera: Formicidae). J. Econ. En
tomol. 82: 1646-1649.
SAS INSTITUTE. 1988. SAS user's guide: statistics, version 6.03 ed. SAS Institute,
Cary NC
VINSON, S. B., AND R. ROBEAU. 1974. Insect growth regulator: effects on colonies of the
imported fire ant. J. Econ. Entomol. 67: 584-87.
VINSON, S. B., R. ROBEAU, AND L. DZUIK. 1974. Bioassay and activity of several insect
growth regulator analogues on the imported fire ant. J. Econ. Entomol. 67: 325
WILLIAMS, D. F., C. S. LOFGREN, J. K. PLUMLEY, AND D. M. HICKS. 1983. Auger-appli
cator for applying small amounts of granular pesticides. J. Econ. Entomol. 76:

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