Title: Florida Entomologist
Full Citation
Permanent Link: http://ufdc.ufl.edu/UF00098813/00061
 Material Information
Title: Florida Entomologist
Physical Description: Serial
Creator: Florida Entomological Society
Publisher: Florida Entomological Society
Place of Publication: Winter Haven, Fla.
Publication Date: 1992
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: VID00061
Source Institution: University of Florida
Holding Location: University of Florida
Rights Management: Open Access

Full Text

(ISSN 0015-4040)


(An International Journal for the Americas)

Volume 75, No. 2 June, 1992


Announcement- 75th Annual Meeting ..................................................... i

Announcement-Solicitation of Applicants for Editor .................................... ii

Research Reports
MAHMOOD, F., AND J. B. ALEXANDER-Immature Stages of Nemapalpus
nearcticus (Diptera: Psychodidae) .................................................... 171
RAID, R. N., AND R. H. CHERRY-Effect of Soil Parameters On Pathogenicity
of the Fungus Metarhizium anisopliae to the Sugarcane Grub Ligyrus sub-
tropicus (Coleoptera : Scarabaeidae)................................................... 179
PUCHE, H., AND J. FUNDERBURK-Intrinsic Rate of Increase of Frankliniella
fusca (Thysanoptera: Thripidae) on Peanuts .................................... 185
HALLMAN, G. J., C. G. MORALES, AND M. C. DUQUE-Biology of Acrosternum
marginatum (Heteroptera: Pentatomidae) On Common Beans ............... 190
CENTER, T. D., AND F. ALLEN DRAY, JR.-Associations Between
Waterhyacinth Weevils (Neochetina eichhorniae and N. bruchi) and
Phenological Stages of Eichhornia crassipes in Southern Florida .......... 196
DOzA-Trap For Capturing and Retaining Rhynchophorus cruentatus
(Coleoptera: Curculionidae) Adults Using Sabal palmetto as Bait ......... 212
MCPHERSON, J. E., S. J. TAYLOR, AND S. L. KEFFER-Evaluation of Charac-
ters to Distinguish Fitchia aptera and F. spinosula (Heteroptera: Red-
uviidae) ........................................ ....................... ................... 222
ROBACKER, D. C.-Effects of Shape and Size of Colored Traps on Attractiveness
to Irradiated, Laboratory-Strain Mexican Fruit Flies ......................... 230
PENA, J. E.-Predator-Prey Interactions Between Typhlodromalus peregrinus
and Polyphagotarsonemus latus: Effects of Alternative Prey and Other Food
R sources ..................................................................................... 241
PORTER S. D.-Frequency and Distribution of Polygyne Fire Ants (Hymenop-
tera: Formicidae) in Florida ........................................................... 248
HALL, H. G.-Suspected African Honeybee Colonies in Florida Tested for Iden-
tifying DNA M arkers ..................................................................... 257
WALDER, J. M., AND C. 0. CALKINS--Gamma Radiation Effects on Ovarian
Development of the Caribbean Fruit Fly, Anastrepha suspense (Loew)
(Diptera: Tephritidae), and Its Relationship to Sterile Fly Identification 267

Scientific Notes

MALO, E. A.-Effect of Bait Decomposition Time on Capture of Anas-
trepha Fruit Flies ....................................................... 272
Continued on Back Cover

Published by The Florida Entomological Society


President .............................................. J. L. Knapp
President-Elect ................................................ .................... D. F. W illiams
Vice-President ......................................... ......................... J. E. Pefia
Secretary .............................. ............ ....... D. G. Hall
Treasurer ................... .... .......................... A. C. Knapp
Other Members of the Executive Committee
J. F. Price J. E. Pefia J. R. Cassani
J. R. McLaughlin O. Liburo F. Oi

J. R. McLaughlin, USDA/ARS, Gainesville, FL ....................................... Editor
Associate Editors
Agricultural, Extension, & Regulatory Entomology
James R. Brown-Disease Vector Ecology & Control Center, NAS, Jacksonville, FL
Richard K. Jansson-Tropical Research & Education Center, Homestead, FL
Michael G. Waldvogel-North Carolina State University, Raleigh, NC
Stephen B. Bambara-North Carolina State University, Raleigh, NC
Biological Control & Pathology
Ronald M. Weseloh-Connecticut Agricultural Experiment Sta., New Haven, CT
Book Reviews
J. Howard Frank-University of Florida, Gainesville
Chemical Ecology, Physiology, Biochemistry
Louis B. Bjostad-Colorado State University, Fort Collins, CO
Ecology & Behavior
Theodore E. Burk-Creighton University, Omaha, NE
John H. Brower-Stored Product Insects Research Laboratory, Savannah, GA
Forum & Symposia
Carl S. Barfield-University of Florida, Gainesville
Genetics & Molecular Biology
Sudhir K. Narang-Bioscience Research Laboratory, Fargo, ND
Medical & Veterinary Entomology
Arshad Ali-Central Florida Research & Education Center, Sanford, FL
J. E. Pefia-Tropical Research & Education Center, Homestead, FL
Systematics, Morphology, and Evolution
Michael D. Hubbard-Florida A&M University, Tallahassee
Howard V. Weems, Jr.-Florida State Collection of Arthropods, Gainesville
Willis W. Wirth-Florida State Collection of Arthropods
Business M manager ............................................................. A. C. Knapp
FLORIDA ENTOMOLOGIST is issued quarterly-March, June, September, and De-
cember. Subscription price to non-members is $30 per year in advance, $7.50 per copy;
institutional rate is $30 per year. Membership in the Florida Entomological Society,
including subscription to Florida Entomologist, is $25 per year for regular membership
and $10 per year for students.
Inquiries regarding membership and subscriptions should be addressed to the Busi-
ness Manager, P. O. Box 7326, Winter Haven, FL 33883-7326.
Florida Entomologist is entered as second class matter at the Post Office in DeLeon
Springs and in Winter Haven, FL.
Manuscripts from all areas of the discipline of entomology are accepted for consider-
ation. At least one author must be a member of the Florida Entomological Society.
Please consult "Instructions to Authors" on the inside back cover.
This issue mailed June 26, 1992


The 75th annual meeting of the Florida Entomological Society will be held August
10-12, 1992 at the Indian River Plantation Resort and Conference Center, 555 N.E.
Ocean Boulevard, Hutchinson Island, Stuart, FL 34996 (407-225-3700) (FAX 407-225-
0003). Registration forms and information will be mailed to members and will appear
in the Newsletter.

Jorge E. Pefia
Program Committee, FES
University of Florida
Tropical Research and Education Center
18905 SW 280th Street
Homestead, FL 33031
(305) 246-7048 or -6340


The Florida Entomological Society is soliciting interested and qualified persons as
applicants for Editor of Florida Entomologist, effective June 1992 or as soon thereafter
as possible. Provided a qualified individual can be located soon, opportunity will be
provided to work with the current Editor in order to facilitate the transition. The Florida
Entomological Society does offer negotiated salary compensation for its Editor, and a
cadre of Associate Editors will continue to work with the Editor to facilitate the review,
acceptance and/or editions of submitted manuscripts. Job descriptions for the Editor
and duties of Associate Editors are provided.

Persons interested in this position should contact Dr. Carl S. Barfield, Professor of
Entomology, Building 970 Hull Road, University of Florida, Gainesville, FL 32611 by
mail or call (904) 392-1901, ext. 141. Fax communications to (904) 392-0190.

Mahmood & Alexander: Immature Nemapalpus nearcticus 171


111500 Summit West Boulevard, 45 D
Temple Terrace, FL 33617 USA

2Fundacion Centro Internacionale de Entrenamiento
e Investigaciones Medicas
Apartado Aereo 5390
Cali Columbia


The external morphology and biology of the immature stages of Nemapalpus
nearcticus, a Psychodid fly from Florida, USA, are described for the first time. Mor-
phological similarities and differences from its near relatives are also discussed.


La morfologia external de los estados inmaduros de Nemapalpus nearticus y su
biologia se described inicialmente. Tambien son descritas y comparadas las differencias
y similitudes morfologicas de estas species y otras relacionadas.

Nemapalpus nearcticus Young is the only known Nearctic representative of
Bruchomyiinae, a subfamily thought to contain the most primitive members of the
Psychodidae. This species was originally collected at Sugarfoot Hammock, a small forest
7 km SW of Gainesville, Florida (Young 1974) and has been recorded only from the type
locality and from Gulf Hammock in Levy County, Florida.
The genus Nemapalpus shows a disjunct geographical distribution, probably indica-
tive of its antiquity (Fairchild 1952). To date a total of 26 species have been described
(Alexander 1987). No published descriptions exist of the immature stages of Nemapalpus
species other than that of Hanson (1968) who illustrated the first and fourth instars of
an unnamed member of the genus in his treatise on immature Panamanian Psychodidae.
Satchell (1953) described the early stages of Bruchomyia argentina Alexander, a species
morphologically comparable to Nemapalpus and belonging to the same subfamily.
Little is known about the biology of the adults or immature stages of Nemapalpus
species, although they often occur in the same habitats as their phlebotomine sand fly
relatives. Unlike sand flies, the females are not hematophagous and it is unclear whether
either sex feeds in the adult stage.
In this paper we describe the immature stages of Nemapalpus nearcticus for the
first time and compare it with other closely related species.


Specimens were collected from the bases of tree trunks during daylight in Gulf
Hammock, Levy County, Florida in June 1986. Adult females were introduced into 120
ml plaster-lined plastic containers (Endris et al. 1982); the plaster bases of which had
been moistened with water prior to use. Drops of KaroR syrup placed on the gauze lids
of the containers provided a carbohydrate source.

172 Florida Entomologist 75(2) June, 1992

Females were allowed to oviposit, after which several eggs were removed for descrip-
tive purposes. The remaining eggs were allowed to develop and when the larvae hatched,
they were provided with rearing diet consisting of an aged 1:1 mixture of dried rabbit
feces and Purina rabbit chow complete diet 5315R (Young et al. 1981).
Before examination larvae were killed and boiled in a 10% NaOH solution for 5
minutes, then transferred to liquid phenol in depression microslides for study.



The egg of Nemapalpus nearcticus is oval and measures 0.57 mm by 0.24 mm (mean
of 4 eggs). The chorion is yellowish and does not darken with age, unlike that of sand
fly eggs. It is covered with irregularly-shaped protuberances arranged in ridges (Fig.
1). These may act as a plastron, allowing oxygen exchange across the egg surface and
the moist habitat in which they are presumably laid in nature. The larva emerges through
a slit made at one end of the chorion that extends for about two-fifth of its length.

There are 4 larval instars, of which the first and fourth instars are illustrated (Fig.
2 & 3). Total length of the single first instar larva examined was 1.93 mm, excluding
the simple paired caudal setae which measured 0.31 mm. The pale brown head capsule
was approximately 0.25 mm long by 0.25 mm wide. The whitish thorax and abdomen
were markedly narrower.
The mature larva is 7.20 mm long, excluding the paired caudal setae, which measure
0.65 mm (Fig. 3). The well-developed, sclerotized head capsule is dark brown and meas-
ures approximately 0.55 mm by 0.45 mm. The pinkish-brown body consists of a 3-seg-
mented thorax and a 9-segmented abdomen. The latter bears a dark brown sclerotized
peritreme on the eighth segment.
The head is eucephalous, hypognathous, with a short coronal suture (co.su) and an
indistinct frontal suture (f.su) enclosing a contiguous frons (fr) and clypeus (Fig. 4 &
5). On the ventral side of the frons, lateral parietals form a narrow hypostomal bridge
(h.y.b) which is split in its center (Figs. 6 & 7). The foramen magnum is bounded by a
heavily-sclerotized post-occipital edge (p.o.e), which also shows posterior tentorial scar
(p.t.s) (Fig. 6). The subgenal edge (s.g.e) is also heavily sclerotized and bounds the
preoral cavity; its two ends meet posteriorly and form the hypostoma (hy) (Figs. 4 &
6). The short, club-shaped, unsegmented antennae (a) project from the posterior end of
the antennal rings. Each bears 4 small, vertically projecting sensilla (Figs. 4 & 5). A
very small ocellus (oc) is located behind each of the antennae (Fig. 5).
There is no demarcation between frons (fr) and clypeus. The clypeus bears a pair of
external clypeal setae (e.c.s) and a pair of internal clypeal setae (i.c.s) that are present
in front and internal to the former (Fig. 5). Whereas, anterior frontal setae (af.s),
post-frontal setae (p.f.s), post-occiptal setae (p.o.s), and ventro-lateral setae are visible
in both lateral and dorsal views (Figs. 4 & 5).
The transparent labrum (lbr) has several setae on its dorsal and ventral sides and
overhangs the preoral cavity (Figs. 4, 5 & 7). A large toothed plate can be seen under-
neath, which Satchell (1953) considered to be part of the hypostome in Bruchomyia.
The mandible (m.d) is sclerotized and bears 4 stout apical teeth and a prostheca (Fig.
8). There is also a penicillus (p.s) and a pair of ventral setae. The penicilli of the two
mandibles project inwardly and touch each other (Fig. 6). The maxilla (m.x) is a flattened
plate at the side of the preoral cavity. It bears a short 1-segmented maxillary palpus
(mx.p), galea (ga), and a lacinia (lc) with 5 small sensory rods at its apex, unfused with

Mahmood & Alexander: Immature Nemapalpus nearcticus 173

41 4
'... !. -k:.I

A r .. '- "





Figs. 1-3. Immature stages of Nemapalpus nearcticus: 1, Egg of Nemapalpus
nearcticus; 2, Dorsal view of the first instar larva; 3, Lateral view of the fourth instar

Florida Entomologist 75(2)






Figs. 4-8. Different parts of the head of Nemapalpus nearcticus: 4, Lateral view of
the head; 5, Dorsal view of the head; 6, Ventral view of the head; 7, Part of the ventral
view of the head showing the different parts of the maxilla; 8, An enlarged view of the

June, 1992

Mahmood & Alexander: Immature Nemapalpus nearcticus 175

Figs. 9-11. Enlarged view of the caudal segment of Nemapalpus nearcticus: 9, Dorsal
view; 10, Ventral view; 11, Lateral view.

Florida Entomologist 75(2)

June, 1992

the reduced galea (ga) (Fig. 7). The labium is reduced to a cushion-shaped hairy lobe
above the hypostome.
The thoracic segments each have two annuli, with the anterior spiracles (sp) on the
posterior annulus of the prothorax. There are no legs or prolegs present on the thorax.
The abdomen has 8 segments and a cylindrical caudal region. Each of the first 7
abdominal segments has three annuli (Figs. 3, 14 & 15). The eighth segment bears a
posterior spiracle (ps) on a tubercle which is surrounded by a dark peritreme (Fig. 9).
The anus lies terminally on the elongated caudal region, from which a pair of delicate
retractable pseudopods (pp) protrude (Figs. 3, 9, 10 & 11). These pseudopods have
ventral unsclerotized swellings (Fig. 10) and a sclerotized, brown dorsal plate (dp) with
minute posteriorly directed spines (Fig. 9). The dorsal plate is bifurcated posteriorly
and carries two vertical processes which themselves each divide to form paired caudal
setae (c.s) (Figs. 9 & 11). These setae are dark brown proximally shading to light brown
and finally to white at their tips. Two slightly sclerotized oval plates (op), that are
devoid of spines, are present behind the setae bearing processes (Fig. 9). Numerous
ventrally directed microspines are present on the ventral surface of the caudal region
(Fig. 10). The base of each pseudopod bears five ventrally directed hooks (hk) (Figs. 10
& 11). The surface of the larval body is covered by two types of microspines, one type
occurring singly and the other in groups (Figs. 2, 3, 12 & 13).
The nomenclature used here for chaetotaxy of the setae follows that of Satchell
(1953). Both simple and brush-like setae are present on larvae of Nemapalpus nearcticus
(Figs. 2, 3 & 9-15). Each brush-like seta consists of a hollow cylinder covered with
imbricated scales, mounted on a distinct plaque. Each seta has a spherical ampulla at
its top (Fig. 12).
The chaetotaxy of the prothorax differs from that of the remaining thoracic and
abdominal segments. The dorsum of the anterior annulus of the prothorax bears a total
of five pairs of brush-like setae, consisting of anterior internal dorsals (a.i.d), anterior
median dorsals (a.m.d), anterior lateral dorsals (a.l.d), anterior external dorsals (a.e.d)
and accessory dorsals (ac.d) (Fig. 14). The posterior annulus of the prothorax bears
three pairs of brush-like setae on its dorsum consisting of posterior internal dorsals
(p.i.d), posterior external dorsals (p.e.d) and posterior lateral dorsals (p.l.d) (Fig. 14).
Two pairs of simple setae are present on the ventral surface of the anterior prothoracic
annulus, together with a pair of accessory ventrals (ac.v) and a pair of anterior internal
ventrals (a.i.v) (Fig. 15). In the middle of the posterior annulus are situated a pair of
posterior internal ventrals (p.i.v), median ventrals (mv) and external ventrals (ev) as
well as a group of pedichaetan setae (pd) (Fig. 15). The posterior internal ventrals (p.i.v)
and pedichaetae (pd) are simple setae, the latter consisting of three separate setae of
different lengths, each with its own separate base. The bases of these setae are clustered
in a small group (Fig. 15). On the sides of the posterior annulus of the prothorax, paired
posterior lateral ventrals (p.l.v) and anterior lateral ventrals (a.l.v) setae are present
(Figs. 14 & 15).
The chaetotaxy of the mesothoracic and metathoracic segments is similar. The an-
terior annuli have paired accessory dorsals (ac.d) and anterior lateral dorsals (a.l.d)
(Fig. 14). The posterior annulus of each segment bears brush-like posterior internal
dorsals (p.i.d) and posterior lateral dorsals (p.l.d) setae (Fig. 14). The ventral surface
of the meso- and meta-thorax each have a pair of paedichaetae (pd), as well as paired
external ventrals (ev), median ventrals (mv), anterior internal ventrals (a.i.v) and an-
terior lateral ventrals (a.l.v) (Fig. 15).
Abdominal segments 1-7 have identical chaetotaxy. The anterior annulus of each
segment bears a pair of accessory dorsal setae (ac.d); the middle annulus has paired
posterior lateral dorsals (p.l.d) and the posterior annulus has paired external dorsals
(p.e.d) (Fig. 14). The ventral surface of anterior annuli has anterior lateral ventrals

Mahmood & Alexander: Immature Nemapalpus nearcticus 177

14 15

Figs. 12-15. Chaetotaxy of thoracic and abdominal segments of Nemapalpus near-
ticus: 12, Brush like seta; 13, Spine like setae; 14, Dorsal view of the thoracic and
abdominal segments; 15, Ventral view of the thoracic and abdominal segments.

178 Florida Entomologist 75(2) June, 1992

(a.l.v) and the posterior annuli has posterior ventrals (p.v) and posterior lateral ventrals
(p.l.v) (Fig. 15). All the setae on abdominal segments 1-7 are brush-like (Figs. 3, 14 & 15).
The eighth abdominal segment differs from the others in having setae only on the
posterior annulus. Both the posterior ventrals (p.v) and posterior lateral ventrals (p.l.v)
of this segment are simple setae (Fig. 10).


The pupae is exarate and adectious. The presumptive antennae, mouth parts, thoracic
legs, and wings can be seen clearly as well as the paired pseudopods on the underside
of the eighth segment. The pupal abdomen terminates in a pair of stout pointed processes.


The number of eggs produced by females of Nemapalpus nearcticus is apparently
less than that laid by Bruchomyia argentina or the phlebotomine sand flies. Eleven
laboratory-reared females laid between 6 and 24 eggs (average 14.36). The larva of
Nemapalpus nearcticus closely resembles Hanson's (1968) description of the unnamed
Panamanian species. Hanson reported that this species had 4 caudal setae, although
these were so closely appressed at the base that they seemed to arise from a single
seta. It may be that in this species there is a single seta which is quadrifurcate distally.
The larva of Nemapalpus nearcticus appears to resemble that of Bruchomyia argen-
tina in most respects. Differences include the presence of simple accessory ventrals on
the ventral surface of the prothorax of N. nearcticus absent in B. argentina. The
anterior and posterior internal ventrals are simple in N. nearticus but branched in B.
argentina. The ventrally directed sensillum present on the antennae of N. nearticus is
absent in B. argentina. Finally, Satchell (1953) recorded the presence of 3 small sensory
setae at the bases of the cylindrical processes on the caudal dorsum of B. argentina,
absent in N. nearcticus which however, possesses nodules at the base of the pseudopods
of the caudal region, not found in the former species.


The authors thank Dr. David G. Young, Department of Entomology and Nematology,
University of Florida Gainesville, Florida for supplying laboratory space and equipment
necessary during the preparation of this manuscript. University of Florida Experiment
Station Journal Series No. R-02021.


ALEXANDER, J. B. 1987. A new species of Nemapalpus (Diptera: Psychodidae:
Bruchomyiinae) from Northeastern Colombia. Florida Entomol. 70: 376-381.
ENDRIS, R. G., P. V. PERKINS, D. G. YOUNG, AND R. N. JOHNSON. 1982. Techniques
for laboratory rearing of sand flies (Diptera: Psychodidae). Mosquito News 42:
FAIRCHILD, G. B. 1952. Notes on Bruchomyia and Nemopalpus (Diptera:
Psychodidae). Ann. Entomol. Soc. America 45: 259-280.
HANSON, W. G. 1968. The immature stages of the subfamily Phlebotominae in Panama
(Diptera: Psychodidae). Ph.D. thesis, Univ. Kansas. 104 p.
SATCHELL, G. H. 1953. On the early stages of Bruchomyia argentina Alexander
(Diptera: Psychodidae). Proc. R. Entomol. Soc. London Ser. A. 28: 1-12.
YOUNG, D. G. 1974. Bruchomyiinae in North America with a description of Nemopalpus
nearticus N. Sp. (Diptera: Psychodidae). Florida Entomol. 57: 109-113.
YOUNG, D. G., P. V. PERKINS, AND R. G. ENDRIS. 1981. A larval diet for rearing
phlebotomine sand flies (Diptera: Psychodidae). J. Med. Entomol. 18: 446.

Raid & Cherry: Metarhizium On Sugarcane Grubs 179


University of Florida
Institute of Food and Agricultural Sciences,
Everglades Research and Education Center,
P. O. Box 8003, Belle Glade, FL 33430


The effect of soil temperature, soil moisture, and inoculum concentration on
pathogenicity of Metarhizium anisopliae to the sugarcane grub, Ligyrus subtropicus
(Blatchley) was tested under laboratory conditions. M. anisopliae killed grubs from 16
to 310 C as defined by the equation Y= -1.04 + 0.06X where Y= percentage grub
mortality and X = soil temperature (C). M. anisopliae also killed grubs at soil moistures
from 0.17 to 1.0 gravimetric weight (weight water/weight oven dry soil) as defined by
the equation Y= -0.42 + 3.59X 2.68X2 where Y= percentage grub mortality and X
= soil moisture. Grubs were killed by M. anisopliae at inoculum concentrations conidiaa/
gr. oven-dry soil) from 10 to 12,500 as defined by the equation Y= 0.108 + 0.00006X
where Y= percentage grub mortality and X= inoculum concentration. The relevance
of the results to field conditions in Florida sugarcane fields and the potential for M.
anisopliae controlling L. subtropicus are discussed.


Se determine en condiciones de laboratorio el efecto de la temperature y la humedad
del suelo y la concentration del inoculo en la patogenicidad de Metarhizium anisopliae
infectando el gusano de la cana de azucar Ligirus subtropicus (Blatchley). M. anisopliae
mato los gusanos del suelo a temperatures entire 160C a 31C como se define en la ecuacion
Y = 1.04 + 0.06X donde Y = porcentaje de mortalidad de los gusanos y X = temperature
del suelo en grades centigrados. M. anisopliae tambi6n mat6 los gusanos del suelo a
una humedad entire 0.17 a 1.0 peso gravimetrico (peso de agua/peso del suelo secado en
el horno) como se define en la ecuacion Y = -0.42 + 3.59X -2.68 x 2 donde Y = porcentaje
de mortalidad de gusanos y X = humedad del suelo. Los gusanos murieron tambien por
el M. anisopliae en concentraciones de 10 y 12,500 (conidia/gr. de suelo secado en el
horno) como define la ecuaci6n Y = 0.108 + 0.00006X donde Y = porcentaje de mortalidad
de gusanos y X = concentraci6n de patogenos. Se dispute la importancia de los resultados
anteriores en condiciones de los campos de cafia de azucar en la Florida y el potential
de control de M. anisopliae sobre L. subtropicus.

Sugarcane is Florida's most valuable field crop and is primarily grown in the
Everglades area of southern Florida. Since 1971, several species of Scarabaeidae have
been observed to cause significant damage to sugarcane in southern Florida. Of these
pests, the white grub, Ligyrus subtropicus (Blatchley) is the primary species of economic
importance (Gordon & Anderson 1981). This grub has been shown to reduce tons of
sugar per ha by 39% in areas of high infestation (Sosa 1984). Currently, no chemical
control is known for these pests and flooding of sugarcane fields is sometimes used for
control (Cherry 1984a).

180 Florida Entomologist 75(2) June, 1992

The fungus, Metarhizium anisopliae (Metschn.) has long been known to be pathogenic
to various insects (Steinhaus 1963). Boucias et al. (1986) found a low incidence of M.
anisopliae in third instar grubs ofL. subtropicus in Florida sugarcane fields. In Australia,
laboratory studies using M. anisopliae to control sugarcane grub species have demon-
strated enough promise to extend the research to field experiments (Anonymous 1986).
More recently, Samuels et al. (1990) conducted field tests in Australia using M. anisopliae
against the sugarcane grub, Antitrogus parvulus Britton. They found that M. anisopliae
survived in the field for 30 months and caused host mortality sufficient to result in
commercially acceptable levels of crop protection. Raid & Cherry (1992) tested M.
anisopliae isolates from grub species found in Florida sugarcane against L. subtropicus
larvae. Only the isolate from L. subtropicus proved to be pathogenic to L. subtropicus.
The high mortalities observed in those pathogenicity tests indicated that the L. sub-
tropicus isolate of M. anisopliae is a good candidate for biological control of that grub
species. This study was conducted to determine the effect of several soil parameters on
pathogenicity of M. anisopliae to L. subtropicus in order to assist in understanding the
interaction of the fungus with its host under field conditions.


The experimental design used in all experiments was essentially the same and is
described here. Specific methods used in individual experiments are described in detail
later. The M. anisopliae used was originally isolated from L. subtropicus and shown to
be pathogenic to this grub (Raid & Cherry 1992). The isolate was grown on Sabouraud
maltose agar in Petri dishes at 270C in darkness. Spore suspensions were made by
dispersing spores in sterile distilled water with a sonicator and using 0.5% Tween-20
as a surfactant. Conidial viability was verified by dilution plating on maltose agar. Spore
concentrations were determined using a stage hemocytometer. Third instars of L. sub-
tropicus were collected by digging under sugarcane plants in sugarcane fields. After
collection, grubs were stored at 190C in covered plastic pans containing muck and raw
carrots for food. Grubs were stored at least 10 days before tests and all damaged or
sick appearing grubs were removed prior to testing. At the start of each test, grubs
were placed in covered plastic pans (18 X 12 X 7 cm high). Each pan contained muck
soil previously fumigated with methyl-bromide, and raw carrots for food. Ten grubs
were placed in each pan. All tests were conducted in darkness in cabinets with temper-
ature tolerances of 0.5C. Grubs were observed after 14 days for M. anisopliae
induced mortality. Mortality is defined by failure of the insect to respond to probing.
Isolations were performed from cadavers not exhibiting the externally borne con-
idiophores characteristic ofM. anisopliae to confirm M. anisopliae as the cause of death.
All experiments were repeated twice.

Soil Temperature

The effect of soil temperature on grub mortality caused by M. anisopliae was deter-
mined at six different temperatures ranging from 16 to 31C. Mean monthly soil temper-
atures at 10 cm deep under Florida sugarcane plants range from 18.3C in January to
28C in July (Cherry & Boucias 1989). Ten centimeters beneath sugarcane plants is the
habitat in which L. subtropicus grubs are found in highest density (Cherry 1984b).
Twenty grubs were tested in soil inoculated with M. anisopliae at each temperature.
A matching number of grubs were controls and were held in soil without inoculum. Pans
with soil were held until reaching test temperatures before grubs were put in the
containers. Inoculum concentration was approximately 5 X 103 conidia/g dry weight soil.
Soil moisture was 250 ml of water/300 g of dry weight soil.

Raid & Cherry: Metarhizium On Sugarcane Grubs

Soil Moisture

The effect of soil moisture on grub mortality caused by M. anisopliae was determined
at six different soil moistures ranging from 0.17 to 1.00 gravimetric weight (weight
H20/weight oven-dry soil. Muck soil at 0.17 is powder-dry and 1.00 is very wet although
drier than field capacity. Inoculum concentration was about 5 X 103 conidia/gram of dry
weight soil. Pans with grubs were held at 250C. Twenty grubs were tested in soil
inoculated with M. anisopliae at each soil moisture and a matching number of grubs
were controls held in soil without the inoculum.

Inoculum Concentration

The effect of inoculum concentration on grub mortality caused by M. anisopliae was
determined at 7 concentrations ranging from 10 to 12,500 conidia/g dry weight soil. A
stock conidia suspension (1 X 105 conidia/ml) was diluted to obtain the different concen-
trations. Each of the 7 aliquots of suspension was brought to a 200 ml final volume using
sterile distilled water. The 200 ml were then added to 300 g of oven dry soil in each
pan. Pans with grubs were held at 250C. Twenty grubs were tested in soil inoculated
with M. anisopliae at each of the 7 concentrations. Grubs held in soil without inoculum
were used as controls.
In all tests, grub mortality due to M. anisopliae was determined by adjusting for
natural mortality in controls using Abbott's formula (1925). Data were analyzed by using
a general linear models (GLM) procedure (SAS Institute 1985). Mortality data were
analyzed by using the percentage of M. anisopliae induced mortality as the dependent
variable and soil temperature, soil moisture, or inoculum concentration as the independ-
ent variable. Mortality data were fitted to both linear and quadratic models and the
final model selected was that model which had the more significant fit to the observed
data. Model fitness was based upon the examination of residuals, statistical significance
levels, regression coefficients, and biological considerations.


M. anisopliae was not found in any of the field collected grubs which were used as
controls in any of the experiments. This finding is consistent with Boucias et al. (1986)
and Raid & Cherry (1992) who also reported low detection levels for M. anisopliae in
L. subtropicus populations in Florida sugarcane fields.

Soil Temperature

M. anisopliae developed and killed L. subtropicus third instars at all temperatures
tested ranging from 160 to 31C (Fig. 1). The percentage of grub mortality was signifi-
cantly (P < 0.05) correlated with soil temperature with a correlation coefficient (r) of
0.93. McCauley et al. (1968) also found that mortality was greater at 250 than 20C in
several species of wireworm larvae infected with M. anisopliae. Our data show that
soil temperatures found in Florida sugarcane fields will allow continual development of
M. anisopliae in L. subtropicus third instars throughout the year.

Soil Moisture

M. anisopliae developed and killed L. subtropicus third instars at all soil moistures
tested ranging from 0.17 to 1.0 gravimetric weight (Fig. 2). The percentage of grub
mortality was significantly (P < 0.05) correlated with soil moisture with a correlation





5 40-

Florida Entomologist 75(2)

Y =-1.04+0.06X


16 19 22 25 28 31
Soil Temperature (C)

Fig. 1. L. subtropicus mortality caused by M. anisopliae at different soil tempera-

coefficient of 0.96. Illingworth (1921) reported that an epizootic of a green "muscardine"
fungus (species unknown) in Australian sugarcane grubs was enhanced by increased soil
moisture. Young (1974) suggested that low soil moisture was a factor in keeping M.
anisopliae at low levels in Tonga. Our data are in general agreement with these studies
in that grub mortality was lowest at the lowest soil moisture and in general increased
with increasing soil moisture.

Inoculum Concentration

M. anisopliae developed and killed L. subtropicus third instars at inoculum concen-
trations ranging from 391 to 12,500 conidia/g dry soil (Fig. 3). The percentage of grub
mortality was significantly (P < 0.05) correlated with inoculum concentration (r = 0.97).
Numerous examples are found in microbial control literature in which host mortality is
directly dependent on pathogen density as our data also show. This matter is discussed
in detail by Tanada & Fuxa (1987).


Interesting questions concerning M. anisopliae arise from data obtained in this
study. First, what factors) are limiting the effectiveness of naturally occurring M.
anisopliae in reducing L. subtropicus populations in Florida sugarcane? Our data show
that M. anisopliae is capable of developing and killing L. subtropicus under a wide
range of soil temperatures, soil moistures, and concentrations. However, a very low
incidence of M. anisopliae in L. subtropicus has been noted in previous studies (Boucias
et al. 1986, Raid & Cherry 1992) as well as in this study. Although this may indeed be
the case, the possibility that M. anisopliae is widespread within sugarcane production
fields and goes undetected due to rapid decomposition of infected cadavers should be

June, 1992

Raid & Cherry: Metarhizium On Sugarcane Grubs


Y =-0.42+3.59X-2.68X2
R =0.92


0.50 0.67
Soil Moisture


Fig. 2. L. subtropicus mortality caused by M. anisopliae at different soil moistures.
Soil moisture = gravimetric weight = weight water/weight oven-dried soil.

Y =0.108+0.00006X
R2- 0.95



7500 10000 12500

Inoculum Concentration

Fig. 3. L. subtropicus mortality caused by M. anisopliae at different inoculum con-
centrations. Inoculum concentration = conidia/gr. oven-dried soil.








0 40


184 Florida Entomologist 75(2) June, 1992

considered. Second, can M. anisopliae be used as inoculative and/or augmentative re-
leases to reduce L. subtropicus populations in Florida sugarcane fields? Vercambre
(1988) reported that a formulation of M. anisopliae was applied at several sites in
sugarcane in Reunion and infected larvae of the grub Hoplochelus marginalis were
found in samples taken later. Samuels et al. (1990) reported significant field control of
the grub, Antitrogus parvulus Britton in Australian sugarcane fields following M.
anisopliae spore applications. In our future research, we shall attempt to determine if
M. anisopliae can be used through inoculative and/or augmentative releases to reduce
L. subtropicus populations.


We are grateful to numerous Florida sugarcane growers for access to their land and
to the Florida Sugar Cane League for grant support. Fla. Agric. Exp. Stn. Journal
Series No. R-01925.


ABBOTT, W. 1925. A method of computing the effectiveness of an insecticide. J. Econ.
Entomol. 18: 265-267.
ANONYMOUS. 1986. New grub method shows good results. Australian canegrower.
8: 18.
BOUCIAS, D., R. CHERRY, AND D. ANDERSON. 1986. Incidence of Bacillus popilliae
in Ligyrus subtropics and Cyclocephala parallel (Coleoptera: Scarabaeidae) in
Florida sugarcane fields. Environ. Entomol. 15: 703-705.
CHERRY, R. 1984a. Flooding to control the grub Ligyrus subtropicus (Coleoptera:
Scarabaeidae) in Florida sugarcane. J. Econ. Entomol. 77: 254-257.
CHERRY, R. 1984b. Spatial distribution of white grubs (Coleoptera: Scarabaeidae) in
Florida sugarcane. J. Econ. Entomol. 77: 1341-1343.
CHERRY, R., AND D. BOUCIAS. 1989. Incidence of Bacillus popilliae in different life
stages of Florida sugarcane grubs. J. Entomol. Sci. 24: 526-530.
GORDON, R., AND D. ANDERSON. 1981. The species of Scarabaeidae (Coleoptera)
associated with sugarcane in south Florida. Florida Entomol. 64: 119-138.
ILLINGWORTH, J. 1921. A study of natural methods of control for white grubs. Austral.
Bureau of Sugar Exp. Sta. (BSES). Div. of Ent. Bull. No. 12.
MCCAULEY, V., R. ZACHARUK, AND R. TINLINE. 1968. Histopathology of green
muscardine in larvae of four species of Elateridae (Coleoptera). J. Invert. Pathol.
12: 444-459.
RAID, R., AND R. CHERRY. 1992. Pathogenicity of Metarhizium anisopliae to the
sugarcane grub Ligyrus subtropicus (Coleoptera: Scarabaeidae). J. Agric. En-
tomol. 9: 11-16.
SAMUELS, K., D. PINNOCK, AND R. BULL. 1990. Scarabeid larvae control in sugarcane
using Metarhizium anisopliae. J. Invert. Pathol. 55: 135-137.
SAS INSTITUTE. 1985. Guide for personal computers. SAS Institute, Cary, N.C.
SOSA, 0. 1984. Effect of white grub (Coleoptera: Scarabaeidae) infestation on sugarcane
yields. J. Econ. Entomol. 77: 183-185.
STEINHAUS, E. 1963. Insect pathology. Academic Press, New York.
TANADA, Y., AND J. FUXA. 1987. The pathogen population. Pages 113-157 in:
Epizootiology of insect diseases. J. Fuxa and Y. Tanada, ed. John Wiley and
Sons, Inc. New York.
VERCAMBRE, B. 1988. Canne a sucre. Lutte centre les insects ravageurs, pp. 39-41
in Rapport annuel 1987, IRAT Reunion, Institut de Recherches Agronomiques
Tropicales, Saint Denis, France.
YOUNG, E. C. 1974. The epizootiology of two pathogens of the coconut palm rhinoceros
beetle. J. Invert. Pathol. 24: 82-92.

Puche & Funderburk: Frankliniella fusca On Peanuts 185


North Florida Research & Education Center
University of Florida
Quincy, Florida 32351


Frankliniellafusca (Hinds) is a thrips species of economic importance that produces
direct injury to leaves and is a vector of the tomato spotted wilt virus. The intrinsic
growth rate of increase and the effect of two initial densities of immature F. fusca on
immature survivorship and adult emergence was determined by enclosing adult or im-
mature thrips on peanut leaflets. The number of emerging immature thrips was 14.5
per female. Immature emergence began in 7 d with a peak of emergence (67%) between
8 and 9 d. No males emerged. Females reached maturity 11.7 d after emergence. The
gross value of the intrinsic growth rate of increase was 0.500 and the value determined
by iteration was 0.405. Total emergence of cohorts with smaller initial numbers of
immatures (32 larvae per bag) was significantly higher than of larger cohorts (117 larvae
per bag). Under high population densities, reduced survivorship may assure that enough
resources are available for successful reproduction of some adults. Our studies reveal
that F. fusca has the potential for high population increases in peanut which may favor
persistence and spread of tomato spotted wilt virus disease.


Frankliniella fusca (Hinds) es un thrips de importancia econ6mica porque produce
dafio foliar y es vector del virus del marchitamiento del tomate. La tasa intrinseca de
crecimiento y el efecto de dos densidades iniciales de larvas sobre la sobrevivencia de
larvas y la emergencia de adults fue determinada, encerrando adults o larvas de thrips
con hojas de mani. El nimero de larvas de thrips emergentes fue 14.5 por hembra. La
emergencia de inmaduros comenz6 en 7 d con un pico de emergencia (67%) entire 8 a 9
d. No emergieron machos. Las hembras llegaron al estado adulto en 11.7 d despues de
la emergencia. El valor bruto de la tasa intrinseca de crecimiento fue 0.500 y el valor
determinado por iteraci6n fue igual a 0.405. La emergencia total en cohortes pequefias
(32 larvas por bolsa) fue significativamente mayor que en cohortes grandes (117 larvas
por bolsa). Bajo condiciones de altas densidades poblacionales, la reducci6n en la sob-
revivencia puede asegurar que suficientes recursos estAn disponibles para la reproducci6n
exitosa de algunos adults. Nuestro studio revela que F. fusca tiene el potential para
un alto incremento poblacional en mani lo cual favorece la persistencia y dispersion del
virus del marchitamiento del tomate.

The tobacco thrips, Frankliniellafusca (Hinds), is an important vector of the tomato
spotted wilt virus in north Florida. This thrips species is the only vector in the region
that commonly inhabits and reproduces in peanuts, and reaches high densities on peanut
leaves (Olson & Funderburk 1986).
Once the population of F. fusca has been established in the peanut crop, symptoms
of the spotted wilt disease may develop and result in the final wilting of the host plant
(Cho et al. 1987, Allen & Broadbent 1986). The tomato spotted wilt virus is acquired
by the vector thrips only during the immature stages and can be inoculated to the host
plant only by the adult thrips (Sakimura 1963, Lewis 1973).

Florida Entomologist 75(2)

June, 1992

The biology and life history of some Frankliniella species have been reported by
several authors (Watts 1934, Bryan & Smith 1956, Lublinkhof & Foster 1977, Robb &
Parrella 1990). However, studies were needed to determine the proportion of immature
thrips capable of reaching the adult stage that later, as adults, could transmit and
disperse the virus among peanut plants. For that purpose, the intrinsic rate of increase
(rm) of F. fusca was estimated under laboratory conditions. To determine rm, estimates
of female thrips fecundity and immature thrips mortality over time were obtained on
peanut leaflets. An additional objective was to evaluate the effect of population density
on immature mortality over time.


Cohorts of eight F. fusca adults of were placed on one peanut leaflet of a potted
plant (Kinzer 1968, Kinzer et al. 1972), and enclosed in a plastic bag (2 x 4 cm). Thrips
populations attain similar densities in peanuts under field conditions (Funderburk &
Gorbet 1991). Thirteen bags were established, and each was sealed around the pedicel
of the peanut leaflet with parafilm paper (Parafilm "M" Laboratory Film, American
National Can, Greenwich, CT 06830). Bags were punctured (about 500 times to allow
for good aeration and to prevent immatures from escaping) with a # 000 insect pin
(Elefant Brand, BioQuip Products, Inc, 17803 La Salle Avenue, Gardena, CA 90248).
Cohorts of adult thrips were maintained in a laboratory with a controlled RH of 60-80
%, a L:D photoperiod of 14:10 h, and a mean air temperature of 270C. After five days,
adults were removed and larval emergence on leaflets enclosed in each bag was recorded
Egg mortality was assumed constant; therefore, total daily number of newly emerged
larvae was considered equivalent to adult thrips age-specific fecundity. From immatures
reaching adulthood, a gross value of rm of the immatures reaching adulthood was calcu-
lated by the equation r = loge Ro/T; an iterative value of rm was estimated from, 1 =
Y e-rmx lx mx (Birch 1948). Ro is the net reproductive rate (Ro= Z lx mx); T is the
generation time (T= Z x lx mx/Ro); lx is the age specific survivorship of the cohort of
immature individuals reaching the adult stage; mx is the age specific fecundity (average
births/female/day); x is a pivotal age; and rm is the potential rate of increase for a cohort
growing in the absence of biotic sources of mortality and for a particular environment.
It was assumed that lx was constant and equal to the average proportion of adult thrips
emerging from all bags and that no mortality occurred during the preoviposition period
of 2.5 days (Watts 1934).
New thrips larvae were separated in cohorts and placed on one peanut leaflet that
was changed every other day. Seven cohorts were started with 32 + 4.3 (x SEM)
first instar larvae per bag, and six cohorts were started with 117 24.3 (x SEM)
first instar larvae per bag. Each cohort was enclosed with a peanut leaflet inside a
plastic bag, as mentioned above. Larval mortality of each cohort was recorded every
other day. A logrank test (Rimm et al. 1980) was used to compare statistically the
survivorship curves of these two experimental densities. The relationship for each cohort
between the proportion alive at the beginning of each age interval and the age in d was
described with linear regression.


After an oviposition period of 5 days, the mean number ( SEM) of emerged immature
thrips per female was 14.5 ( 2.5), ranging from 11.3 to 18.1. The emergence of immature
thrips began 7 d after female thrips were placed in plastic bags, and extended for 4
more days. After this period, no more thrips larvae were found. Peanut leaflets were

Puche & Funderburk: Frankliniella fusca On Peanuts

checked every other day for one additional week. Immature emergence and/or age-spe-
cific fecundity followed a normal distribution with a peak of emergence (67%) between
8 and 9 d. (Fig 1.). Watts (1934) reported a similar incubation period for F. fusca in
South Carolina. He reported an average of 55.5 immatures per female, but did not
indicate the number of females or the cohort size of females used to determine the
None of the emerging adults in this experiment were males. No information has
been published about factors affecting sex ratios. Under field conditions females fre-
quently are much more prevalent in peanuts than males (unpublished data), but no
explanation is suggested until additional studies are conducted. Mean time to reach
maturity ( SEM) of the surviving females was 11.7 d ( 2.1). Using the assumed
average age-specific survivorship (lx) and the age-specific fecundity (mx), the gross
value of rm was 0.500. When calculated by iteration, the value of rm was 0.405.
Mortality over time of the immature stages of F. fusca is given in Fig. 2. Survival
to adult of cohorts with an average of 32 larvae per bag was 9.3%, while survival of
cohorts with an average of 117 larvae per bag was 1.4%. This difference was significant
according to a logrank test (X2 = 55.49; P < 0.05). Overcrowding is known to reduce
the rate of development and oviposition of some insects, e.g., Locusta (Gunn & Hunter-
Jones 1952). We hypothesize that interference from overcrowding resulted in reduced
survivorship of immature thrips at the greater population density. We also suggest, as
a practical consequence of these findings, that overcrowding be avoided in the rearing
and maintenance of laboratory colonies.
The shapes of the survivorship curves for both population densities tested were
similar, with high losses early in life followed by much lower and relatively constant
losses (Fig. 2.). Mortality was greatest during the first 4 d after emergence when most
of the larvae were first or second instars (Watts 1934). The relationship between the


C) 100

1 -.60-


20 -1

0 2 4 6 8 10
Time (d)
Fig. 1.The mean number ( SEM) of immature Frankliniella fusca emerging per
female over time. The arrow indicates the time when adult female thrips were removed
from the plastic bags.

Florida Entomologist 75(2)

June, 1992

0 3 5 7 9 11 13 15
I I oo-- I I
First Second
Instar Instar Prepupa Pupa

Time (d)
Fig. 2. Survivorship curves through time and stage of insect development for imma-
ture Frankliniella fusca at average densities of 32 and 117 immatures per bag. The
linear regression relationships for both densities are described by the model y = a +
10b(x); y is the proportion alive at the beginning of each age interval and x is the age in d.

transformed average immature survival and age (d) was significantly linear for cohorts
of both 32 and 117 larvae per bag.
In our study, the intrinsic rate of increase and the effects of overcrowding of F.
fusca developing on healthy peanut leaflets were determined. Robb & Parrella (1990)
noted that thrips larvae exposed to leaves infected with tomato spotted wilt virus
exhibited shorter developmental times and increased mortality. We suggest that addi-
tional studies are needed to determine the intrinsic rate of increase and the effects of
overcrowding of F. fusca developing on infected peanut plants. The vectors acquire the
virus only as larvae and infect host plants as adults (Reddy et al. 1983). Much of the
tomato spotted wilt virus disease incidence in peanut fields results from viruliferous
thrips migrating in from outside sources. However, additional spread of the disease in
a field can occur by the larvae developing on infected plants, and controlling these larvae
with insecticides is suggested as a way of reducing the amount of disease in a field. Our
findings suggest that increases in mortality of larvae developing on leaves infected with
tomato spotted wilt virus may be, at least in part, compensated by increases in fecundity
and larval survival due to less overcrowding.


The authors thank A. Grajal and J. H. Frank for reviewing the early drafts of this
manuscript. This is Florida Agricultural Experiment Station Journal Series No. R-01293.








Puche & Funderburk: Frankliniella fusca On Peanuts 189


ALLEN, W. R., AND A. B. BROADBENT. 1986. Transmission of tomato spotted wilt
virus in Ontario greenhouses by Frankliniella occidentalis. Canadian J. Plant
Pathol. 8: 33-38.
BIRCH, L. C. 1948. Experimental background to the study of the distribution and
abundance of insects. I. The influence of temperature, moisture and food on the
innate capacity for increase of three grain beetles. Ecology. 34: 109-122.
BRYAN, D. E., and R. F. SMITH. 1956. The Frankliniella occidentalis (Pergande)
complex in California (Thysanoptera: Thripidae). Univ. California. Pub. in En-
tomol. 10: 359-410.
CHO, J. J., W. C. MITCHELL, L. YUDIN, AND L. TAKAYAMA. 1987. Ecology and
epidemiology of tomato spotted wilt virus (TSWV) and its vector, Frankliniella
occidentalis. Phytopathology. (Abstract) 74: 866.
FUNDERBURK, J. E., AND D. W. GORBERT. 1991. Thrips control in seedling peanuts
in North Florida 1990. Insect. & Acar. Tests. 16: 206-207.
GUNN, D. L., AND P. HUNTER-JONES. 1952. Laboratory experiments on phase differ-
ences in locusts. Anti-Lo cust Bull. 12: 1-29.
KINZER, R. E. 1968. Mass rearing the tobacco thrips, Frankliniella fusca (Hinds),
and laboratory technique for testing peanut resistance to thrips. MS thesis, Ok-
lahoma State University. 13p.
KINZER, R. E., S. YOUNG, AND R. R. WALTON. 1972. Rearing and testing tobacco
thrips in the laboratory to discover resistance in peanuts. J. Econ. Entomol. 65:
LUBLINKHOF, J., AND D. E. FOSTER. 1977. Development and reproductive capacity
of Frankliniella occidentalis (Thysanoptera: Thripidae) reared at three temper-
atures. J. Kansas Entomol. Soc. 50: 313-316.
LEWIS, T. 1973. Thrips, their biology, ecology and economic importance. Academic
Press, London. 349p.
OLSON, S. M., AND J. E. FUNDERBURK. 1986. New threatening pest in Florida -
western flower thrips, pp 43-51 in W. M. Stall [ed.], Proceedings Florida Tomato
Institute, University of Florida Extension Report VEC 86-1, Gainesville.
AND J. M. TRESH. 1983. Epidemiology and control of groundnut bud necrosis
and other diseases of legume crops in India caused by tomato spotted wilt virus.
ICRISAT. Pub. 1983: 93-102.
HOFFMAN. 1980. Basic biostatistics in medicine and epidemiology. Appleton-
Century-Crofts; New York.
ROBB, K. L., AND M. P. PARRELLA. 1990. Biology of the western flower thrips and
the effect of tomato spotted wilt virus on thrips development. USDA TSWV
Workshop. Apr 18-19, 1990. Beltsville Agricultural Research Center, Beltsville,
SAKIMURA, K. 1963. Frankliniellafusca, an additional vector for the tomato spotted
wilt virus, with notes on Thrips tabaci, another vector. Phytopathology 53: 412-
WATTS, J. G. 1934. A comparison of the life cycles of Frankliniella tritici (Fitch), F.
fusca (Hinds), and Thrips tabaci Lind. (Thysanoptera-Thripidae) in South
Carolina. J. Econ. Entomol. 37: 1158-1159.

190 Florida Entomologist 75(2) June, 1992


Centro Internacional de Agricultura Tropical,
Apartado Aereo 6713, Cali, Colombia


The biology of Acrosternum marginatum (Palisot de Beauvois) (Heteroptera: Pen-
tatomidae) on common bean, Phaseolus vulgaris L., was studied at a mean temperature
of 24C (range 19-320C) near Palmira, Colombia. Total mean ( SEM) developmental
time from egg to adult was 42.1 4.2 d. Head capsule widths of the five instars did not
overlap. The female lived a mean of 44.4 2.84 d and laid 96.29.72 eggs in 7.50.74
masses. The mean preovipositional period was 10.1 1.31 d. The time interval between
consecutive ovipositions fit the geometric distribution with P (the probability that a
female will oviposit on any day) = 0.19 (included first oviposition) or P = 0.22 (excluded
first oviposition). The mean number of eggs per mass was 12.8, with a marked peak at
14. There were no significant correlations between the number of days between consecu-
tive ovipositions nor age of female versus number of eggs per mass. Mean emergence
from egg masses was 80 0.77%. Percentage emergence was not significantly related
to the number of days since the previous oviposition nor age of the female.


Se estudi6 la biologia de Acrosternum marginatum (Palisot de Beauvois) (Heteropt-
era: Pentatomidae) alimentado con frijol, Phaseolus vulgaris L., cerca de Palmira,
Colombia. El period promedio de desarrollo fue de 42.1 4.2 dias a la temperature
promedia de 24C. Las anchuras de las cApsulas cefAlicas de los cinco instares no se
sobrecruzaron. Las hembras vivieron un promedio de 44.4 2.84 dias y pusieron un
promedio de 96.2 9.72 huevos en 7.5 + 0.74 posturas. El period promedio de
preoviposici6n fue de 10.1 1.31 dias. El intervalo entire oviposiciones consecutivas se
ajust6 a la distribuci6n geometrica con P (la probabilidad que una hembra oviposite
cualquier dia) = 0.19 (incluye la primera oviposici6n) o P = 0.22 (excluye la primera
oviposici6n). El nfmero promedio de huevos por masa fue de 12.8, con un pico marcado
a los 14 huevos por postura. No hubo relaciones significativas entire el nimero de dias
entire oviposiciones consecutivas ni entire la edad de la hembra versus el nfmero de
huevos por postura. La emergencia promedia de las posturas fue del 80 0.77%. El
porcentaje de emergencia no se relacion6 significativamente con el namero de dias desde
la oviposici6n previa ni la edad de la hembra.

Several species of Pentatomids (Heteroptera) are considered important pests of the
common bean, Phaseolus vulgaris L., in the Neotropics. Saunders et al. (1983) listed
the following Pentatomids attacking beans in Central America: Acrosternum mar-
ginatum (Palisot de Beauvois), Chlorochroa ligata (Say), Edessa confusionata Breddin,
E. rufomarginata (de Geer), Mormidea pictiventris Stal, M. ypsilon (L.), Murgantia
histrionica (Hahn), Nezara viridula (L.), Thyanta antiguensis (Westwood), and T.
perditor (Fabricius). The relationship between bean yield and Pentatomid population
levels was studied by Costa et al. (1980) (N. viridula), Costa et al. (1981) (Piezodorus

'Current address: USDA, ARS, Miami, Florida, 33158.


Hallman et al.: Biology of Acrosternum marginatum 191

guildinii (Westwood)), and Hallman et al. (1985) (A. marginatum). In addition to reduc-
tion in yield caused by direct feeding on the pods, some species are implicated in trans-
mission of Nematospora coryli Peglion, the pathogen of yeast spot disease (Costa et al.
1980). This study reports on aspects of the biology of A. marginatum (Palisot de
Beauvois), the most common species of Acrosternum in Central America. It is found
from the southwestern United States to Venezuela and Ecuador and throughout the
Caribbean from Florida to Guadalupe (Rolston 1983). Besides beans, A. marginatum
is also a pest of soybeans (Waldbauer 1977, Temerak & Whitcomb 1984).


A colony of A. marginatum was established in a screen house using bugs collected
from common bean near Palmira, Colombia. Mean temperature within the house was
24C (range 19-320C), and the relative humidity (RH) averaged 80% (range 45-100%).
Twenty-five F2 generation egg masses were taken from the colony over a period of
two weeks and placed individually in petri dishes with a piece of moist cotton until
eclosion. First instars were transferred on the day of emergence to potted common
beans (line BAT 41) with young pods and placed inside screen cages (59 by 29 by 50
cm) inside the screen house. Field collected pods were also placed in the cages. The
bugs were reared to adults, and date of ecdysis and head-capsule width for each instar
were recorded.
Twenty-one F4 generation general females collected over a period of 12 d were placed
individually in cages on potted common bean (BAT 41) supplemented with field collected
pods. Two males were maintained in each cage. Longevity of the adult female, dates of
oviposition, numbers of eggs per mass, and percentage eclosion of the eggs were re-
corded. Observed discrete frequency distributions were fit to theoretical mathematical
models where appropriate (Parzen 1960, Gates & Ethridge 1972). Linear regression was
used to examine the relationship between pairs of variables (SAS Institute 1985, 183-260).


Egg masses collected from the colony were found in the following locations: leaf
underside, 59%; leaf upper surface, 14%; stems, 8%; pods, 4%, and cage 14%. Seventy-six
percent of the eggs hatched. Developmental time and head-capsule width of the various
stages are summarized in Table 1.
On eclosion, the nymphs were pale yellow, but darkened after 1-2 h to a mottled
black and white color that persisted throughout their nymphal life. The newly hatched
nymphs remained on the chorion for approximately 1 d. The second instars fed on the
foliage for 3-6 d before becoming basically pod feeders. Young pods were preferred by
early instars. The gregarious habit characteristic of immature pentatomids persisted
throughout the nymphal period, although the groups became progressively smaller.
There was no overlap in the ranges of head-capsule widths between the instars (Table
Teneral adults were a pale yellow-green, becoming green after 1-2 h. The green
color darkened as the adults aged. Adult females were commonly observed to mate
more than once.
Females lived a mean ( SEM) of 44.4 2.84 d and laid 96.2 9.72 eggs in 7.5 + 0.74
masses. Eggs were laid in masses of 2-3 rows. A female would complete one row of 4-8
eggs before turning around and laying another row alongside the first, but in reverse
order. The preovipositional period was 10.0 + 1.31 d and ranged from 1 (three females)
to 23 d. The number of days between consecutive ovipositions fit the geometric distribu-
tion (Table 2; Fig. 1). P is the probability that a female A. marginatum will oviposit

Florida Entomologist 75(2)

June, 1992


Head capsule width,
Mean duration,
Stage n d SEM Mean, mm SEM Range, mm

Egg 327 6.8 0.053 -
1 99 4.5 + 0.14 0.71 0.0023 0.67-0.76
2 76 6.5 + 0.14 0.97 0.0047 0.92-1.04
3 72 6.3 0.16 1.42 0.0048 1.35-1.45
4 50 6.0 t 0.33 1.95 0.0082 1.85-2.05
5 17 12.0 + 1.2 2.51 0.017 2.40-2.60
Eggto adult 17 42.1 4.2

on any day. When the preovipositional period was not included in the analysis the data
still fit the geometric (Table 2). The data did not fit the binomial, zero-truncated Poisson,
or log zeroes distributions, other discrete frequency distributions which could be used
on this type of data (Table 2). No female was observed to lay more than one egg mass
on the same day.
The mean number of eggs per mass was 12.8 (range 3-28) (Fig. 2). There was a
marked peak at 14 eggs per mass, which may reflect the number of ovarioles (14)
possessed by A. marginatum. The data did not fit the Poisson, negative binomial,
binomial, Thomas double Poisson, Neyman type A, or Poisson-binomial distributions
regardless of whether all of the data were included in the analysis or only those data
up to and including 14 eggs per mass.
The correlations between days between consecutive ovipositions and the size of the
following egg mass or between age of the female and number of eggs per mass were
not statistically significant. For the correlation of days between oviposition and number


Value of previous column for'

Parameter, Excludes first
Distribution X2 value, or df All data oviposition

Geometric P 0.19 0.22
SX2 26.84 29.86
df 22 22
Binomial X2 666.44** 224.60**
df 12 8

Zero-truncated Poisson
Theta 0.848 0.846
Lamda 5.02 3.96
X2 295.93** 50.65**
df 10 11
Log zeroes Theta2
'** Significantly different from distribution model at 1% level.
'Theta too small to permit fit to log zeroes distribution.


Hallman et al.: Biology of Acrosternum marginatum



Days between oviposition

Fig. 1. Frequency of number of days between consecutive ovipositions (includes
preovipositional period) of A. marginatum on common beans. Expected values are for
geometric distribution with P = 0.19.

of eggs per mass, F = 0.78, P > F = 0.72, df = 19, and r2 = 0.097; for the correlation
between age of female and number of eggs per mass, F = 1.32, P > F = 0.12, df =
55, and r2 = 0.42.
Mean (SEM) percentage emergence of egg masses was 80 0.77%. Thirty-one
percent of the masses had an emergence rate of 100%. Six of 139 masses had a frequency
of emergence of s 30%; the lowest emergence was 10%. Percentage emergence was not
significantly related to the number of days since the previous oviposition (F = 0.74, P
> F = 0.76, df = 18, r2 = 0.10) nor age of the female (F = 1.24, P > F = 0.19, df =
54, r = 0.44).


Beans are a short season (2-3 months) crop, and the life cycle of A. marginatum is
relatively long. The susceptible (pod-producing) stage of bush bean varieties at an am-
bient temperature of 24C and a 12 h day length lasts about one month, insufficient time
for A. marginatum to complete more than one generation in beans, even if the eggs
were laid before flowering. With climbing beans the susceptible stage may last another
week or two, still insufficient time for producing more than one generation. Hallman et
al. (1985) showed that with infestation levels as low as one late instar A. marginatum
per 0.6 m2 of beans, significant yield loss occurred. Therefore, high infestation levels of
A. marginatum are not necessary for it to cause significant damage.
The second and third egg masses of Nezara viridula (L.) (Heteroptera: Pentatomidae)
were smaller than the first, and egg mass size gradually increased after the third egg
mass (Kiritani 1963). In our research with A. marginatum the opposite tendency was
observed, although differences in mass size were small. Mean numbers of eggs per mass

expected values

- observed values

*e -,

*e -|

*. 0oe

Florida Entomologist 75(2)

June, 1992

3 5 10 15 20 25

Number of eggs per mass

Fig. 2. Frequency of number of eggs per mass of A. marginatum reared on common

( SEM) for the first six egg masses (of those 14 females which laid six or more masses),
consecutively, were: 12.7-0.75, 15.9-+1.4, 13.81.4, 13.81.2, 12.6+0.71, and
11.8 0.74.
The fit of the number of days between successive ovipositions to the geometric
distribution indicates that daily attempts to oviposit are independent and succeed with
a probability of 0.19 (0.22 when the first egg mass laid per female is not included).
Kiritani & Hokyo (1965) present frequency distribution data on the number of days
between successive ovipositions of three pentatomids. It is obvious that their data do
not fit the geometric distribution; the geometric distribution expects the highest prob-
ability at the lowest frequency with probabilities diminishing as the frequency increases.
Their data appear to fit the normal distribution.
Kiritani (1963) found that the preovipositional period of N. viridula was 2-3 times
greater than the time interval between ovipositions after the third egg mass. In our
study with A. marginatum, the mean preovipositional period was 10.0 1.31 d, and the
mean time interval between consecutive ovipositions after the first was 4.43 0.38 d.
The difference is of the same order of magnitude as that observed by Kiritani (1963)
for N. viridula.

Hallman et al.: Biology of Acrosterum marginatum 195

100 -** ***,* ******* * ** -t*** ** *
100 % *%* 0 **- *% eee *0 Oo@0 Oee

90 0 * *C 0
eg *** *** * *
** *
80 -* .
0 .0
70 *
S.. ** .
g 60 ** *
0s* *
S50 .

S40 *

30 -

20 -
10 *

0 I I I I I
0 10 20 30 40 50 60 70

Age of female (days)

Fig. 3. Age of female A. marginatum versus percentage emergence of egg mass.

Because the number of eggs per mass was unrelated to the time interval between
ovipositions, a female could lay average or greater sized egg masses on consecutive
days. For example, one female laid a mass of 13 eggs followed by masses of 12 and 11
eggs at 2 d intervals, then masses of 16 and 12 eggs at one day intervals. Another female
laid masses of 14, 14, and 8 eggs on consecutive days, followed by a mass of 14 eggs 2
d later and a mass of 25 eggs the day after that.
Although percentage emergence was not significantly correlated to age of the female,
the tendency for higher rates of emergence to be associated with younger females cannot
be ignored (Fig. 3). In only one instance (of 31) did an egg mass from a female younger
than 15 d have < 75% emergence. Thirty-six percent of the egg masses from females
> 40 d old showed emergence of < 75%.


We thank Jim Matis, Department of Statistics, Texas A&M University, for reviewing
the manuscript and suggesting improvements.


COSTA, E. C., D. LINK, AND J. L. MARIO. 1980. Dafios causados por Nezara viridula
(L.) em feijoeiro (Phaseolus vulgaris L.). Revista do Centro Ciencias Rurais
10(4): 335-341.
COSTA, E. C., D. LINK, AND J. L. MARIO. 1981. Efeitos de niveis de Piezodorus
guildinii (Westwood, 1837) sobre feijoeiro (Phaseolus vulgaris L.) cultivar Rio
Tibagi. Revista do Centro Ciencias Rurais 11(4): 251-256.
GATES, C. E., AND F. G. ETHRIDGE. 1972. A generalized set of discrete frequency
distributions with FORTRAN program. Mathematical Geology 4(1): 1-24.
taci6n por el chinche verde de frijol Acrosternum marginatum (Palisot de
Beauvois) sobre rendimiento de Phaseolus vulgaris L.: su efecto. Turrialba 36:

196 Florida Entomologist 75(2) June, 1992

KIRITANI, K. 1963. Oviposition habit and effect of parental age upon the post-em-
bryonic development in the southern green stink bug, Nezara viridula. Japanese
J. Ecol. 13: 88-96.
KIRITANI, K., AND N. HOKYO. 1965. Variation of egg mass size in relation to the
oviposition pattern in Pentatomidae. Kontyfi 33: 427-433.
PARZEN, E. 1960. Modern Probability Theory and its Application. John Wiley & Sons,
Inc. New York. 509 pp.
ROLSTON, L. H. 1983. A revision of the genus Acrosternum Fiber, subgenus Chinadia
Orian, in the Western Hemisphere (Hemiptera: Pentatomidae). J. New York
Entomol. Soc. 91(3): 97-176.
SAS INSTITUTE. 1985. SAS/STAT guide for personal computers, version 6 edition.
SAS Institute, Cary, N. C.
SAUNDERS, J. L., A. B. S. KING, AND C. L. VARGAS. 1983. Plagas de Cultivos en
Am6rica Central: Una Lista de Referencia. Centro Agron6mico Tropical de Inves-
tigaci6n y Ensefanza Boletin T6cnico No. 9. Costa Rica. 90 pp.
TEMERAK, S. A., AND W. H. WHITCOMB. 1984. Parasitoids of predaceous and phy-
tophagous pentatomid bugs in soybean fields at two sites of Alachua County,
Florida. Zeits. Angew. Ent. 97: 279-282.
WALDBAUER, G. P. 1977. Damage to soybean seeds by South American stink bugs.
Anais da Soc. Entomol. do Brasil 6: 224-229.


'Aquatic Plant Management Laboratory
Agricultural Research Service
United States Department of Agriculture

2Fort Lauderdale Research and Education Center
Institute of Food and Agricultural Sciences
University of Florida
3205 College Ave.
Fort Lauderdale, FL 33314


We studied Neochetina eichhorniae and N. bruchi populations at 22 sites in southern
Florida during mid-summer 1985. Species composition, reproductive females (as a per-
centage of total females), and flight muscle development of N. eichhorniae populations
varied greatly among sites. N. bruchi populations were heterogeneous among sites for
these parameters and for sex ratio. The prevalent waterhyacinth (Eichhornia crassipes)
phenostage and degree of biocontrol stress accounted for much of this variation. A higher
proportion (74%) of N. eichhorniae females contained functioning ovaries when collected
from mature, healthy plants, as opposed to developing colonies (57%), or declining plant
populations (34%). The proportion of N. bruchi females with functioning ovaries was
similar on mature, healthy plants (61%) and developing colonies (54%), but less (24%)
on declining plants. More females from stressed plant populations possessed fully-de-

Center & Dray: Waterhyacinth Weevil in Florida

veloped indirect flight muscles (N. eichhorniae 9%, N. bruchi 16%) than those collected
from healthy plants (1% and 6% respectively). The proportion of the total weevil popu-
lation (i.e. both species) composed of reproductively active females of either species was
higher on mature, healthy plants than on declining plants, but N. eichhorniae populations
consistently exceeded N. bruchi. Reproductive potential declined and vagility increased
in both species on waterhyacinth suffering from weevil-induced stress. N. bruchi seemed
more sensitive to plant quality than N. eichhorniae. Both species were ubiquitous, but
N. eichhorniae was more abundant, comprising 70% of mid-summer weevil populations.


Se estudiaron las poblaciones de Neochetina eichhorniae y N. bruchi en 22 sitios en
el sur de Florida durante el medio del verano de 1985. Composici6n de species, propor-
ci6n sexual, hembras reproductivas (como un porcentaje del total de las hembras) y
desarollo de mfisculos de vuelo de las hembras de N. eichhorniae variaban much entire
los sitios. Poblaciones de N. bruchi eran heterogeneas para estos parametros y para la
proporci6n sexual entire estos sitios. El fenostadio prevalente del jacinto acuatico (Eic-
hhornia crassipes) y el grado de biocontrol causaron la mayor parte de esta variaci6n.
El estado fenol6gico del jacinto acuatico, Eichhornia crassipes, y el grado de control
biol6gico contribuyeron a esta variaci6n. Un porcentaje mas alto (74%) de N. eichhorniae
contenfan ovarios funcionables cuando fueron colectados de plants maduras y sanas en
comparaci6n a colonies en desarollo (57%) o poblaciones de plants en declive (34%). La
proporci6n de hembras de N. bruchi con ovarios funcionables fue similar sobre plants
sanas y maduras (61%) y colonies en desarollo (54%) pero menos sobre plants en declive
(24%). Mas hembras provenientes de plants de mala condici6n tenian musculos de vuelo
indirectos completamente desarollados (N. eichhorniae 9%, N. bruchi 16%) que las
colectadas desde plants sanas (1% y 6% respectivamente). La proporci6n de la poblaci6n
total de los gorgojos (i.e., ambas species) compuesta de hembras reproductivamente
activas de una de las dos species fu4 mas alto sobre plants maduras y sanas que sobre
plants en declive pero las poblaciones de N. eichhorniae consistentemente sobrepasaban
N. bruchi. El potential reproductive disminuy6 y la tendencia de dispersi6n aument6
en ambas species de jacinto acuatico las cuales sufrieron los efectos de los gorgojos.
N. bruchi parecia mas sensitive a la calidad de la plant que N. eichhorniae. Ambas
species fueron ubicuos pero N. eichhorniae fue mAs abundante, constituyendo 70% de
las poblaciones del medio verano.

Waterhyacinth was introduced into Florida during the late 19th century when surplus
plants from a garden pond were discarded into the St. Johns River near Palatka. These
infestations spread rapidly to invade water bodies throughout the state. This attractive
garden plant soon interfered with water management and use to a magnitude that
required the creation of several federal, state, and local bureaucracies to deal with the
problem (Buker 1982).
Waterhyacinth control was limited to herbicidal and mechanical means until the 1970s
when biological controls became available. In 1972, the weevil species Neochetina eic-
hhorniae Warner was introduced from Argentina. Introduction of the closely related
Neochetina bruchi (Hustache), also from Argentina, followed in 1974. Ample evidence
(Center et al. 1990) suggests that these weevils and the pyralid moth Sameodes albigut-
talis (Warren), which was released in 1977, provide substantial control, but consistent,
reliable reductions at all sites have not resulted. This variability in the performance of
the biological control agents seems at least partially due to interference from herbicidal
control (TDC pers. obs.), but may also be due to variation in plant quality. In general,
weevil population growth has seemed more satisfactory on higher-quality plants. How-
ever, high quality plants are often associated with eutrophic conditions and exhibit rapid


198 Florida Entomologist 75(2) June, 1992

growth rates. Even though weevil populations fare well under these circumstances,
their impact may be lessened by profuse plant growth. Obviously, size of weevil popu-
lations and degree of biological control are not necessarily correlated. Rather the severity
of the impacts depends upon complicated interactions among aquatic nutrient loads,
proximate composition of the plant tissue, and the physiology of the biological control
During the past few years we have been examining the relationship between plant
quality and selected characteristics of waterhyacinth weevil populations. As a prelimi-
nary step towards understanding these relationships we compared weevil populations
at several sites in southern Florida within a 3-week period during summer 1985. Our
assumption was that if these characteristics were influenced primarily by climatic factors,
then waterhyacinth weevil populations should be relatively homogeneous throughout
the area. Plant quality, which should be affected by site-specific characteristics such as
water quality, flow rates, and nutrient subsidies, should vary greatly among sites. If
plant quality also influenced the weevil populations then the latter should also vary
greatly. The characteristics that we compared included species proportions (N. eichhor-
niae vs. N. bruchi), sex ratios, reproductive capacity (frequency of females with func-
tional ovaries), and dispersive ability (frequency of weevils with fully-developed indirect
flight muscles).


Adult N. eichhorniae and N. bruchi were collected by hand from waterhyacinth at
22 sites in southern Florida during 31 July to 22 August 1985 (Table 1). We attempted
to collect between 100 and 150 weevils (both species) per sample, but the number actually
obtained varied according to their relative abundance or scarcity. The weevils were
immediately placed in a vial of 70% isopropanol then stored for one to several days until
they could be sorted (by species and sex) and dissected. The weevils were air-dried
prior to sorting because colors and markings on the vestiture are difficult to discern on
wet specimens.
The weevils were sorted and then submersed in isopropanol and dissected following
techniques described by Buckingham & Passoa (1985). Each female was classified by
reproductive status and flight muscle condition. Males were classified by flight muscle
condition alone. Females with eggs in their oviducts or with at least one fully-developed
follicle were considered actively reproductive (egg-bearing). This possibly included some
late-stage nulliparous individuals and excluded parous individuals with degenerative
ovaries. Males or females with fully-developed indirect flight muscles (Muda et al. 1981)
were considered vagile (capable of dispersing via flight).
Plant populations at each of the 22 sites were classified according to their stage of
development (phenostage) or decline status. Populations in Group A (developing colonies)
consisted of small plants with bulbous petioles and many offsets, and/or plants in which
petioles were beginning to elongate but offsets were still abundant. Group B was com-
prised of mature stands of tall, healthy plants with few offsets. Group C represented
stressed, mature stands that were declining as a result of weevil attack. We tested for
heterogeneity of proportions among sites and among phenostages using Chi2 analysis
(Sokal & Rohlf 1981, p. 731).


Among-Site Distributions

All of the waterhyacinth populations that we investigated harbored both N. eichhor-
niae and N. bruchi (Coleoptera: Curculionidae) (Table 2). N. eichhorniae constituted

Center & Dray: Waterhyacinth Weevil in Florida



No. Date Locality

8 Aug 1985
8 Aug 1985
19 Aug 1985
20 Aug 1985
20 Aug 1985
20 Aug 1985
20 Aug 1985

aSmall, healthy plants with bulbous petioles and many offsets, and/or plants on which petioles were beginning to
elongate but offsets were still abundant.
bMature stands of tall, healthy plants with few offsets.
'Mature stands stressed by weevil attack, usually with abundant feeding spots on curling leaf blades, tough
leathery texture, spindly petioles, few or no offsets, generally shorter than phenostage B plants but not so short as
phenostage A.

Phenostage A Plantsa

Tamiami Canal (US 41) near Ochopee, Collier Co.
Tamiami Canal (US 41) at Bridge 103, Collier Co.
Alligator Alley (US 75/SR 84), Broward Co.
Blue Cypress Lake, Indian River Co.
Lake Kissimmee, Osceola Co.
Shell Creek Slough, Charlotte Co.
Lake Myakka, Myakka River State Park, Sarasota Co.

Phenostage B Plantsb

Myakka Slough, Manatee Co.
Manatee River, Manatee Co.
Pierson Rd. canal at US 441, Palm Beach Co.
M-Canal (north shore), Palm Beach Co.
M-Canal (south shore), Palm Beach Co.
Loxahatchee Recreation Area, Palm Beach Co.
Conservation Area 3A, near 26-mile bend, Broward Co.

Phenostage C Plants'

Conservation Area 3A, Loop canal, Broward Co.
Lake Trafford, Collier Co.
L. Okeechobee, near Horse Island, Okeechobee Co.
Rainey Slough, near Tasmania, Glades Co.
Lake Hicpochee, Glades Co.
Okeechobee Farms Canal, Palm Beach Co.
Canal near Fellsmere, Indian River Co.
Lake Marian, Osceola Co.

20 Aug 1985
20 Aug 1985
15 Aug 1985
15 Aug 1985
15 Aug 1985
15 Aug 1985
19 Aug 1985

31Jul 1985
8 Aug 1985
13 Aug 1985
13 Aug 1985
13 Aug 1985
15 Aug 1985
20 Aug 1985
20 Aug 1985

Florida Entomologist 75(2)

June, 1992


Females females Vagile females Vagile males
% Total (% species) (% females)a (% species) (% species)

Site N Ne Nb Ne Nb Ne Nb Ne Nb Ne Nb

1 45 75.6 24.4 35.3 27.3 25.0 33.3 8.3 0 4.5 12.5
2 116 93.1 6.9 50.9 0 29.1 -- 0 -- 0 0
3 96 65.6 34.4 52.4 27.3 81.8 66.7 0 0 0 0
4 129 55.8 44.2 63.9 28.1 67.4 0 0 6.3 7.7 4.9
5 109 85.3 14.7 44.1 25.0 58.5 100.0 4.9 0 7.7 0
6 159 59.7 40.2 41.1 39.1 66.7 80.0 0 4.0 0 0
7 105 73.3 26.7 49.4 39.3 60.5 54.5 0 9.1 0 0
8 171 51.5 48.5 47.7 34.9 81.0 62.1 0 3.4 4.3 3.7
9 118 34.7 65.3 53.7 51.9 68.2 40.0 4.5 12.5 0 21.6
10 173 26.0 74.0 37.8 51.6 82.4 66.7 0 7.6 7.1 41.9
11 123 90.2 9.8 41.4 50.0 82.6 66.7 0 0 0 16.7
12 120 97.5 2.5 60.7 33.3 67.6 100.0 0 0 0 0
13 139 78.4 21.6 43.1 43.3 78.7 84.6 2.1 0 4.8 0
14 112 54.5 45.5 41.0 43.1 60.0 63.6 0 9.1 0 0
15 98 29.6 70.4 41.4 49.3 58.3 38.2 16.7 5.9 17.6 14.2
16 114 88.6 11.4 49.5 15.4 26.0 0 2.0 0 2.0 0
17 142 82.4 17.6 37.6 28.0 34.1 0 6.8 0 4.1 0
18 138 87.0 13.0 42.5 11.1 56.9 50.0 5.9 0 0 0
19 111 96.4 3.6 45.8 50.0 20.4 50.0 6.1 0 3.4 0
20 116 62.1 37.9 54.2 27.3 25.6 16.7 46.2 66.7 24.2 25.0
21 189 82.0 18.0 52.9 35.3 30.5 8.3 8.5 8.3 20.5 9.1
22 165 89.7 10.3 48.0 23.5 36.6 0 0 25.0 0 0

Mean 126.7 70.9 29.1 47.0 33.4 54.4 46.7 5.1 7.5 4.9 6.8

aThe proportion of females of either species that had eggs or mature follicles.
bMeans represent percentages averaged across sites and may differ slightly from some percentages reported in
the text that are calculated from totals summed across sites.

70.4% of the 2,788 adult weevils examined. Although N. eichhorniae was frequently
predominant, weevil populations at several sites were composed chiefly of N. bruchi
(Table 2). Consequently, species composition of weevil populations varied greatly
(Chi2= 618, p<0.001, df=21) from site to site.
Females constituted 47.5% of the N. eichhorniae populations when summed across
sites, as compared to 52.5% males (932 vs. 1031 weevils, respectively). The sexual
composition of N. eichhorniae populations was not consistent among sites (Chi2 = 36.6,
df= 21, p = 0.018). Of the N. bruchi collected, 320 (38.8%) were female and 505 (61.2%)
were male. Thus, sexual composition differed from site to site (Chi2 =45.6, p=0.001,
df= 21) for this species as well.
When we compared the relative abundances of N. eichhorniae and N. bruchi females,
we found considerable variability among sites (Chi2 = 380.0, p<0.001, df= 21). The ratio
of females of the two species could merely reflect the underlying distributional pattern
for species proportions, however. To adjust for this, we generated expected frequencies
based upon the proportion of the total weevil population represented by each species
at each site. For example, we collected 34 N. eichhorniae at site 1 (Tables 1 and 2)
which represented 1.2% of all weevils examined. Because we collected 649 female weevils,
we expected 8 (1.2% of 649) N. eichhorniae females from site 1. Based on this, the
relative frequencies of N. eichhorniae and N. bruchi females changed substantially from


Center & Dray: Waterhyacinth Weevil in Florida 201

site to site (total Chi2 =53.85, P<0.001, df=22). The relatively large pooled Chi2 (9.81,
P<0.005, df= 1) verifies our general observation that female N. eichhorniae tended to
be more abundant than female N. bruchi, but this trend was not universal (heterogeneity
Chi2=44.04, P<0.005, df=21).
Reproductive females constituted only 163 (19.7%) of the 825 N. bruchi and 486
(24.8%) of the 1963 N. eichhorniae. These represented 50.9% and 52.1% of the total
females of each respective species. The proportional representation of the two species
was extremely heterogeneous (Chi2= 198.9, p<0.001, df=21) among sites and varied
from 0% to 76% N. bruchi.
We felt that the underlying distributional pattern for females might mask the true
distributional pattern of reproductive. Consequently, we generated expected frequen-
cies for reproductive females based upon the proportions for females of each species at
each site in the manner described above. The resultant tests indicated that the proportion
of reproductive females of N. eichhorniae was not consistent among sites (Chi2 = 119.3,
P<0.001, df= 22). When the data were pooled across sites, the distribution of reproduc-
tives reflected a pattern that was similar to the overall distribution of females (pooled
Chi2 = 0.1, P>0.50, df= 1), but this resulted from the predominance of reproductive N.
eichhorniae at some sites while reproductive N. bruchi were more prevalent at others
(heterogeneity Chi= 119.2, P<0.001, df=21).
Individuals with fully-developed flight muscles made up only 6.1% (171) of the 2,788
weevils we dissected. Abundances for weevils with fully developed flight muscles (i.e.
those that were vagile) were similar for the two species: N. eichhorniae populations
harbored 88 vagile individuals, N. bruchi populations harbored 83. Vagile weevils con-
stituted a greater proportion (10.1%) of the N. bruchi population, however, than of the
N. eichhorniae population (4.5%).

Influence of Phenostage on Species Composition

We examined the influence of waterhyacinth phenostage on the composition of
Neochetina populations, and found that species proportions varied substantially among
phenostages (Table 3). The distribution of N. bruchi among phenostages contributed
most to the overall Chi2 value. N. bruchi frequencies were higher than expected on
mature healthy plants (384 observed vs. 283 expected, Chi2 = 36.1) and lower than ex-
pected on stressed plants (224 observed vs. 318 expected, Chi2= 30.2). Species propor-
tions on young waterhyacinth colonies were very representative of overall proportions
and contributed little to the overall Chi2.
We also tested the hypothesis that species proportions approximated a 1:1 ratio.
Results of this test (Table 3) show that N. eichhorniae was consistently over-represented
in each of the three phenostages. The magnitudes of the deviations varied considerably,
however. This is reflected in the large heterogeneity component of the analysis. Although
N. eichhorniae were consistently more abundant than N. bruchi, their proportions were
most similar on mature healthy plants.

Influence of Phenostage on Sexual Composition

Sex ratios for N. eichhorniae varied little among phenostages (Table 4), so we tested
the data for conformity to a theoretical 1:1 sex ratio. The results were ambiguous, but
suggested that it was not unreasonable to assume that sex ratios approximated 1:1 in
populations from all phenostages. None of the Chi2 values for the three respective
phenostages exceeded the critical value of 3.84 for significance at the 5% level. The total
Chi2 and the heterogeneity Chi2 again indicated that sex ratios were homogeneous among
plant phenostages and conformed to a 1:1 ratio. However, the pooled Chi2 of 5.0 was

Florida Entomologist 75(2)

June, 1992

o C)

C 10 t-
0 < L6
3q (Dt t~












C I)





51 II


I 1I

- 4'

l -4 '- I CIN



M cc
10 10 tO
t- 0a 0

4 100

S t 0-

CP7 Q o



2 -r



r- l -4 -4 CO -l4

Center & Dray: Waterhyacinth Weevil in Florida

44 00 ;:L4

o r- CO Lo0 O





















00C Cl -
- 0 to
CI Co q

'C M 0



' A

Cl i- Co C0
0C 0 CD

-4 O I I

-E 0

4. 0 00


Lt- Om COD
Cl Cl C 8

Florida Entomologist 75(2)

June, 1992

significant suggesting that the total sex ratio calculated by summing across phenostages
did deviate from 1:1. Hence, it seemed that the sex ratio ofN. eichhorniae was independ-
ent of the plant phenostage and was generally constant at approximately a 1:1 ratio.
The trend of increased Chi2 values with progressively later phenostages suggests that
the sex ratio deviated from 1:1 as plants became stressed, but our data failed to clearly
demonstrate this.
N. bruchi sex ratios differed considerably among phenostages (Table 5), but over
half of the contribution to Chi2 came from mature, healthy plants (which harbored 46%
females). When we tested for conformity to a 1:1 sex ratio, we found good agreement
for populations from healthy, mature plants, but not from young plants, or from stressed
plants. Consequently, sex ratios varied greatly among phenostages. Hence, where N.
eichhorniae sex ratios seemed unaffected by plant type, N. bruchi populations seemed
quite sensitive to it. Populations of N. bruchi favored males when from young plants
or stressed plants (68.7% and 66.5% males respectively), but tended toward a 1:1 sex
ratio on healthy, mature plants (only 53.9% males).
The ratio of female N. eichhorniae to female N. bruchi varied according to plant
phenostage (Table 6). When we calculated expected values in a manner that eliminated
the underlying distributional patterns of the two species, this ratio was still highly
variable among phenostages. Female N. eichhorniae were predominant on all phenos-
tages. However, on mature, healthy plants the representation of females of either species
(21.6% N. eichhorniae and 14.1% N. bruchi) reflected overall species proportions on
this phenostage (20.5% and 13.8%, respectively). Otherwise, the ratios were distinctly
different, with N. bruchi females less abundant than predicted on young waterhyacinth
(97 expected vs. 68 observed: Chi2= 10.8, p<0.001, df = 1) and on stressed plants (100
expected vs. 75 observed: Chi2=7.1, p<0.01, df= 1).

Influence of Phenostage on Reproductive Status of Females

Relative abundances of reproductively active N. eichhorniae and N. bruchi females
exhibited considerable disparity among the three phenostages (Table 7), with mature,
healthy plants hosting the highest frequencies of N. bruchi. Reproductive females rep-
resent a sub-set of the total females, however, so the above results could merely reflect
the distribution of females. We therefore compared the distributions of reproductive
females among phenostages, by calculating expected values from the percentages that
females of each species in each phenostage contributed to the overall female weevil
population. Populations on mature, healthy plants deviated from predicted frequencies,
mainly due to an overabundance of N. eichhorniae (201 observed vs. 139 expected).
Both species were less numerous on stressed plants than predicted (N. eichhorniae: 135
observed vs. 206 expected, N. bruchi: 18 observed vs. 39 expected). Proportions for
reproductive females collected from young plants (23.1% N. eichhorniae and 5.7% N.
bruchi) closely paralleled the corresponding populations of total females (21.1% and 5.4%

Influence of Phenostage on Flight Muscle Development.

The low incidence of fully-developed flight muscles resulted in mean values for either
species from each phenostage that seldom differed from zero or from one another.
Although we wanted to test for heterogeneity among sites, results were difficult to
interpret due to the high frequency of cells with expected counts of less than 5. We
obtained a valid test only by combining sexes and cross-tabulating the data by species
and phenostage. The results indicated that the relative numbers of flight-muscled weevils
of either species were not the same on the three phenostages (Table 8).


Center & Dray: Waterhyacinth Weevil in Florida

m 1-14



Pi -iO-


(0 .i 1
10 ti r-

r-4 l 1- N C l

m :


.2 -


-4 (0 I 1-

0 00 I 0

cq CO N CO

L-4 C

< l -4



Florida Entomologist 75(2)

June, 1992

P. P6

00 e0 00 -
0 t- 0 OS 0



,# O. I I ,t
CF 1 t-I -

-4 1 I C


(N L-- CO
c' I:- c- i

i-l 4

0 N
C11 N M

Center & Dray: Waterhyacinth Weevil in Florida

&< LCL

m t- -r.
*-1 0 10 0 CO
IN c C(0

,-4 '- '- CO Nt(


















< a) -s /

c4 -

a )



s? ^




-I -4 ,-4


+. 0L

I:- C0 0
CID 4 -

) 10 CO
*-4 N r-



00 m L O cN

Florida Entomologist 75(2)

June, 1992

c 4 -
t- M 00








CL I--:
^ 5

a l

>m m

" II




e s


a d



00 0

1- c-

o 00
- I I eq




QLO O t-

C) 0

cC Ci 0


C ci

0o 00

< M o





(0 0 1-
i- eq

.-4 1-4 1-4 M C (

Center & Dray: Waterhyacinth Weevil in Florida 209

This was not surprising, however. Flight muscle incidence could be expected to
reflect underlying distributional patterns of species proportions, which were heterogene-
ous. We therefore generated expected frequencies based upon the percentage of all
weevils comprised by each species on each phenostage. Based on this, we recovered
more vagile N. bruchi (83) and fewer vagile N. eichhorniae (88) than expected (51 and
120 respectively, see also Table 8). Both species were under-represented on young plants
(N. eichhorniae: 10 observed vs. 33 expected, N. bruchi: 6 observed vs. 13 expected),
as was N. eichhorniae (9 observed vs. 35 expected) on mature, healthy waterhyacinth.
N. bruchi was more abundant than predicted on mature, healthy plants (50 observed
vs. 24 expected) and on stressed plants (27 observed vs. 14 expected). Stressed plants
also harbored more vagile N. eichhorniae (69) than expected (52).
Despite the overabundance of vagile weevils of both species on stressed plants, the
difference between the two species was most pronounced on mature plants. Within each
species, the incidence of developed flight muscles differed from the distribution of that
species among phenostages. These differences were more pronounced in N. eichhorniae
than in N. bruchi populations. While 28%, 29%, and 43% of the N. eichhorniae were
found on young, mature, and stressed plants, respectively, the corresponding proportions
for flight-muscled individuals were 11%, 10%, and 78% (Chi2=70.7, p<0.001, df=2).
Likewise, while 26%, 47%, and 27% of the N. bruchi were found on the three respective
plant types, the corresponding proportions for flight- muscled individuals were 7%, 60%,
and 32% (Chi = 15.7, p<0.001, df=2). In the former case, the greatest contribution to
Chi2 (73%) was from stressed plants where observed numbers greatly exceeded expected
numbers (69 vs. 38). In the latter case the greatest contribution (72%) was from young
plants where expected numbers exceeded observed numbers (27 vs. 6).
Relatively high numbers of flight-muscled N. bruchi were found on both mature and
stressed plants. In contrast, N. eichhorniae showed relatively high frequencies on only
the latter phenostage. The highest percentages of flight-muscled N. bruchi males oc-
curred on mature plants (2.0%, 17.9%, and 10.1% on young, mature and stressed plants,
respectively), but highest percentages were on stressed plants for females (4.4%, 7.3%,
and 16.0%, respectively). Highest percentages of flight-muscled N. eichhorniae occurred
on stressed plants for both sexes (males: 2.5%, 2.3%, and 7.1%; females: 1.1%, 0.7%,
and 9.3%, respectively). This suggests that male N. bruchi are capable of dispersing
earlier than N. eichhorniae or female N. bruchi.


N. eichhorniae was first released in Florida on 23 August 1972. After initial field
colonies were successfully established, several agencies helped distribute it to additional
sites. This labor-intensive effort was deemed necessary because this species (and its
sibling, N. bruchi) was thought incapable of flight, thus tending to be philopatric. As a
result, our records indicate that this species was released at 199 sites in Florida (unpub-
lished data). When N. bruchi later became available, these agencies didn't mount a
similar dissemination campaign. As a result, this insect was released at only 21 sites
statewide (unpublished data). We found, as a result of our survey, that both species of
weevils are now ubiquitous in southern Florida, despite the disparity in degree of
human-assisted dispersal. This suggests that N. bruchi, at least, is less philopatric than
originally presumed.
In a seminal paper, Buckingham & Passoa (1985) explained the apparently flightless
nature of these weevils. They discovered that both Neochetina species undergo reversible
flight muscle generation and degeneration. They theorized that flight muscle generation
alternates with reproductive periods whereby the ovarian follicles are resorbed as flight
muscles are produced, and vice-versa. They also demonstrated that temperature influ-
ences flight muscle development.

Florida Entomologist 75(2)

June, 1992

In our study, the fact that site conditions accounted for a significant portion of the
variation in weevil populations suggests that plant quality might also influence odgenesis,
flight muscle generation, and population vitality (the proportion of the overall weevil
population comprised of reproductively active females of either species) at least during
mid-summer. The relative abundance of the two species seemed particularly influenced
by site characteristics. The species exhibited differential responses to site conditions
that were manifested by altered sex ratios, reproductivity, and flight muscle develop-
ment. Plant phenostage and degree of biocontrol stress were two site characteristics
that influenced these attributes. N. bruchi populations seemed especially sensitive to
these plant differences and presumably reacted more strongly.
The most reproductively active N. bruchi populations occurred on mature, unstressed
plants, those that we would classify as the highest "quality" plants. This impression
was reinforced by the fact that waterhyacinth mats harboring the three most vigorous
N. bruchi populations abutted agricultural areas. Plants in these mats were obviously
flourishing, likely as a result of nutrient runoff. This suggests that subjective evaluations
of plant quality reflect nutrient supplies available to the plants and, in turn, to the
weevils. Also, plants growing in crowded conditions, when subjected to high levels of
weevil attack, can exhibit reduced tissue nitrogen concentrations (Center & Van 1989).
N. bruchi then may have a higher nutritional requirement than N. eichhorniae, a fact
that would be consistent with its lesser tolerance for heavily attacked plants. A higher
nitrogen requirement also might be attributable to the presumed greater fecundity of
N. bruchi as compared to N. eichhorniae (DeLoach & Cordo 1976).
Plant traits influenced whether N. eichhorniae adults were reproductive or disper-
sive. This effect was heightened for N. bruchi. Whereas reproductively active N. eichhor-
niae females were 45% less abundant on stressed plants relative to mature healthy ones,
reproductively active N. bruchi females were 80% less abundant. The data suggest that
this differential was due to the early dispersal ofN. bruchi as plant quality deteriorated.
Also, whereas N. eichhorniae population vitality was reduced primarily on stressed
plants, N. bruchi population vitality was lowered on both young, developing plant colonies
and on stressed, older colonies. The reason for this is unclear but it may be related to
herbicidal control efforts. During summer most young waterhyacinth colonies represent
regrowth from earlier herbicide treatments. The insects on these may represent residual
populations that survived despite the ensuing plant degradation caused by herbicide
application. N. eichhorniae, by being less sensitive to declining plant quality, may
therefore be more persistent at treated sites than N. bruchi and thereby be more
available to reinfest the plants as they recover.
In summary, we theorize that waterhyacinth and weevil growth cycles are inter-re-
lated. Incipient waterhyacinth populations are colonized by dispersive individuals or by
residual populations depending upon the history of the infestation and recency of aquatic
plant control operations. As the plant infestation matures, both weevil species build
active, vigorous populations. With sufficient time, the weevils may become excessive,
causing plant quality to deteriorate. The reduced plant quality, in turn, perhaps contrib-
utes to follicular resorption and generation of flight muscles. N. bruchi, being more
sensitive than N. eichhorniae, disperse earlier. As a result, the proportions of the two
species in the total weevil population increasingly favor N. eichhorniae. Sexual compo-
sition then increasingly favors males on older, highly stressed waterhyacinth stands
(see also Center & Durden 1986). N. eichhorniae, by being more persistent, is probably
the better biological control agent despite the presumed higher fecundity of N. bruchi.


We wish to thank N-Y. Su and M.J. Grodowitz for reviewing this manuscript, and
F.W. Howard for translating the abstract. W.C. Durden and M.D. Bouhadana assisted


Center & Dray: Waterhyacinth Weevil in Florida

with the research and V. Chew provided statistical assistance. This study was conducted
through Specific Cooperative Agreement No. 58-7B30-3-586 between the United States
Department of Agriculture, Agricultural Research Service and the University of Florida,
Institute of Food and Agricultural Sciences.


BUCKINGHAM, G., AND S. PASSOA. 1985. Flight muscle and egg development in
waterhyacinth weevils, pp. 497-510 in Delfosse, E.S. [ed.), Proc. VI Int.
Symp. Biol. Control Weeds, 19-25 August 1984, Vancouver, British Columbia.
Agriculture Canada, Ottawa.
BUKER, G. E. 1982. Engineers vs. Florida's green menace. Florida Hist. Q. 1982
(April): 413-27.
CENTER, T. D. 1982. The waterhyacinth weevils Neochetina eichhorniae and N.
bruchi. Aquatics 4(2): 8-19.
control of aquatic and wetland weeds in the Southeastern United States, pp.
239-62 in Delfosse, E.S. [ed.], Proc. VII Int. Symp. Biol. Contr. Weeds,
6-11 March 1988, Rome, Italy. Ist. Sper. Patol. Veg. (MAF).
CENTER, T. D., AND W. C. DURDEN. 1986. Variation in waterhyacinth/weevil interac-
tions resulting from temporal differences in weed control efforts. J. Aquat. Plant
Manage. 24: 28-38.
CENTER, T. D., AND T. K. VAN. 1989. Alteration of water hyacinth (Eichhornia
crassipes (Mart.) Solms) leaf dynamics and phytochemistry by insect damage and
plant density. Aquat. Bot. 35: 181-95.
DELOACH, C. J., AND H. A. CORDO. 1976. Life cycle and biology of Neochetina
bruchi, a weevil attacking waterhyacinth in Argentina, with notes on N. eichhor-
niae. Ann. Entomol. Soc. Am. 69: 643-52.
MUDA, A. R. B., N. P. TUGWELL, AND M. B. HAIZLIP. 1981. Seasonal history and
indirect flight muscle degeneration and regeneration in the rice water weevil.
Environ. Entomol. 10: 685-90.
SOKAL, R. R., AND F. J. ROHLF. 1981. Biometry. W.H. Freeman and Co., New
York, N.Y. 859 pp.

212 Florida Entomologist 75(2) June, 1992


'University of Florida, Institute of Food and Agricultural Sciences,
Fort Lauderdale Research and Education Center
3205 College Ave., Fort Lauderdale, FL 33314

2Centro Agronomico Tropical de Investigacion y Ensefanza,
Turrialba, Costa Rica


Freshly felled cabbage palmettos, Sabal palmetto (Walter), were attractive to Rhyn-
chophorus cruentatus (Fabricius) adults for at least 35 d. Chopped S. palmetto crown
tissue placed in 19-liter plastic buckets attracted weevils, but for less than 35 d. Captured
weevils did not always remain in the harborage of chopped tissue and were able to
escape from uncovered buckets. Baffles constructed from 5 cm long polyvinyl chloride
tubes glued together longitudinally and placed over bucket openings effectively pre-
vented escape of R. cruentatus adults. Buckets covered by baffles and baited with S.
palmetto tissue were effective at trapping and preventing escape of weevils. Using this
trap design, we determined in laboratory and field assays that S. palmetto crown tissue
is most attractive to R. cruentatus adults within 72 h after harvest from a healthy tree.
The optimal time for collection and analysis of volatile compounds from palm tissue
which are attractive to weevils appeared to be 24-72 h after palm harvest. The trap
described here has several potential applications for the biological study, detection, and
management of Rhynchophorus spp.


Palmetos de repollo [S. palmetto (Walter)] recientemente tumbados atraian a los
adults de Rhnychophorus cruentatus (Fabricius) durante por lo menos 35 dias. Tejido
picado de hojas colocado en baldes plasticos de 19-litros atrala a los picudos, pero durante
menos de 35 dias. Picudos capturados no siempre se quedaban en el tejido picado y eran
capaces de escapar de los baldes. Deflectores construidos de tubos de polivinilo de 5 cm
de longitud pegados a lo largo y puestos sobre las aperturas de los baldes prevenian
efectivamente el escape de adults de R. cruentatus y R. palmarum (L.). Baldes cubier-
tos por deflectores y utilizando tejidos de palmas como cebo capturaban y prevenian el
escape de R. cruentatus en el campo. Utilizando este disefio de trampa, determinamos
en ensayos de laboratorio y de campo que el tejido de las hojas de S. palmetto es de
mayor atracci6n a los adults de R. cruentatus dentro de 72 horas despues de estar
cortado de una palma sana. Los resultados de este studio indican que el period 6ptimo
para colecci6n y analysis de los compuestos volatiles los cuales emanan de los tejidos de
palmas y son atractivos a R. cruentatus es de 24-72 horas despues de cortarlos de la
palma. El disefio de la trampa descrita tiene varias aplicaciones para el studio biol6gico,
detecci6n y control de Rhynchophorus spp.

The largest weevil in the continental United States is the Palmetto weevil, Rhyn-
chophorus cruentatus (Fabricius) (Woodruff 1967). This relatively rare insect breeds in
a variety of palms that are stressed or have been recently transplanted (Giblin-Davis

Weissling et al.: Palmetto Weevil Trap


& Howard 1989). Rhynchophoruscruentatus is confined to the southeastern U.S. (Wat-
tanapongsiri 1966), sympatric with the native cabbage palmetto [Sabal palmetto (Wal-
ter)] (Woodruff 1967).
The closely related American palm weevil (R. palmarum [L.]) has been reported
from southern Texas and California (based on the collection of two specimens), but its
distribution extends primarily southward from Mexico and Cuba into South America
and the West Indies (Wattanapongsiri 1966). Rhynchophorus palmarum is the primary
vector of Bursaphelenchus cocophilus (Cobb). This nematode is responsible for red ring
disease, an important malady of coconut (Cocos nucifera L.) and African oil (Elaeis
guineensis Jacq.) palms in the Neotropics (see Giblin-Davis 1991 for distribution of red
ring disease).
It has been suggested that R. cruentatus could be a capable vector for the red ring
nematode if the nematode were ever introduced into southern Florida (Gerber et al.
1990, Giblin et al. 1987). Esser (1969) noted that red ring disease could easily be imported
into Florida from infested areas. Considering the close proximity of this disease to the
southeastern U.S. (< 400 km from its closest confirmed point in the Yucatan peninsula
of Mexico) it is important to learn as much about the biology, detection, and management
of Rhynchophorus spp. as possible.
Rhynchophorus adults are strongly attracted to stressed or dying palms (Wat-
tanapongsiri 1966) and also appear to rely on semiochemicals for aggregation (Rochat
et al. 1991a, 1991b, Oehlschlager et al. 1991). Trapping with coconut stem tissue as bait
has been successful for monitoring the seasonal population dynamics of R. palmarum
in coconut (Hagley 1963) and oil palm (Schuiling & van Dinther 1981) plantations. Griffith
(1987) suggested that traps baited with Lannate-treated coconut palm tissue would
effectively control R. palmarum adults and red ring disease, although quantitative
support of this method is not available. Similarly, Chittenden (1902) suggested that
stressed S. palmetto and Phoenix canariensis Hortorum ex Chabaud tissue is attractive
to R. cruentatus adults, therefore, tissue-baited traps could be used to monitor this
The objectives of this study were to verify that R. cruentatus adults are attracted
to stressed S. palmetto tissue, to develop traps to capture and retain live R. cruentatus
and R. palmarum adults, and to use an optimized trap to temporally quantify attractive-
ness of cut S. palmetto tissue to R. cruentatus adults.


Attraction to Felled Trees

Five mature S. palmetto trees (4-7 m tall) were felled within a grove near West
Palm Beach, Palm Beach Co., Florida and periodically observed for R. cruentatus adults
attracted to cut surfaces. The trees (spaced approximately 10 m apart) were cut on 12
July 1990 at about 0.5 m above the ground. The crown (distal growing portion) was then
severed and dragged 0.5 m from the felled stem (trunk), resulting in four exposed tissue
cross-sections (stump, basal stem, distal stem, and basal crown). Leaves trimmed from
the crown were used to cover cut surfaces and provide shelter for weevils. Adult R.
cruentatus on each surface were removed, counted, and sexed at 1, 4, 6, 8, 11, 14, 19,
and 35 d after felling. Data were subjected to /(X + 0.5) transformation and analyzed
by the Statistical Analysis System general linear models procedure (SAS Institute 1985)
for differences between surfaces within each collection period. Least significant differ-
ence tests (SAS Institute 1985) were used for mean separation where significant (P -
0.05) effects occurred. Untransformed means are presented.

214 Florida Entomologist 75(2) June, 1992

Trapping Studies

Nineteen-liter black plastic buckets baited with 3.7 kg of chopped S. palmetto crown
tissue were used to trap R. cruentatus adults. Two 5-mm diam. holes drilled in the
bottom of each bucket provided drainage. On 23 August 1990, three trap buckets (uncov-
ered) were set on the ground approximately 10 m apart in a S. palmettogrove at the
West Palm Beach site. All R. cruentatus adults in the buckets were collected and sexed
1, 4, 6, 10, and 15 d after installation.
To determine if closures could be used to retain captured weevils, an empty bucket
covered with parallel steel wires (3-mm diam.) spaced 2.5 cm apart was compared with
an empty, uncovered bucket. In the first test, 10 R. cruentatus adults were placed in
each bucket and observed for 30 min. for the following responses: successful flight out
of trap, successful walking out of trap, attempted but failed flight out of trap, and
attempted but failed walking out of trap. This test was repeated except that the inside
surface of the top 10 cm of each trap was coated with a thin layer of petroleum jelly,
11 weevils were placed in each bucket, and responses were observed for 1 h. The
temperature during both tests was 32C.
Based on the results of the above experiment, a different trap top was designed to
retain captured weevils. Tubes made from polyvinyl chloride (PVC) with inside diameters
(ID) of 1.8, 2.4, 2.6, and 2.9 cm were cut into 5 cm lengths (two for each size) and used
to determine what diameter could be used to prevent weevils from walking out. For
each size, the inside bore of one tube of the pair was coated with a thin layer of a 1:1
mixture of petroleum jelly and mineral oil. Tubes were vertically oriented, placed side
by side, and a R. cruentatus adult was placed in each, posterior first. Tubes were
observed for escaping beetles for 30 min. This test was repeated six times.
Baffles designed to prevent captured weevils from escaping were constructed by
gluing PVC tubes (2.4 cm ID) together longitudinally to form 32-cm diam. lids that
covered 19-liter buckets (Fig. la). The inside bore of each tube of each baffle was treated
with the 1:1 grease mixture, described above, and used for all subsequent tests with
R. cruentatus.
A bucket covered with a baffle was compared with an uncovered bucket for prevention
of escape by flight of R. cruentatus. A coating of the grease mixture on the top 10 cm
of the inside surface of both buckets prevented weevils from walking out. Traps contain-
ing 10 field-collected weevils were placed in a room maintained at 32 + 2 C with constant
light. After 24 h, the number of weevils remaining in each bucket was determined. This
test was replicated four times. Differences between percent of weevils escaping from
each trap type were analyzed by the Kruskal-Wallis test (X2 approximation) (SAS Insti-
tute 1985).
To determine if the baffle design was effective for capturing R. cruentatus adults,
four traps baited with 2.5 kg chopped S. palmetto crown tissue were placed in the Big
Cypress National Preserve (Collier Co.) in an area interspersed with mature S. palmetto
and saw palmetto [Serenoa repens (Bartr.)]. Each baffled trap was treated with the 1:1
grease mixture and suspended from a holder constructed of a 0.95 cm (diam.) metal rod
1.2 m high with a 0.6 m side arm welded to the top. The longer vertical metal rod was
cemented into a cinder block (Fig. lb). Traps were separated by 20 m and baited on 14
May 1991. Weevils in each trap were collected after 7 d and sexed.

Tissue Age Bioassays

Baffled traps baited with S. palmetto crown tissue and placed in S. palmetto groves
at two locations in southern Florida (one in Davie, Broward Co., and one in Big Cypress
National Preserve, Collier Co.) were used to quantify weevil response to aging tissue

Weissling et al.: Palmetto Weevil Trap 215

a. b.

25 em above the bottom. Palm tissue (2.5 kg/trap) was placed beneath this barrier, thus
preventing weevil contact with, and possible chemical modification of tissue. Moist paper
towels were placed above the mesh barrier to provide refugia for trapped beetles. For
each replicate (two per location), four traps were suspended from holders (described
above) 20 m apart and captured weevils were counted, sexed, and released (50 m from
the trap site) daily for 6 days following installation. Tests were conducted in 1991 from
29 May to 4 June and 14 to 20 June in Collier Co., and 7 to 13 June and 23 to 29 June
in Broward Co. Data were analyzed by the SAS general linear models procedure for
differences in mean weevil capture between days and means were separated by the
least significant difference test (P s 0.05, SAS Institute 1985).
Two bioassay arenas (2.4 x 1.2 x 1.2 m: L x W x H) were constructed and used
to quantify R. cruentatus adult response to aging palm tissue under controlled conditions.
Rigid arena sides and bottoms were painted white and removable tops were constructed
of 1-mil clear polyethylene sheeting. In addition, the inside of each arena was lined with
1-mil polyethylene sheeting which was periodically replaced to prevent carry-over de-
sorption from volatile chemicals. Air-flow in each arena (8 cm/sec; measured by movement
of smoke) was achieved by placing a 25.4 cm in-line duct fan in a 25.4 cm duct coupled
to one 1.2 m2 outlet. Flexible tubing (10 cm diam.) was used to vent air passing through
arenas to the outside of the laboratory. The air-inlet end was covered with 14-mesh
fiberglass screen.
One baited and one unbaited (check) baffled trap were placed in each arena at the
air inlet end 20 cm from walls and each other. For each of the six replications of the
experiment, baited traps received 2.5 kg of fresh shredded S. palmetto crown tissue
dissected from 7.5 cm above the growing point to 7.5 cm into the woody stem. A mesh

216 Florida Entomologist 75(2) June, 1992

barrier and moist paper towel was placed in each trap. The outside surface of the bottom
half of each bucket was treated with a thin layer of the grease mixture to prevent
weevils from walking up buckets. For each test, either 15 male or 15 female laboratory-
reared (Giblin-Davis et al. 1989) or recently field-collected (Hendry Co., Florida) weevils
were placed in a clear plastic release container (31 x 23 x 10 cm: L x W x H) located
in each arena near the air outlet end. The inside walls of each release container were
greased to insure an exit by flight. After 24 h, weevils were removed from each trap
and arena and allowed to "recover" for 24 h with an apple slice and moist paper towels,
and were then placed back in the arena. During the "recovery" period another group
of weevils was placed within each arena as above. Weevils were reused because a limited
number were available. Attractancy bioassays were conducted with palm tissue aged
24, 48, 72, 96, 120, and 144 h after dissection from trees. During tests, temperature
was maintained at 32 2 C, and relative humidity ranged from 40-70%. Continual light
was provided by fluorescent fixtures (1.0 Lux).
For each 24-h period, mean percent response of weevils to baited and unbaited traps
was calculated as the number caught in each trap per total number flying (total number
released number remaining in release container). The t-test procedure (P < 0.05, SAS
Institute 1985) was used to compare mean percent response between baited and unbaited
traps during each 24-h period.


Attraction to Felled Trees

Rhynchophorus cruentatus adults were collected from one or more cut surfaces of
S. palmetto trees for at least 35 d after felling (Table 1). There was a significant difference
in the number of weevils collected from surfaces on the first (F = 4.95; df = 3,12; P =
0.018) and fourth (F = 3.94; df = 3,12; P = 0.048) day after trees were felled. On day
one and four, significantly more weevils were collected from the distal stem than any
other surface.


Mean' per cut surface SEM
after Stump Basal Stem Distal Stem Basal Crown
cutting No. :3d No. 9 : No. :&: No. :d

1 0.00.0b 0.00.0b 1.20.5a 1:1 0.20.2b 1:0
42 1.0-0.0b 1:1 1.00.5b 3:2 3.40.9a 9:8 0.60.6b 1:2
6 3.21.6a 1:1 0.00.0a 1.00.6a 3:2 1.40.6a 3:4
8 0.00.0a 0.20.2a 0:1 0.4-0.2a 3:1 0.80.5a 3:1
11 0.20.2a 1:0 0.00.0a 0.00.0a 0.60.4a 1:2
14 0.20.2a 0:1 0.00.0a 0.20.2a 0:1 0.00.0a -
19 0.00.0a 0.0-0.0a 0.80.6a 3:1 0.00.0a -
35 0.00.0a 0.00.0a 0.60.6a 0.00.0a

'Means within a row followed by the same letter are not significantly different (least significant difference, P s 0.05).
2N = 2 for stump.
3Weevils burrowed into tissue and could not be sexed.

Weissling et al.: Palmetto Weevil Trap 217

Trapping Studies

Rhynchophorus cruentatus adults were caught in uncovered buckets baited with
chopped S. palmetto crown tissue that were placed in the field for 15 days (mean no.
per trap per d + SEM, females: 0.7 0.2; males: 0.6 0.3).
Observation of weevils placed in an ungreased and uncovered bucket or a similar
bucket covered with parallel steel wires spaced 2.5 cm apart demonstrated that weevils
could escape from both. Nine of ten weevils escaped from the uncovered bucket within
30 min.; 8 flew out and 1 walked out. Three weevils escaped from the wire-covered
bucket within 30 min. by walking out but none flew out despite 24 attempts. Application
of the 1:1 petroleum jelly/mineral oil mixture to the inside surface of buckets eliminated
escape by walking out. However, within 1 h, seven and five weevils flew out of the
uncovered and covered greased buckets, respectively.
Rhynchophorus cruentatus adults were able to walk out of ungreased individual
PVC tubes of 1.8, 2.4, and 2.6 cm ID, and greased tubes of 1.8 and 2.6 cm ID. Baffles
for subsequent tests were constructed from greased 2.4 cm ID PVC tubes because this
size prevented escape by walking while minimizing tube diameter and decreasing the
possibility of escape by flight (wing span range 2.1 3.4 cm).
After 24 h, no R. cruentatus adults escaped from buckets covered by a greased baffle
while 85 5 percent of the weevils placed in uncovered buckets escaped. This difference
was significant (X2 = 6.40, df = 3, P = 0.01).
Rhynchophorus cruentatus adults were caught in baited buckets covered with baffles
that were placed in the field for 7 d (mean number per trap per d + SEM, males = 0.6
0.1, females = 0.8 0.3).

Tissue Age Bioassay

When tested in the field, the number ofR. cruentatus adults caught in traps baited
with chopped S. palmetto crown tissue varied through time (F = 2.46; df = 5,87; P =
0.04). Significantly more weevils were caught in traps on day two than at any other
time (Table 2).
In the laboratory, most R. cruentatus males and females were caught in baffled
traps baited with chopped S. palmetto crown tissue from 24 to 72 h after harvesting
(Fig. 2). Weevil response to baited traps was significantly greater than to unbaited
traps between 24 and 72 h (males, 24 h: t = 4.00; df = 5; P = 0.010, 48 h: t = 3.59; df
= 5; P = 0.016, 72 h: t = 5.06; df = 5; P .= 0.004, females, 24 h: t = 2.86; df = 5; P =
0.036, 48 h: t = 2.76; df = 5; P = 0.040, 72 h: t = 2.72; df = 5; P = 0.042). In addition,
response of females to traps baited with 120 h aged crown tissue was significantly greater
than to unbaited traps at 120 hours (t = 2.61; df = 5; P = 0.048).


Days after Mean' No./Trap
cutting (n=4) SEM 9:6

1 0.1 0.1b 0:1
2 0.6 0.3 a 1:8
3 0.2 0.1 b 2:1
4 0.1 0.1b 0:1
5 0.1 0.1b 0:1
6 0.0 O.Ob

'Means followed by the same letter are not significantly different (least significant difference, P < 0.05).

Florida Entomologist 75(2)



24 48 72 96 120 144


Fig. 2. Percentage of R. cruentatus adults flying in arenas and caught in unbaited
traps and traps baited with 2.5 kg chopped S. palmetto crown tissue aged 24, 48, 72,
96, 120, and 144 h. (*) indicates significant difference between baited and unbaited traps,
paired t-test, P- 0.05.

Weevils in the genus Rhynchophorus attack stressed and damaged palms (Wat-
tanapongsiri 1966). It was suggested by Chittenden (1902) that fermented sap exuding
from dead or wounded palms acts as a strong attractant for R. cruentatus adults. Results
of this study substantiate earlier observations that R. cruentatus adults are attracted
to volatile compounds emitted by stressed S. palmetto trees and suggest that olfaction














June, 1992

Weissling et al.: Palmetto Weevil Trap 219

plays an important role in orientation of adults to their host plants. The chemical ecology
of Rhynchophorus spp. has, until recently, received little attention. Hagley (1965) em-
pirically tested several organic compounds for their ability to attract R. palmarum
adults in an olfactometer. He found that malt extract, skatole, and isoamyl acetate were
attractive in the laboratory, and that a mixture of these compounds was more attractive
in the field than coconut stem tissue. This mixture was, however, an ineffective R.
cruentatus attractant when tested in the field (unpub. data). Rochat (1987) reported
that R. palmarum females responded strongly (in an olfactometer) to fermented oil
palm sap and that males responded similarly to the palm sap and to a skatole, isoamyl
acetate, and ethanol mixture. Furthermore, Moura et al. (1989) and Rochat et al. (1991a)
provided evidence that R. palmarum males produce an aggregation pheromone attrac-
tive to both males and females. Rochat et al. (1991b) identified the major component of
the pheromone as (2E)-6-methyl-2-hepten-4-ol, however, the chirality of this compound
is now known to be 4(S) (Oehlschlager et al. 1991). Both the racemic and 4(S) chiral
isomer of the pheromone are highly attractive in the field when used in conjunction with
food sources (Oehlschlager et al. 1991). Preliminary olfactometer results suggest that
R. cruentatus males also produce an aggregation or arresting pheromone (unpub. data)
but at this time, it is not known how or if this compound interacts with palm-produced
attractants to influence trap catch.
When placed in traps, chopped S. palmetto crown tissue was most attractive to R.
cruentatus adults within the first 72 h after harvesting. Results of these tests are
consistent with studies conducted to determine when most R. palmarum adults would
be captured at traps after they were baited with palm tissue. Rochat (1990) and Morin
et al. (1986) reported that oil palm tissue was most attractive to R. palmarum adults
in the first two and four days, respectively, after excision from trees. These results
suggest that the optimal time to collect volatiles from palm tissues for identification of
compounds attractive to Rhynchophorus spp. is 24 to 72 h after cutting.
Quantification of R. cruentatus attraction to compounds identified from palm tissue
will require an effective trap that retains weevils after capture. A number of traps have
been described for monitoring and management of R. palmarum in the field that have
ranged from piles of palm tissue (Morin et al. 1986) and hollowed pieces of sugarcane
(Raigosa 1974) to various manufactured devices baited with palm tissue (Maharaj 1965,
Mireles 1984, Moura et al. 1989, Raigosa 1974, Rochat 1990, Rochat et al. 1991a). While
these systems appear to be effective for attracting weevils, their ability to retain insects
after capture was not studied. It has long been assumed that palm tissue used to attract
weevils would also provide harborage and dissuade trapped insects from escaping (Chit-
tenden 1902). However, observation of marked R. cruentatus adults placed in uncovered
buckets baited with chopped S. palmetto tissue indicated that weevils are not retained
by harborage alone (R.M.G.-D. & R.H.S. pers. obs.). In this study, R. cruentatus were
able to fly out of uncovered buckets suggesting that escape may be an important element
in evaluating trap effectiveness. Insecticides have been used in conjunction with traps
to kill weevils before they can escape but this technique may require frequent reappli-
cation of toxin. Maharaj (1965) described a trap designed to physically prevent captured
R. palmarum adults from flying out by placing wires spaced 2.5 cm apart over the top
of the trap. While data indicated that more weevils were caught in wired traps baited
with palm tissue than in exposed piles of palm tissue, no data were presented to suggest
that weevils did not escape once captured. We have tested several traps in the laboratory
and field designed to retain weevils once captured. Using the design similar to that
described by Maharaj (1965), we found that R. cruentatus adults were able to fly between
wires if properly oriented, or were able to grasp wires during flight and pull themselves
through. This observation indicated that a more efficient baffle was necessary to interfere
with weevil flight out of a trap. Baffles made from 2.4 cm ID PVC tubes did not appear
to interfere with R. cruentatus entry into buckets (T.J.W. pers. obs.) and when greased,

Florida Entomologist 75(2)

June, 1992

proved to be very effective at retaining captured weevils. In addition, greased baffles
constructed from 2.9 cm ID PVC tubes and placed on baited 19-liter buckets were
effective for capturing and retaining R. palmarum adults in Costa Rica (unpub. data).
The trap design presented in this manuscript has several potential applications as a
tool for the biological study, detection, and management of Rhynchophorus spp. We
have been using this trap to collect live R. cruentatus adults for use in laboratory
studies, to monitor seasonal population changes, and to quantify weevil attraction to
chemicals identified from S. palmetto tissues and conspecifics. Traps baited with C.
nucifera tissue, or identified attractants could be used to detect R. palmarum in areas
of Florida where introduction is most probable. Because these traps are effective at
retaining captured weevils, no toxicants need to be added and weevil health is maintained.
This is important if weevil dissections are to be performed to determine the presence
or absence of red ring nematodes. This trap design, in conjunction with attractants,
may also prove to be an effective alternative to insecticide treated traps for capturing
R. palmarum adults and subsequently reducing red ring disease incidence.


We thank John Cangiamila for technical assistance, Joan Perrier for preparation of
Fig. 1, Carlos Chinchilla of United Fruit Company, Coto 47, Costa Rica, for his assistance
in collecting R. palmarum, A. Duda and Sons for donating S. palmetto trees, and A.C.
Oehlschlager of Simon Fraser University for sharing his knowledge of R. palmarum
trapping systems. We are also grateful to Ted Center and Bill Howard for critically
reviewing the manuscript and to James Snyder of the Big Cypress National Preserve
for his cooperation. This research was supported in part by a USDA Special Grant in
Tropical and Subtropical Agriculture CRSR-90-34135-5233 to R.M.G.-D., R.H.S. and
J.P. Toth. This manuscript is Florida Agricultural Research Station Journal Series


CHITTENDEN, F. H. 1902. The palm and palmetto weevils. USDA Entomol. Bull.
38: 28.
ESSER, R. P. 1969. Rhadinaphelenchus cocophilus a potential threat to Florida palms.
Nematol. Circ., Div. Plant Industry, Florida Dept. Agric., No. 9.
GERBER, K., R. M. GIBLIN-DAVIS, AND J. ESCOBAR-GOYES. 1990. Association of
the red ring nematode, Rhadinaphelenchus cocophilus, with weevils from
Ecuador and Trinidad. Nematropica 20: 39-49.
GIBLIN-DAVIS, R. M. 1991. The potential for introduction and establishment of the
red ring nematode in Florida. Principes 35: 147-153.
GIBLIN-DAVIS, R. M., AND F. W. HOWARD. 1989. Vulnerability of stressed palms to
attack by Rhynchophorus cruentatus (Coleoptera: Curculionidae) and insecticidal
control of the pest. J. Econ. Entomol. 82: 1185-1190.
GIBLIN, R. M., K. GERBER, AND R. GRIFFITH. 1987. Comparison of Rhynchophorus
species as vectors of the red ring nematode, Rhadinaphelenchus cocophilus. J.
Nematol. 19: 524.
GIBLIN-DAVIS, R. M., K. GERBER, AND R. GRIFFITH. 1989. Laboratory rearing of
Rhynchophorus cruentatus and R. palmarum (Coleoptera: Curculionidae).
Florida Entomol. 72: 480-488.
GRIFFITH, R. 1987. Red ring disease of coconut palm. Plant Disease. 71: 193-196.
HAGLEY, E. A. C. 1963. The role of the palm weevil, Rhynchophorus palmarum, as
a vector of red ring disease of coconuts. I. results of preliminary investigations.
J. Econ. Entomol. 56: 375-380.


Weissling et al.: Palmetto Weevil Trap 221

HAGLEY, E. A. C. 1965. Test of attractants for the palm weevil. J. Econ. Entomol.
58: 1003.
MAHARAJ, S. 1965. A new design of trap for collecting the palm weevil, Rhynchophorus
palmarum (L.). Trop. Agric., Trinidad, 42: 373-375.
MIRELES, H. C. 1984. El picudo negro del cocotero en Tabasco. Institute Nacional
Investigaciones Agricolas, Desplegable para Productores Nuim. 1, 5pp.
chophorus palmarum control using traps made from oil palm cubes. Oleagineux
41: 57-62.
SENDE. 1989. Estudo do comportamento olfativo de Rhynchophorus palmarum
(L.) (Coleoptera: Curculionidae) no campo. Anais da Sociedade Entomol6gica do
Brasil. 18: 267-273.
JIRON, C. M. CHINCHILLA, AND R. G. MEXZAN. 1991. The chirality and field
activity of the aggregation pheromone of the American palm weevil, Rhyn-
chophorus palmarum (L.). Naturwissenschaften 78: (in press).
RAIGOSA, J. 1974. Nuevos disefos de trampas para control de plagas en cafia de azucar
(Saccharum officinarum L.), pp. 5-23 in Congreso de la Sociedad de Entomologia
ROCHAT, D. 1987. Etude de la communication chimique chez un Coleoptere Cur-
culionidae: Rhynchophorus palmarum L. DEA Report, Universit6 Paris VI. 30
ROCHAT, D. 1990. Rhynchophorus palmarum L. (Coleoptera: Curculionidae): nuevos
datos sobre el comportamiento del insecto y su control por trampeo olfativo.
Revista Palmas 11: 69-79.
1991a. Evidence for male-produced aggregation pheromone in American palm
weevil, Rhynchophorus palmarum (L.) (Coleoptera: Curculionidae). J. Chem.
Ecol. 17: 1221-1230.
AND C. DESCOINS. 1991b. Male-produced aggregation pheromone of the Amer-
ican palm weevil, Rhynchophorus palmarum (L.) (Coleoptera: Curculionidae):
collection, identification, electrophysiological activity, and laboratory bioassay.
J. Chem. Ecol. 17: 2127-2140.
SAS INSTITUTE. 1985. SAS user's guide: statistics, 5th ed. SAS Institute, Cary, N.C.
SCHUILING, M., AND J. B. M. VAN DINTHER. 1981. "Red ring disease" in the
Paricatuba oil palm estate, Para, Brazil. Z. ang. Entomol. 91: 154-169.
WATTANAPONGSIRI, A. 1966. A revision of the genera Rhynchophorus and Dynamis
(Coleoptera: Curculionidae). Department of Agriculture Science Bulletin,
Bangkok 1: 1-328.
WOODRUFF, R. E. 1967. A giant palm weevil, Rhynchophorus cruentatus (Fab.), in
Florida (Coleoptera: Curculionidae). Florida Dept. Agr. Div. Plant Ind. Entomol.
Circ. No. 63.

222 Florida Entomologist 75(2) June, 1992


Department of Zoology
Southern Illinois University
Carbondale, Illinois 62901


Presence or absence of spines, and number of spines (4), on the posterior margin of
the pronotum were found to be an unreliable character to distinguish Fitchia aptera
and F. spinosula. The morphology of the male parameres and female terminalia, and
the color of the abdominal spiracular peritreme, are diagnostic. Other characters were
found that, as a group and in combination with presence of spines, are useful in distin-
guishing the two species but are not diagnostic. These include the lateral anteocular
length, lengths and ratios of antennal segments, and width of the lateral abdominal
stripe. Brachypterous and macropterous adults are found in both species; brachyptery
is more common in F. spinosula than in F. aptera.


La presencia o ausencia de espinas, el nimero de espinas (4), en el margen posterior
del pronotum fue un character poco seguro para distinguir Fetchia aptera y F. spinosula.
Se utilizaron como caracteres de diagnosis la morfologia de los parametros del macho y
la terminalia de la hembra y el color del peritreno espiracular abdominal. Otros caracteres
en grupo y en combination con la presencia de espinas son utiles en la diferenciacion de
dos species, pero no se usan en la diagnosis. Estos incluyen la longitud del anteocular
lateral, longitud y radio de los segments de las antenas y el grosor de la franja lateral
abdominal. Adultos brachipteros y macropteros se encuentran en ambas species; com-
unmente se encuentran mas especimenes brachipteros en la especie F. spinosula que
en F. aptera.

Stal described the genus Fitchia in 1859 (p. 370-371) and included the single species,
F. aptera (p. 371), apparently based on a single female specimen. He added a second
species in 1872, F. spinosula (p. 79), also apparently based on a single female specimen.
One of the primary characters given by StAl was the presence of spines in spinosula
and, in fact, he divides the species into two groups, "Thorace inermi" (aptera) and
"Thorace spinoso" (spinosula). Subsequent authors have used the presence or absence
of these spines alone (e.g., Fracker 1912, Readio 1927, Slater & Baranowski 1978,
Torre-Bueno 1923) or in combination with head shape (e.g., Blatchley 1926, Froeschner
1944) to distinguish these species.
Blatchley (1926) provided the most complete descriptions of the two species including
the absence (aptera) or presence (spinosula) of spines on the posterior margin of the
prothorax (two submedial and one on each humeral angle) and the shape of the posterior
lobe of the head (gradually tapering from eyes in aptera, suddenly constricted behind
middle in spinosula); both characters are mentioned only in his key. Other characters

'Current address: Department of Entomology, NHB-127, National Museum of Natural History, Smithsonian
Institution, Washington, D.C. 20560.

McPherson et al.: Evaluation of Characters of Fitchia spp. 223

included coloration and markings, comparative lengths of antennal segments 3 and 4,
presence or absence of obtuse ridges on the posterior lobe of the pronotum, and overall
body length. During a recent survey of the Reduviidae of Illinois (unpublished), the
senior author found that the presence or absence of spines, and the number of spines
(4), to separate the two species was unreliable. This paper discusses problems with the
use of the spine character, evaluates several additional characters, and identifies those
characters we find diagnostic.


We examined specimens from several collections (see Acknowledgments). Characters
evaluated included the presence or absence of spines, and number of spines, on the
posterior margin of the pronotum, lateral anteocular length, lengths and ratios of antennal
segments, wing form (macropterous or brachypterous), width of lateral abdominal stripe,
color of spiracular peritreme, and the morphology of the male pygophore and female
terminalia. Pygophores were extracted from dried specimens after the specimens were
softened in warm water. Drawings of the pygophore and female terminalia were prepared
using a camera lucida. The terminology for the figures of genitalia and terminalia follows
Davis (1966) for males and Scudder (1959) for females.
SAS procedures (SAS Institute 1988) were used in all statistical analyses. Compari-
sons of spine pattern and wing morph were made using Chi-square contingency tables.
Differences between species in lateral anteocular length, antennal segment lengths,
ratios of antennal segment lengths, width of abdominal stripe, and, in females, the ratio
of length to width of the genital plates, were tested using Student's t test. Stripe width
was measured as a ratio of its width on the third visible abdominal segment (generally
its widest point) with the lateral length of the same segment. Level of significance was
0.05 for all tests.


Examination of 221 specimens from several states showed that the presence or
absence of pronotal spines, and the number of spines (4), were unreliable as diagnostic
characters. For example, specimens were found that matched the general description
ofF. spinosula but lacked the pronotal spines. Spines varied in size from simple swellings
to well developed and in number from one to four. Most frequently, they were represented
by groups of two or four, i.e., they were present only submedially on the posterior
margin (2 spines), or submedially and on each lateral humerall) angle. Thus, it was
necessary to search for characters that were less variable. We selected individuals that
matched the general description of spinosula (elongate body; StAl 1872, Blatchley 1926)
and had four well developed spines and individuals that matched the general description
of aptera (less elongate body) and lacked spines, and examined the male external genitalia
and female terminalia of both groups. We found distinct dimorphism within sex between
species (Fig. 1A-D). The parameres of aptera are slender and digitate, do not extend
beyond the posterior rim of the pygophore, and are not visible in posteroventral view
when the pygophore rests against the tergum. Those of spinosula are distinctly clavate,
extend beyond the posterior rim of the pygophore, and are visible in posteroventral
view. In females, the ninth tergum of aptera is depressed medially and tumescent
laterally as a prominent ridge. That of spinosula is flat to slightly convex medially.
Once the two species could be reliably separated, it was possible to reexamine the
spine characters (presence or absence, and number) and additional ones to determine
their value as diagnostic characters.

Florida Entomologist 75(2)

0.5 mm 0.5 mm

0.5 0.5
0.5 mm 0.5 mm

Fig. 1. Male and female genitalia of Fitchia aptera and F. spinosula. A. male,
pygophore and claspers, posterior view, F. aptera. B. female terminalia, ventroposterior
view, F. aptera. C. male, pygophore and claspers, posterior view, F. spinosula. D.
female terminalia, ventroposterior view, F. spinosula. Abbreviations: 1 GPO, first
gonapophysis; 1 GX, first gonocoxa; MP, medial process; P, parameres; T9, tergum 9.

We found that spines were present or absent in both species, and that there were
several spine patterns (Table 1). Spines were usually present submedially and on each
lateral humerall) angle (denoted in tables as 1-1-1-1) or only submedially (0-1-1-0), or
infrequently in other patterns (e.g., 1-0-0-1, 1-0-0-0); the infrequent patterns accounted
for 7.4% of spined individuals and 3.2% of all individuals (7 of 95 and 220, respectively).
The two species differed significantly in the proportion of individuals with and without
spines but the most dramatic difference was in the percent of these with four spines,
i.e., 4.7% in aptera, 36.0% in spinosula (Table 1). This may explain why earlier authors
relied so heavily on the presence or absence of four spines. Also, if only the two most
frequent spine patterns are examined, i.e., 1-1-1-1 and 0-1-1-0, aptera specimens usually
had two spines rather than four whereas spinosula specimens usually had four (Table
1). There was no significant difference in the proportion of males and females with spines


June, 1992

McPherson et al.: Evaluation of Characters of Fitchia spp. 225


Species Spine Pattern' WingMorph2 Sex N df x2 Prob.

aptera 1-1-1-1
spinosula 1-1-1-1
aptera 1-1-1-1
spinosula 1-1-1-1
aptera present3


spinosula present3


aptera present3
spinosula present3

M+B d + Y 5
M+B 6 + 9 5
M+B 6 13
9 25
6 33
9 35
M+B S 36
9 21
6 15
9 42
M d + 7
B 31
M + ? 20
B 37

2 34.795 <0.001

1 41.015 <0.001

1 2.035 0.154

1 15.647 <0.001

1 17.250 <0.001

1 15.741 <0.001

'1-1-1-1 = spines present submedially and on humeral angles; 0-1-1-0 = spines present submedially; other = spines
present in various patterns excluding 1-1-1-1 (e.g., 0-1-1-0, 1-0-0-0-, 0-0-1-0); 0-0-0-0 = spines absent.
"M = macropterous, B = brachypterous.
present includes 1-1-1-1, 0-1-1-0, and other spine patterns.

in aptera but a higher proportion of males had spines than females in spinosula (Table
1). Within each species, there was a significant difference in wing morph in proportion
of individuals with spines but the proportion differed within each species (Table 1). In
aptera, only 14.6% of macropterous individuals and 53.5% of brachypterous individuals
had spines whereas in spinosula, 87.0% of macropterous individuals and 40.7% of
brachypterous individuals had them.
Within each species, there was no significant difference in the frequency of macrop-
terous and brachypterous morphs by sex (Table 2). However, between species, there
was a significant difference. A higher percentage of adults were macropterous in aptera
(45.3%) than in spinosula (20.2%) (Table 2). This was also true within each sex (Table 2).
All other characters and ratios examined differed significantly in mean values between
the two species but all ranges overlapped (Table 3). Thus, they cannot be used as
diagnostic characters.
Our conclusion, then, is that although there are significant differences in the charac-
ters between species listed in Tables 1-3, these differences are not diagnostic because
of the overlap of their ranges or proportions. Thus, they must be used only as confirming
characters. To this can be added general body shape which falls into two broad categories
mentioned or implied in Stil's 1872 descriptions (more elongate in spinosula).

Florida Entomologist 75(2)

June, 1992


Males +
Species Wing Morph Males Females Females df x2 Prob.

aptera macropterous 23 25 -
brachypterous 23 35 1 0.730 0.393
spinosula macropterous 13 10 -
brachypterous 38 53 1 1.619 0.203
aptera macropterous 48
brachypterous 58
spinosula macropterous 23
brachypterous 91 1 15.842 <0.001
aptera macropterous 23 -
brachypterous 23 -
spinosula macropterous 13 -
brachypterous 38 1 6.225 0.013
aptera macropterous 25 -
brachypterous 35 -
spinosula macropterous 10 -
brachypterous 53 1 10.043 0.002

In addition to the structure of the male parameres and the female terminalia, we
found the color of the abdominal spiracular peritremes to be diagnostic. Spiracular
peritremes in aptera, though very narrow, varied from brown to black in both sexes;
those in spinosula varied from concolorous with the surrounding yellowish sternite, to
occasionally pale brown.
The geographical distribution of the two species is shown in Figure 2. Even using
only those records from the present study, it is apparent that the species are sympatric
or nearly so.


1. Spines usually absent on posterior margin ofpronotum but if present, then usually
only two submedially; abdominal spiracular peritremes varying from brown to black;
males with parameres slender, digitate, not visible in posteroventral view (Fig. 1A);
females with ninth tergum depressed medially, tumescent laterally as prominent ridge
(Fig. 1B ) ......................................................................................... aptera
1'. Spines present or absent on posterior margin of pronotum but if present, then
usually two submedially and one on each humeral angle; abdominal spiracular perit-
remes usually concolorous with surrounding yellowish sternum, occasionally pale
brown; males with parameres distinctly clavate, visible in posteroventral view (Fig.
1C); females with ninth tergum flat to slightly convex medially, not tumescent later-
ally (Fig. 1D ) .............................................................................. spinosula


Fitchia aptera StAl 1859, p. 371.
Body elongate, slender, wider behind middle. Color yellow, head and margins of
anterior lobe of pronotum with dark brown tinge; abdomen with terga medially and
sterna medially and laterally with brown to dark brown stripe, sternal stripes usually


McPherson et al.: Evaluation of Characters of Fitchia spp. 227


S l- (0 L0 tO O 0 c- O- 0O
-Q O O- 0 .L9 h O L9 hO Co Co c! C<
r1 .-I I

z C o oR

m V mo LCsO Lo 0 t-
Lm Co-4 Co0 1" IV 0N0
QO i ) & & m

w CI Cq m -V C4 -4 .r I '4i D'I 0m

.. . . .."
11 +1 +1+1+1+1 +1+1 +1+1+1+1+1+1++1 +1 +1 + 1 +1+1 +1+1+1

Z a) 00oI -I0C1Co -4c'0r-4-40 0, ,-40 Z> -4,--q 4=
[4 44 000 00 LO 0 m t-o Lm tmo00Co

CL4 Cd %

a) 0 0 0 0 00
- --'- -W

Z 5 41 -C -41 41 41 -d
O o + +1

C o .o No<

E k tI

ma e~ e i e % e e S aH ea e e c >g
a f^ d* ^ ^i A ^ F^ "

Florida Entomologist 75(2)

June, 1992



Fig. 2. Geographical distribution in U.S. of Fitchia as given by Froeschner (1988)
and of specimens examined in present study. A. F. aptera. B. F. spinosula.


McPherson et al.: Evaluation of Characters of Fitchia spp. 229

wider than in F. spinosula (i.e., width of lateral stripe on third abdominal sternite
averaging 0.59 x length of lateral margin of same sternite [range = 0.38-0.82, N = 97]);
spiracular peritreme, though narrow, varying from brown to black.
Head with anteocular distance generally shorter than in F. spinosula (X = 0.73 mm,
N = 106). Antennae with ratio of segment lengths one to four about 2.43:1.00:1.56:1.28;
segment four about 0.87x length of three (range = 0.54-1.06, N = 56).
Spines usually absent on posterior margin of pronotum but if present, then usually
only two submedially. Macropterous and brachypterous individuals present but macrop-
terous form more common in males than females (d = 50.0%, N = 46; 9 = 41.7%, N = 60).
Males with parameres slender, digitate (Fig. 1A), not visible in posteroventral view.
Females with ninth tergum depressed medially, tumescent laterally as prominent ridge
(Fig. 1B).
Length: 9.0-14.0 mm.
Distribution: AL, CO, CT, D.C., FL, IL, IN, KS, MA, ME, MI, MO, NC, ND, NH,

Fitchia spinosula StAl 1872, p. 79.
Similar in appearance to F. aptera but averaging longer and relatively narrower.
Color similar to F. aptera, abdominal sternal stripes usually narrower (i.e., width of
lateral stripe on third abdominal sternite averaging 0.35x length of lateral margin of
same sternite [range, 0.18-0.70, N = 107]); spiracular peritreme usually concolorous
with surrounding yellowish sternum, occasionally pale brown.
Head with anteocular distance generally longer than in F. aptera (X = 0.92 mm, N
= 114). Antennae with ratio of segment lengths one to four about 2.67:1.00:1.95:1.23;
segment four about 0.64x length of three (range = 0.48-0.76, N = 61).
Spines present or absent on posterior margin of pronotum but if present, then usually
two submedially and one on each humeral angle. Macropterous and brachypterous indi-
viduals present but brachypterous more common in both sexes (d = 74.5%, N = 51; 9
= 84.1%, N = 63).
Males with parameres distinctly clavate (Fig. 1C), visible in posteroventral view.
Females with ninth tergum flat to slightly convex medially (Fig. 1D).
Length: 11.5-16.0 mm.
Distribution: AL, AR, AZ, CO, CT, Dakota, FL, GA, IL, KS, MA, MI, MO, NC, NE,
NM, NY, PA, TX, UT, WA, WY (Fig. 2B) (Mexico).


We thank the following individuals for the loan of specimens from their respective
institutions: J. K. Bouseman, Illinois Natural History Survey, Champaign; R. W. Brooks,
Snow Museum, University of Kansas, Lawrence; R. L. Fischer, Michigan State Univer-
sity, East Lansing; R. C. Froeschner, National Museum of Natural History, Washington,
D. C.; M. F. O'Brien, University of Michigan, Ann Arbor; N. D. Penny, California
Academy of Science, San Francisco; A. V. Provonsha, Department of Entomology,
Purdue University, West Lafyette, IN; R. T. Schuh, American Museum of Natural
History, NY; and R. W. Sites, Department of Entomology, Texas Tech University,


BLATCHLEY, W. S. 1926. Heteroptera or true bugs of eastern north America with
especial reference to the faunas of Indiana and Florida. Nature Pub. Co., In-
dianapolis. 1116 pp.

230 Florida Entomologist 75(2) June, 1992

DAVIS, N. T. 1966. Contributions to the morphology and phylogeny of the Reduvioidea
(Hemiptera:Heteroptera). Part III. The male and female genitalia. Ann. Entomol.
Soc. Amer. 59: 911-924.
FRACKER, S. B. 1912. A systematic outline of the Reduviidae of North America. Proc.
Iowa Acad. Sci. 19: 217-252.
FROESCHNER, R. C. 1944. Contributions to a synopsis of the Hemiptera of Missouri,
Pt. III. Lygaeidae, Pyrrhocoridae, Piesmidae, Tingididae, Enicocephalidae,
Phymatidae, Ploiariidae, Reduviidae, Nabidae. Amer. Midland Natur. 31: 638-
FROESCHNER, R. C. 1988. Family Reduviidae, pp. 616-651, in T. J. Henry and R. C.
Froeschner [eds.]. Catalog of the Heteroptera, or true bugs, of Canada and the
continental United States. E. J. Brill, New York. 958 pp.
READIO, P. A. 1927. Studies on the biology of the Reduviidae of America North of
Mexico. Univ. Kansas Sci. Bull. 17: 5-291.
SAS INSTITUTE. 1988. SAS/STAT user's guide, version 6, 4th ed. Vols. 1 & 2. SAS
Institute, Cary, North Carolina.
SCUDDER, G. G. E. 1959. The female genitalia of the Heteroptera: morphology and
bearing on classification. Trans. Roy. Entomol. Soc. London 111(14): 405-467.
SLATER, J. A., AND R. M. BARANOWSKI. 1978. How to know the true bugs (Hemiptera
Heteroptera). Wm. C. Brown Co. Pub., Dubuque. 256 pp.
STAL, C. 1859. Till kannedomen om Reduvini. Ofversigt af Kongliga Svenska Vet-
enskaps-Akademiens Firhandlingar 16: 175-204, 363-386.
STAL, C. 1872. Enumeratio Hemipterorum. Bidrag till en f6rteckning Ofver alla hittills
kanda Hemiptera, jemte systematiska meddelanden. 2. Kongliga Svenska Vet-
enskaps-Akademiens Handlingar 10(4): 1-159.
TORRE-BUENO, J. R. DE LA. 1923. Family Reduviidae, pp. 677-692, in W. E. Britton.
Guide to the insects of Connecticut. Part IV. The Hemiptera or sucking insects
of Connecticut. Connecticut State Geol. Natur. Hist. Surv. Bull. 34: 1-807.


Crop Quality & Fruit Insects Research
2301 S. International Blvd.
Weslaco, Texas 78596


Twenty-four trap types representing all combinations of 4 colors, 3 shapes and 2
sizes were evaluated for visual attractiveness to irradiated, laboratory-reared Mexican
fruit flies, Anastrepha ludens (Loew), released into a grapefruit orchard when 1-3 days
old. Spheres and vertically oriented rectangular panels were not significantly different
in attractiveness summed over spring, summer and autumn seasons. Horizontally
oriented rectangles were much less attractive than spheres and vertical rectangles.
Larger rectangles (13 by 18 cm) and spheres (13 cm diam.) were more attractive than
smaller rectangles (10 by 13 cm) and spheres (8 cm diam.). Yellow, green and red were
equally attractive summed over trap shapes, sizes and seasons. Red was less attractive

Robacker: Responses of A. ludens to Trap Shapes

than green in spring and summer but more attractive in autumn. Relative attractiveness
of yellow compared to green was less affected by season. Overall, vertical rectangles
were more attractive than spheres in spring while spheres were more attractive in
autumn. In spring, red spheres were more attractive than vertical red rectangles while
yellow and white rectangles were more attractive than yellow and white spheres. Traps
in trees with mature grapefruit generally captured more flies than those in trees with
only small, immature fruit. Small spheres were more attractive than small vertical
rectangles to females in trees with small, immature fruit but were less attractive than
small vertical rectangles in trees with mature fruit. Overall, the best traps were the
large sizes of the 6 combinations of yellow, green and red spheres and vertical rectangles
that captured 8 times as many flies as the least attractive traps.


El actractivo visual de 24 tipos de trampas que representaban todas las combinaciones
posibles de cuatro colors, tres formas y dos tamafos diferentes; fue evaluada con moscas
mexicanas de la fruta, Anastrepha ludens (Loew), que fueron criadas en el laboratorio,
irradiadas como pupas, y liberadas de uno a tres dias de edad en una huerta de toronjas.
La atractividad total de trampas esfericas y panels rectangulares orientados vertical-
mente no fue significativamente diferente a trav6s de las temporadas de primavera e
invierno. Trampas de panels rectangulares orientados horizontalmente fueron much
menos atractivas. Trampas rectangulares (13 x 18 cm) y esfericas (13 cm diam.) de
mayor tamaho fueron mas atractivas que las trampas rectangulares (10 x 13 cm) y
esfericas (8 cm diam.) mas pequefias. Trampas pintadas de amarillo, verde y rojo, fueron
igualmente atractivas independientemente de la forma, tamaiio y estaci6n del afo. Las
trampas de color rojo fueron menos atractivas que las de color verde durante la primavera
y verano, pero mas atractivas durante el otofio. La atractividad de trampas de color
amarillo fue menos afectada por las estaci6nes del afio. Trampas de rectangulos verticales
fueron mas atractivas que las trampas esfericas durante la primavera, mientras que las
trampas esfericas fueron mAs atractivas durante el otofio. Durante la primavera las
trampas esfericas de color rojo fueron mAs atractivas que las trampas rectangulares
verticales del mismo color, mientras que las trampas de rectangulos verticales de colors
amarillo o blanco fueron mas atractivas que las trampas esf4ricas de estos mismos colors.
Trampas localizadas en Arboles con toronjas maduras generalmente capturaron mas
moscas que aquellas localizadas en Arboles con fruta inmadura pequefia. Trampas esfericas
pequefias fueron mAs atractivas que trampas pequefias de rectangulos verticales, para
moscas hembra que infestaron Arboles con fruta inmadura pequefia; pero fueron menos
atractivas que las trampas de rectangulos verticales en Arboles con fruta madura. En
general, las trampas mas efectivas fueron las trampas esfericas y de rectAngulos verticales
de tamafio grande, en sus seis combinaciones posibles con colors amarillo, verde y rojo.

Mexican fruit fly (Anastrepha ludens Loew) is a serious pest of citrus and mango in
Mexico and threatens the citrus industry in Texas where it occurs sporadically and in
California, Florida and many countries throughout the world where potentially it could
become established. Approximately 1 million dollars annually are spent on its detection
and management in Texas alone. One of the critical problems is the lack of an effective
and inexpensive trapping system. For this reason, the California Department of Food
and Agriculture currently funds research in excess of $150,000 per year to develop a
better trap and lure for the Mexican fruit fly than the bulky and fragile McPhail trap
(Baker et al. 1944) that is currently in use.
The work described in this paper was conducted as part of a continuing effort to
develop a trap that is visually attractive to the Mexican fruit fly and that can be combined
with an olfactory lure to replace the McPhail trap. To accomplish that goal, my first
objective is to establish baseline attractiveness of visual traps without complicating and

Florida Entomologist 75(2)

June, 1992

possibly overriding effects of olfactory attractants. In earlier experiments directed to-
ward the first objective, Robacker et al. (1990) demonstrated that green and yellow
were the most attractive colors to the Mexican fruit fly, when presented as rectangular
panels in laboratory tests or as colored solutions in McPhail traps in orchard tests.
Further, that study showed that the attractiveness of red relative to green doubled as
seasons progressed from spring to autumn. Effects of trap size and shape were not
The specific objectives of this paper were to evaluate the importance of trap size
and shape for the most attractive colors discovered in the previous work (Robacker et
al. 1990), to investigate interactions of trap size and shape with trap color, and to
determine the most attractive combinations of color, size and shape for male and female
Mexican fruit flies. Twenty-four trap types comprising all combinations of 4 colors, 3
shapes and 2 sizes were evaluated for attractiveness to irradiated, laboratory-colony
flies released into a grapefruit orchard.



Flies were from a culture maintained for at least 85 generations with no wild-fly
introductions. Laboratory flies were used due to lack of natural populations. Flies were
irradiated, to comply with quarantine laws, with 70-92 Grays (Cobalt 60) 1 to 2 days
before adult eclosion. They were released in test plots 1 to 3 days after eclosion, still
sexually immature. Flies were fed sugar and water between eclosion and release.


Twenty-four trap types consisting of all combinations of 4 colors, 2 sizes and 3
shapes/orientations were evaluated. Colors were green, yellow, red and white Benjamin
Moore (New York, N.Y.) Impervo high gloss enamels mixed to match the colors of PVC
rectangles previously used to test color preferences ofA. ludens (Robacker et al. 1990).
Green and yellow were chosen because they were the most attractive colors in previous
experiments, red because of its interaction with season (Robacker et al. 1990), and white
to serve as a control. Colored traps and reflectance data for the traps were provided
by Great Lakes IPM (Vestaburg, Mich.). Their data showed that the PVC rectangles
used by Robacker et al. (1990) and the traps used in the present study were similar in
reflectance intensity and hue. Mean difference in reflectance between 400 and 700 nm
(20 nm increments) was 1.7% with the greatest difference (9.4%) occurring at 500 nm
for green.
Two basic shapes were evaluated: spheres and two-dimensional rectangles. Rectang-
les were hung either with the surface oriented horizontally (parallel to the ground) or
vertically. Two sizes of rectangles were used: 10 x 13 cm and 13 x 18 cm. Two sizes
of spheres were 8 cm and 13 cm diameters. Total surface areas were: small rectangle
= 260 cm2; large rectangle = 470 cm2; small sphere = 200 cm2; and large sphere = 530
cm2. The 24 traps are listed in Table 1.

Experimental Procedure

Experiments were conducted in a 5-ha grapefruit, Citrus paradise MacFadyen, or-
chard near Weslaco, Tex., for 3 weeks during May-June (spring season), 3 weeks during
Aug-Sept (summer), and 2 weeks during Nov-Dec (autumn), 1989. Four rectangular
plots of 24 trees each were used. Two plots were located in a section of the orchard


Robacker: Responses of A. ludens to Trap Shapes




Rank Color Size Shape Mean' SEM CV2

1 yellow large vertical 7.1 a 1.0 0.8
2 green large vertical 6.9 a 1.0 0.8
3 yellow large sphere 6.7ab 1.1 0.9
4 red large sphere 6.5 ab 1.2 1.0
5 green large sphere 6.4ab 1.1 1.0
6 red large vertical 6.1ab 1.7 1.6
7 white large vertical 5.2 be 0.9 0.9
8 yellow small vertical 5.1bc 0.9 1.0
9 white large sphere 4.3 cd 0.7 0.9
10 green small sphere 4.1 cd 0.8 1.1
11 green small vertical 3.8 cde 0.6 0.9
11 yellow small sphere 3.8 cde 0.6 0.9
13 red small vertical 3.2 def 0.7 1.2
14 red small sphere 2.9 defg 0.5 0.9
15 yellow large horizontal 2.2 efgh 0.4 1.0
16 white small sphere 1.8fgh 0.3 1.9
17 green large horizontal 1.7 fgh 0.3 1.1
17 white small vertical 1.7 fgh 0.4 1.4
19 red large horizontal 1.5gh 0.3 1.3
20 yellow small horizontal 1.3gh 0.3 1.1
20 white large horizontal 1.3gh 0.2 1.0
22 white small horizontal 1.2h 0.4 1.8
23 green small horizontal 0.9h 0.2 1.4
24 red small horizontal 0.8h 0.3 1.9

'Flies/trap/week. Means followed by the same letter are not significantly different from each other at the 5% level
by LSD.
"Coefficient of variation (s/mean).

where fruit had been harvested before the spring season experiments began. Fruit was
not harvested from trees in the other 2 plots. Thus, during spring, 2 plots had trees
laden with ripe grapefruit and 2 had only very small fruit. By summer, fruit differences
among plots were less and by autumn all plots had large fruit.
One each of the 24 trap types was hung in the 24 trees in each plot, 1 to a tree.
Traps were coated on all external surfaces with a thin layer of Tangle-Trap (Tanglefoot
Company, Grand Rapids, Mich.) and were placed 1-2 m aboveground on the northwest
sides of trees since these locations were optimum in previous work (Robacker et al.
1990). Traps were replaced and flies counted after each week. Flies were released at
10 sites per plot, 200-250 flies per site, every 3-4 days beginning several days before
each replication (week). Locations of the 24 traps in each plot were randomized for each

Statistical Analyses

The experimental design for each season (spring, summer, autumn) was a randomized
complete block (plots = blocks) with plots nested within color, size and shape treatments.
Color, size and shape main effects and all interactions of these factors were tested using
the 'plots nested within the color by size by shape interaction' as the error term. Separate
analyses of spring data were also conducted to test effects of fruit vs no fruit on trees.
Fruit main effect was tested using the 'plot nested within fruit interaction' as the error

234 Florida Entomologist 75(2) June, 1992

term and interactions of fruit with color, size and shape were tested using the 'color by
size by shape by plot interaction nested within fruit' as the error term. For analyses
involving seasons, the 'plot by season interaction nested within color, size and shape'
was used as the error term to test season main effect and interactions involving season
with the other factors. Analyses were conducted using the ANOVA procedure of PC
SAS (SAS Institute 1988). Because this procedure can handle only completely balanced
designs, data used for analyses involving seasonal effects were means of the 3 weeks
of spring, the 3 of summer, and the 2 of autumn.



Attractiveness of colors was similar to that reported by Robacker et al. (1990).
Summed over sizes, shapes and seasons, mean captures were: yellow traps 4.4 flies/trap/
week 0.3 SEM; green traps 4.0 0.3; red traps 3.5 + 0.4; and white traps 2.6
0.2. Captures by yellow and green were significantly greater (p < 0.05) than those
by white by LSD (F = 3.7; df = 3, 72; p < 0.05). No differences in responses of males
and females were evident.


Large traps were consistently more attractive than small traps (F = 24.0; df = 1,
72; p < 0.001). Summed over colors, shapes and seasons, mean captures were: large
traps 4.6 flies/trap/week 0.3 SEM; and small traps 2.6 0.1. No differences in
responses of males and females were evident.
Haniotakis (1986) reported that larger rectangular traps baited with pheromone
attracted more Dacus oleae Gmelin than smaller ones when fly density was high but
size had no effect at low population density. Katsoyannos (1987) found that Ceratitis
capitata (Wiedemann) preferred spheres of 7 cm diameter to smaller or larger ones.
Effects of low population density and very large sphere size have not been evaluated
in A. ludens.


Summed over colors, sizes and seasons, vertically oriented rectangles were equally
attractive as spheres and both were much more attractive than horizontally oriented
rectangles (F = 29.7; df = 2, 72; p < 0.001). Mean captures were: vertical rectangles -
4.9 flies/trap/week 0.4 SEM; spheres 4.6 + 0.3; and horizontal rectangles 1.4
0.1. No differences in responses of males and females were evident. That horizontally
oriented rectangles were relatively unattractive was surprising since they would seem
to be good leaf mimics.


Summed over all traps, captures during spring were higher (p < 0.05 by LSD) than
those in summer and autumn (F = 26.5; df = 2, 144; p < 0.001). Mean captures were:
spring 4.9 flies/trap/week 0.2 SEM; summer 2.9 0.2; and autumn 2.7 0.3.
No differences in responses of males and females were evident. Main effects of season
are probably trivial since they depend on the number of flies released as well as orchard
conditions and trap parameters.


Robacker: Responses of A. ludens to Trap Shapes


Color by Season Interaction

A color by season interaction similar to that reported by Robacker et al. (1990) was
observed (F = 2.1; df = 6, 144; p = 0.06). The interaction was an increase in the
attractivenesses of red and yellow relative to green as the seasons progressed from
spring to autumn. Captures on red doubled relative to green: red in autumn (3.4 flies/trap/
week)/ green in autumn (2.3) divided by red in spring (4.6)/ green in spring (5.9) = 1.9.
Captures on yellow increased less dramatically relative to green: yellow in autumn (3.5)/
green in autumn (2.3) divided by yellow in spring (6.1)/ green in spring (5.9) = 1.5. The
corresponding ratios from Robacker et al. (1990) were 1.8 and 1.6 for red and yellow,
respectively. No differences in responses of males and females were evident. The biolog-
ical explanation for this effect is not known but a possibly related phenomenon was
observed in Rhagoletis pomonella (Walsh) (Kring 1970), and is discussed in the next
The color by season interaction was not influenced by the presence or absence of
fruit in plots during the spring. This conclusion was reached by comparing responses
to colors in plots with or without fruit during the spring using data shown in Fig. 3.
For each color, the number of flies captured was greater by similar proportions in plots
with fruit and less by similar proportions in plots without fruit. Thus, the seasonal
change in response of flies to yellow and red relative to green appears independent of
fruit effects in the plots during the spring.

Shape by Season Interaction

This interaction was highly significant (F = 4.9; df = 4, 144; p < 0.001) because,
summed over all traps and plot types (fruit vs. no fruit in spring), vertical rectangles
were slightly more attractive than spheres during the spring and summer but spheres
were slightly more attractive during the autumn (Fig. 1). However, closer scrutiny of
the data reveals a more complex response pattern. Most of the preference for vertical
rectangles during spring and summer was due to male responses. Females generally
preferred spheres to vertical rectangles during both spring and autumn, with one excep-
tion that influenced the overall data in favor of vertical rectangles during spring. Fig.
4 shows the exception occurred in the case of small traps in trees with fruit in which
females greatly preferred vertical rectangles. This size by shape by fruit interaction is
discussed below. During summer, females slightly preferred vertical rectangles to
spheres, suggesting that females seek oviposition sites less during the hot summer
months than at other times. Attractiveness of horizontal rectangles was low during all
3 seasons.
Although statistically significant, these results must be viewed with caution since
effects of seasons were not replicated over years. However, it is noteworthy that Kring
(1970) reported that attractiveness to wild R. pomonellaof red spheres increased relative
to yellow rectangles as season progressed from July to August. Color and shape were
confounded in Kring's (1970) report so the cause could not be isolated.

Size by Season Interaction

Although this effect was statistically significant (F = 3.2; df = 2, 144; p < 0.05), the
change in response to size from spring to autumn was subtle and perhaps not biologically
significant. Larger traps captured more flies than smaller ones during all 3 seasons, but
the ratios of captures on large traps to captures on small traps varied from 2.2/1 in the
summer and 1.8/1 in the spring to 1.5/1 in the autumn. No differences in responses of
males and females were evident. Unlike the shape by season interaction, this effect was
not influenced by the presence or absence of fruit in plots during the spring.

Florida Entomologist 75(2)

June, 1992








Fig. 1. Mean captures (flies/trap/week) (- SEM) of sterile laboratory-culture Mex-
ican fruit flies released into a grapefruit orchard, by vertically oriented rectangular
panels and spheres during 3 seasons.

Color by Shape Interaction
Color and shape interacted significantly for captures of males during the spring
season (F = 2.3; df = 6, 72; p < 0.05). The effect was similar for females but was not
significant. In each case red spheres captured more flies than vertical red rectangles
whereas yellow and white rectangles captured more flies than yellow and white spheres
(Fig. 2A). For comparison, Fig. 2B shows that the color by shape interaction was nearly
opposite during autumn as spheres generally were preferred over rectangles except
that red rectangles captured a few more flies than red spheres. The interaction was not
significant during autumn possibly owing to insufficient replications of the color by shape
treatments. Although it appears that season affected the color by shape interaction, the
season by color by shape interaction was not significant.
Color by shape interactions have been demonstrated in other tephritidae. In C.
capitata, males and females preferred yellow rectangles over other colors of rectangles,
but females preferred red and black spheres over other colors of spheres while males
generally were not attracted to any spheres (Cytrynowicz et al. 1982). However, in the
same study, A. fraterculus Wiedemann males and females preferred yellow over green
and red for both rectangles and spheres. Johnson (1983) reported a markedly different
color by shape interaction in R. pomonella. In this species, red spheres were more
attractive than yellow rectangles to males while the 2 trap types were about equally


Robacker: Responses of A. ludens to Trap Shapes

a. 5
cc 4-




Fig. 2. Mean captures (flies/trap/week) ( SEM) of sterile laboratory-culture Mex-
ican fruit flies released into a grapefruit orchard, by vertically oriented rectangular
panels (V) and spheres (S) of 4 colors during 2 seasons.

attractive to females. As another example, Riedl & Hislop (1985) showed that both sexes
of R. complete Cresson were most attracted to yellow rectangles but preferred green
spheres. The results of these studies indicate a general preference for yellow and green
colors of rectangles but preferences for sphere colors seem to be species specific, possibly
related to host-fruit color for each species.

Size by Shape Interaction

Size and shape interacted significantly during the spring for females (F = 3.3; df =
2, 72; p < 0.05) and during the summer for males (F = 3.8; df = 2, 72; p < 0.05). In
the spring (summed over all plots), females preferred large spheres over large vertical
rectangles but small vertical rectangles over small spheres while males preferred vertical
rectangles over spheres of either size. This effect on females was influenced by fruit in
plots as is discussed below (size by shape by fruit) interaction). In summer, males
preferred large vertical rectangles over large spheres and responded equally to small
rectangles and spheres while females preferred vertical rectangles over spheres of either
size. In autumn, spheres of either size generally were preferred over vertical rectangles
by both sexes. Size by shape by season interactions were not significant. Although 2
size by shape interactions were significant as discussed above, the effects were small
and inconsistent and will not be treated further.

Fruit (Spring Only)

Plots with fruit had significantly more captures of males (F = 22.8; df = 1, 2; p <
0.05) than plots without fruit, summed over all traps. The effect was similar for females
but was not significant. Overall mean captures (males + females) were: plots with fruit
- 5.9 flies/trap/week 0.5 SEM; and plots without fruit 4.0 0.3. This same finding
was reported previously (Robacker et al. 1990).

Size by Fruit Interaction (Spring Only)

A significant interaction was observed for females (F = 8.1; df = 1, 46; p < 0.01)
because large traps in plots with fruit captured twice as many females as small traps
whereas in unfruited plots large traps captured only 1.4 times as many females as small
traps. The effect did not occur in males.


LU 4
u 2


Florida Entomologist 75(2)

Color by Size by Fruit Interaction (Spring Only)

This interaction was significant for males (F = 3.5; df = 3, 46; p < 0.05). The effect
was similar for females but was not significant. The effect is illustrated in Fig. 3 for
males + females. For red and white traps, large traps consistently captured more flies
than small traps in both fruited and fruitless plots, and those in fruited plots generally
captured more flies than those in fruitless plots (Fig. 3C, D). However, for green traps,
large traps captured more than twice as many flies as small traps in fruited plots but
captured only as many as small traps in plots without fruit (Fig. 3A). Conversely, for
yellow traps, large traps captured nearly twice as many flies as small traps in fruitless
plots but only about equal numbers of flies in fruited plots (Fig. 3B). The biological
significance of these results is not known.

Size by Shape by Fruit Interaction (Spring Only)

No interaction occurred for males for which vertical rectangles were consistently
more attractive than spheres regardless of trap size or whether plots had fruit (Fig.
4A, B). A large interaction occurred for females (F = 5.0; df = 2, 46; p = 0.01). Females
slightly preferred large spheres over vertical rectangles in both fruited and fruitless
plots (Fig. 4D) and preferred small spheres over vertical rectangles in fruitless plots
(Fig. 4C). However, females greatly preferred small vertical rectangles over small










0 -



Fig. 3. Mean captures (flies/trap/week) ( SEM) of sterile laboratory-culture Mex-
ican fruit flies released into a grapefruit orchard, by small and large traps of various
colors in plots with or without mature grapefruit on trees during the spring season.


June, 1992

Robacker: Responses of A. ludens to Trap Shapes









Fig. 4. Mean captures (flies/trap/week) ( SEM) of sterile laboratory-culture Mex-
ican fruit flies released into a grapefruit orchard, by small and large vertically oriented
rectangular panels and spheres in plots with or without mature grapefruit on trees
during the spring season.

spheres in plots with fruit (Fig. 4C). A possible explanation is that spheres generally
were more attractive than rectangles because most females were searching for fruit. In
the case of small spheres (8 cm diameter) in competition with mature grapefruit (9-10
cm diameter), the small spheres were poor fruit mimics and therefore unattractive. The
rectangles on these fruited trees were more attractive than the small spheres perhaps
because they attracted females searching for foliage stimuli. Horizontal rectangles were
not included in this analysis because they were much less attractive than both spheres
and vertical rectangles under all conditions.

Trapping Considerations

Table 1 expresses the ranks of the 24 trap types summed over all seasons. The best
6 traps, which are not significantly different from each other, were the large sizes of
yellow, green and red spheres and vertically oriented rectangles. These traps captured
about 8 times as many flies as the least attractive traps (bottom of Table 1). The data
suggest that yellow, green and red vertical rectangles are as attractive as the same
colors of spheres. However, test flies used here were released into the orchard at age
1-3 days when they were sexually immature. Although many of these flies probably
remained in the test plots until they were sexually mature at age 7-12 days old, most


Florida Entomologist 75(2)

flies trapped were probably 1-4 days old owing to mortality and emigration of older
flies. As evidence, Robacker & Wolfenbarger (1988) found that the number of Mexican
fruit flies that could be trapped 4 days after releases similar to those used in the present
work was only about half of the numbers trapped during each of the first 3 days. These
1-4 day old flies probably were attracted to traps that mimiced foliage and fruit as they
searched for food on both leaf and fruit surfaces. Possibly, sexually mature, gravid
females would have a greater preference for spheres than we observed here. This
possibility is suggested by results of Landolt et al. (1988) who showed that sexually
mature, unmated papaya fruit flies (Toxotrypana curvicauda Gerstaecher) preferred
green spheres over green panels by a 2 to 1 margin. Landolt et al. (1988) found that
attractiveness of spheres and panels to mature males was not significantly different.
Another important consideration is the size of spherical traps. The results of this
study indicate that spherical traps should be larger than the fruit they mimic to be
competitive with those fruit for attracting females. However, it is also possible that
spheres that are too large may be less attractive than smaller ones if they cease to
represent a fruit stimulus to the flies. As discussed above, such a finding was reported
for C. capitata by Katsoyannos (1987).
Finally, note that the 24 test traps contained no olfactory lures. As such, only 2%
of the released flies were captured on the traps. Addition of suitable lures would undoub-
tedly increase captures on the traps. However, the possibility of interactions of lure
types with visual parameters of traps is a possibility that will require investigation when
new lures become available.


I thank Jose Garcia and Miguel Diaz for technical assistance, Sammy Ingle for insects,
A. W. Guenthner (USDA-APHIS, Mission, TX) for irradiation of flies, H. Del Var
Peterson (USDA-ARS, College Station, TX.) for statistics, and Karen Robacker and
Mike Firko for computer assistance. I also thank D. L. Chambers, D. 0. McInnis and
D. A. Wolfenbarger for critical reviews of the manuscript. Special thanks are extended
to Jim Hansell (Great Lakes IPM, Vestaburg, MI) for his tremendous efforts to make
the traps to my exacting specifications. This work was partially supported by a grant
from the California Department of Food and Agriculture (89-0440). Mention of a pro-
prietary product does not constitute an endorsement or recommendation for its use by
the USDA.


BAKER, A. C., W. E. STONE, C. C. PLUMMER, AND M. MCPHAIL. 1944. A review
of studies on the Mexican fruit fly and related Mexican species. USDA Miscellane-
ous Publication 531.
CYTRYNOWICZ, M., J. S. MORGANTE, AND H. M. L. DE SOUZA. 1982. Visual re-
sponses of South American fruit flies, Anastrephafraterculus, and Mediterranean
fruit flies, Ceratitis capitata, to colored rectangles and spheres. Environ. En-
tomol. 11: 1202-1210.
HANIOTAKIS, G. E. 1986. Effect of size, color and height of pheromone baited sticky
traps on captures of Dacus oleae flies. Entomol. Hellenica. 4: 55-61.
JOHNSON, P. C. 1983. Response of adult apple maggot (Diptera: Tephritidae) to Phero-
con A. M. traps and red spheres in a non-orchard habitat. J. Econ. Entomol.
76: 1279-1284.
KATSOYANNOS, B. I. 1987. Some factors affecting field responses of Mediterranean
fruit flies to colored spheres of different sizes, pp. 469-473 in A. P. Economopoulos
(ed.), Fruit flies. Elsevier, Amsterdam, The Netherlands.


June, 1992

Pe~a: Predator Prey Interactions

KRING, J. B. 1970. Red spheres and yellow panels combined to attract apple maggot
flies. J. Econ. Entomol. 63: 466-469.
1988. Sex pheromone-based trapping system for papaya fruit fly (Diptera: Tep-
hritidae). J. Econ. Entomol. 81: 1163-1169.
RIEDL, H., AND R. HISLOP. 1985. Visual attraction of the walnut husk fly (Diptera:
Tephritidae) to color rectangles and spheres. Environ. Entomol. 14: 810-814.
trap color, height, and placement around trees on capture of Mexican fruit flies
(Diptera: Tephritidae). J. Econ. Entomol. 83: 412-419.
ROBACKER, D. C., AND D. A. WOLFENBARGER. 1988. Attraction of laboratory-
reared, irradiated Mexican fruit flies to male-produced pheromone in the field.
Southwest. Entomol. 13: 75-80.
SAS INSTITUTE. 1988. SAS/STAT user's guide, release 6.03 edition. SAS Institute,
Cary, North Carolina.


Tropical Research and Education Center
University of Florida, IFAS
18905 S.W. 280 Street
Homestead, Florida 33031


Studies on predation and feeding habits of Typhlodromalus peregrinus (Muma) were
conducted in the laboratory and greenhouse. Throughout the study, the broad mite,
Polyphagotarsonemus latus (Banks), the citrus rust mite, Phyllocoptruta oleivora
(Ashmead), commercial bee pollen and pollen from Schinus terebinthifolius, Parthenium
hysterophorus and Bidens bipinata were used as predator food.
Typhlodromalus peregrinus consumed P. latus eggs, immatures and adults. T. pereg-
rinus consumed 23-75% of the prey population in 6 days. T. peregrinus did not reject
P. oleivora as prey, but favored P. latus when the latter was present. T. peregrinus
developed on bee pollen, bee pollen and P. latus, and on S. terebinthifolius pollen.


Se realizaron studios de los habitos alimenticios y predacion de Typhlodromalus
peregrinus (Muma) en el laboratorio y en el invernadero. Se suministro como alimentos
el acaro blanco Polyphagotarsonemus latus Banks, el acaro tostador de los citricos
Phyllocoptruta oleivora (Ashmead), pollen de abejas, pollen de Schinus terebinthifolius,
de Parthenium hysterophorus, y de Bidens bipinata. El acaro predador T. peregrinus
consumio huevos, estados larvarios, y adults de P. latus. T. peregrinus consumio el
23-75% de la presa en 6 dias. T. peregrinus prefirio P. latus como presa y no rechazo
P. oleivora como presa, pero cuando P. latus se encontro present prefirio a este ultimo.

242 Florida Entomologist 75(2) June, 1992

T. peregrinus se desarrollo en pollen de abejas, en una mezcla de pollen de abejas y P.
latus, y en pollen de Schinus terebinthifolius.

The broad mite, Polyphagotarsonemus latus (Banks) is one of the major pests of
limes in southern Florida (Wolfenbarger 1974, Campbell 1979). Broad mite (BM) is also
a pest of ornamentals, row and fruit crops in tropical, subtropical and temperate areas
(Jeppson et al. 1975). Several predaceous mites have been observed feeding on this pest
in California (McMurtry et al. 1984, Badii & McMurtry 1984) and in Florida (Pefia et
al. 1989).
Typhlodromalus peregrinus (Muma) is the most common species of phytoseiid mites
in lime groves in southern Florida (Pefia et al. 1989). The effectiveness of this predator
against different prey species was investigated by Muma (1971). He observed that T.
peregrinus is a facultative predator able to feed on purple scale, Lepidosaphes beckii
(Newman), Florida red scale, Chrysomphalus aonidum, (L.) sooty mold, and mites but
had little impact on mite populations. Little is known about the food preferences of T.
peregrinus in the lime ecosystem. The objectives of this study were to investigate the
potential of T. peregrinus as a predator of P. latus; to determine if T. peregrinus fed
on an alternative eriophyid prey species, Phyllocoptruta oleivora (Ashmead); and finally,
to determine if it utilizes pollen of three common plant species found near most lime
groves, Schinus terebinthifolius, Spanish needles, Bidens bipinata and Parthenium


Experiment 1.

All experiments were performed in the laboratory at a temperature of 26-+1lC and
RH of 55%-65%. The T. peregrinus used in the test were reared in the insectary on bee
pollen and broad mites following a similar procedure reported by McMurtry and Scriven
(1975) for Amblyseius hibisci (Chant).
In tests where broad mites or citrus rust mites (P. oleivora) were offered as prey,
the predaceous mites were placed on young lime fruits ca. 3.5 cm in diameter. To confine
predator and prey to the stylar area (less than 19 cm2) the basal half of each fruit was
ringed with a Tanglefoot barrier. The fruit was supported by a water-saturated plastic
foam pad in an 80 ml plastic cup filled with distilled water. Adult and larval stages of
broad mites were transferred by brushing them from infested fruits to the fruit arenas.
To provide eggs as prey, female adults were confined to the fruit arenas and removed
after oviposition 24 h later. After 2-6 days, when eggs, larvae or adults were present,
one adult female T. peregrinus mite was placed on each of the fruits. Each treatment
was replicated 10 times.
In tests where pollen was offered as food, the sources were ground commercial bee
pollen, and locally collected pollen from S. terebinthifolius, P. hysterophorus and B.
bipinata. Four to 10 predaceous female mites were placed in each of 10 experimental
arenas as described by McMurtry and Scriven (1975). Pollen grains were spread with
a fine camel's hair brush over the arena. Fresh pollen was added daily. The number of
predators was counted daily for 24 days.

Experiment 2.

To determine the efficacy of T. peregrinus as a predator of broad mites under
greenhouse conditions, different predator-prey ratios were tested. First, pinto beans,

Pee a: Predator Prey Interactions 243



z 0

Z 0

cl 1 +1 +1


z 01 +1 +1
S+1 c+

I0 t 1
+1 +1 +1

S +1 oo
s .

3 c +1 +1+1
+1 t-o
Ix -


I c

o Zo

1 ^

g -< i +i ^ i

Florida Entomologist 75(2)

June, 1992

Phaseoulus vulgaris L. were planted 1.5-2 cm deep and were infested with bean stock
leaves containing only P. latus. The mites moved to plants in 2 days and the dried stock
leaves were removed after 2-3 days. Plants were rated as having a low broad mite
density (lower than 24 mites/trifoliate leaf) or high broad mite density (greater than 240
mites/trifoliate leaf).


Experiment 1.

Typhlodromalus peregrinus attacked all stages of broad mites. There was no differ-
ence (P = 0.05) among the mean number of different stages eaten by T. peregrinus at
24, 48, and 72 h (Table 1). Only 14% of the initial egg population was eaten during the
first hour. Feeding increased in the next 24 h. Sixty-one percent of the larval prey was
eaten during the first 24 h and consumption increased up to 83% at 72 h. A similar
feeding rate was observed when P. latus adults were exposed to T. peregrinus females.
Forty-six percent of the adult prey was consumed during the first 24 h and feeding





0 1 2 3 4 5
Days After Exposure

Fig. 1. Number of broad mites (BM) and rust mites (CRM) with and without T.
peregrinus. 1A: broad mites (cross hatch) and rust mites (diagonal hatch) after exposure
to T. peregrinus. 1B: Rust mites with (cross hatch) and without (open bar) T. peregrinus.
1C: control, broad mites (solid bar), rust mites (diagonal hatch).


Pefia: Predator Prey Interactions


+1 +1I

O t- CcO-4
0i coe'i 000
+ + +1 1 +1 +1

o ,... c

E F-




a +1 +1 +1 +1 +1
z 3 0.0 %v CD Y
Q 0 d
0 ,
t jC)


| |
co C

Florida Entomologist 75(2)

June, 1992

M=oj oo


l 1 L 10 t- I* C LO Cc NM
i=O P P C M 3O
tot-tOdd t I.m6
1-1 ~

0 1d 0 10 10
NM 00 t Cc
C(M i-4 r-1

.0 cd NO.0

(M ni-

od 0 3 e 3d 0 od M cd Gd
Ous 1 V- t- tI-

C'] L- N LO NI
CN C' 1

- C1 1" 00

rl C' II 00








C) C

0e B



Pefia: Predator Prey Interactions

continued 96 h after initiation of the experiment. There was a significant linear relation-
ship between predation and time during the first seven hours (r2 = 0.90) (Predation
rate = 18.77 -5.47 hours). In the absence of prey, most T. preregrinus females died in
1-6 days.
Polyphagotarsonemus latus appeared to be more suitable as prey than P. oleivora
(Fig. 1A). When both species were present in an arena, the number of P. latus was
maintained at a relatively lower level (5% level by Student's t-test). T. peregrinus only
reduced P. oleivora density when P. latus was not present in the arena. When P.
oleivora was the only prey species offered (Fig. 1C) there was a significant reduction
(P < 0.05) 5% level by Student's t-test in its density 2 days after exposure. This may
indicate that 7'. peregrinus may consume P. oleivora only when other preferred prey
is scarce.
The number of T. peregrinus immatures and eggs produced/day increased over 24
days when 1 mg of bee pollen was supplied daily (Table 2). When broad mites were
offered as prey and pollen was supplied, the number of eggs and immatures did not
increase significantly 7-15 days after exposure (Table 2).
Pollen from S. terebinthifolius was the best for the predator allowing a better survival
and a higher rate of egg laying than the other types of pollen (Table 2). The capability
of T. peregrinus females to prolong their survival to 24 days on pollen would enable
such predators to exist in the field when prey is lacking at unfavorable conditions. The
present data confirms Muma's (1971) observations and suggest that the survival of T.
peregrinus depends on the availability of both prey and alternative food.

Experiment 2.

Number of P. latus were significantly higher on the control plants than on T. pereg-
rinus treated plants. Further, at highest host densities, P. latus were reduced in higher
numbers (58-73%) than at lower host densities (0.6-8%) (Table 3). At lower host densities
(n = 17 25) more P. latus were destroyed when the number of predators per plant
was 8. At higher host densities (n = 204 311) numbers were destroyed whether at low
or high predator densities. These differences between lowest and highest host densities
could be the result of different factors. First, the availability of low prey density may
force the predator to increase its searching activity, resulting in a reduction for the
predator to capture more prey in a given period of time. Predators were not forced to
stay on the leaf substrates, thus, their rates of dispersal may increase than when they
are confined in fruit arenas. Whether high host density also increased more chances for
the predators to capture broad mites is an open question. This data suggests that T.
peregrinus is a general predator limited in its ability to regulate broad mite population.


I thank D. Buisson and Y. Lee-Sin for technical assistance, H. Denmark for identifi-
cation of the predator species, H. Nadel and L. Osborne for criticism of the manuscript,
and the Florida Lime Committee for financial support of this project. This is Florida
Agricultural Experiment Station Journal Series No. 9754.


BADII, M. H., AND J. A. MCMURTRY. 1984. Feeding behavior of some phytoseiid
predators on the broad mite, Polyphagotarsonemus latus (Acar.:Phytoseiidae,
Tarsonemidae) 29: 49-53.

Florida Entomologist 75(2)

June, 1992

CAMPBELL, C. W. 1979. Tahiti lime production in Florida. Fla. Coop. Extension Serv.
Institute of Food and Agric. Sci. Bull 187. 45p.
JEPPSON, L. R., H. KEIFER, AND E. W. BAKER. 1975. Mites Injurious to Economic
Plants. Univ. Calif. Press. Los Angeles. 614p.
MARTIGNONI, M. E., AND E. STEINHAUS. 1961. Insect Microbiology and Insect
Pathology. Burgess Publishing Co., Minneapolis. 75 p.
MCMURTRY, J. A., ANDG. T. SCRIVEN. 1975. Insectary production of phytoseiid
mites. J. Econ. Entomol. 58: 282-284.
MCMURTRY, J. A., M. H. BADII, AND H. G. JOHNSON. 1984. The broad mite,
Polyphagotarsonemus latus, as a potential prey for phytoseiid mites in California.
Entomophaga 29: 83-86.
MUMA, M. H. 1971. Food habits of Phytoseiidae (Acarina:Mesostigmata) including com-
mon species on Florida citrus. Florida. Entomol. 54: 21-34.
WOLFENBARGER, D. 0. 1974. Incidence and control of the broad mite on limes in
Florida. Ceiba 18: 70-74.

Se e a a a a a a -a a


Medical and Veterinary Entomology Research Laboratory
Agricultural Research Service, USDA
P.O. Box 14565, Gainesville, Florida 32604, U.S.A.


In order to determine the frequency and distribution of polygyne and monogyne fire
ants (Solenopsis invicta) in Florida, preselected sites were surveyed from Key West to
Tallahassee. Polygyne colonies were found at 15% of infested sites-a frequency similar
to other states in the southeastern United States, but much less than in Texas. Polygyny
was most common in the region around Marion county, but smaller populations were
also scattered across the state. The density of mounds at polygyne sites was more than
twice that at monogyne sites (262 versus 115 mounds/ha), although mound diameters
were about 20% smaller. Polygyne and monogyne queens averaged the same size (1.42
mm, head width), but monogyne queens were much heavier (24.3 mg versus 14.4 mg)
due to their physogastry. As expected, workers in polygyne colonies were considerably
smaller than those in monogyne colonies (0.28 mg versus 0.19 mg, dry fat-free).


Para determinar la frecuencia y distribuci6n de "hormigas bravas" (Solenopsis in-
victa) poliginas y mon6ginas en Florida, se muestrearon sitios preseleccionados desde
Key West hasta Tallahassee. Se encontraron colonies poliginas en 15% de los lugares
infestados, una frecuencia similar a la de otros estados en el sureste de los EE.UU.,
pero much menor que la frecuencia de Texas. Las hormigas poliginas resultaron mas
comunes en el condado Marion pero tambien se encontraron pequefias poblaciones espar-
cidas a lo largo del Estado. La densidad de los tfimulos en lugares con poliginia fue mis
del double que en lugares con monoginia (262 versus 115, tumulos/ha), a pesar de que el
diametro medio de los tumulos fue 20% menor. Las reinas poliginas y mon6ginas fueron
del mismo tamafio (1.42 mm, ancho de cabeza) pero debido a su fisogastria, las reinas


Porter: Polygyne Fire Ants in Florida

mon6ginas fueron much mas pesadas (24.3 mg versus 14.4 mg). Como se esperaba, las
obreras de colonies poliginas fueron considerablemente menores que las de colonies
mon6ginas (0.28 mg versus 0.19 mg, peso seco, libre de grasa).

High densities of imported fire ants (Solenopsis invicta Buren) can cause substantial
ecological and economic difficulties (Lofgren 1986, Porter & Savignano 1990). Polygyne
or multiple-queen fire ants are of particular concern because they occur in densities that
are often several times those of the more familiar monogyne form (Porter et al. 1991,
1992). High densities of polygyne fire ants are apparently a consequence of interconnected
supercolonies that lack normal territorial boundaries (Bhatkar & Vinson 1987).
Polygyne S. invicta colonies were first discovered near Mobile, Alabama, where they
appeared to be a curious anomaly (Glancey et al. 1973). Polygyne populations were
subsequently found at scattered locations throughout the Southeastern United States
(Mirenda & Vinson 1982, Fletcher 1983, Lofgren & Williams 1984). Recent studies
indicate that polygyne colonies are much more common than previously suspected (Porter
et al. 1991, 1992) and may even be spreading (Glancey et al. 1987).
Current genetic and morphological evidence indicates that both forms are the same
species (Ross & Fletcher 1985, Trager 1991). Biochemical analyses of venom alkaloids
and cuticular hydrocarbons indicate a close relationship, although slight differences have
been detected in Texas (Greenberg et al. 1990). The origin of the polygyne form in the
United States is unknown, but an independent introduction from South America does
not seem likely (Ross et al. 1987). At least five other species and a hybrid in the genus
Solenopsis also have polygyne populations (Summerlin 1976, Glancey et al. 1989,
Jouvenaz et al. 1989, MacKay et al. 1991, unpublished data).
This study will document the frequency, distribution, and abundance of polygyne
fire ants in Florida. It will also establish a baseline for monitoring future population
changes of polygyne and monogyne fire ants.


In the primary survey, 85 sites were preselected along rural roadsides from Key
West to Tallahassee, Florida (Fig. 1). Sampling began in south Florida on 20 March
1990 and concluded a month later near Jacksonville.
The following is a summary of environmental data obtained from primary sample
sites. Adjacent habitat was 49% forests and shrubs, 19% residential or urban, 17%
pasture or grassland, and 15% miscellaneous. Grass along roadsides averaged 12 6
cm (SD) high with 88% of the ground vegetated. Soil moisture was dry at 38% of sites,
damp at 41%, moist at 19%, and wet at 2%. The average soil temperature at 5 cm was
27 5 C. Right of ways were 8 4 m wide.
Mound densities were determined from four belt transects, two on either side of the
road (Porter et al. 1991). One transect on each side of the road was along the outer
border of the right-of-way while the other was on the inner border adjacent to the road.
Each transect was 70 paces long. All active mounds within reach of a 1.2 meter stick
were tallied into one of six categories according to their diameter: -15, s30, 546, -61,
s76, and >76 cm. The pace and reach of each investigator were determined and used
to calculate the area sampled. On average, transects were 55 m long and 2.4 m wide.
A supplemental survey of 113 sites was conducted between February and April 1991
to more precisely delineate the extent and distribution of the polygyne populations in
and around Marion County (Fig. 1 and 2). Mound densities were not measured at
supplemental sites.

Florida Entomologist 75(2)

Fire Ants
0 Confirmed Monogyne
o Probable Monogyne



" 9

Fig. 1. Abundance and distribution of polygyne and monogyne imported fire ants
(Solenopsis invicta) at 85 preselected roadside sites in Florida. The native fire ant,
Solenopsis geminata, was found at sites marked with a "g". A plus symbol indicates
one site where S. invicta was present but not sufficiently abundant to occur in transects.
Sites without either species are indicated by a null set symbol "0". Asterisks indicate
counties where polygyny was confirmed by Glancey et al. (1987).


June, 1992

Porter: Polygyne Fire Ants in Florida

Fig. 2. Distribution of polygyne and monogyne fire ants (Solenopsis invicta) at 113
supplemental sites in north central Florida (14 additional sites are included from Fig.
1). Sites which appeared to have both forms are shown with overlapping symbols. Sites
with Solenopsis geminata are marked with a "g". Major cities are indicated with an
open star.

Polygyne and monogyne colonies were detected in both surveys by removing several
shovelfuls of soil from a mound and scattering them across a plastic sorting sheet.
Wingless queens in the supplemental survey were held on ice for several hours until
they could be weighed (nearest 0.1 mg). A wedge micrometer (Porter 1983) was used
to measure their head widths. In the primary survey, 4-5 mounds were inspected per
site when possible. In the supplemental survey, we also sampled up to 5 mounds, but
sampling was terminated after confirming one monogyne or two polygyne colonies.
Preserved queens were later dissected to determine if their spermathecae were filled
with sperm. Sites were declared to be polygyne if at least one colony contained two or
more inseminated queens. Confirmed monogyne sites contained at least one colony with
a single highly physogastric queen. Probable monogyne sites were those where sexual
production (Vargo & Fletcher 1987) and the worker size distribution (Greenberg et al.
1985) were characteristic of monogyne colonies, but no inseminated queens were recov-
Several hundred workers were collected from mounds at each site by placing vials
(20 ml) in mounds for several minutes. The inside rims of these vials were coated with
talcum powder to prevent the escape of workers falling inside. Fifty workers were
randomly chosen from selected colonies. This number allowed mean worker weight to
be estimated within about 16% of the true colony mean 95% of the time. Workers were
initially preserved in 70% isopropyl alcohol and then soaked in ether for 3-4 days to

Florida Entomologist 75(2)

remove fats and reduce weight variability due to this factor. Preservation in alcohol
reduced dry weights by 11% and soaking in ether by an additional 25% (n = 4 colonies).
Species identifications were made by the author and voucher samples have been
placed in the Florida State Collection of Arthropods, Florida Department of Agriculture
and Consumer Service, Division of Plant Industry, Gainesville, Florida, U.S.A. Means
are presented one standard error unless otherwise indicated. Statistical analyses were
done with Statview II software (Abacus Concepts 1987). Most statistical differences
were determined using two-tailed unpaired t-tests. In order to normalize distributions,
mound densities were log-transformed and mound areas were squareroot-transformed.
A one-way ANOVA was used with Scheff6's S test to discriminate mean worker weights.
A chi-square test was used to compare distributions of S. invicta and S. geminata (F.).


Primary Survey

Solenopsis invicta occupied 87% of the sites (74/85) in the primary survey (Fig. 1).
Polygyne colonies were found at 15% of the infested sites (11/74). Most of the polygyne
sites (7/11) were in Marion and Putnam counties. Polygyny was relatively sparse in
other counties, but localized populations are evidently scattered throughout the state.
A previous sample of 10 sites in the Florida panhandle (Porter et al. 1992) failed to find
polygyny in this part of the state.
Single highly physogastric queens confirmed monogyny at 14 sites in this survey.
An additional 49 sites were designated as probable monogyne based on the large size
of workers, the relative abundance of sexual brood, and the absence of multiple insemi-
nated queens. Several monogyne colonies may have been present at two of the polygyne
sites based on the size of workers in these colonies.
Mound densities of S. invicta at polygyne sites averaged 262 mounds/ha versus 115
mounds/ha for monogyne sites (Table 1; t = 2.79, df = 70, P < 0.007). The average
diameter of mounds of polygyne sites was about 20% smaller than those of monogyne
sites (Table 1; t = 2.29, df = 69, P = 0.025). The total surface areas of polygyne and
monogyne mounds at sites in this survey were 1.2 and 0.9 m2/site, respectively (Table
1; t = 1.15; df = 69; P = 0.26).

Supplemental Survey

The supplemental survey revealed that polygyne colonies predominate in an area of
about 4600 km2 (-1800 mi2), including most of Marion county and large parts of Alachua,
Levy, and Putnam counties (Fig. 2). Of the 113 supplemental sites, 40 were confirmed
as polygyne, 33 were confirmed as monogyne, and 22 were probable monogyne. Six
sites appeared to contain both monogyne and polygyne colonies and the remaining 12
sites contained only S. geminata. Seven of the previous sites also contained both S.
invicta and S. geminata.
Solenopsis invicta was absent from a total of 23 sites in the primary and supplemental
surveys (Figs. 1 and 2). Solenopsis geminata was found at 83% of the sites without S.
invicta, but only 7% of the sites with S. invicta (19/23 versus 14/178; x2 = 82.9; df = 1;
P < 0.0001).
Queens from polygyne and monogyne colonies were the same size; mean head widths
for both groups were 1.42 0.03 mm (SD, n = 61 and 37, respectively; each queen was
from a separate colony). Queens from monogyne colonies were much more physogastric
than queens from polygyne colonies (Fig. 3A; t = 21.0, df = 183, P < 0.0001), and there
was little overlap in their weight distributions. Ninety-five percent of queens greater


June, 1992

Porter: Polygyne Fire Ants in Florida 253


Survey Location2
Louisiana to
Florida Georgia3 Texas4
(11,63) (9,43) (288,232)

Mound Density (mounds/ha)
Polygyne 262 65 544 115 680 + 28
Monogyne 115+ 10 170 17 295 16
Mound Diameter (cm)
Polygyne 32 + 3 32 3 36 1
Monogyne 40 1 42 1 39 1
Total Mound Area (m2/site)
Polygyne 1.2 0.2 3.0 0.9 4.4 0.2
Monogyne 0.9 0.1 1.4 0.1 2.2 0.2

'Data are means SE.
2Numbers of polygyne and monogyne sites sampled are shown in parentheses.
3From Porter et al. 1992.
'From Porter et al. 1991.

than 20 mg belonged to monogyne colonies compared to less than 2% of queens from
polygyne colonies.
Fat-free weights of workers from monogyne colonies were 50% heavier on average
than workers from confirmed polygyne colonies (Fig. 3B; F = 31.9, df = 2, 90, P =
0.0001). The mean weight of workers from probable monogyne colonies did not differ
significantly from their confirmed counterparts (P = 0.86, Scheff6's S Test) suggesting
that very few polygyne colonies were included by accident.

10 15 20 25 30
Weight (mg)


} j34


Probable Confirmed Polygyne
Monogyne Monogyne

Fig. 3. A) Weight distributions of inseminated queens from 70 polygyne colonies
(n = 145 queens) and 39 monogyne colonies (n = 39). B) Average dry weight (fat-free)
of workers in polygyne and monogyne colonies. Standard deviations (narrow lines),
standard errors (wider lines), and the number of colonies sampled are shown for each

Florida Entomologist 75(2)

June, 1992


The frequency of polygyny in Florida (15%) was similar to that in other southeastern
states (18%, Porter et al. 1992), but only a third of that reported in Texas (53%, Porter
et al. 1991). Polygyny is apparently absent from most of the Mato Grosso region of
Brazil (Porter et al. 1992), although several populations have been found in northern
Argentina (Trager 1991, S.D.P. unpublished data).
The geographic pattern of polygyny in north central Florida (Fig. 2) was similar to
the mosaic pattern observed in Texas (Porter et al. 1991). The reasons why polygyny
is more abundant in one region than another (Fig. 1) are still not known. Habitat and
climatic variables do not appear to be related to its presence or absence in the United
States (Porter et al. 1991, 1992). Likewise, there does not appear to be a pattern
associated with the invasion process, although very little is known about how polygyne
populations change over time. Perhaps the mosaic distribution of polygyny in the United
States is largely the haphazard result of expanding fire ant populations (Porter et al.
Polygyny was first reported in Florida by Lofgren & Williams (1984). Three years
later, Glancey et al. (1987) confirmed polygyny in eight counties from Broward to Duval
(Fig. 1). This survey confirmed polygyny in an additional six counties (Fig. 1, Fig. 2).
The gradual expansion of polygyne populations in Florida and the United States seems
likely (Glancey et al. 1987, Porter et al. 1991); however, current evidence of expansion
is still mostly anecdotal. Surveys across time will be necessary to document the actual
rate and extent of population changes.
The absence of S.. invicta from sites in the Florida Keys and the Everglades is
attributable to the fact that S. invicta has only recently invaded this area, and that
sawgrass and mangrove swamps make very poor habitat for both native and imported
fire ants. The absence of S. invicta from three sites in northern Florida (Fig. 1) is more
of a puzzle because S. invicta has been in this area for several decades. The native fire
ant, S. geminata, is known to persist in drier undisturbed habitats of northern Florida
(Tschinkel 1988a), but the habitat at these three sites was moist, disturbed and otherwise
appeared ideal for S. invicta. Perhaps, an abundant population of S. geminata at these
sites has diminished the success of founding S. invicta queens and delayed the invasion
process (Porter et al. 1988).
The high frequency of the native fire ant, S. geminata, at sites without S. invicta
indicates that the imported species has displaced native species from most of Florida's
roadsides. S. geminata has met a similar fate almost everywhere the two species have
met (Porter et al. 1988). In contrast with Texas (Porter et al. 1991), S. geminata
continues to persist with the imported species at a few sites in Florida (Figs. 1-2). It
will be interesting to determine if such coexistence will be maintained permanently.
Polygyne colonies are unambiguously identified by the presence of multiple insemi-
nated queens with degenerate wing muscles. Almost all mounds at polygyne sites ap-
peared to be polygyne based on the small size of their workers and their low sexual
production. We were able to collect multiple inseminated queens from more than 60%
of mounds inspected at polygyne sites; thus the probability of not confirming polygyny
at a polygyne site was generally less than 5% when 4-5 mounds were inspected under
appropriate weather conditions.
The presence of a single highly physogastric queen (>20 mg) identifies a monogyne
colony with a high degree of probability (Fig. 3A). The weights of monogyne queens
were similar to those reported by Tschinkel (1988b). The average weight of polygyne
queens was about 2 mg more than that reported by Vargo & Fletcher (1989) probably
because colonies in Florida had fewer queens than those studied in Texas. Worker
weight alone was a relatively poor predictor of polygyny for a single colony (Fig. 3B),

Porter: Polygyne Fire Ants in Florida

but the average size of workers from 6-7 mature colonies would identify pure monogyne
or polygyne sites with about 95% accuracy. The overlapping standard deviations of
worker weights probably resulted because young monogyne colonies have small workers
(Tschinkel 1988c) and polygyne colonies with only a few queens may produce larger
workers than those with many queens. Care should also be taken when using worker
size or weight as an indicator of polygyny (Greenberg et al. 1985) because these characters
can fluctuate seasonally (Tschinkel 1988b).
The average densities of mounds at roadside polygyne sites is usually 2-3 times those
at monogyne sites (Table 1). Mound densities in Florida were less than those in other
states. The cause of lower densities is uncertain, but they could reflect differences in
climate, habitat, weather, or road management. Time since invasion is probably not a
factor for most sites because S. invicta has occupied most of Florida for 15-20 years and
mound densities usually stabilize within 4-5 years of the initial introduction.
Higher mound densities do not necessarily mean that polygyne sites have proportion-
ately more fire ants. Polygyne mounds in this survey and in the survey from Louisiana
to Georgia (Porter et al. 1992) were about 20% smaller in diameter than mounds of
monogyne sites (Table 1). Polygyne mounds in Texas were also smaller, but the difference
was less than 10% (Porter et al. 1991).
If total mound area at a site is used as an index of ant abundance, then ant densities
at polygyne sites in Florida were only about 30% greater than those at monogyne sites.
The total surface area of mounds at polygyne sites in the Louisiana to Georgia survey
and the Texas survey were twice as large as totals for monogyne sites (Table 1). These
data indicate that polygyne fire ant populations are consistently larger than monogyne
populations; however, additional studies are needed which directly compare the number
and biomass of workers in polygyne and monogyne populations.


Special thanks are extended to C. K. Porter, P. C. Porter, J. K. Porter, and R. A.
Porter for again donating their spring break to counting ants. T. Krueger completed
the survey, dissected the queens, and organized the data for analysis. W. A. Banks, J.
L. Stimac, and D. F. Williams read the paper and provided a number of helpful comments.
J. Briano kindly translated the abstract into Spanish.


ABACUS CONCEPTS. 1987. Statview II. Abacus Concepts, Berkeley, Ca.
BHATKAR, A. P., AND S. B. VINSON. 1987. Colony limits in Solenopsis invicta Buren,
pp. 559-600 in J. Eder and H. Rembold [eds.], Chemistry and biology of social
insects. Verlag J. Peperny, Munich.
FLETCHER, D. J. C. 1983. Three newly-discovered polygynous populations of the fire
ant, Solenopsis invicta, and their significance. J. Georgia Entomol. Soc. 18: 538-
GLANCEY, B. M., R. K. VANDER MEER, AND D. P. WOJCIK. 1989. Polygyny in hybrid
imported fire ants. Florida Entomol. 72: 632-636.
GLANCEY, B. M., C. H. CRAIG, C. E. STRINGER, AND P. M. BISHOP. 1973. Multiple
fertile queens in colonies of the imported fire ant, Solenopsis invicta. J. Georgia
Entomol. Soc. 8: 237-238.
C. T. ADAMS. 1987. The increasing incidence of the polygynous form of the red
imported fire ant, Solenopsis invicta (Hymenoptera: Formicidae), in Florida.
Florida Entomol. 68: 162-168.


256 Florida Entomologist 75(2) June, 1992

GREENBERG, L., D. J. C. FLETCHER, AND S. B. VINSON. 1985. Differences in worker
size and mound distribution in monogynous and polygynous colonies of the fire
ant Solenopsis invicta Buren. J. Kansas Entomol. Soc. 58: 9-18.
GREENBERG, L., H. J. WILLIAMS, AND S. B. VINSON. 1990. A comparison of venom
and hydrocarbon profiles from alates in Texas monogyne and polygyne fire ants,
Solenopsis invicta, pp. 95-101 in R. K. Vander Meer, K. Jaffe, and A. Cedeno
[eds.], Applied myrmecology: a world perspective. Westview Press, Boulder,
JOUVENAZ, D. P., D. P. WOJCIK, AND R. K. VANDER MEER. 1989. First observation
of polygyny in fire ants, Solenopsis spp., in South America. Psyche 96: 161-165.
LOFGREN, C. S. 1986. The economic importance and control of imported fire ants in
the United States, pp. 227-255 in S. B. Vinson [ed.], Economic impact and control
of social insects. Praeger, New York.
LOFGREN, C. S., AND D. F. WILLIAMS. 1984. Polygynous colonies of the red imported
fire ant, Solenopsis invicta(Hymenoptera: Formicidae) in Florida. Florida En-
tomol. 67: 484-6.
REBELES, AND S. B. VINSON. 1991. A comparison of monogyne and polygyne
populations of the tropical fire ant, Solenopsis geminata (Hymenoptera: For-
micidae), in Mexico. J. Kansas Entomol. Soc. 63: 611-615.
MIRENDA, J. T., AND S. B. VINSON. 1982. Single and multiple queen colonies of
imported fire ants in Texas. Southwestern Entomol. 7: 135-141.
PORTER, S. D. 1983. Fast, accurate method of measuring ant head widths. Ann. En-
tomol. Soc. Am. 76: 866-867.
PORTER, S. D., AND D. A. SAVIGNANO. 1990. Invasion ofpolygyne fire ants decimates
native ants and disrupts arthropod community. Ecology 71: 2095-2106.
PORTER, S. D., H. G. FOWLER, AND W. P. MACKAY. 1992. A comparison of fire ant
densities in North and South America (Hymenoptera: Formicidae). Econ. En-
tomol. (in press)
PORTER, S. D., B. VAN EIMEREN, AND L. E. GILBERT. 1988. Invasion of red imported
fire ants (Hymenoptera: Formicidae): microgeography of competitive replace-
ment. Ann. Entomol. Soc. Am. 81: 913-918.
Distribution and density of polygyne fire ants (Hymenoptera: Formicidae) in
Texas. J. Econ. Entomol. 84: 866-874.
Ross, K. G., AND D. J. C. FLETCHER. 1985. Comparative study of genetic and social
structure in two forms of the fire ant Solenopsis invicta (Hymenoptera: For-
micidae). Behav. Ecol. Sociobiol. 17: 349-356.
Ross, K. G., E. L. VARGO, AND D. J. C. FLETCHER. 1987. Comparative biochemical
genetics of three fire ant species in North America with special reference to the
two social forms of Solenopsis invicta (Hymenoptera:Formicidae). Evolution 41:
SUMMERLIN, J. W. 1976. Polygyny in a colony of the southern fire ant. Ann. Entomol.
Soc. Am. 69: 54.
TRAGER, J. C. 1991. A revision of the fire ants, Solenopsis geminata group,
(Hymenoptera: Formicidae, Myrmicinae). J. New York Entomol. Soc. 99: 141-
TSCHINKEL, W. R. 1988a. Distribution of the fire ants Solenopsis invicta and S.
geminata (Hymenoptera: Formicidae) in northern Florida in relation to habitat
and disturbance. Ann. Entomol. Soc. Am. 81: 76-81.
TSCHINKEL, W. R. 1988b. Social control of egg-laying rate in queens of the fire ant
Solenopsis invicta. Physiol. Entomol. 13: 327-50.
TSCHINKEL, W. R. 1988c. Colony growth and the ontogeny of worker polymorphism
in the fire ant, Solenopsis invicta. Behavior. Ecol. Sociobiol. 22: 103-15.
VARGO, E. L., AND D. J. C. FLETCHER. 1987. Effect of queen number on the produc-
tion of sexual in natural populations of the fire ant, Solenopsis invicta. Physiol.
Entomol. 12: 109-116.

Hall: DNA Markers In Florida Honeybees


VARGO, E. L., AND D. J. C. FLETCHER. 1989. On the relationship between queen
number and fecundity in polygyne colonies of the fire ant, Solenopsis invicta.
Physiol. Entomol. 14: 223-232.


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


Fourteen honeybee colonies, present in the vicinity of Florida ports and suspected
to be of African descent, were analyzed for distinguishing mitochondrial and nuclear
DNA markers. Four of the colonies, previously identified as Africar on the basis of
morphometrics, were found to have African mitochondrial DNA. Nuclear DNA markers
characteristic of African bees were found predominantly in these four colonies.


Catorce colonies de abejas localizadas cerca a puertos en Florida, y las cuales se
sospechaba tenian ascendencia africana, fueron analizadas por marcadores de DNA de
la mitocondria y el nucleo. Se encontr6 que cuatro colonies, previamente identificadas
como africanas en base a morfometria, tenfan DNA mitocondrial africano. Se encontr6
predominando en estas cuatro colonies marcadores de DNA nuclear, caracteristico de
las abejas africanas.

The American honeybee population represents a number of introduced subspecies
of Apis mellifera L. The bees imported by early settlers were mostly of the west
European race A.m. mellifera L. Over the past hundred years, the more docile east
European types, A.m. ligustica Spinola, A.m. carnica Pollmann, and A.m. caucasia
Gorbachev, have become preferred for beekeeping and have largely replaced managed
bees from the earlier importations (Pellet 1938, Oertel 1976, Sheppard 1989). European
swarms that escaped from apiaries established self-sustaining feral populations in tem-
perate regions but not in the tropics.
African honeybees are better adapted to tropical conditions. For this reason, a sub-
species from South-Central Africa, A.m. scutellata Lepeletier (Ruttner 1988), was intro-
duced into Brazil in 1956 with the hope of increasing commercial honey production (Kerr
1967). Unlike the European bees, the African bees established a large feral population
that has expanded through much of the neotropics over the last 35 years (Michener
1975, Taylor 1977, 1985). Within the past year, the front of the African bee population
has spread into Texas. Temperate climatic and ecological conditions are expected to
limit African bees to the southern tier of the United States. A hybrid zone may be

258 Florida Entomologist 75(2) June, 1992

formed between them and the resident European bee population to the north (Taylor
& Spivak 1984, Taylor 1985).
Although African bees are better adapted to the tropics, several characteristics make
them less amenable to commercial beekeeping (Michener 1975, Taylor 1985). The most
notable problem is the African bees' extremely defensive stinging which has been respon-
sible for the deaths of animals and humans (Taylor 1986). Within a few years after
African bees entered South and Central American countries, their beekeeping industries
collapsed. In the United States, the value of the bee for pollination exceeds by almost
a hundred fold the value for honey production (McDowell 1984, Robinson et al. 1989).
If beekeepers in this country go out of business, the greatest economic damage will be
from the loss of crops requiring pollination (Taylor 1985, 1988).
Due to a combination of factors, Florida will be the state most severely affected by
the African bee. Because of the subtropical climate, the feral population will occupy the
entire state and may reach saturating densities (Taylor & Spivak 1984). Florida has a
large beekeeping industry, and many crops within the state require pollination. Annually,
thousands of colonies from Florida are transported to meet pollination needs out of the
state. Sensationalized publicity surrounding stinging incidents could generate an exag-
gerated perception of danger from African bees and seriously impact tourism.
The front of the expanding African bee population may reach the panhandle of Florida
by 1995 (Taylor 1985). Over the last several years, swarms have repeatedly entered
Florida ports as stowaways aboard ships arriving from South and Central American
countries. The Florida Department of Agriculture and Consumer Services (FDACS)
and the United States Department of Agriculture Animal and Plant Health Inspection
Service (USDA APHIS) maintain pheromone-baited hives in the proximity of the ports
to intercept this type of introduction. These agencies identify the colonies suspected to
be of African descent using morphometrics (reviewed by Daly 1991).
Several colonies identified as either African or European were examined here for
distinguishing nuclear and mtDNA markers. Because of limitations of morphometric
identification (Daly 1988, 1991), DNA testing will likely become an increasingly necessary
tool for regulatory agencies (Hall 1986, 1991), especially as the African bees begin their
en masse entry into Florida. The results presented here demonstrate the value and
drawbacks of current DNA tests for identification of African bees. Needed improvements
are discussed.


Honeybee colony samples were collected in the vicinity of Florida ports by agents
of the FDACS and the USDA APHIS. The samples used for this study were randomly
provided from a larger number collected. Samples came from swarms caught in wooden
or fiber-pot bait hives, from swarm clusters on ships or containers, and from colonies
in apiaries. Brood and some adult samples were transported on ice and then frozen.
Other adult samples were killed, left at ambient temperatures for an unknown number
of hours, put into 95% ethanol, and stored refrigerated thereafter.
The FDACS initially screened the samples using forewing length, a component of a
shortened morphometric method (Rinderer et al. 1987). Samples that scored a 90% or
higher probability of being African were then tested by the USDA Bee Research
Laboratories in either Baton Rouge, LA, or Beltsville, MD, using more thorough mor-
phometric analyses. The fourteen colonies tested in this study are listed in Table 1 which
gives the dates and locations of the collections. Where a ship was involved, the previous
port of call is given. The probability that the sample was African, based on forewing
length, is also provided. These scores, obtained by Dr. Lionel Stange and provided by
Mr. Laurence Cutts, Chief Apiary Inspector, FDACS, are part of the public record.

Hall: DNA Markers In Florida Honeybees 259

As of September 1991, seven colonies in Florida had been identified as African. Four
of these were included in this study.
The DNA isolation protocols for adults and brood were as previously described (Hall
1990). DNA was isolated from a mixture of workers from each colony. Samples 1 through
9 were from larvae or pupae; samples 10 through 14 were from adults. Twenty individuals
were used in each isolation. Restriction endonuclease digestions, electrophoretic separa-
tion, blotting, probe preparation, probe labelling, and hybridizations were as previously
described (Hall 1986, 1990). The polymerase chain reaction (PCR) and digestion of the
amplified segments were performed as previously described (Hall & Smith 1991) except
that the PCR reaction profile was modified as follows: 95C for 1 min, 35 cycles of 93C
for 30 sec, 62C for 90 sec, 720C for 1 min, and a final 72C for 15 min.


The fourteen honeybee colonies listed in Table 1 were tested for mitochondrial and
nuclear DNA polymorphisms that either completely or partially distinguish African from
European honeybees. Four of the colonies, 3, 12, 13 and 14, had been previously identified
as African by morphometrics.

Mitochondrial DNA

To identify the honeybee mtDNA type, a rapid, accurate procedure, employing the
polymerase chain reaction (PCR) was utilized (Hall & Smith 1991). Two regions contain-
ing informative polymorphisms were amplified. One region includes a segment between


Date Location
Collected Port Previous ports Score2

1. Apr30, 1987 Panama City Swarm in tree 0.96
2. Apr30, 1987 Panama City Swarm in tree 0.92
3. 1Apr 24, 1987 Panama City On ship Marco trader 1.00
from Guatemala
4. Jun28, 1988 Fernandina Beach Bait hive 0.70
5. May31, 1989 Tampa Bait hive N.A.
6. Sep21, 1989 Palmetto Apiary colony 0.98
7. Sep21, 1989 Palmetto Apiary colony 0.70
8. Oct06, 1989 Tampa Bait Hive 0.86
9. Oct06, 1989 Tampa Bait Hive 0.86
10. Jul01, 1987 Tampa In pipe at terminal 0.21
11. Jan03, 1988 Tampa Apiary colony 0.11
12. 'May 06,1988 Ft. Lauderdale On ship Senator, 1.00
from Guatemala
and Honduras
13. 'Apr 24, 1989 Miami On container ship 1.00
Werner from Suriname
14. 'Feb07, 1990 Ft. Lauderdale On ship Water Stoker, 1.00
from Guatemala

'Colonies determined to be African based on more extensive morphometric analysis by the USDA.
'Score is the probability of African descent based on forewing length, determined by the Florida Department of
Agriculture and Consumer Services.

Florida Entomologist 75(2)

June, 1992

the cytochrome c oxidase subunit I and II genes, that has an insert of several sizes in
west European and African honeybees. An insert of approximately 70 base pairs (bp)
is most common among neotropical African bees. An insert of about 270 bp is present
at a lower frequency among neotropical African bees but is the most common class
among South African and west European bees (Hall & Smith 1991). Figure 1 shows an
electrophoretic agarose gel of this amplified fragment from the different samples. Sam-
ples 3, 12 and 13 each carried mtDNA with the 70 bp insert, and sample 14 carried
mtDNA with both the 70 and 270 bp inserts. The second amplified region was within
the cytochrome c oxidase subunit I gene that has a HincII restriction site in west
European but not in east European or African bees. From all the samples tested here,
this amplified fragment was not cleaved with HincII (not shown). Thus, colonies 3, 12,
13, and 14 had African mtDNA, and the other colonies had east European mtDNA.
MtDNA is maternally inherited, and, with a few exceptions, the mtDNA molecules
of an individual animal are identical (reviewed in Wilson et al. 1985). Since workers
comprising a honeybee colony are progeny of a single queen or, occasionally, also of a
coexisting daughter queen, the mtDNA within a colony is identical. However, as seen
in Figure 1, sample 14 had two African mtDNA size classes. When individual bees were
tested, each carried one size class or the other but not both. Thus, the presence of two
mtDNA classes was not due to heteroplasmy. Rather, the swarm apparently represented
an amalgamation of bees from more than one colony, a common occurrence among African
bees (Kigatiira 1988). This swarm was described as being unusually large (L. Cutts,
personal communication).

1 2 3 4 5 6 7 8 9 10 11 12 13 14


-me 0.87

Fig. 1. Samples of Florida honeybee colonies, suspected to be of African descent,
tested for a distinguishing mtDNA length polymorphism.
A region of the mitochondrial genome that includes part of the cytochrome c oxidase
subunit I and II genes and the intervening segment was amplified by the PCR. The
amplified region from each sample was separated by electrophoresis and stained with
ethidium bromide. From samples 3, 12, 13 and 14, an amplified fragment of about a 918
bp (base pairs) in length was obtained, about 70 bp longer than the 848 bp amplified
fragment from the other samples. From sample 14, an additional amplified fragment of
about 1118 bp in length was obtained, about 270 bp longer. The inserts are characteristic
of both west European and African bees. The mtDNA of these samples were determined
not to be west European, because it lacked a HincII restriction site in another amplified
region within the cytochrome c oxidase subunit I gene. Molecular size standard in outer
lanes is bacteriophage QX174 disgusted with HAEIII. Sizes given in Kilobases.


Hall: DNA Markers In Florida Honeybees

Nuclear DNA

Using standard RFLP (restriction fragment length polymorphism) analyses, the
samples were tested with four cloned nuclear DNA probes. The clones carried random
inserts of honeybee genomic DNA and, thus, likely represented four separate loci (Hall
1986). In the following paragraphs, a brief description is given of the alleles detected
by each probe and their population distributions. More detail can be found in the reports
cited or in preparation.
In AluI digests, probe P130 detects an allele, E, that is characteristic of the Italian
race, A.m. ligustica, (100% frequency in a small number of samples) and perhaps of
other east European races. Allele E is found at about a 75% frequency in United States
honeybee populations. The alternate allele, 0, is fixed in west European and African
bees (Hall 1990). Figure 2A shows that samples 3, 12, 13 and 14 lacked allele E. This
finding was consistent with the mtDNA and morphometric results identifying these
colonies as African, although the absence of allele E could have reflected a west European
In AluI digests, probe 2A2 detects an allele, B, present at a high frequency (about
80 to 90%) in South African bees, present at a lower frequency (about 25%) in west
European bees and absent in east European bees. So far, among North American colonies
of European descent, allele B has been found at a low frequency (about 6%) only among
feral colonies in northern Mexico (prior to African bee invasion) (Hall 1992). Figure 2B
shows that allele B was present in samples 3, 12, 13 and 14. Similar to the findings with
probe P130, this result was consistent with the morphometric and mtDNA results
identifying an African ancestry, but allele B could have come from a west European
lineage. Sample 9 also carried the allele. The other results did not indicate that this
sample had an African background. Nevertheless, the presence of allele B may have
resulted from hybridization with African drones.
In Mbol digests, probe P271 detects an allele, S, found only in African-derived
populations. The frequencies of this allele are approximately 25% to 45% in Old and
New World African populations (McMichael & Hall, manuscript in preparation). Figure
2C shows that samples 12, 13 and 14 carried the allele, again consistent with the mor-
phometric and mtDNA results. Due to the specificity of this marker compared to those
described above, its presence indicated an African background with greater certainty.
This allele was absent in sample 3 which was an African matriline. A faint indication of
the African allele was seen in sample 11. As with sample 9 carrying allele 2A2-B, the
other results did not point to an African background for sample 11. However, the results
did not preclude that this colony had some progeny of African paternity. This colony
was originally suspect because of its excessive defensive stinging but, on the basis of
forewing length, had a low probability of being African (11%).
In MspI digests, probe P178 detects several alleles specific for African bees, collec-
tively referred to here as the A alleles, characterized by the presence of a 1.1kb fragment.
The A alleles, as a group, are present at about a 15% frequency in South African bees
and at a 15 to 25% frequency in neotropical populations (McMichael & Hall, manuscript
in preparation). Figure 2D shows that samples 13 and 14 carried A alleles, reliably
indicating an African background, consistent with the morphometric and mtDNA results.
The A alleles were absent in samples 3 and 12 which had an African maternal origin.
In MspI digests, probe P178 detects another allele, M, present in almost all bees
tested of the west European race A.m. mellifera but absent in east European bees. It
is also found at about a 5% frequency in South African bees (McMichael & Hall, manuscript
in preparation). Figure 2D shows that samples 1, 3, 4, 5, 6, 8, 9, 11 and 14 carried this
allele. The strong presence of the P178-M allele in samples 3 and 14, which had African
mtDNA but not the P130-E nuclear allele characteristic of the east European A.m.


Florida Entomologist 75(2)

June, 1992

1 2 3 4 5 6 7 B 9 10 11 12 13 14
P130 Au I

. .6 [01

S1.2 [El

S 2A2 Akul

-6,6W .a [si

1.1 [Al

Fig. 2. The honeybee samples tested for nuclear DNA RFLPs. Digested DNA from
a mixture of sibling workers from the different colonies is in the separate lanes. Different
intensities of bands in the lanes are due to varying quantities and qualities of the isolated
DNA loaded on the gels.
A. An RFLP generated by AluI and detected with probe P130: samples 3, 12, 13
and 14, and only these samples, lacked the 1.2 kb (kilobase) fragment characteristic of
the east European A.m. ligustica allele E. The 1.6 kb fragment is characteristic of the
alternate allele 0. In the original study (Hall 1990), another fragment between the
arrows was also detected but not with this preparation of the probe.
B. An RFLP generated by AluI and detected with probe 2A2: samples 3, 9, 12, 13
and 14 carry the 2.6 kb fragment characteristic of allele B. The larger fragment is 2.8
kb in length and is characteristic of allele A. Allele B is present at a high frequency in
African bees but is also present in west European bees.
C. An RFLP generated by MboI and detected with probe P271: samples 11, 12, 13
and 14 carry the 0.9 kb fragment present in allele S. This allele is specific to African bees.
D. Two RFLPs generated by MspI and detected with probe P178: samples 13 and
14 carry the 1.1 kb fragment characteristic of several alleles specific to African bees,
collectively referred to here as the A alleles. The 1.8 kb fragment indicated by the open
arrow is characteristic of allele M found predominantly in west European bees.



Hall: DNA Markers In Florida Honeybees

ligustica, suggested that these African matrilines had hybridized with west, but not
east, European bees.


Among the suspected African honeybee colonies found in Florida and tested in this
study, there was a strong correlation among the mtDNA types, the nuclear DNA
markers and the previous morphometric identifications. The same four colonies, 3, 12,
13 and 14, identified as African by morphometrics were determined here to have African
mtDNA. All four colonies carried two nuclear DNA markers alleless P130-0 and 2A2-B)
common to African bees, although the markers are also present in west European bees.
Sample 12 also carried one African-specific marker (allele P271-S), and samples 13 and
14 also carried two African-specific markers (allele P271-S and P178-A). These alleles
are absent in the majority of African bees, which may be the reason that sample 12
lacked one of the two alleles and sample 3 lacked both. Alternatively, the absence of
these African alleles could have been due to hybridization with European bees. In
addition to the four samples that showed consistent indications of an African identity,
sample 11 carried a very low level of the P271-S allele, and sample 9 carried the 2A2-B
allele. The presence of these markers may have been due to some African paternal
introgression, although the latter may have reflected a west European lineage.
The results from the three types of identification methods strongly reinforce each
other. Separately, they provide a less certain identification, reflecting inherent limita-
tions in the different types of analyses or the individual markers.
The morphometric method provides a probability that bees are either African or
European but does not distinguish hybrids. Intermediate morphometric scores were
found to be more common among bees in the tropical-temperate transition zone in South
America, a result taken as evidence for African-European bee hybridization (Sheppard
et al. 1991). However, intermediate probabilities mean that the bees cannot be confidently
identified. Previous DNA studies have revealed little hybridization of feral neotropical
African colonies with European bees (Hall 1990). The limited admixture has probably
helped retain the effectiveness of the morphometric method and may be responsible for
the good correlation with the DNA results reported here. With increased African-Euro-
pean hybridization expected as African bees approach the temperate climates of the
United States (Taylor and Spivak 1984, Taylor 1985, Lobo et al. 1989), the reliability
of morphometric identification is likely to suffer.
Only samples with the highest probability (1.00) of being African, on the basis of
forewing length, were ultimately determined to be African using more complete mor-
phometric tests and DNA analyses. However, most of the other samples had scores of
high African probability (Table 1) but were found not to be African upon further analysis.
Thus, scores less than 1.00, obtained by the abbreviated method, appear to be question-
able as useful indicators of African background.
The identification of African honeybee mtDNA is precise and unambiguous, but it
only determines the matrilineal origin. Colonies found with African mtDNA, as the four
samples reported here, are derived as continuous maternal lineages from the bees intro-
duced from Africa. Previous studies have demonstrated that the feral neotropical African
population is comprised almost exclusively of African matrilines (Hall & Muralidharan
1989, Smith et al. 1989, Hall & Smith 1991). Therefore, in testing neotropical colonies
with mtDNA alone, an African maternal ancestry can be recognized in almost all feral
colonies and in most managed colonies. However, an African paternity in European
matrilines, which can result in significant defensive behavior, would not be recognized.
Although hybrids cannot be identified using only mtDNA, valuable insight can be ob-


Florida Entomologist 75(2)

trained using mitochondrial and nuclear DNA together to distinguish maternal and pa-
ternal gene flow (Hall 1990).
The presence or absence of introgression between populations can be demonstrated
using frequency differences among nuclear alleles, even if the alleles are not specific.
A few honeybee allozymes have been useful for this purpose (Lobo et al. 1989, Smith
et al. 1989). However, to identify certain African or European paternal ancestry in
individual hybrids bees, nuclear alleles specific to each group are required (Page &
Erickson 1985). The most useful markers are those present at higher frequencies, thereby
representing more members of the group. Alleles that are diagnostic represent virtually
all members (Ayala & Powell 1972). Because segregation results in the loss of alleles in
hybrids, extensive admixture reduces the effectiveness of any single marker as an
identifier. Thus, it is important to have specific alleles representing a number of loci.
Most of the useful honeybee nuclear DNA polymorphisms discovered, like two em-
ployed in this study, distinguish African from east European bees. Fewer markers have
been found that distinguish African from west European bees. The other two markers
used here appear to have the needed African specificity but are not present at high
frequencies. The multiple paternity of honeybee colonies increases the probability that
even low frequency African alleles would be present among the members of an African
colony. Despite the limited specificities and frequencies of these nuclear DNA markers,
they are effective when used together, as shown by the high correlation obtained here
among the different African identifiers. Expanding the collection of markers, particularly
with specific alleles at higher frequencies, will provide greater ability to characterize
African-European hybridization in the United States.
Testing of a number of individuals is necessary to include the multiple patrilines of
a honeybee colony. To facilitate the screening of colonies, individuals can be mixed
together as one sample, as in this study. However, to quantitate more accurately the
presence of the DNA alleles and to ascertain the genotypes of the queen and of the
drones with which she has mated, individuals must be tested separately (Hall 1990).
For this purpose, the samples must be well-preserved. That level of accuracy was not
sought here, and many of the samples that had been provided as adults were not
adequately preserved to allow recovery of sufficient intact DNA from individuals for
these analyses. Other useful markers are available but are only effective in testing
individuals. These markers represent multiple alleles characterized by differences in
several restriction fragments (Hall 1992, McMichael & Hall, unpublished data). Because
the different alleles share fragments in common, allele identities can be obscured in
heterozygous individuals and especially in mixed sibling samples.
A major limitation of standard restriction fragment analyses is the expense, time
and labor involved, especially if a number of individuals from each colony must be tested.
Furthermore, low quality DNA preparations from poorly preserved specimens can pre-
clude analysis. This study serves to contrast a newer PCR-based method used to identify
the mtDNA (Hall & Smith 1991) with the standard method used to analyze the nuclear
DNA. With the new method, very low quantities of DNA are needed, and the prepara-
tions can be more crude. The entire identification procedure can be accomplished in
about 12 hours and is relatively inexpensive. The standard approach involves more than
a week and is much more costly. The application of the PCR to the analysis of honeybee
nuclear DNA polymorphisms is forthcoming. Implementation of this technology will
enhance the search for additional markers and make nuclear DNA analyses feasible for
regulatory identification and breeding stock certification.
Morphometric identification of honeybees is limited by the few known subtle mor-
phological differences. MtDNA already allows for precise and rapid distinction but will
always be limited to matriline identification. With the large amount of nuclear DNA,
many more distinguishing polymorphisms can be found, and the application of new

June, 1992


Hall: DNA Markers In Florida Honeybees


technologies will greatly facilitate its use. Thus, the use of nuclear DNA for identifying
bees of African ancestry has the greatest potential for continued improvement.


I thank Laurence Cutts and the apiary inspectors of the State of Florida and the
USDA APHIS for providing the samples and relevant information. I thank Margaret
McMichael for her help with the analyses. I am grateful to Andrew Cockburn, Paul
Shirk and Karen Kilgore for critically evaluating the manuscript. This work was sup-
ported by the USDA Competitive Research Grants Office. This is Florida Experiment
Station Journal Series No. R-02072.


AYALA, F., AND J. POWELL. 1972. Allozymes as diagnostic characters of sibling
species of Drosophila. Proc. Natl. Acad. Sci. U.S.A. 69: 1094-1096.
DALY, H. V. 1988. Overview of the identification of Africanized honey bees, p. 245-249
in G. R. Needham, R. E. Page, M. Delfinado-Baker, and C. E. Bowman [eds.].
Africanized honey bees and bee mites. Ellis Horwood Limited, Chichester, United
DALY, H. V. 1991. Systematics and identification of Africanized honey bees, p. 13-44
in M. Spivak, M. Breed, and D.J.C. Fletcher [eds.]. The "African" honey bee.
Westview Press, Boulder, Colorado.
HALL, H. G. 1986. DNA differences found between Africanized and European hon-
eybees. Proc. Natl. Acad. Sci. U.S.A. 83: 4874-4877.
HALL, H. G. 1990. Parental analysis of introgressive hybridization between African
and European honeybees using nuclear DNA RFLPs. Genetics 125: 611-621.
HALL, H. G. 1991. Genetic characterization of honey bees through DNA analysis, p.
45-73 in M. Spivak, M. Breed, and D.J.C. Fletcher [eds.]. The "African" honey
bee. Westview Press, Boulder, Colorado.
HALL, H. G. 1992. Further characterization of nuclear DNA RFLP markers that
distinguish African and European honeybees. Arch. Insect Biochem. Physiol. 19:
HALL, H. G., AND K. MURALIDHARAN. 1989. Evidence from mitochondrial DNA
that African honey bees spread as continuous maternal lineages. Nature 339:
HALL, H. G., AND D. R. SMITH. 1991. Distinguishing African and European honey
bee matrilines with amplified mitochondrial DNA. Proc. Natl. Acad. Sci. U.S.A.
88: 4874-4877.
KERR, W. E. 1967. The history of the introduction of African bees to Brazil. South
African Bee J. 39: 3-5.
KIGATIIRA, K. I. 1988. Amalgamation in tropical honey bees, p. 62-71 in G. R.
Needham, R. E. Page, M. Delfinado-Baker, and C. E. Bowman [eds.]. Africanized
honey bees and bee mites. Ellis Horwood Limited, Chichester, United Kingdom.
LOBO, J. A., M. A. LAMA, AND M. A. MESTRINER. 1989. Population differentiation
and racial admixture in the Africanized honeybee (Apis mellifera L.). Evolution
43: 794-802.
MCDOWELL, R. 1984. The Africanized honey bee in the United States. What will
happen to the U.S. beekeeping industry? USDA Agricultural Economic Report
No. 519.
MICHENER, C. D. 1975. The Brazilianbee problem. Annu. Rev. Entomol. 20: 399-416.
OERTEL, E., 1976. Bicentennial Bees. Early records of honey bees in the eastern
United States. American Bee J. 116: 70, 71, 114, 128, 156, 157, 214, 215, 260,
261, 290.
PAGE, R. E., AND E. H. ERICKSON. 1985. Identification and certification of Africanized
honeybees. Ann. Ent. Soc. America 78: 149-158.

266 Florida Entomologist 75(2) June, 1992

PELLETT, F. C. 1938. History of American Beekeeping. Collegiate Press, Ames, Iowa.
p. 1-5, 69-67.
BERT, A. M. COLLINS, AND R. L. HELLMICH. 1987. Improved simple tech-
niques for identifying Africanized honey bees. Apidologie 18: 179-197.
ROBINSON, W. S., R. NOWOGRODZKI, AND R. A. MORSE. 1989. The value of honey
bees as pollinators of U.S. crops. Part II. American Bee J. 129: 477-487.
RUTTNER, F. 1988. Biogeography and taxonomy of honey bees. Springer, Berlin,
SHEPPARD, W. S. 1989. A history of the introduction of honey bee races into the
United States. American Bee J. 129: 617-619, 664-667.
SHIMANUKI. 1991. Gene flow occurs between African- and European-derived
honey bee populations in Argentina. Nature 349: 782-784.
SMITH, D. R., O. R. TAYLOR, AND W. W. BROWN. 1989. Neotropical Africanized
honey bees have African mitochondrial DNA. Nature 339: 213-215.
TAYLOR, 0. R. 1977. The past and possible future spread of Africanized honeybees
in the Americas. Bee World 58: 19-30.
TAYLOR, O. R. 1985. African Bees: Potential impact in the United States. Bull. Ent.
Soc. America 31: 14-24.
TAYLOR, O. R. 1986. Health problems associated with African bees. Ann. Internal
Med. 104: 267-268.
TAYLOR, O. R. 1988. Ecology and economic impact of African and Africanized honey
bees, p. 29-41 in G. R. Needham, R. E. Page, M. Delfinado-Baker, and C. E.
Bowman [eds.]. Africanized honey bees and bee mites. Ellis Horwood Limited,
Chichester, United Kingdom.
TAYLOR, O. R., AND M. SPIVAK. 1984. Climatic limits of tropical African honeybees
in the Americas. Bee World 65: 38-47.
SAGE, AND M. STONEKING. 1985. Mitochondrial DNA and two perspectives
on evolutionary genetics. Biol. J. Linnean Soc. 26: 375-400.

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