Title: Florida Entomologist
Full Citation
Permanent Link: http://ufdc.ufl.edu/UF00098813/00040
 Material Information
Title: Florida Entomologist
Physical Description: Serial
Creator: Florida Entomological Society
Publisher: Florida Entomological Society
Place of Publication: Winter Haven, Fla.
Publication Date: 1999
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: VID00040
Source Institution: University of Florida
Holding Location: University of Florida
Rights Management: Open Access

Full Text

Wetterer et al.: Dominance by Alien Ant in Biosphere 2


1Honors College, Florida Atlantic University, 5353 Parkside Drive
Jupiter, FL 33458

2Biosphere 2 Center, P.O. Box 689, Oracle, AZ 85623

'International Centre for Insect Physiology & Ecology, Box 30772, Nairobi, Kenya

4Department of Entomology, Smithsonian Institution, Washington, DC 20560

5Department of Entomology, University of Arizona, Tucson, AZ 85721

6Department of Zoology, Arizona State University, Tempe, AZ 85287

'Department of Entomology, University of California, Riverside, CA 92521

'Institute of Ecotechnics c/o 7 Silver Hills Road., Santa Fe, NM 87505

'1365 Jacobs Place, Dixon, CA 95620


Tramp ants are invading disturbed ecosystems worldwide, exterminating untold
numbers of native species. They have even invaded Biosphere 2, a 1.28-hectare closed
greenhouse structure built in the Arizona desert as a microcosm for studying ecolog-
ical interactions and global change. Invertebrate surveys within Biosphere 2 from
1990 to 1997 have revealed dramatic changes in faunal composition, including an al-
most complete replacement of the ant fauna by a single tramp ant species.
In 1990-91, surveys in Biosphere 2 found no one ant species dominant. By 1993,
populations of the crazy ant, Paratrechina longicornis (Latreille), a tramp species not
found in 1990-91, had increased to extremely high levels. In 1996, virtually all ants
(>99.9%) coming to bait were P. longicornis. We observed P. longicornis foragers feed-
ing almost exclusively on the sugary excretions (honeydew) produced by vast num-
bers of Homoptera, primarily scale insects and mealybugs, found on many of the
plants. High densities of ants were associated with high densities of homopterans. In
1997, soil and litter surveys found that the only invertebrates thriving in Biosphere
2, besides P. longicornis and homopterans, were either species with effective defenses
against ants (well-armored isopods and millipedes) or tiny subterranean species that
can escape ant predation (mites, thief ants, and springtails). A convergent pattern of
biodiversity occurs in disturbed tropical and subtropical ecosystems dominated by
tramp ants.

Key Words: ants, invasive species, microcosms, Paratrechina longicornis

Florida Entomologist 82(3) September, 1999


En todo el mundo, hormigas vagabundas han estado invadiendo ecosistemas per-
turbados, eliminando grandes numeros de species nativas. Han invadido hasta la
"Bi6sfera 2", una estructura cerrada estilo invernadero de 1.28 hectareas que fue
construida en el desierto de Arizona como un microcosmo para el studio de interac-
ciones ecol6gicas y de cambios globales. Censos de los invertebrados dentro de "Bi6s-
fera 2" de 1990 a 1997 han revelado cambios dramaticos en la composici6n faunistica,
incluyendo un reemplazo casi total de la fauna de las hormigas por una sola especie
de hormiga vagabunda.
En 1990-91, censos dentro de "Bi6sfera 2" indicaron que ni una sola especie de hor-
miga era dominant en sus numerous. En 1993, la poblaci6n de la hormiga Paratre-
china longicornis (Latreille), una especie vagabunda que no se habia encontrado en
1990-91, habia aumentado a niveles extremadamente altos. En 1996, virtualmente
todas las hormigas (>99.9%) que habian visitado los cebos eran de la especie P. longi-
cornis. Observamos que estas hormigas se alimentaban casi exclusivamente de las se-
creciones azucaradas (miel de rocio) producidas por grandes numeros de Homoptera,
principalmente escamas y cochinillas harinosas, que se encontraban en muchas de las
plants. Densidades altas de hormigas estaban asociadas con densidades altas de ho-
m6pteros. En 1997, se descubri6 en censos de suelo y de hojarasca que los unicos in-
vertebrados que estaban prosperando en "Bi6sfera 2", ademas de P. longicornis y
hom6pteros, eran species con defenses eficaces en contra de las hormigas (is6podos
y milpies con buenas armaduras) o pequefisimas species subterraneas que escapan
depredaci6n por las hormigas (acaros, hormigas robadoras y colembolos). Un patron
convergente de biodiversidad existe en ecosistemas tropicales y subtropicales pertur-
bados y dominados por hormigas vagabundas.

Biosphere 2 is a 1.28-hectare closed greenhouse structure built in the Arizona
desert north of Tucson as a model for a self-contained space colony and used as a mi-
crocosm for understanding ecosystem dynamics and global change. Biosphere 2 con-
tains not only human residences and an agricultural area, but also "wilderness" areas
designed to model natural biomes, including a desert, a savanna, a rain forest, a
marsh, and an ocean (Alling et al. 1993, Nelson et al. 1993). Eight people lived sealed
in Biosphere 2 for two years, from 1991 to 1993, and 7 more lived inside for 6.5 months
in 1994. These experiments indicated that maintaining a balance of atmospheric
gases suitable for human life was difficult in this closed system. For example, in the
fifteen months after initial closure in September 1991, the oxygen level dropped from
21 percent to 14 percent, requiring oxygen to be added to the system (Broecker 1996,
Cohen & Tilman 1996). In addition to atmospheric changes, many ecological changes
within Biosphere 2 were noted (Broecker 1996, Cohen & Tilman 1996).
The plants and animals living in Biosphere 2 at the time of initial closure came
from many sources (Alling et al. 1993, Nelson et al. 1993). Planners imported hun-
dreds of species of plants and animals and integrated them into Biosphere 2 in an at-
tempt to recreate self-sustaining artificial ecosystems. Many species were selected to
fill specific community functions, e.g., bees and hummingbirds were introduced to
serve as pollinators. Other species, such as chickens and bananas, were included as
sources of human food. As Biosphere 2 was being constructed, many local Arizona
plant and animal species also took up residence in the structure (Alling et al. 1993,
Nelson et al. 1993). In addition, accidental species arrived in building material and in
soil and plant samples.

Wetterer et al.: Dominance by Alien Ant in Biosphere 2 383

It has been widely reported anecdotally that the crazy ant, Paratrechina longicor-
nis (Latreille), has taken over Biosphere 2 and that little else of the original fauna re-
mains. For example, Cohen & Tilman (1996) made the widely repeated but
undocumented observation that in Biosphere 2, "the majority of the introduced in-
sects went extinct, leaving crazy ants (Paratrechina longicornus) (sic.) running every-
where, together with scattered cockroaches and katydids." Indeed, at present, P.
longicornis swarm in exceedingly high numbers in many parts of Biosphere 2, but
other insects occur in very high densities as well. For example, plant-feeding Ho-
moptera, primarily scale insects (Families Coccidae, Diaspididae, and Asterolecani-
idae) and mealybugs (Family Pseudococcidae), are extremely common in Biosphere 2.
Homopterans feed by piercing plant tissue with their hypodermic-like mouthparts
and ingesting the phloem liquid. Homopterans pass large volumes through their di-
gestive tracts, excreting a sugary liquid called honeydew. Homopterans are major ag-
ricultural pests, in both the field and in greenhouses, causing damage both through
sapping plants of nutrients and by increasing the occurrence of diseases, including vi-
ral and fungal infections (Power 1992, Miller & Stoetzel 1997). Many species of ants,
including P. longicornis, feed on the honeydew produced by homopterans (Holldobler
& Wilson 1990). In turn, ants protect homopterans from attack by predators and par-
asites, boosting homopteran population levels. Homopteran exudate appears to be an
important source of food for many ants. Due to the tremendous economic impact of ho-
mopterans in agricultural systems, a vast literature exists on the relationship be-
tween homopterans and their attendant ants (Holldobler & Wilson 1990).
In the present study, we documented P. longicornis' rise to dominance in Biosphere
2 and its relationship with homopterans. Further, we surveyed the current inverte-
brate biodiversity. These analyses may provide novel insights into ecological dynam-
ics in highly-disturbed "natural" ecosystems.


Ant and Homopteran Introductions

We sought all records of intentional introductions of ants and homopterans to Bio-
sphere 2, searching internal reports of Biosphere 2.

Ant Surveys

We conducted surveys of ants on numerous occasions between 1990 and 1997. Be-
tween August 1990 and March 1991, D. E. W. conducted visual surveys of Biosphere
2, hand collecting ants. In August 1990, C. A. 0. conducted surveys using hand collect-
ing, sweep nets, and pitfall traps. From April to August 1993, during the first human
enclosure experiment (Mission 1), M. N. ran ~25 pitfall traps once a month through-
out the wilderness biomes. From 27 September to 4 October 1993, immediately after
the end of Mission 1, S. E. M, D. A. P., J. L., and S. Buchmann conducted surveys using
hand collecting, aerial and sweep netting, Malaise traps, ultraviolet lights, ant baits
(honey, peanut butter and tuna-based cat food), pitfall traps, Berlese funnel samples
of soil and leaf litter, and yellow pan traps (for ground-dwelling and low-flying in-
sects). In January 1994, C.A.O. surveyed the desert area. In November 1996, J. K. W,
I. W. A., K. R. H., M. P., E. L. H., and J. G. made visual ant surveys and collected ants
at 174 bait stations (tuna in oil and "Keebler Pecan Sandies" cookies) set out at 5 m
intervals along every trail throughout the wilderness areas.

Florida Entomologist 82(3) September, 1999

Soil and Leaf Litter Surveys

In June 1997, J. K. W., A. G. H., M. M. Y, and C. E. D. collected 18 1-liter soil and
leaf litter samples from 18 widely separated parts of "wilderness" areas on Biosphere
2 (6 each from the desert, savanna, and rain forest biomes). We placed the samples in
Berlese funnels for 48 hours to extract ants and other invertebrates.

Homopteran-tending by Ants

Between September 1996 and March 1998, J. K. W. searched for the food source of
foraging P. longicornis in Biosphere 2 during numerous visits, tracing hundreds of for-
aging trails to their origins.
In November 1996, J. K. W, I. W. A., M. P., E. L. H., and J. G. counted ants foraging
on 24 large Thalia geniculata L. leaves, each on a different plant in order to document
the association between ants and homopterans. The number of ants present on each
leaf was assessed 3 times at 10 minute intervals on one day, and 3 times again 2 days
later. The leaves were visually ranked according to density of scale insects, and all
scale insects were counted on the 17 lowest-ranked leaves and the one highest. Ant
visitation rate was compared for the 8 low-, 8 medium-, and 8 high-ranked leaves.


Ant and Homopteran Introductions

Between August 1990 and August 1991, before Biosphere 2 was sealed, W. Scott re-
corded 1 or more colonies of 11 ant species intentionally introduced by S. Schneider
and others: Camponotus sexguttatus (Fabr.), Camponotus ustus (Forel), Camponotus
festinatus (Buckley), Camponotus mina Forel, Camponotus sp., Dorymyrmex insanus
(Buckley) (formerly Conomyrma insana), Crematogaster steinheili (Forel), Cremato-
gaster sp., Cyphomyrmex minutus Mayr, Forelius pruinosus Roger, and Pheidole sp. 1.
These species were selected, based on advice from J. Longino and others, to serve a va-
riety of vital functions such as seed-dispersal and recycling of dead plant and animal
material. None of these deliberately-introduced ant species was recorded in any sub-
sequent surveys (Table 1).
We found no records of intentional introductions of homopterans into Biosphere 2.

Ant Surveys

Between August 1990 and March 1991, invertebrate surveys conducted to docu-
ment pre-closure conditions recorded 9 ant species (A in Table 1), including local na-
tive ants, as well as "tramp" ants, i.e., species dispersed worldwide by human
commerce and associated with human disturbance (H6lldobler & Wilson 1990). No
single ant species showed clear dominance.
Between April 1993 and January 1994, surveys conducted during and immedi-
ately after the first human enclosure experiment recorded 6 ant species (B in Table 1),
including 2 tramp ant species not recorded before closure in 1991. One of these spe-
cies, Paratrechina longicornis, now occurred in high numbers. The five other ant spe-
cies occurred in very low numbers.
In November 1996 and June 1997, surveys recorded 10 ant species (C in Table 1),
including 7 species not recorded previously and only 1 species found in surveys before

Wetterer et al.: Dominance by Alien Ant in Biosphere 2 385

TABLE 1. ANTS OF BIOSPHERE 2. SURVEY: A = 1990-91; B = 1993-94; C = 1996-97. STA-

Species Survey Status

Odontomachus clarus Roger A N
Crematogaster opuntiae Buren A N
Linepithema humile (Mayr) A T
Paratrechina bourbonica (Forel) A T
Pheidole sp. 2 A ?
Pheidole hyatti Emery AB N
Solenopsis xyloni (MacCook) AB N
Cardiocondyla ectopia Snelling AB T
Hypoponera opaciceps (Mayr) ABC T
Paratrechina longicornis (Latreille) BC T
Tetramorium bicarinatum (Nylander) BC T
Solenopsis molesta Say C N
Forelius mccooki (McCook) C N
Cardiocondyla wroughtoni (Forel) C T
Monomorium floricola (Jerdon) C T
Tapinoma melanocephalum (Fabr.) C T
Paratrechina vividula (Nylander) C T
Strumigenys rogeri Emery C T

initial closure in 1991. Among the extremely large numbers of ants observed foraging
above ground, every one was P. longicornis, except small numbers of Monomorium flo-
ricola (Jerdon) on 5 plants, and Cardiocondyla wroughtoni (Forel), Tapinoma melano-
cephalum (Fabr.), Tetramorium bicarinatum (Nylander), and Paratrechina vividula
(Nylander) each on a single plant. Of the 28,827 ants collected at baits, all were P. lon-
gicornis, except 1 T bicarinatum specimen.

Soil and Leaf Litter Surveys

Soil and leaf litter surveys in June 1997 found that only a few types of inverte-
brates predominate in Biosphere 2, though many other invertebrates occurred in low
numbers. Of the 1193 ants extracted from the 18 soil and leaf litter samples, there
were 987 P. longicornis, 200 Solenopsis molesta Say, 3 Hypoponera opaciceps (Mayr),
1 M. floricola, 1 Strumigenys rogeri Emery, and 1 Forelius mccooki (McCook). In addi-
tion to ants, we extracted 1271 mites (Acari), 323 Isopoda, 74 millipedes (Diplopoda),
39 springtails (Collembola), 12 earthworms (Oligochaeta), 8 Homoptera, 7 spiders
(Araneae), 7 webspinners (Embioptera), 5 snails (Gastropoda), 4 termites (Isoptera),
3 cockroaches (Blattaria), 3 beetles (Coleoptera), 2 wasps (Hymenoptera), 1 caterpil-
lar (Lepidoptera), 1 lacewing (Neuroptera), and 1 Diplura.

Florida Entomologist 82(3) September, 1999

Although we found few specimens in our soil and leaf litter samples, cockroaches
remain conspicuous throughout Biosphere 2, especially at night. The only other large
insect to maintain a noticeable population in Biosphere 2 was the katydid Scudderia
mexicana (Saussure) (Orthoptera: Tettigoniidae), which we consistently observed in
low numbers on vegetation in the savanna and desert areas.

Homopteran-tending by Ants

Trails of foraging P. longicornis on plants in Biosphere 2 invariably led to ho-
mopterans. High densities of ants on plants were always found to be tending high den-
sities of homopterans, such as scale insects that heavily encrusted the trunks,
branches, and leaves of many Piper trees, and mealybugs that covered the branches
of many mangrove trees. Ants returning from these sources to their nests were
bloated with liquid.
Surveys of ants on Thalia geniculata L. leaves demonstrated a strong positive as-
sociation between ants and scale insects. Leaves with few scales (3 to 23) averaged 0.5
+ 0.8 ants per survey, leaves with an intermediate number of scales (27 to 314) aver-
aged 3.9 + 3.5 ants per survey, and leaves with the most scales (534 to 5755 scales) av-
eraged 51.8 + 28.5 ants per survey (Kruskal-Wallis Test for differences between
means, H = 24.0, df = 2, p < 0.0001).


In the few years since arriving in Biosphere 2, the crazy ant, Paratrechina longi-
cornis, has quickly become the ecologically dominant ant species. Paratrechina longi-
cornis is called the "crazy" ant because of its fast, jerky movements. It is of Old World
origin, but is now found throughout the tropics and subtropics, usually associated
with disturbance, e.g., in highly-disturbed natural environments, in cities, in green-
houses, and even on ships (Weber 1940, Smith 1965, Wilson & Taylor 1967, Trager
1984, H6lldobler & Wilson 1990, Yamauchi & Ogata 1995, Wetterer 1998). One strik-
ingly parallel example of ecological dominance by P. longicornis occurs in the Dry Tor-
tugas, the highly-exposed, outermost islands of the Florida Keys (H6lldobler & Wilson
1990). Here, P. longicornis "is an overwhelmingly abundant ant and has taken over
nest sites that are normally occupied by species of Camponotus and Crematogaster,
which are absent from the Dry Tortugas; and open soil, normally occupied by crater
nests of Conomyrma (now Dorymyrmex) and Iridomyrmex (now Forelius), which gen-
era are also absent from the Dry Tortugas" (p. 433; H6lldobler & Wilson 1990).
Besides P. longicornis, the only common ant now in Biosphere 2 is a tiny subterra-
nean thief ant belonging to the S. molesta species group. In California, S. molesta is
the most common native ant persisting in areas invaded by the tramp Argentine ant,
Linepithema humile (Mayr) (formerly Iridomyrmex humilis) (Human & Gordon
1997). Thief ants also persist in areas invaded by the highly-destructive red imported
fire ant (Solenopsis invicta Buren) (Stein & Thorvilson 1989).
In Biosphere 2, P. longicornis foragers appear to obtain almost all their carbohy-
drates from honeydew produced by homopterans. The extremely high densities of ho-
mopterans in Biosphere 2 appear to be essential for maintaining the high populations
of ants. The same may be true in many "natural" ecosystems. Tramp ants most com-
monly invade and dominate disturbed communities (Wilson & Taylor 1967, H6ll-
dobler & Wilson 1990, Yamauchi & Ogata 1995, Wetterer 1998), possibly because they
depend on honeydew produced by plant-feeding homopterans that thrive on second-
ary-growth plants (Tennant & Porter 1991).

Wetterer et al.: Dominance by Alien Ant in Biosphere 2 387

Although earlier anecdotal reports only noted cockroaches and katydids among
the insect species remaining in Biosphere 2, this observation probably reflected not so
much the numbers, but the visibility of these large insects. In contrast, the vast num-
bers of Homoptera covering the plants and mites in the soil are much less conspicu-
ous. The present invertebrate diversity in Biosphere 2 is strikingly skewed in a way
that suggests heavy influence of ants. Although a wide diversity of invertebrates per-
sist in low numbers in Biosphere 2, the only ones that now thrive are species that are:
1) ant mutualists homopteranss tended by ants), 2) ant resistant (well-armored iso-
pods and millipedes), or 3) or escape ant attack by being very small and subterranean
(mites, thief ants, and springtails). Researchers have documented similar patterns of
invertebrate diversity in disturbed tropical and subtropical ecosystems dominated by
tramp ants (Stein & Thorvilson 1989, Porter & Savignano 1990, Cole et al. 1992, Wil-
liams 1994, Human & Gordon 1997). Several of us were struck by how much the char-
acter of the plant and animal community within Biosphere 2 resembled that of
disturbed lowland areas in Hawaii. Thus, Biosphere 2, a 1.28-hectare habitat island
surrounded by desert, appears to be a fairly good ecological analog for a small, highly-
disturbed, subtropical island. As we understand more of the ecological dynamics in
Biosphere 2, the results may provide important insights into the workings of simpli-
fied ecosystems that are useful for the conservation and restoration of Earth's increas-
ingly disturbed habitats. Future research in Biosphere 2 can examine more closely
the interactions among ants, homopterans, and plants and study the impact of biolog-
ical introductions, particularly introductions of natural enemies of ants and ho-


We thank M. Wetterer, A. Wetterer, H. Young, J. Thornton, C. Raxworthy, L. Mor-
rison, J. C. Morales, K. Ingram, D. Gordon, G. Ferraz, and M. Cushman for comments
on this manuscript; S. Cover and R. Selling for identifying ants; J. Longino for pro-
cedural suggestions; W. Broecker, W. Harris, and N. Majer for administrative help; W.
Scott, S. Schneider, S. Buchmann, D. Preston, S. Swift, P. Warshall, A. Weitzman, J.
Allen, M. Augustine, M. Bierner, L. Leigh, B. Marino, S. Silverstone, and J. Poynter for
various assistance; Biosphere 2 Center, Columbia University, the Bishop Museum,
and the University of Arizona for financial support; and E. Bass for financing Bio-
sphere 2.


CALLING, A., M. NELSON, AND S. SILVERSTONE. 1993. Life Under Glass: The Inside
Story of Biosphere 2. Biosphere Press, Oracle, AZ.
BROECKER, W. 1996. The biosphere and me. GSA Today 6: 1-7.
COHEN, J. E., AND D. TILMAN. 1996. Biosphere 2 and biodiversity: the lessons so far.
Science 274: 1150-1151.
COLE, F. R., A. C. MEDEIROS, L. L. LOOPE, AND W. W. ZUEHLKE. 1992. Effects of the
Argentine ant on arthropod fauna of Hawaiian high-elevation shrubland. Ecol-
ogy 73: 1313-1322.
HOLLDOBLER, B., AND E. 0. WILSON. 1990. The Ants. Harvard University Press, Cam-
bridge, MA.
HUMAN, K. G., AND D. M. GORDON. 1997. Effects of Argentine ants on invertebrate
biodiversity in Northern California. Conserv. Biol. 11: 1242-1248.
MILLER, G. L. AND M. B. STOETZEL. 1997. Aphids associated with chrysanthemums in
the United States. Florida Entomol. 80: 218-239.

Florida Entomologist 82(3) September, 1999

WALFORD, AND J. P. ALLEN. 1993. Using a closed ecological system to study
Earth's biosphere: Initial results from Biosphere 2. BioScience 43: 225-236.
POWER, A. G. 1992. Host plant dispersion, leafhopper movement and disease trans-
mission. Ecol. Entomol. 17: 63-68.
PORTER, S. D., AND D. A. SAVIGNANO. 1990. Invasion of polygyne fire ants decimates
native ants and disrupts arthropod community. Ecology 71: 2095-2106.
SMITH, M. R. 1965. House-infesting ants of the eastern United States. U.S. Dept. Ag-
ric. Tech. Bull. 1326: 1-105.
STEIN, M. B., AND H. G. THORVILSON. 1989. Ant species sympatric with the red im-
ported fire ant in southeastern Texas. Southwest. Entomol. 14: 225-231.
TENNANT, L. E., AND S. D. PORTER. 1991. Comparison of diets of two fire ant species
(Hymenoptera: Formicidae): solid and liquid components. J. Entomol. Sci. 26:
TRAGER, J. C. 1984. A revision of the genus Paratrechina (Hymenoptera: Formicidae)
of the continental United States. Sociobiol. 9: 51-162.
WEBER, N. A. 1940. Ants on a Nile River steamer. Ecology 21: 292-293.
WETTERER, J. K. 1998. Non-indigenous ants associated with geothermal and human
disturbance in Hawai'i Volcanoes National Park. Pac. Sci. 52: 40-50.
WILLIAMS, D. F. (Ed.) 1994. Exotic ants. Biology, impact, and control of introduced spe-
cies. Westview Press, Boulder, CO.
WILSON, E. 0., AND R. W. TAYLOR. 1967. Ants of Polynesia. Pac. Insects Monogr. 14:
YAMAUCHI, K., AND K. OGATA. 1995. Social structure and reproductive system of
tramp versus endemic ants (Hymenoptera: Formicidae) of the Ryukyu Islands.
Pac. Sci. 49: 55-68.


Everglades National Park, Research, 40001 SR 9336, Homestead, Fl. 33034

Current address: 120 W 45th St., 39th Floor, New York, NY 10036


Brazilian pepper (Schinus terebinthefolius Raddi) is the focus of a major restora-
tion project in Everglades National Park, and here I have attempted to broaden our
understanding of this plant and the phenomenon of its invasion. Brazilian pepper leaf
litter fauna (especially ants but also beetles, wasps, centipedes, millipedes, isopods,
and collembola) was compared to that of hammock and pineland. The abundance of
certain species in the different habitats was consistent with previous records of hab-
itat preference and confirmed Brazilian pepper leaf litter as a high-moisture habitat.
In addition, two species collected here were not previously known from the Ever-
glades: Strumigenys rogeri Emery and S. lanuginosa Wheeler, W. M. Hammock leaf
litter had more ant species, more beetle families, and more wasp families than Bra-
zilian pepper. Rank-abundance plots for all three habitats were either log-series or
log-normal distributions, but the sample size was too small to discriminate. K-series
plots showed ants in Brazilian pepper to be more diverse than in pineland but slightly

Clouse:Ants in Brazilian Pepper Leaf Litter

less diverse than in hammock. Brazilian pepper is more likely to harbor exotic ants
than native habitats (52% versus 36% of species were exotic respectively), and exotic
ants are more likely to be found in both hammock and Brazilian pepper than in only
one or the other (67% of species found in both habitats were exotic).

Key Words: Everglades, ants, Brazilian pepper, Schinus, Strumigenys, exotic, Formi-


La pimienta de Brasil (Brazil pepper; Schinus terebinthefolius Radii) es el foco de
atenci6n de un amplio proyecto de restauraci6n en el Everglades National Park. En
este trabajo he intentado ampliar nuestro conocimiento de esta plant y el fen6meno
de su invasion. La fauna que habitat en la hojarasca de la pimienta de Brasil (princi-
palmente hormigas, pero tambien escarabajos, avispas, cienpies, milpies, is6podos y
collembola) se compare con aquella de los habitats tipo "hammock" y "pineland". La
abundancia de ciertas species en los diferentes habitats fue consistent con registros
previous de la preferencia de habitat y confirmaron a la hojarasca de la pimienta de
Brasil como un habitat de alta humedad. Ademas, se colectaron dos species del Ever-
glades no conocidas previamente: Strumigenys rogeri Emery y S. lanuginosa Wheeler,
W. M. La hojarasca del hammock tuvo mayor cantidad de species de hormigas, mas
families de escarabajos y mas families de avispas que la de la pimienta de Brasil. Ma-
pas de abundancia por jerarqufa de los tres habitats tuvieron distribuciones logarit-
micas en series o normales, aunque el tamano de muestra empleado fue demasiado
pequeno para diferenciar entire los dos tipos. Mapas de tipo series K (k-series) mostra-
ron que las hormigas en la pimienta de Brasil tuvieron mayor diversidad que en el ha-
bitat tipo pineland, pero que su diversidad fue menor en comparaci6n con el habitat
tipo hammock. La pimienta de Brasil es mas propensa a albergar hormigas ex6ticas
que los habitats nativos (52% vs. 36% de las species fueron ex6ticas, respectiva-
mente) y es mas probable encontrar hormigas ex6ticas en los habitats tipo hammock
y de pimienta de Brasil combinados que en uno solo de ellos (67% de las species en-
contradas en ambos habitats fueron ex6ticas).

Exotic plants and ants in South Florida demonstrate some of the more disheart-
ening cases of exotic invasion in the world, inhabiting significant areas of the Ever-
glades and Florida Keys ecosystem. Melaleuca quinquenervia Blake (Melaleuca) has
turned open prairie into impenetrable forest, Casaurina spp. (Australian pine) in-
vades beaches, and Schinus terebinthifolius Raddi (Brazilian pepper) dominates 2000
hectares of former agricultural land on Long Pine Key and germinates regularly in
native habitat (Craighead 1971, Ewel et al. 1982). As for ants, most of the species en-
countered in and around houses and gardens are common tropical exotics, most no-
ticeably the exotic fire ants Solenopsis invicta Buren and Wasmannia auropunctata
(Roger). Even worse, exotic ants are significant invaders beyond the confines of lawns
and roadsides, compromising 26% of ant species in Ferster and Prusak's survey of Ev-
erglades National Park (1994) and 33% of Deyrup et al.'s collection from the Florida
Keys (1988), the highest percentage for anywhere in the continental U.S. (Having no
native ants, 100% of Hawaii's ants are exotic (Wilson & Taylor 1967).)
The proliferation of Brazilian pepper has attracted a strong management response
in Everglades National Park. Brazilian pepper is an ornamental introduced in South
Florida during the latter half of the nineteenth century (Bennett & Habeck 1991,
Doren et al. 1990). Being in the family Anacardiaceae, Brazilian pepper produces ir-
ritant toxins that have an allelopathic effect (Morton 1978); this and its ability to

Florida Entomologist 82(3) September, 1999

grow into dense thickets prevents most other plants from entering Brazilian pepper
thickets. A significant Brazilian pepper stand in the Everglades (commonly called the
"Hole-in-the-Donut") consists of about half of 4000 hectares of abandoned agricultural
land on Long Pine Key. This area was farmed from the early 1900s, and after rock
plowing was introduced in the 1950's, various parts were farmed and abandoned up
to the 1970s. The restoration plan for the Hole-in-the-Donut consists mainly of razing
the Brazilian pepper down to unplowed limestone; test plots have shown that this
technique produces wetland prairie relatively free of Brazilian pepper seedlings
(Jones & Doren 1997).
My goal in this study was to understand the Brazilian pepper area in more detail
than previously known. This may seem odd, given that the Hole-in-the-Donut is
scheduled to be destroyed and does not constitute native habitat, but there are theo-
retical and site-specific reasons for such a study. First, the extent to which the Hole-
in-the-Donut area is dominated by Brazilian pepper is extreme in size, intensity, and
persistence. Profiling this area and understanding how it interacts with its surround-
ing native habitat may lend insight to other areas of exotic invasion and even native
habitats and species. Second, the outcome of the Hole-in-the-Donut restoration, while
likely to be wetland prairie, cannot be predicted with complete certainly. There may
be subtle differences between the restored area and native prairie that persist long af-
ter the project is completed, such as the presence or absence of species that were
present or absent when it was dominated by Brazilian pepper. In this study I analyze
the leaf-litter fauna (primarily ants) from the Hole-in-the-Donut, and I compare it to
native leaf litter. I address three questions: (1) What is the microfauna of Brazilian
pepper leaf litter, (2) What does this fauna indicate about the nature of Brazilian pep-
per leaf litter, and (3) What is the relationship between native and exotic ants and na-
tive and Brazilian pepper habitats?


All work was done on Long Pine Key in Everglades National Park, Dade Co., FL
between 13 August and 21 December 1996. Samples were processed at the Dan Beard
Research Center in Everglades National Park, and voucher specimens are currently
property of the Everglades Museum.
The main goal was to compare the abundance and diversity of ants and other small
arthropods in Brazilian pepper, hardwood hammock, and to lesser degree, rocky pine-
land leaf litter. I also wanted to describe the nature of Brazilian pepper and hammock
leaf litter, including moisture content. These sampling efforts were divided into two
phases: a General Survey and Quantitative Sampling.
The General Survey was conducted by taking Burlese funnel samples from Brazil-
ian pepper, hardwood hammock, and rocky pineland. Each sample consisted of a full
one-gallon plastic salable bag of leaf litter. On each date four samples were taken,
with each sample taken 1 to 2 meters from the others. An effort was made on subse-
quent visits to the same habitat to use different access roads and paths than before,
until most areas within each habitat on Long Pine Key had been visited. Each sample
was sifted through a coarse screen, which removed rocks, large sticks, and whole
leaves. Sticks were broken open and sifted before discarding. The remaining debris
was placed in a Burlese funnel and dried for approximately 24 hours. After drying, the
extracted arthropods were identified and pinned. Four samples were taken on each of
the following dates: Brazilian pepper-August 14, 23, 26, 28, 31, September 6, 23, Oc-
tober 2 and 3; Hammock-August 17, 23, 27, September 19 and 26; Pineland-August
13, 20, 21, 29, September 2, 4, and 11.

Clouse:Ants in Brazilian Pepper Leaf Litter

Quantitative Sampling methods resembled the General Survey in that all samples
consisted of a one-gallon bag of leaf litter, and an attempt was made to collect from as
many different areas within each habitat on Long Pine Key as possible (but not in
pineland). The goal of this phase was to compare Brazilian pepper and hammock leaf
litter quantitatively, so the samples were paired to minimize the effect of superfluous
variables, such as weather, and they were analyzed in more detail after each collec-
tion. During each sampling trip, two samples were taken from two different Brazilian
pepper locations, and two were taken from two different hammock locations. Ten trips
were made, yielding 40 separate samples. An effort was made to sample at least three
meters into the Brazilian pepper stands and as far from roads as possible into ham-
mocks. The samples were weighed to the nearest gram and placed, unsifted, into Bur-
lese funnels. After 24 hours, the dried samples were removed, weighed again, and the
number of ants (identified to species), beetles (identified to family), parasitic wasps
(identified to family), centipedes (Chilopoda), millipedes (Diplopoda), isopods, and col-
lembola were counted.

Diversity and Abundance

Rank-abundance and k-series plots were plotted from the following two data sets:
(1) a combination of General Survey and Quantitative Sampling for pineland, Brazil-
ian pepper, and hammock in which the appearance of a species in a sample consti-
tuted one incidence datum (because the numbers of individuals per species per
sample were not counted in the General Survey), and (2) the Quantitative Sampling
data. Rank-abundance plots juxtapose the abundance of each species (on a log scale)
against the abundance rank of each species (most abundant first). As discussed by
Magurran (1988), rank-abundance plots usually conform to one of four distributions:
geometric, log-series, log-normal, or "broken stick." In this study the sample size was
too small to perform goodness-of-fit tests, but hypothetical geometric and "broken
stick" rank-abundance curves were drawn for the Quantitative Sampling data for
comparison; they represent the two extremes between a harsh or empty habitat
where incoming species occupy the same percentage of remaining niche space as pre-
vious settlers (most geometric distributions) and situations where closely related spe-
cies evenly share a single resource (most "broken stick" distributions). Hypothetical
geometric curves in this paper are based on the assumption that arriving species use
50% of the available niche space. K-series plots show the cumulative abundance ver-
sus the rank (on a log scale); higher k-series curves indicate a smaller number of spe-
cies and/or a greater dominance by the most abundant species than lower curves. In
addition, combined data on hymenopteran and coleopteran families were compared.

Density and Clustering

Quantitative Sampling data were used to calculate density measurements (num-
ber of individuals per 100 grams of dry leaf litter) for ants, millipedes, centipedes, iso-
pods, and collembola. Density measurements were compared using t-tests, and
density data were plotted versus moisture content of leaf litter to search for correla-
tions between density and moisture for both ants and non-insects. These data were
analyzed by comparing the density in samples above the median moisture content
with the density in samples below the median moisture content using Mann-Whitney
U tests, done separately for data from hammock and Brazilian pepper habitats.
The average number of individuals of each species per sample was divided into the
variance to produce a measure of clustering for each ant species and non-insects.

Florida Entomologist 82(3) September, 1999

(Variance/mean ratios less than one indicate that individuals are maximizing the dis-
tance between themselves, and greater than one indicates clustering.) The variance/
mean measurement assumes equal volumes for all samples, which, since the dry
weight could not be determined until after the sample was processed, was impossible;
thus I report only those species that were the most clustered or evenly distributed. All
variance/mean ratios greater than 1.78 were significantly clumped at c = 0.002 (Bon-


Leaf-Litter Characteristics

Brazilian pepper leaf litter consisted more of sticks than leaves. It was about three
to four centimeters deep in most areas, and it sat on a layer of hard-packed soil mixed
with broken limestone. A dense network of pencil-sized roots was in the soil. Most
sticks in the litter were smaller than 1 cm in diameter, but larger sticks (1 to 4 cm di-
ameter) were in deeper piles. Larger pieces of debris (to 7 cm diameter) were found at
the base of the largest Brazilian pepper trees. In some areas the sticks seemed to be
hollowed before falling, and termite galleries could be seen in dead twigs still at-
tached to the plants. Brazilian pepper leaf litter contrasted sharply with hammock
leaf litter, which was thick and soft. In the hammock large logs were also common,
and the range in leaf sizes was, of course, much greater. Brazilian pepper leaf litter
collected during the quantitative sampling had a higher water content than hammock
leaf litter (40.22% versus 33.70%; p = 0.04; t-test). Brazilian pepper litter moisture
also tended to be more variable (0.018 versus 0.0094; p = 0.09; F-test).

Diversity and Abundance

At the end of both the General Survey and Quantitative Sampling, hammock leaf
litter had a larger number of ant species (23), followed by Brazilian pepper (20) and
then pineland (9) (Table 1). There were no ant species unique to the pineland, but four
native species were found only in the pineland and the hammock. Seven species were
found only in the hammock, and eight were only in the Brazilian pepper. Pheidole mo-
erens Wheeler was significantly more abundant in the Brazilian pepper than the ham-
mock at the Bonferoni-corrected alpha value of 0.0016 (Chi-squared, df = 1).
Solenopsis abdita Thompson was more common in the hammock than the Brazilian
pepper. Despite the larger number of species collected from hammock habitat in gen-
eral, there was no clear difference in the evenness of these species when compared to
Brazilian pepper.
Rank-abundance plots for Brazilian pepper and hammock are very similar, and
neither habitat has a geometric or "broken stick" distribution (Figs. 1 & 2). Histo-
grams based on these distributions could represent log-series curves or log-normal
curves with small sample sizes (Fig. 2). The k-series curves-which incorporate both
the number and evenness of species-shows Brazilian pepper to be more diverse than
pineland and slightly less diverse than hammock (Fig. 3).
For the combined General Survey and Quantitative Sampling data, 52% of all ant
species were exotics. For just native habitat (hammock and pineland), 36% of the spe-
cies were exotic. Of species found in both native habitat and Brazilian pepper, 67%
were exotic, and of species found only in Brazilian pepper, 55% were exotic.
I collected more families of parasitic wasps and beetles from the hammock than
the pineland or Brazilian pepper. The hammock had five families of wasps: Aphelin-

Clouse:Ants in Brazilian Pepper Leaf Litter

idae, Ceraphronidae, Diapriidae, Encrytidae, and Eulophidae. The pineland and Bra-
zilian pepper samples had wasps only from the family Scelionidae. Hammock also had
more families of beetles (13) than pineland (9) or Brazilian pepper (8). Representa-
tives of the following families of beetles were found exclusively in the hammock: Cry-
tophagidae, Helodidae, Nitidulidae, Scolytidae, Scymaendiae, and Tenebrionidae. The
pineland produced the only Chrysomelids, and the Brazilian pepper had the only
Dytiscids. Hydraenids, and Hydrophilids were found exclusively in the pineland and
Brazilian pepper.

Density and Clustering

The average density of ants in Brazilian pepper samples (16 + SD25 per 100g dry
litter; n = 20) was not significantly different from ant density in hammock (27 + SD32;
n = 20; p > 0.05; t-test). Broken down by species, the number of individuals per 100
grams of hammock leaf litter varied from 0.066 (Brachymyrmex minutus Forel) to 7.4
(Solenopsis abdita), and in Brazilian pepper leaf litter it ranged from 0.056
(Paratrechina bourbonica (F..... I to 5.9 (Paratrechina guatemalensis i F....- I The
densities for ant species were not significantly correlated with the water content of
the leaf litter. Still, Hypoponera opaciceps (Mayr) and Strumogenys rogeri were gen-
erally found in higher-moisture leaf litter (41% and 39% moisture respectively) than
Cyphomyrmex rimosus (Spinola), Solenopsis tenneseesis Smith, M. R., and Brachy-
myrmex minutus (all 25%).
Millipede density did not differ between Brazilian pepper and hammock, but it
was more variable in the hammock (p < 0.0001; F-test; Table 2). Centipedes were more
dense (p < 0.0017; t-test) and more variable (p < 0.0001; F-test) in the hammock. Mil-
lipedes tended to be denser in high-moisture leaf litter in both the hammock and Bra-
zilian pepper (p < 0.025; Mann-Whitney U test; Bonferoni-corrected alpha = 0.0125).
Centipedes did not vary significantly with moisture content in Brazilian pepper or
hammock. Isopods tended to be denser in wetter hammock leaf litter (p < 0.05), but
collembola were not significantly correlated with water content in the hammock or
Brazilian pepper.
Most ants and non-ant arthropods were clumped. In the hammock, those species
most evenly distributed were Aphaenogaster miamiana Wheeler, Brachymyrmex
minutus, Cyphomyrmex rimosus, Monomorium floricola (Jerdon), and Paratrechina
longicornis (Latreille). In Brazilian pepper, Hypoponera opaciceps, Odintomachus
ruginodus Wheeler, Paratrechina bourbonica, and the centipedes showed the most
even distributions. In the hammock, the most clumped species were Crematogaster
minutissima Mayr (variance/mean ratio = 58) and Solenopsis abdita (57). In the Bra-
zilian pepper, the most clumped species were Paratrechina guatemalensis (30) and
Strumigenys rogeri (21).


Two items deserve clarification before going further. First, it is important to re-
member that this study pertains to leaf-litter inhabitants only, since some common
Everglades ants such as Pseudomyrmex gracllis (Fab.), Crematogaster arkinsonii
Wheeler, W. M., and C. ashemeadi Mayr, etc. are noticeably absent. A study of insects
on Brazilian pepper foliage in Lee Co., FL by Cassani (1986) documents an ant fauna
that, while also highly exotic (>25%), contains species not found in this study. Second,
the pineland samples are shown only briefly in the results, for they are difficult to


Species Native? H P B p Ferster & Prusak 19941 Deyrup et al. 1988

Odontomachus ruginodus no 2 3 5 disturbed areas open, disturbed areas
Paratrechina guatemalensis2 no 48 9 23 hammock pineland, densely wooded areas
Pheidole moerens no 11 3 40 ** hammock, slough, wet moist, wooded, and
prairie disturbed
Solenopsis abdita yes 41 3 8 ** hammock and pineland not collected
Solenopsis geminata yes 2 6 2 hammock and pineland sandy soils and open areas
Brachymyrmex depilis' yes 9 6 0 sloughs, mangroves, closely-related sp. in
hammock shaded, rotten wood and
deep leaf litter
Leptothorax torrei yes 2 3 0 wetland prairie only new record for the U.S.
Paratrechina ,./- yes 5 3 0 pineland and wetland upland forest

Ferster and Prusak (1994) categorized collections by seven habitat types: hardwood hammock, rocky pineland, coastal prairie, mangrove, wetland prairie, disturbed, and freshwater
slough. Their notes on habitat preference are repeated here, some of which are originally from Koptur (1992), Smith (1930 and 1933) and Wheeler (1932).
Trager (1984) reports it from semlopen and secondary habitats.
Taxonomically unresolved (Deyrup etal. 1988).
Trager (1984): lives in a variety of habitats.
5Known to invade native and disturbed habitats swiftly.
'Twig and vine nester.
7One dealate found in this study.
'Trager (1994) reports it from flooded, disturbed areas devoid of other ants.
'Aggressive invader of moist areas (Deyrup & Trager 1984).


Species Native? H P B p Ferster & Prusak 19941 Deyrup et al. 1988

Solenopsis tennesseensis yes 11 6 0 pineland and slough wooded area leaf litter
Aphaenogaster miamiana yes 9 0 3 all habitats but pineland nests in rotten wood
Cardiocondyla wroughtonii no 2 0 2 hammock only nests in hollow twigs
Cyphomyrmex rimosus no 14 0 3 pineland not collected
Pheidole floridana3 yes 2 0 8 pineland, hammock, dry shaded areas
disturbed areas
Quadristruma emmae no 11 0 3 hammock and slough disturbed leaf litter
Strumigenys eggersi yes 18 0 7 hammock only shady to partly-shaded ar-
Wasmannia auropunctata' no 18 0 2 all habitats but wetland dominant ant in some na-
prairie or mangrove tive habitats
Brachymyrmex minutus yes 9 0 0 hammock not collected
Crematogaster minutissima yes 5 0 0 hammock and slough deep leaf litter

Ferster and Prusak (1994) categorized collections by seven habitat types: hardwood hammock, rocky pineland, coastal prairie, mangrove, wetland prairie, disturbed, and freshwater
slough. Their notes on habitat preference are repeated here, some of which are originally from Koptur (1992), Smith (1930 and 1933) and Wheeler (1932).
Trager (1984) reports it from semlopen and secondary habitats.
Taxonomically unresolved (Deyrup etal. 1988).
Trager (1984): lives in a variety of habitats.
Known to invade native and disturbed habitats swiftly.
'Twig and vine nester.
'One dealate found in this study.
'Trager (1994) reports it from flooded, disturbed areas devoid of other ants.
'Aggressive invader of moist areas (Deyrup & Trager 1984).


Species Native? H P B p Ferster & Prusak 1994' Deyrup et al. 1988

Hypoponera opacior yes 5 0 0 hammock only soil and rotten wood
Momomorium floricola6 no 2 0 0 pineland, disturbed slough disturbed areas
Paratrechina longicornis no 2 0 0 hammock, coastal prairie, disturbed
and disturbed
Strumigenys lanuginosa' no 2 0 0 not previously recorded not collected
Tapinoma melanocephalum no 2 0 0 hammock and pineland nest in crevices and under
bark o
Brachymymex obscurior3 yes 0 0 2 coastal prairie, slough lawns and beaches
Camponotus abdominalis3 yes 0 0 2 not mangrove and coastal pinelands, hammock edges g
Cardiocondyla nuda no 0 0 2 not coastal prairie, dis- open grassy places 0o
turbed areas, or slough
Hypopnera opaciceps yes 0 0 22 hammock and slough moist, disturbed areas

PFerster and Prusak (1994) categorized collections by seven habitat types: hardwood hammock, rocky pineland, coastal prairie, mangrove, wetland prairie, disturbed, and freshwater (fl
slough. Their notes on habitat preference are repeated here, some of which are originally from Koptur (1992), Smith (1930 and 1933) and Wheeler (1932).
Trager (1984) reports it from semlopen and secondary habitats.
Taxonomically unresolved (Deyrup etal. 1988).
Trager (1984): lives in a variety of habitats.
Known to invade native and disturbed habitats swiftly.
'Twig and vine nester. C
'One dealate found in this study.
'Trager (1994) reports it from flooded, disturbed areas devoid of other ants.
'Aggressive invader of moist areas (Deyrup & Trager 1984). D



Species Native? H P B p Ferster & Prusak 19941 Deyrup et al. 1988

Paratrechina bourbonica" no 0 0 3 not pineland or coastal disturbed areas only
Solenopsis invicta no 0 0 5 not slough moist shaded areas a.
Strumigenys louisianae yes 0 0 5 pineland and slough moist, shaded leaf litter
Strumigenys rogeri' no 0 0 22 not previously recorded not collected

PFerster and Prusak (1994) categorized collections by seven habitat types: hardwood hammock, rocky pineland, coastal prairie, mangrove, wetland prairie, disturbed, and freshwater
slough. Their notes on habitat preference are repeated here, some of which are originally from Koptur (1992), Smith (1930 and 1933) and Wheeler (1932).
Trager (1984) reports it from semiopen and secondary habitats.
Taxonomically unresolved (Deyrup etal. 1988). C
'Trager (1984): lives in a variety of habitats.
5Known to invade native and disturbed habitats swiftly.
'Twig and vine nester. .
7One dealate found in this study.
'Trager (1994) reports it from flooded, disturbed areas devoid of other ants.
'Aggressive invader of moist areas (Deyrup & Trager 1984).

Florida Entomologist 82(3) September, 1999


--2 Hammock
S--- Brazilian pepper
M -4- Pineland

U 10.00


1.00 . .
1 2 3 4 5 6 7 8 9 1011121314151617181920212223
Abundance Rank

Fig. 1. Rank-abundance curves for combined General Survey and Quantitative
Sampling ant data.

standardize and not sources of large amounts of data. This is because "leaf litter" in
the rocky pinelands is often non-existent. (I finally began taking litter from around so-
lution holes and where it had accumulated near fallen trees. This is why the beetles
from the pinelands include both very common and large families as well as ones that
are found mostly near water.)

Brazilian pepper vs. Hammock Leaf Litter

Brazilian pepper leaf litter seems more likely to become saturated and dried than
hammock leaf litter. This could be caused by the greater openness of the Brazilian
pepper canopy relative to hammock. Also, it seems likely that because Brazilian pep-
per leaf litter lies directly on hard-packed soil and rock, it is more likely to be sub-
merged during rainstorms than hammock leaf litter, which sits on top of more
absorbent material.
Overall, Brazilian pepper ants and their relative abundance are consistent with
high moisture. The only ant species found in both hammock and Brazilian pepper but
found significantly more frequently in Brazilian pepper, Pheidole moerens Wheeler, is
known to favor mesic forests. The only two species found frequently in Brazilian pep-
per and not at all in hammock, Hypoponera opaciceps, and Strumigenys rogeri, are
known inhabitants of wet areas. The brazilian pepper beetle list is also consistent
with high moisture: the only beetle family found in exclusively in Brazilian pepper
was Dytiscidae, commonly known as "predaceous diving beetles." In addition, the only
two beetle families found here in both the Brazilian pepper and pineland are also wa-
ter-loving families: Hydraenidae and Hydrophilidae.
Although no ant densities correlated with water content, millipedes showed a
strong preference for wetter leaf litter, and isopods preferred moisture content be-

Clouse:Ants in Brazilian Pepper Leaf Litter

sA. Hammock
10 -

0.01 -.. ..

0.001 0
1-4 5-4 9-16 17-32 33-44 65-121

0.0001 Number of Individmuls


0.0001 Number o B Individuals
0 5 10 15 20
Abundance Rank
Fig. 2. Rank-abundance curve for hammock Quantitative Sampling ant data (II)
for Hammock (A) and Brazilian Pepper (B) leaf litter. The same number of individuals
and species were used to construct hypothetical distributions that conform to the
"broken stick" (I) and the geometric (III) model. The inset histogram shows the num-
ber of species found in different abundance categories (log base 2). The histogram
would be a normal curve if the distribution were a log-normal curve with a large sam-
ple size.

tween Brazilian pepper and hammock. Centipedes were almost totally absent from
Brazilian pepper leaf litter, and the lack of any correlation between centipede density
and moisture content in the hammock suggests that they avoid Brazilian pepper for
some reason other than moisture. Collembola showed no trends at all, but the fact
that I did not identify collembola any more specifically may hide important trends at
the family or species level.
Being a social animal, it is not surprising that ants are usually clumped. The few
low variance-to-mean ratios probably indicate rare colonies and far-ranging foragers.
Uncommon species with large, roaming workers (Aphaenogaster miamiana, Odon-
tomachus ruginodus, Hypoponera opaciceps, and Paratrechina bourbonica), had small
variance to mean ratios, as did centipedes.

Florida Entomologist 82(3) September, 1999



S80 --- -

E 70--- ----- -----

560 .

E -*-Pineland

5 40
40 Brazilian ~ Brazilian
pepper pepper

30 Hammock Hammock

1 10 100 10 100

Abundance Rank
Fig. 3. K-series curves for (A) combined General Survey and Quantitative Sam-
pling ant data and (B) Quantitative Sampling ant data alone.

Ant Anomalies

Strumigenys rogeri and S. lanuginosa have not previously been found in Ever-
glades National Park. Strumigenys rogeri is a common tropical exotic from Africa, and
its restriction to wet areas may confine it to Brazilian pepper at Everglades National
Park. Deyrup and Trager (1984) report S. rogeri in extremely dense populations in
moist bayheads at Archbold Biological Station, and I have seen it in high densities
along marshy trails on the Eastern Florida coast. An aggressive invader, S. rogeri
shares its moist habitat preference with S. louisianae, a native species. Why S. rogeri
was not found in the moist litter around pineland solution holes is an interesting
question. As for S. lanuginosa, it has not been collected in South Florida with enough
frequency to establish a habitat correlation.
Two species collected here made somewhat surprising appearances in hammock,
Odontomachus ruginodus and Monomorium floricola (Jerdon). In the Everglades, nei-
ther species has previously been collected in hammock, and 0. ruginodus has not been

Clouse:Ants in Brazilian Pepper Leaf Litter


mean variance

Brazilian Brazilian
Hammock pepper t-test Hammock pepper F-test

1.20 4.50 0.13 0.05 0.84 <0.0001*

1.81 0.14 0.0017* 0.04 0.00 <0.0001*

6.00 4.70 0.29 0.48 0.41 0.38

72.40 97.60 0.29 38.00 74.40 0.08

collected in pineland before (Ferster & Prusak 1994). Collecting 0. ruginodus in pine-
land is somewhat consistent with its habit of frequenting open areas, although often
those areas are also disturbed (e.g., roadsides). It is difficult to discern any pattern to
M. floricola's habitat, being that it has now been collected in slough, pineland, ham-
mock, and disturbed habitats. As an exotic, its ability to occupy several different Ev-
erglades habitats with ease is disturbing.


More extensive ant, beetle, and wasp lists for hammocks make the native forests
appear more diverse than Brazilian pepper. Inasmuch as k-series plots encapsulate
diversity, in this study they show that hammock ants are slightly more diverse than
Brazilian pepper. The crossing of the curves seems minor enough to ignore, especially
since it happens at the appearance ofParatrechina guatemalensis, the first- or second-
most abundant species in each habitat. However, the Brazilian pepper fauna is not
merely a subset of that of the native habitat, for some species seem to prefer or only
survive in Brazilian pepper leaf litter. This phenomenon is also documented by Cur-
nutt (1989) in a study of breeding birds in the Hole-in-the-Donut: even though Cur-
nutt lists a smaller number of species in Brazilian pepper relative to the number in
native habitat, the Red-winged Blackbird and Common Yellowthroat are far more
common in the Brazilian pepper. It could also be argued that the reason why more ant
species were found in the hammock is simply because I found more ants there, and the
rank-abundance curves further emphasize the similarity between hammock and Bra-
zilian pepper. My inability to discriminate between a log-series and log-normal curve
for each habitat is probably no great loss of information: the biological meaning of
these two distributions is debated, and together they seem to represent the majority
of stable, complex communities (Magurran 1988).
Regardless of diversity measures, however, Brazilian pepper is clearly an inhospi-
table habitat for native ants. Even though I found only two fewer native species in the
Brazilian pepper than in the hammock during Quantitative Sampling, I found only 16

Florida Entomologist 82(3) September, 1999

native individuals in the Brazilian pepper (as opposed to 194 in the hammock). Even
if one found a larger number of native species in the Brazilian pepper, the fact any one
ant has a 5% chance of being native in the Brazilian pepper (as opposed to 54% in the
hammock) is cause for concern.
The rise in diversity caused by the inclusion of exotic species has been called the
"mall effect" by Deyrup (pers. comm.) and "Macdonaldization" by Lovel (1997). The
analogy of the "mall effect" is that a small town may have a sudden increase in the di-
versity of retail stores by building a mall, but they are the same stores one might see
anywhere else in the country, like popular fast-food outlets. Similarly, the invasion of
exotic ants enhances diversity in many Everglades habitats. L6vel shows that this
process is homogenizing the biosphere, and he points out that even though exotics
have led to small-scale increases in diversity (e.g., in New Zealand) exotic invasion
could cause a 50% decrease in global biodiversity. Thus part of the problem with exotic
invasion is that although exotics may maintain or increase local diversity (as they are
added to and in some cases replace existing species), they will ultimately lower the
number of species counted on a global scale.
The more fundamental problem with exotic invasion is that regardless of how it
changes the total number of species and their evenness, the fact that the historical as-
semblage of species is altered has a destabilizing effect. The main role of Brazilian
pepper in this process is equivalent to a keystone species (Deyrup, pers. comm.), a
term (like "diversity") that usually carries positive connotations but is actually neu-
tral. Under Brazilian pepper branches thrives a microfauna more diverse than that
found in pineland and not much less diverse than hammock. However, the Brazilian
pepper community is not as native as in hammock, and the Hole-in-the-Donut area
serves as a safe refuge for exotics that would otherwise not have such a strong foot-
hold in the heart of native habitat. It is easy to imagine a scenario in which natural
Everglades disturbances (e.g., hurricanes and fire) are followed by quicker and more
thorough invasions by Brazilian pepper and its associates than if they were not al-
ready entrenched in the park.
The important question now is not whether exotic invasions pose a major threat to
endemic communities and global biodiversity, but how can we stop them in the face of
increasing human mobility. It is difficult to imagine a solution to the ant invasion (or
other invasions among fish, reptiles, etc.) that is surgical enough to remove only the
exotics but does not involve the resource-consuming process of finding a specific nem-
esis for each pest species. Moreover, as shown here, some exotics establish whole new
species assemblages, which are also quite exotic.


This work was supported by Everglades National Park, which provided equip-
ment, and Florida International University, which provided financial support in coop-
eration with the National Park Service. Dr. Betty Ferster helped collect samples, and
Dr. Mark Deyrup helped with identifications and understanding the "ant scene" in
South Florida. The comments of Lloyd Davis, Jr. are greatly appreciated. All speci-
mens were collected under permit #960062 and are stored in the South Florida Re-
search Center (Everglades Museum).


BENNETT, F. D., AND D. H. HABECK. 1991. Brazilian Peppertree-Prospects for Bio-
logical Control in Florida. p. 23-33 in T. D. Center et al. [eds.]. Proceedings of

Clouse:Ants in Brazilian Pepper Leaf Litter

the Symposium on Exotic Pest Plants. U.S. Department of the Interior, Wash-
ington, D.C.
CASSANI, J. R. 1986. Arthropods on Brazilian peppertree, Schinus terebinthifolius
(Anacardiaceae) in South Florida. Florida Entomol. 69: 184-196.
CURNUTT, J. L. 1989. Breeding bird use of a mature stand of Brazilian Pepper. Florida
Field Naturalist. 17: 53-76.
CRAIGHEAD, F. C., Sr. 1971. The Trees of South Florida. University of Miami Press,
Coral Gables, FL. 212 pp.
DEYRUP, M., AND J. TRAGER. 1984. Strumigenys rogeri, an African dacetine ant new
to the U.S. (Hymenoptera: Formicidae). Florida Entomol. 67: 512-516.
DEYRUP, M., J. TRAGER, N. CARLIN, AND G. UMPHREY. 1988. A review of the ants of the
Florida Keys. Florida Entomol. 71: 163-176.
DOREN, R. F., L. D. WHITAKER, G. MOLNAR, AND D. SYLVIA. 1990. Restoration of
Former Wetlands within the Hole-in-the-Donut in Everglades National Park.
p. 16-35 in F. J. Webb, Jr. [ed.]. Proceedings of the Seventh Annual Conference
on Wetlands Restoration and Creation. Hillsborough Community College,
Tampa, Florida.
EWEL, J. J., D. S. OJIMA, D. A. KARI, AND W. F. DEBUSK. 1982. Scinus in successional
ecosystems of Everglades National Park. South Florida Research Center Re-
port T-676. 141 pp.
FERSTER, B., AND Z. PRUSAK. 1994. A preliminary checklist of the ants (Hymenoptera:
Formicidae) of Everglades National Park. Florida Entomol. 77: 508-512.
JONES, D., AND DOREN, R. 1997. The distribution, biology, and control of Scinus tre-
binthefolius in Southern Florida, with special reference to Everglades National
Park. p. 81-93 in J. H. Brock et al. [eds.]. Plant Invasions: Studies from North
America and Europe. Backhuys Publishers, The Netherlands.
KOPTUR, S. 1992. Plants with extrafloral nectaries and ants in Everglades habitats.
Florida Entomol. 75: 38-50.
LOVEL, G. L. 1997. Global change through invasion. Nature. 388: 627-628.
MAGURRAN, A. E. 1988. Ecological Diversity and Its Measurement. Princeton Univer-
sity Press, Princeton. 179 pp.
MORTON, I. F. 1978. Brazilian pepper-its impact on people, animals and the environ-
ment. Econ. Botany 32: 353-359.
SMITH, M. R. 1930. A list of Florida ants. Florida Entomol. 14: 1-6.
SMITH, M. R. 1933. Additional species of Florida ants, with remarks. Florida Entomol.
17: 21-26.
TRAGER, J. C. 1984. A revision of the genus Paratrechina (Hymenoptera: Formicidae)
of the continental United States. Sociobiol. 9: 51-113.
WHEELER, W. M. 1932. A list of the ants of Florida with descriptions of new forms. J.
New York Entomol. Soc. 40: 1-17.
WILSON, E. 0., AND R. W. TAYLOR. 1967. Ants of Polynesia. Pac. Insects Monogr. 14:

Florida Entomologist 82(3) September, 1999


Department of Zoology and Brackenridge Field Laboratory, University of Texas
Austin, TX 78712 USA


We tested the host specificity of two South American Pseudacteon phorid flies,
P. obtusus Borgmeier and P. borgmeieri Schmitz, on North American colonies of the
red imported fire ant, Solenopsis invicta Buren, and the tropical fire ant, S. geminata
(F.). Sequential host specificity tests conducted in the laboratory indicated that P. ob-
tusus was highly specific to S. invicta. In individual trials, 20 females attacked S. in-
victa but none attacked S. geminata. Pseudacteon borgmeieri females, in contrast,
attacked both Solenopsis species. Six of the eighteen known South American Pseudac-
teon species have now been tested for host specificity, and 4 of the 6 reveal a high de-
gree of specificity to S. invicta, thus representing good biocontrol candidates.

Key Words: biological control, parasitism, Solenopsis geminata, Solenopsis invicta


Se prob6 la especificidad de dos moscas phorid Sudamericanas, P. obtusus Borg-
meier y P. borgmeieri Schmitz, en colonies norteamericanas de la hormiga roja de
fuego importada, Solenopsis invicta Buren, y la hormiga de fuego tropical S. geminata
(F.). Pruebas secuenciales de especificidad de hospederos efectuadas en laboratorio in-
dicaron que P. obtusus fue altamente especifica a S. invicta. En pruebas individuals,
20 hembras atacaron a S. invicta pero ninguna atac6 a S. geminata. En contrast,
hembras de Pseudacteon borgmeieri atacaron a las dos species de Solenopsis. Hasta
la fecha se ha probado la especificadad de huespedes en seis de las 18 species Pseu-
doacteon sudamericanas conocidas. Cuatro de las seis han mostrado un alto grado de
especificidad a S. invicta, por lo cual se les consider buenos candidates para el control
biol6gico de la hormiga roja de fuego importada.

Phorid flies in the genus Pseudacteon are biocontrol candidates of pest Solenopsis
ant species (Feener and Brown 1992, Orr et al. 1995, Porter et al. 1995b). An impor-
tant component in any biocontrol program is the assessment of host specificity in the
proposed biocontrol agents (Simberloff & Stiling 1996, Samways 1997). All known
Pseudacteon species are ant parasitoids, and all Pseudacteon species that parasitize
Solenopsis ants are apparently restricted to that genus (Disney 1994, Porter et al.
Previously, we reported on the host specificity patterns of 4 South American Pseu-
dacteon species that are known to parasitize workers in the S. saevissima complex
(which includes the red imported fire ant, S. invicta Buren and the black imported fire

Morrison & Gilbert: Host Specificity in Pseudacteon

ant, S. richteri Forel) (Gilbert and Morrison 1997). In laboratory tests, three of the
four species (P. litoralis Borgmeier, P. tricuspis Borgmeier, and P. wasmanni
[Schmitz]) rarely attempted to parasitize the tropical fire ant, S. geminata (F.), a spe-
cies in the geminata complex and native to the southern U.S., but readily attacked
populations of S. invicta. In contrast, P. curvatus Borgmeier attacked S. geminata rel-
atively frequently.
At least 18 species of Pseudacteon phorid flies attack saevissima complex fire ants
in South America (Porter 1998a), and are potential candidates for fire ant biocontrol
in the U.S. Because variation does exist among species in their relative host specificity
(Gilbert and Morrison 1997), it is critical that each species under consideration be
screened before release in the U.S. This article describes the host specificity of two
Pseudacteon species, P obtusus Borgmeier and P. borgmeieri Schmitz.


Pseudacteon obtusus females were collected from a residential area near the Uni-
versity of Campinas (UNICAMP), Sao Paulo State, Brazil (n = 18) and near the
Reserva Ecol6gica Costanera Sur in Buenos Aires, Argentina (n = 2). Pseudacteon
borgmeieri females were obtained from roadsides in the foothills near Jundiai, Sao
Paulo State, Brazil (n = 3), and near the Reserva Ecol6gica Costanera Sur in Buenos
Aires, Argentina (n = 2). Solenopsis mounds were disturbed and aspirators were used
to collect female phorids when they attempted to parasitize exposed worker ants. All
female phorids were transferred to the containment facility at The University of
Texas Brackenridge Field Laboratory where host specificity tests were conducted. Po-
lygyne (multiple queen) colonies of S. invicta were obtained from the grounds at
Brackenridge Field Laboratory and Circle C Ranch, Travis County, Texas. Polygyne
colonies of S. geminata were obtained from Circle C Ranch and along Barton Hills
Drive in Austin, and along the Colorado River in Mills County, Texas.
The methodology of host specificity testing was exactly the same as the sequential
choice procedure described in Gilbert and Morrison (1997) for 4 other Pseudacteon
species previously tested, and complete details of the procedure can be found there.
Briefly, each female was evaluated individually by introducing it to a flight box of S.
invicta, then to a flight box ofS. geminata, and finally back to a flight box ofS. invicta.
If the female did not demonstrate a motivation to oviposit on S. invicta (i.e., did not
attack within 20 minutes), the trial was aborted. If the female did begin to attack S.
invicta within 20 minutes, the attack rate was recorded over a 5 minute period, begin-
ning with the first attack, and then the female was immediately transferred into a
flight box of S. geminata. After 20 minutes in the S. geminata flight box, the female
was immediately transferred back into the original S. invicta flight box. By this
method we could be certain that females that ignored S. geminata were motivated to
oviposit before and after exposure to S geminata. In each case attack rates were based
on the time between the first attack and the removal of the phorid. Once a phorid be-
gan to attack, it usually did not stop until it was removed from the tray.
The orientation time, or time elapsed between introduction of a phorid into the
flight box and the first attack on an S. invicta worker, was compared for first and sec-
ond exposures to S. invicta for P. obtusus females. A Wilcoxon matched-pairs signed-
ranks test (Daniel 1990) was used.
Tests were conducted at intervals from June 1995 to June 1998. Voucher speci-
mens of Pseudacteon spp. have been retained by the authors and deposited in the Nat-
ural History Museum of Los Angeles County, CA and the Museu de Hist6ria Natural,
UNICAMP, Sao Paulo State, Brazil.

Florida Entomologist 82(3) September, 1999


Pseudacteon obtusus revealed a very high degree of host specificity for S. invicta.
Of 20 females that attacked S. invicta, none attempted to oviposit on S. geminata
when transferred to a flight box containing only S. geminata. Many P. obtusus females
hovered low over S. geminata for a few minutes, but then turned away and flew near
the top of the flight box or rested on the side. Eighteen of the 20 females that initially
attacked S. invicta were transferred back to the S. invicta flight box after exposure to
S. geminata, and 14 of these 18 resumed attacking S. invicta. The attack rate per in-
dividual female on initial exposure to S. invicta was 0.94 + 0.66 attacks per minute
(mean + SD), compared with 0.98 + 0.77 attacks per minute on subsequent exposure
to S. invicta.
Total time spent with S. invicta (after the first attack) on the initial exposure for
all 20 P. obtusus females considered collectively was 129.2 minutes, in which time 111
attacks occurred, for an overall attack rate of 0.86 attacks per minute. On the subse-
quent exposure, 18 P. obtusus females collectively spent a total time of 310.4 minutes
with S. invicta and attacked 199 times, for an overall attack rate of 0.64 attacks per
minute. Twenty P. obtusus females collectively spent a total of 386.1 minutes with
S. geminata without attacking.
The orientation time for P. obtusus was 9.5 + 6.0 (mean + SD) minutes on the first
exposure, compared with 5.1 + 4.4 minutes on the second exposure. Although this dif-
ference was not significant (P = 0.20; Wilcoxon matched-pairs signed-ranks test), it
follows the pattern observed in the 4 previously tested Pseudacteon species of having
shorter orientation times on the second exposure to S. invicta (Gilbert & Morrison
In contrast, P. borgmeieri readily attacked S. geminata. Of the 5 P. borgmeieri fe-
males that attacked S. invicta, 4 (2 from Brazil and 2 from Argentina) also attacked
S. geminata. Individual attack rates were higher on exposure to S. invicta than on
subsequent exposure to S. geminata (0.63 + 0.21 [mean + SD] versus 0.22 + 0.25 at-
tacks per minute, respectively). Because P. borgmeieri frequently attacked S. gemi-
nata, this phorid species was not exposed to S. invicta a second time. Considering all
females collectively, P. borgmeieri spent 39.5 minutes with S. invicta and attacked 24
times, for an overall attack rate of 0.61 per minute. Pseudacteon borgmeieri spent 93.4
minutes with S. geminata and attacked 16 times, for an overall attack rate of 0.17 per


The rationale for the design of the testing procedure and the conservative nature
of the tests were described in detail in Gilbert & Morrison (1997). Previously, in sim-
ilar tests with other Pseudacteon species, we found that P. litoralis, P. tricuspis, and
P. wasmanni attacked S. invicta almost exclusively, with only a few oviposition at-
tempts observed on S. geminata (Gilbert & Morrison 1997). Porter & Alonso (1999) ob-
tained similar results in independent laboratory tests. A fourth species tested only by
Gilbert & Morrison (1997), P. curvatus, was found to attack S. geminata relatively fre-
quently. Complementary tests conducted in the field in Brazil confirmed the high de-
gree of host specificity of P. litoralis and P. tricuspis, both of which parasitized
saevissima complex species but ignored S. geminata (Porter 1998b).
The data reported here are directly comparable with the data presented for P. lito-
ralis, P. tricuspis, P. wasmanni, and P. curvatus in Gilbert & Morrison (1997). Overall,
P. obtusus had a lower rate of attack on S. invicta, both on an individual and collective
basis, than the 4 species previously tested. It was also characterized by a longer ori-

Morrison & Gilbert: Host Specificity in Pseudacteon

entation time, in both the initial and subsequent exposures to S. invicta. Attack rates
of P. borgmeieri were the lowest of the six species tested, although the sample size for
this species was small. Like P. curvatus, P. borgmeieri often attacked S. geminata, but
attacked S. invicta more frequently (Gilbert & Morrison 1997). A summary of the rel-
ative host specificity of the 6 Pseudacteon species tested so far is presented in Table
1. The attack rate ratio illustrates, for each Pseudacteon species, the amount of para-
sitism pressure exerted on S. geminata relative to that exerted on S. invicta.
Although P. obtusus did not oviposit as frequently as the other S. invicta-specific
species tested, it may still exert significant parasitism pressure on S. invicta, and
therefore be useful in a biocontrol program, for two reasons: First, we have found a
positive correlation to exist between size of Pseudacteon species and size of worker
ants attacked (Morrison et al. 1997). Pseudacteon obtusus is a small phorid in most ar-
eas, with an average mesonotum width of 0.35 mm (Orr et al. 1997). The other three
Pseudacteon species that are known to be relatively host specific to S. invicta (P. lito-
ralis, P. tricuspis, and P. wasmanni) have average mesonotum widths ranging from
0.47 to 0.57 mm (Morrison et al. 1997, Orr et al. 1997) and are all significantly larger
than P. obtusus (Orr et al. 1997). Although we did not quantify the size of worker ants
attacked by P. obtusus, it was qualitatively obvious that P. obtusus preferred smaller
workers than the larger Pseudacteon species. Second, P. obtusus was found to be more
common along Solenopsis foraging trails than disturbed mounds in South America,
while P. litoralis, P. tricuspis, and P. wasmanni were all found predominantly at
mounds and only rarely along foraging trails (Orr et al. 1997).
Although relatively few P. borgmeieri were tested, preliminary results indicate
that this species may attack S. geminata in nature. We did not attempt to rear off-
spring from S. geminata workers parasitized by P. borgmeieri. It is possible that some
Pseudacteon species may attempt to oviposit on S. geminata but are unsuccessful (or
less successful) in inserting an egg, or offspring may not complete development in this
host. Porter & Alonso (1999), however, were able to get P. tricuspis females to attack
S. geminata workers by surrounding them with S. invicta colony odors, and reared a
single P. tricuspis to the adult stage in a S. geminata worker. Pseudacteon curvatus


Number of % of females
females attacking also attacking Attack rate
Species' S. invicta S. geminata ratio2

P. obtusus 20 0 0
P. tricuspis 25 4.0 0.001
P. litoralis 23 8.7 0.013
P. wasmanni 18 11.1 0.066
P. curvatus 20 65.0 0.301
P. borgmeieri 5 80.0 0.282

'Species are listed in order of decreasing specificity to S. inuicta. Data onP tricuspis, R litoralts, R wasmanni,
andFR curvatus are from Gilbert & Morrison (1997).
Attack rate ratio is the overall attack rate (= cumulative number of attacks/cumulative amount of time all
females of the species spent in the flight box) on S. geminata / overall attack rate on S. inuvcta (on initial expo-
sure). (For example: an attack rate ratio of 0.3 would indicate that, over the same period of time, 3 S. geminata
workers were attacked for every 10 S. inuvcta workers.)

Florida Entomologist 82(3) September, 1999

has also been reared to the adult stage in S. geminata (S. D. Porter, USDA-ARS,
Gainesville, pers. comm.). Thus development in S. geminata is possible for some, if not
all, South American Pseudacteon species.
Even if Pseudacteon species that attempt to oviposit on S. geminata do not fre-
quently complete development in S. geminata, the fact that these Pseudacteon species
are attracted to S. geminata and will attempt oviposition indicates the need for cau-
tion in the selection of species for biocontrol agents. This is because in general Pseu-
dacteon phorids appear to infect relatively few Solenopsis ants and thus have a very
small direct effect on mortality of their hosts (Jouvenaz et al. 1981, Morrison et al.
1997). The mere presence of Pseudacteon phorids, however, has been documented to
induce behavioral changes in Solenopsis workers, disrupting foraging efficiency and
decreasing the competitive ability of Solenopsis species relative to other ants in the
community (Feener & Brown 1992, Orr et al. 1995, Porter et al. 1995b, Morrison:
1999). Thus the major impact of Pseudacteon phorids on Solenopsis ants appears to be
mediated through behavioral changes in workers when flies are present, and exotic
Pseudacteon species that would potentially harass and attempt oviposition on S. gem-
inata could have a negative impact on S. geminata populations even if the flies do not
frequently complete development in the novel host.
Eighteen Pseudacteon species that attack S. saevissima complex workers in South
America are now known (Porter 1998a). The four Pseudacteon species that show a
strong preference for S. invicta (P. tricuspis, P. litoralis, P. wasmanni, and P. obtusus)
all have trilobed ovipositors, whereas the two that are less specific (P. curvatus and
P. borgmeieri) have an unlobed ovipositor. It remains to be seen whether this associa-
tion will hold true for all species. In limited laboratory observations of P. solenopsidis,
which has an unlobed ovipositor, we did not observe any oviposition attempts on
S. geminata (unpublished data). This species, however, exhibits an attack behavior
quite different than other South American Pseudacteon species that have been stud-
ied in this respect, and is characterized by a low attack rate (Orr et al. 1997). Further
laboratory testing of host specificity preferences will be necessary for this species. The
majority (13 of 18) of described South American Pseudacteon species have trilobed ovi-
positors (Porter 1998a).
If Pseudacteon phorid flies do prove to be effective biocontrol agents on imported
fire ants, a potential strategy would be to introduce Pseudacteon species of different
body sizes, to increase the range of worker sizes attacked (Morrison et al. 1997), and
to introduce species that attack workers at both mounds and along foraging trails
(Orr et al. 1997). Pseudacteon obtusus would be a welcome addition to the known 'ar-
senal' of larger, disturbed mound Pseudacteon species that are highly specific to S. in-
victa (P. litoralis, P. tricuspis, and P. wasmanni).


R. Guerra and E. A. Kawazoe (both University of Texas at Austin) assisted with
this project in Texas. W. Benson, C. G. Dall'Aglio-Holvorcem, S. Seike (all Univer-
sidade Estadual de Campinas), and M. Orr (San Francisco State University), helped
collect Pseudacteon spp. in Brazil. P. Folgarait collected Pseudacteon spp. in Argen-
tina. We thank B. Brown (Natural History Museum of Los Angeles County) for veri-
fying species identifications. CNPq and IBAMA (Brazil) and USDA APHIS supplied
the necessary permits. This research was funded by grants from The Fondren Foun-
dation, Ewing Halsell Foundation, R. J. Kleberg and H. G. Kleberg Foundation, the
Houston Livestock Show and Rodeo Educational Fund, and the State of Texas Fire
Ant Research and Management Committee (FARMAC).

Morrison & Gilbert: Host Specificity in Pseudacteon


DANIEL, W. W. 1990. Applied Nonparametric Statistics. 2nd Ed. PWS-Kent, Boston.
635 pp.
DISNEY, R. H. L. 1994. Scuttle flies: The Phoridae. Chapman & Hall, London. 467 pp.
FEENER, D. H., JR., AND B. V. BROWN. 1992. Reduced foraging of Solenopsis geminata
(Hymenoptera: Formicidae) in the presence of parasitic Pseudacteon spp.
(Diptera: Phoridae). Ann. Entomol. Soc. Am. 85: 80-84.
GILBERT, L. E., AND L. W. MORRISON. 1997. Patterns of host specificity inPseudacteon
parasitoid flies (Diptera: Phoridae) that attack Solenopsis fire ants (Hy-
menoptera: Formicidae). Environ. Entomol. 26: 1149-1154.
JOUVENAZ, D. P., C. S. LOFGREN, AND W. A. BANKS. 1981. Biological control of im-
ported fire ants: A review of current knowledge. Bull. Entomol. Soc. Am. 27:
MORRISON, L. W. 1999. Indirect effects of phorid fly parasitoids on interspecific com-
petition between fire ants (in press). Oecologia.
tion behavior and development of Pseudacteon flies (Diptera: Phoridae), para-
sitoids of Solenopsis fire ants (Hymenoptera: Formicidae). Environ. Entomol.
26: 716-724.
ORR, M. R., S. H. SEIKE, W. W. BENSON, AND L. E. GILBERT. 1995. Flies suppress fire
ants. Nature 373: 292.
ORR, M. R., S. H. SEIKE, AND L. E. GILBERT. 1997. Foraging ecology and patterns of di-
versification in dipteran parasitoids of fire ants in south Brazil. Ecol. Entomol.
22: 305-314.
PORTER, S. D. 1998a. Biology and behavior of Pseudacteon decapitating flies (Diptera:
Phoridae) that parasitize Solenopsis fire ants (Hymenoptera: Formicidae).
Florida Entomol. 81: 292-309.
PORTER, S. D. 1998b. Host-specific attraction of Pseudacteon flies (Diptera: Phoridae)
to fire ant colonies in Brazil. Florida Entomol. 81: 423-429.
PORTER, S. D., AND L. E. ALONSO. 1999. Host specificity of fire ant decapitating flies
(Diptera: Phoridae) in laboratory oviposition tests. J. Econ. Entomol. 92:110-
PORTER, S. D., H. G. FOWLER, S. CAMPIOLO, AND M. A. PESQUERO. 1995a. Host speci-
ficity of several Pseudacteon parasites of fire ants in South America (Diptera:
Phoridae; Hymenoptera: Formicidae). Florida Entomol. 78:70-75.
FOWLER. 1995b. Solenopsis (Hymenoptera: Formicidae) fire ant reactions to at-
tacks of Pseudacteon flies (Diptera: Phoridae) in Southeastern Brazil. Ann. En-
tomol. Soc. Am. 88: 570-575.
SAMWAYS, M. J. 1997. Classical biological control and biodiversity conservation: what
risks are we prepared to accept? Biodiversity & Conserv. 6: 1309-1316.
SIMBERLOFF, D. S., AND P. STILING. 1996. Risks of species introduced for biological
control. Biol. Conserv. 78: 185-192.

Florida Entomologist 82(3) September, 1999


1University of Florida, Citrus Research and Education Center
700 Experiment Station Road, Lake Alfred, Florida 33850

'Indian River Research and Education Center, 2199 South Rock Road
Fort Pierce, Florida 34945-3138


Four field experiments were conducted during 1995 and 1996 to evaluate in-
creased fruit set of navel and 'Valencia' oranges or 'Marsh' grapefruit varieties by con-
trolling Frankliniella bispinosa (Morgan) (Thysanoptera: Thripidae) on citrus flowers
during bloom. In three experiments, Chlorpyrifos 4EC at 4.68 liters/ha provided 5, 5,
and 4 days of effective residual control ofF. bispinosa adults on citrus flowers and 3
or more days at 3.16 liters/ha in the remaining experiment. Corresponding lengths of
larval thrips control were 12, 12, and 8 days, respectively. The Chlorpyrifos 4EC treat-
ments at 3.16 liters/ha applied on 25 March in the fourth experiment provided 11 or
more days of larval thrips control. Use of a single, properly timed insecticide applica-
tion between maximum bud swell and full bloom, when F. bispinosa or F. kelliae
adults are increasing, will increase fruit set. Preventing establishment of larval
thrips populations appears to be important.

Key Words: citrus; pest management; flower thrips; flowering


Durante 1995 y 1996 se efectuaron cuatro experiments de campo para incremen-
tar el amarre de fruta en naranja 'navel' y 'Valencia' y en toronja 'Marsh' mediante el
control del thrips Frankliniella bispinosa (Morgan) (Thysanoptera: Thripidae) en las
flores. En tres experiments, Chlorpyriphos 4EC a 4.68 litros/ha demostr6 un control
residual efectivo de adults en las flores por 5, 5, y 4 dias. En el cuarto experiment,
Chlorpyriphos 4EC a 3.16 litros/ha tuvo mostr6 control por 3 o mas dias. El control de
larvas con la dosis alta de Chlorpyriphos fue de 12, 12, y 8 dias, mientras que la dosis
baja, aplicada en Marzo 25, mostr6 control de larvas por 11 dias. Una aplicaci6n sen-
cilla del insecticide entire las fases de yema hinchada y floraci6n complete, cuando las
poblaciones de F. bispinosa yF. kelliae van en aumento, incrementa el amarre de fruta.
La prevenci6n del establecimiento de poblaciones de larvas se consider important.

Three species of Scirtothrips (Thysanoptera: Thripidae) are recognized as eco-
nomic pests on citrus in different parts of the world (Lewis et al. 1997). They feed on
citrus leaves and developing citrus fruits, and include S. aurantii Faure in South Af-
rica (Samways et al. 1987, Gilbert 1990), S. dorsalis Hood in Japan and Africa (Takagi
1981, Gilbert 1986), and S. citri (Moulton) in California and Arizona (Tanigoshi &
Nishio-Wong 1982, Arpaia & Morse 1991). More recently, the effects of flower and
fruitlet drop on navel orange caused by Frankliniella bispinosa (Morgan) and
F. kelliae Sakimura (Thysanoptera: Thripidae) were evaluated in Florida (Childers et

Childers & Bullock: Flower Thrips Control on Citrus

al. 1990, Childers & Achor 1991a, Childers & Beshear 1992, Childers et al. 1994). The
results demonstrated that thrips feeding damage (Childers 1992) and a fungal patho-
gen Colletotrichum acutatum J. H. Simmonds (Sonoda & Pelosi 1988, Brown et al.
1996) were associated with postbloom fruit drop in Florida.
Increased fruit set occurred where thrips and disease were controlled separately
or together on navel orange (Childers 1992). In some years, control of thrips with in-
secticides on navel orange during flowering provided higher fruit set based on late
June-early July frame counts compared with 1 or 2 fungicide benomyll 50WP) appli-
cations for postbloom fruit drop disease control alone. Although the pathogen was
present in these groves, disease pressure was not assessed.
Field cage studies followed by scanning and transmission electron microscopy
have shown that larvae and adults of F. bispinosa feed on the ovary, style, floral disk,
petals, and anthers of swollen buds and open flowers of navel orange (Childers &
Achor 1991a, b). Thrips oviposition and subsequent eclosion were recorded from the
pistil, calyx, petals, and filaments of both flowers and swollen buds. Population sup-
pression of F. bispinosa during the flowering cycle and increased fruit set of navel or-
anges was shown by Childers (1992).
One important aspect of managing F. bispinosa is determining when to apply
(proper timing) an insecticide to increase fruit set. Also, additional field experimenta-
tion was desired to determine if fruit set increase could be obtained on'Valencia' orange
or grapefruit varieties by controlling flower thrips. Therefore, field experiments were
conducted during 1995 and 1996 to evaluate: (1) timing of insecticides for thrips control,
and (2) differences in fruit set on navel and 'Valencia' oranges and grapefruit varieties.


Experimental Design

Four experiments were conducted during 1995 and 1996 in 3 different grove sites
with previous or suspected problems of postbloom fruit drop. Experiment 1 was per-
formed at the Duda farm in the Felda vicinity, Hendry County during March 1995.
Navel orange trees were 5 to 6.1 m tall with a 5.2 by 7.3 m tree spacing (264 trees/ha).
Plot size consisted of 12 rows by 15 trees per row (180 trees/plot). Treatments were ap-
plied in 1.169 kliters/ha with an airblast sprayer. Experiment 2 was conducted at the
Butler farm in the LaBelle vicinity in Hendry County during March-April 1995. 'Va-
lencia' orange trees were 3.6 to 6.1 m tall with a 4.6 by 7.6 m tree spacing (287 trees/
ha). Treatments were assigned to 8 rows by 20 to 30 trees within the row (160-240
trees per plot). Treatments were applied in 2.3 kliters/ha. Experiment 3 was per-
formed at the Cracker Trail farm in western St. Lucie County about 4 km north of
highway 68 and within 4 km of the Okeechobee County line. 'Marsh' grapefruit trees
were 3.7 to 5.3 m tall with a 5.5 by 7.3 m tree spacing (249 trees/ha). Plot size con-
sisted of 6 rows by 16 trees within the row or 12 rows by 9 trees within the row (96-
108 trees per plot). Treatments were applied in 1.17 kliters/ha. Experiment 4. The 'Va-
lencia' orange trees were the same block as in experiment 3 and conducted during
March-May 1996. Plot size consisted of 8 rows by 20 to 40 trees within the row (160-
320 trees per plot). Treatments were applied 25 March in 1.6 kliters/ha. Other treat-
ments were applied on 29 March or 9-10 May in 2.3 kliters/ha.


A standard grove spray program for disease, insect, and mite control was main-
tained throughout the season after application of the experimental treatments during

Florida Entomologist 82(3) September, 1999

the flowering cycle at each grove site. The only difference was the pesticide treatments
applied during flowering and 0-6 wk immediately following petal fall. Pesticide treat-
ments are shown as formulated weights or volumes per hectare and were applied to
plots in a randomized complete block design with 6 replicates (Table 1). In experiment
4, Mattch 12% aqueous flowable (AF) (Mycogen Corp., San Diego, CA) consisting of a
blend of 127.2 g/liter of delta endotoxins of Bacillus thuringiensis Berliner encapsu-
lated in killed Pseudomonas fluorescens (Trevisan) Migula was applied at 4.68 liters/
ha in treatments 2 and 3 on 9 May.


Populations of F. bispinosa and F. kelliae were sampled at various time intervals
before and after spray applications by collecting a maximum of 4 open flowers at ran-
dom from around the canopy of each of 5 trees in the center of each treatment repli-
cate. Each flower of the 20 flower sample was immediately immersed in one jar
containing 70-80% ethanol and a label indicating collection site, date, and treatment
replicate. Following petal fall, small fruitlets were collected in the same way as open
flowers. In the laboratory, flowers and alcohol from each jar were poured into a dish,
floral parts were separated, and adult and larval thrips were counted and removed
using a stereomicroscope at 10-12x magnification.
Thrips were recorded as Frankliniella spp. and "other" thrips. Frankliniella bispi-
nosa was the dominant thrips species on infested floral buds and open flowers of citrus
accounting for 92% of the slide-mounted specimens subsampled from over 80 citrus
growing areas in Florida (Childers et al. 1990, Childers & Beshear 1992).
Fruit set differences between treatments were determined in each test by using a
1 m3 frame constructed of 12.5-mm diameter pvc pipe with an attached vertical pipe
1.37 m in length. This supported the frame at a fixed height. This procedure was mod-
ified from Stout (1962) for determining the bearing surface of a tree and to estimate
the number of fruits per tree (Childers 1992). The frame was placed into the canopy
at approximately 450 angles from the tree row at each of the 4 corners in each of 10
trees per plot (treatment replicate) in late June or early July during 1995 and 1996.
The majority of premature fruit drop usually occurred shortly after petal fall and dur-
ing June (Lima & Davies 1981). All fruits except obviously very small late-bloom fruit
within each frame per treatment replicate were counted and recorded.

Statistical Analysis

In all experiments, data were subjected to analysis of variance; Waller-Duncan k-
ratio procedures were used to separate treatment means when the ANOVA provided
a significant F value (P > 0.05) (SAS Institute 1991). Both adult and larval thrips
counts per 20 flowers or fruitlets and number of fruits per m3 from the frame counts
were subjected to Logl0(X + 1) transformations for statistical analysis. Untransformed
means are shown in all tables.


[Experiment 1] Duda, 1995

Low densities of both adult and larval thrips were observed in all plots prior to ini-
tiation of treatment applications (Table 2). Chlorpyrifos applied on 22 March (Table 1)
provided at least 5 d residual control of adults while larval densities per 20 flowers re-


Treatment Treatment Rate per Application
Experiment no. no. regimes Formulation hectare Phenology1 date

1. Duda 1995

2. Butler 1995

3. Cracker Trail 1995

1 Chlorpyrifos
2 Benomyl
3 Untreated
1 Chlorpyrifos
2 Chlorpyrifos
3 Formetanate
4 Chlorpyrifos
5 Benomyl
6 Untreated
1 Chlorpyrifos
2 Chlorpyrifos
3 Formetanate
4 Chlorpyrifos
5 Benomyl

4 EC
50 WP

4 EC
92 SP
4 EC
92 SP
4 EC
92 SP
50 WP
50 WP

4 EC
92 SP
4 EC
92 SP
4 EC
92 SP
50 WP

4.68 liters 50-70% BS
2.25 kg Onset F

4.68 liters
1.12 kg
4.68 liters
1.12 kg
4.68 liters
1.12 kg
2.25 kg
2.25 kg

4.68 liters
1.12 kg
4.68 liters
1.12 kg
4.68 liters
1.12 kg
2.25 kg

50-70% BS
50-70% BS
Onset F

50-70% BS
50-70% BS
Onset F

22 Mar
10 Mar

21 Mar
29 Mar
21 Mar
29 Mar
29 Mar
1 Apr
14 Mar
29 Mar

27 Mar
4 Apr
27 Mar
4 Apr
3 Apr
11 Apr
20 Mar


Treatment Treatment Rate per Application
Experiment no. no. regimes Formulation hectare Phenology' date

Benomyl 50 WP 2.25 kg FB 3-4 Apr
6 Untreated -
4. Butler 1996 1 Chlorpyrifos 4 EC 4.68 liters Maximum BS 25 Mar
2 Chlorpyrifos 4 EC 4.68 liters Maximum BS 25 Mar
Mattch 12% AF 4.68 liters 5-6 wk PB 9 May
3 Chlorpyrifos 4 EC 4.68 liters FB 29 Mar
Mattch 12% AF 4.68 liters 5-6 wk PB 9 May
4 Chlorpyrifos 4 EC 4.68 liters Maximum BS 25 Mar
Carbaryl XLR 4 L 5.85 liters 5-6 wk PB 10 May
5 Chlorpyrifos 4 EC 4.68 liters 5-6 wk PB 10 May
6 Untreated -

Childers & Bullock: Flower Thrips Control on Citrus

mained significantly lower than the check through 31 March or per 20 fruitlets on 3
April (12 days after treatment = 12 DAT). Benomyl applied on 10 March did not pro-
vide control of either F. bispinosa or F. kelliae adults or larvae between 21 March and
3 April. Petal fall occurred some time between 31 March and 3 April.
Chlorpyrifos resulted in significantly higher fruit set compared to the untreated
trees based on frame counts taken on 25 July (Table 2). The single benomyl applica-
tion did not significantly increase fruit set per m3 of canopy.

[Experiment 2] Butler, 1995

Low population pressure of adult thrips was recorded prior to insecticidal applica-
tions on 21 and 29 March (Table 3). Significantly lower adult densities per 20 flowers
occurred on 27 March and 3 April for treatment 1 (Table 1) compared to the untreated
trees while treatment 2 resulted in reduced numbers of adults per 20 flowers on 27
March only. Treatment 3 failed to significantly reduce adult densities on 3 April com-
pared with treatment 1. Treatment 4 provided comparable suppression of adult thrips
on 3 April compared with treatment 1. Treatment 5 had significantly lower adults per
20 flowers on 3 April compared with the untreated check trees.
Larval thrips numbers were significantly reduced in treatment 1, treatment 2, and
treatment 4 between 23 March and 6 April compared with the untreated trees (Table
3). Petal fall occurred some time between 3-6 April. Larval densities in treatments 3
and 5 were significantly lower only on 3 April compared with the untreated check
Treatment 1, treatment 2, treatment 3, and treatment 4 all resulted in signifi-
cantly higher fruit sets compared with either treatment 5 or the untreated check trees
(Table 3). Treatment 3 did not provide a measurable decrease in either adult or larval
thrips counts in this experiment when compared with treatment 5.

[Experiment 3] Cracker Trail, 1995

Low densities of both adult and larval thrips were present in all treatment repli-
cates on 21 March prior to spray applications (Table 4). Treatment 1 (Table 1) pro-
vided control of adults on 28 March and significantly lower densities of larvae
between 28 March and 4 April compared with the untreated check trees. Treatment
2 had significantly lower numbers of adults on 28 March compared with the untreated
check trees. The addition of formetanate in treatment 1 did not result in improved
thrips control compared with chlorpyrifos alone in treatment 2. Treatments 3 and 5
did not reduce either adult or larval thrips numbers between 28 March and 4 April
compared with the untreated check trees. Treatment 4 resulted in reduced adult
thrips densities on 4 April only and larval thrips counts were not significantly reduced
between 21 March and 4 April. None of the treatments in this experiment resulted in
increased fruit set compared with the untreated check trees (Table 4).

[Experiment 4] Butler, 1996

Low populations of adult and larval thrips were present on the citrus flowers in all
treatment replicates prior to onset of spray applications (Table 5). Treatment 1, treat-
ment 2, and treatment 4 all provided significant reductions in adult thrips densities
on 27 and 28 March (Table 1). Residual control of adult thrips in all 3 treatments was
lost by 1 April. Treatment 3 provided significant reductions in adult thrips populations
between 1-10 April compared with the untreated check trees. Larval thrips numbers


Pre-treatment Mean
means' Post treatment means' Fruitlets no. of
Rate per Spray oranges
Treatment Formulation hectare 12 Mar dates 21 Mar 24 Mar 27 Mar 31 Mar 3 Apr per m3

Adults/20 flowers
1. Chlorpyrifos 4 EC 4.68 liters 29 a2 22 Mar 25 a 3 b 4 b 150 a 3 a 13.5 a
2. Benomyl 50 WP 2.25 kg 33 a 10 Mar 21 a 9 ab 24 a 172 a 2 a 12.8 ab
3. Untreated 29 a 21 a 15 a 27 a 155 a 4 a 11.7 b

Larvae/20 flowers
1. Chlorpyrifos 4 EC 4.68 liters 2 a2 22 Mar 11 a 0 b 0 b 13 b 1 c -
2. Benomyl 50 WP 2.25 kg 1 a 10 Mar 12 a 4 a 7 a 68 a 17 a -
3. Untreated 1 a 7 a 2 ab 13 a 54 a 5b -

Means within a column followed by the same letter are not significantly different by ANOVA followed by Waller-Duncan K ratio (ifANOVA P s 0.05).
Mean number of adult or larval thrips per 20 flowers, 20 fruitlets or oranges per m' of canopy volume.


Pre-treatment Mean
means' Post treatment means' Fruitlets no. of
Rate per Spray oranges
Treatment Formulation hectare 21 Mar dates 23 Mar 27 Mar 3 Apr 6 Apr per m3

Adults/20 flowers
1. Chlorpyrifos 4 EC 4.68 liters 21 Mar -
Formetanate 92 SP 1.12 kg 59 14 a2 29 Mar 2+ la 4+ lb 115 16 cd 2 1 ab 21.8 a
2. Chlorpyrifos 4EC 4.68 liters 52 17 a 21 Mar 2 0 a 4 1 b 263 69 ab 1 -lb 22.2 a
3. Formetanate 92 SP 1.12 kg 76 20 a 29 Mar 23 + 10 a 26 + 7 a 321 87ab 1+ lb 21.6 a
4. Chlorpyrifos 4 EC 4.68 liters 29 Mar -
Formetanate 92 SP 1.12 kg 51 12 a 1Apr 17+ 9 a 16+ 3 a 73 + 10 d 0 0 b 20.7 a
5. Benomyl 50 WP 2.25 kg 14 Mar -
Benomyl 50 WP 2.25 kg 35 + 9 a 29 Mar 27 + 16 a 15 + 4 a 166 + 14 bc 4 + 1 a 17.8 b
6. Untreated 50 + 11 a 9 + 4 a 18 + 6 a 410 + 87 a 2 + 1 ab 16.5 b

Larvae/20 flowers
1. Chlorpyrifos 4 EC 4.68 liters 21 Mar -
Formetanate 92 SP 1.12 kg 6 + 3 a 29 Mar 1 + 0 b 0 0 b 1 + 0 cd 0 + 0 bc -
2. Chlorpyrifos 4EC 4.68 liters 7 +3 a 21 Mar 0 0 b 0+ 0b 8 + 4bc 0 0bc -
3. Formetanate 92 SP 1.12 kg 5 + 1 a 29 Mar 5 + 3 ab 19 + 12 a 8 + 5 bcd 2 + 1 abc -

Means within a column followed by the same letter are not significantly different by ANOVA followed by Waller-Duncan k ratio (ifANOVAP 0.05).
Mean number of adult or larval thrips per 20 flowers, 20 fruitlets or oranges per m' of canopy volume.


Pre-treatment Mean
means' Post treatment means' Fruitlets no. of
Rate per Spray oranges
Treatment Formulation hectare 21 Mar dates 23 Mar 27 Mar 3 Apr 6 Apr per m3

4. Chlorpyrifos 4 EC 4.68 liters -29 Mar -
Formetanate 92 SP 1.12 kg 5 2 a 1Apr 3 1 ab 64 4a 0 0d 0 0c -
5. Benomyl 50 WP 2.25 kg -14 Mar -
Benomyl 50 WP 2.25 kg 4 + 1 a 29 Mar 12 + 8 a 4 2 ab 10 + 4 b 2 1 ab
6. Untreated 8+5a 2 1 ab 4+2ab 119 54a 4 2a

'Means within a column followed by the same letter are not significantly different by ANOVA followed by Waller-Duncan k ratio (ifANOVA P s 0.05).
'Mean number of adult or larval thrips per 20 flowers, 20 fruitlets or oranges per m' of canopy volume.


Pre-treatment Mean
means' Post treatment means' number of
Rate per Application grapefruit
Treatment Formulation hectare 21 Mar dates 28 Mar 31 Mar 4 Apr per m3

Adults/20 flowers
1. Chlorpyrifos 4 EC 4.68 liters 27 Mar -
Formetanate 92 SP 1.12 kg 37 8 a2 4 Apr 4- 1 b 51 3 c 213 52 a 11.7 c
2. Chlorpyrifos 4 EC 4.68 liters 43 + 12 a 27 Mar 5+ lb 70 + 17 bc 143 + 31 a 13.7 ab
3. Formetanate 92 SP 1.12 kg 37 + 6 a 4 Apr 18 5 a 119 + 23 a 207 + 57 a 13.5 abc
4. Chlorpyrifos 4 EC 4.68 liters 3 Apr -
Formetanate 92 SP 1.12 kg 55 + 11 a 11 Apr 11 + 3 ab 87 + 19 abc 59 + 10 b 15.2 a
5. Benomyl 50 WP 2.25 kg 20 Mar -
Benomyl 50 WP 2.25 kg 53 + 7 a 3-4 Apr 13 + 3 a 97 + 9 ab 160 + 33 a 12.4 bc
6. Untreated 50 + 9 a 10 + 4 ab 76 + 12 abc 147 + 27 a 13.3 abc

Larvae/20 flowers
1. Chlorpyrifos 4 EC 4.68 liters 27 Mar -
Formetanate 92 SP 1.12 kg 5 + 2 a 4 Apr 2 + 1 c 1 + 0 b 5 2b -
2. Chlorpyrifos 4 EC 4.68 liters 3 + 2 a 27 Mar 3 + 1 bc 2 + 0 b 4 0b -
3. Formetanate 92 SP 1.12 kg 6 +2 a 4 Apr 7 1 a 23 + 9 a 82 38 a -

'Means + standard errors within a column followed by the same letter are not significantly different by ANOVA followed by LSD (ifANOVAP 0.05).
'Mean number of adult or larval thrips per 20 flowers or grapefruit per m' of canopy volume.


Pre-treatment Mean
means' Post treatment means' number of
Rate per Application grapefruit
Treatment Formulation hectare 21 Mar dates 28 Mar 31 Mar 4 Apr per m3

4. Chlorpyrifos 4 EC 4.68 liters -3 Apr
Formetanate 92 SP 1.12 kg 9 3 a 11 Apr 9 3 ab 28 7 a 16 5 a
5. Benomyl 50 WP 2.25 kg -20 Mar -
Benomyl 50 WP 2.25 kg 9 2 a 3-4 Apr 7 2 abc 22 6 a 29 a 7 a
6. Untreated 22 10 a 12 3 a 15 5 a 30 9 a

'Means + standard errors within a column followed by the same letter are not significantly different by ANOVA followed by LSD (ifANOVAP a 0.05).
'Mean number of adult or larval thrips per 20 flowers or grapefruit per m' of canopy volume.

Childers & Bullock: Flower Thrips Control on Citrus

were significantly reduced in treatments 1 through 4 between 1-5 April when com-
pared with treatment 5 or the untreated check trees. Treatments 1 and 4 had signifi-
cantly higher numbers of fruit set compared with the untreated check trees (Table 5).


In experiment 1 at the Duda site, maximum thrips numbers averaged 172 adults
per 20 flowers on 31 March or 8.6 adults per flower compared with 3.4 larvae per
flower in the same set of samples (Table 2). Maximum measurable differences in fruit
set were 15% in treatment 1 while treatment 2 had 9%.
In experiment 2, the maximum number of adult and larval thrips per 20 flower
sample was 410 on 3 April at Butler (Table 3). This averaged 20 adult thrips per flower
compared with 6 larvae per flower. Petal fall occurred between 3-6 April and low den-
sities of both adult and larval thrips were recorded on the small fruitlets on 6 April.
All 4 insecticide treatments resulted in significant increases in fruit set in this exper-
iment compared with the fungicide only and untreated check trees. Results of this ex-
periment indicate that adult thrips densities approaching 20 or more per flower prior
to petal fall resulted in reduced fruit set for this grower. Economic benefit was ob-
tained with at least a 17-26% range of increase in fruit set resulting from 1 or 2 insec-
ticidal applications during flowering in this 'Valencia' orange grove. A single
insecticide treatment applied between 21 and 29 March was adequate to provide com-
parable fruit set increase compared with treatments that received 2 insecticide appli-
cations between 21 March and 1 April.
In experiment 3 on 'Marsh' grapefruit, maximum adult thrips numbers averaged
213 adults per 20 flowers in treatment 1 on 4 April (Table4). Maximum larval thrips
numbers averaged 82 per 20 flowers in treatment 3 on the same date. Residual activ-
ity of adults in treatments 1 and 2 was 4 days while larval suppression was sustained
between 28 March and 4 April compared with the untreated check. However, no ben-
efit in increased fruit set was shown.
In experiment 4 on 'Valencia' orange at Butler, treatments 1 through 4 provided
significant reductions in larval thrips numbers between 1 and 10 April (Table 5). Only
2 of the 4 treatments (1 and 4) resulted in significant increases in fruit set. However,
mean numbers of fruit/m3 were not significantly different in the 4 insecticide treat-
ments. An approximate 6-8% gain in fruit set on 'Valencia' orange trees was obtained
by applying chlorpyrifos during maximum bud swell in 2 of 4 treatments (1 and 4).
Based on these experiments when Frankliniella (primarily F. bispinosa) adult den-
sities approached 20 per flower and when larval densities were 3 to 5 in the same flow-
ers, then reduced fruit set injury occurred on navel or 'Valencia' oranges in Florida in
3 field trials. Previously published field experiments using insecticides on navel or-
ange demonstrated increased fruit set in earlier trials, too (Childers 1992). For exam-
ple, maximum adult and larval densities per flower were 55 and 5, respectively, near
petal fall during 1990 in the untreated check trees compared with maximum densities
of 25 adults and 1 larva per flower in 2 insecticide treatments. Sustained suppression
of both adults and larvae was achieved between 1-7 March in 2 treatments receiving
formetanate 92 SP at 1.40 kg/ha on 1 March. Both treatments resulted in significant
increases of 43 and 42% in fruit set, respectively, compared to the untreated check
trees in that experiment. Use of a single, properly timed insecticide application be-
tween maximum bud swell and full bloom, when populations of Frankliniella bispi-
nosa and F. kelliae adults are increasing on flowers, will increase fruit set on navel and
'Valencia' oranges in Florida. Preventing establishment of larval thrips populations
from feeding and developing on citrus flowers appears to be important. Use of formet-


Pre-treatment Mean
means' Post treatment means no. of
Formula- Rate per Spray oranges
Treatment tion hectare 14 Mar dates 27 Mar 28 Mar 1 Apr 2 Apr 4 Apr 5 Apr 10 Apr per m8

Adults/20 flowers
1. Chlorpyrifos 4 EC 3.16liters 103 10 a2 25 Mar 4 1 b 2- 1 be 31- 5 bc 61 7 a 185- +20 a 321 23 a 406 86 a 26.4 a
2. Chlorpyrifos 4 EC 3.16 liters 25 Mar -
Mattch(Bt) AF 4.68 liters 83- 10 a 9 May 4 1 b 3- 1 be 43 + 5 abc 72 + 9 a 214- 34 a 330- 29 a 379- +54 a 25.0 ab
3. Chlorpyrifos 4 EC 4.68 liters 29 Mar -
Mattch (Bt) AF 4.68 liters 120+ 25 a 9 Apr 29s 13 b 34s 3 a 31 s 13 c 35+ 7b 94-+ 18b 182+ 37 b 163+ 74b 24.8 a
4. Chlorpyrifos 4 EC 3.16 liters 25 Mar -
Carbaryl 4L 11.70 liters 101 17 a 10 May 3 +2b 1 0 c 42 + 9 abc 73 15 a 221 +35 a 401- +25 a 436- +67 a 26.8 a
5. Chlorpyrifos 4 EC 4.68 liters 90_ 14 a 10 May 14 lib 21 16 ab 71_ 14 a 83_ 16 a 269_ 38 a 385_ +68 a 333_ 39 a 24.0 b
6. Untreated 129 19 a 32 14 a 31 14 a 57- 14 ab 69- 11 a 264- +51 a 306- 31 a 336- +55 a 22.8 b
Larvae/20 flowers
1. Chlorpyrifos 4 EC 3.16 liters 10s 4 a' 25 Mar 1 1 c 3 2 a 0 +0b 0 +0b 1 1 c 4: lb 191 50b -
2. Chlorpyrifos 4 EC 3.16 liters 25 Mar -
Mattch (Bt) AF 4.68 liters 14 + 7 a 9 May 1 + 0 bc 2 1 a 0 + 0 b 0 + 0 b 3 1 bc 4 1 b 182 +30b -
3. Chlorpyrifos 4 EC 4.68 liters 29 Mar -
Mattch (Bt) AF 4.68 liters 7 +3 a 9 Apr 16 11 a 7 +3a 1 +0b 0 +0b 0 0 c 0 0 c 33s6c -
4. Chlorpyrifos 4 EC 3.16 liters 25 Mar -
Carbaryl 4L 11.70 liters 5+la 10 May 0 0 c 2l a 0 +0b l1+0b 1 0c 5 +2b 202+ 37b -
5. Chlorpyrifos 4 EC 4.68 liters 7 + 3 a 10 May 4 + 3 bc 8s 4 a 1 s 1 ab 5 + 2 a 15 + 7 ab 18s 3 a 410 +85 a -
6. Untreated 124a 6+4ab 104a 8+5a 96a 17s+10a 16+5a 261+31ab -

'Means within a column followed by the same letter are not significantly different by ANOVA followed by Waller-Duncan K ratio (ifANOVAP a 0.05).
'Mean number of adult or larval thrips A SE per 20 flowers or oranges per m' of canopy volume.

Childers & Bullock: Flower Thrips Control on Citrus

anate is less desirable due to the highly toxic nature of this insecticide to the predatory
mites Euseius mesembrinus (Dean) and Typhlodromalus peregrinus (Muma) (Acari:
Phytoseiidae) (Childers unpublished data). An effective scouting system must be de-
veloped to simplify monitoring of adult flower thrips populations in the field.


We gratefully acknowledge J. W. Noling, P. A. Phillips, and anonymous reviewers
for their constructive reviews of this manuscript. Funding was provided by the Flor-
ida Citrus Production Research Advisory Council. Florida Agricultural Experiment
Station Journal Series No. R-06659.


ARPAIA, M. L., AND J. G. MORSE. 1991. Citrus thrips Scirtothrips citri (Moulton)
(Thysanoptera: Thripidae) scarring and navel orange fruit quality in Califor-
nia. J. Appl. Entomol. 111: 28-32.
BROWN, A. E., S. SREENIVASAPRASAD, AND L. W. TIMMER. 1996. Molecular character-
ization of slow-growing orange and key lime anthracnose strains of Colletotri-
chum from citrus as C. acutatum. Phytopathology 86: 523-527.
CHILDERS, C. C. 1992. Suppression of Frankliniella bispinosa (Thysanoptera: Thripi-
dae) and the fungal pathogen Colletotrichum gloeosporioides, with pesticides
during the bloom cycle and improved fruit set on navel orange in Florida.
J. Econ. Entomol. 85: 1330-1339.
CHILDERS, C. C., AND D. S. ACHOR. 1991a. Feeding and oviposition injury to flowers
and developing floral buds of navel orange by Frankliniella bispinosa (Thysan-
optera: Thripidae) in Florida. Ann. Entomol. Soc. 84: 272-282.
CHILDERS, C. C., AND D. S. ACHOR. 1991b. Structure of the mouthparts of Franklin-
iella bispinosa (Morgan) (Thysanoptera: Thripidae). p. 71-94 in B. L. Parker,
M. Skinner and T. Lewis [eds.]. Towards understanding Thysanoptera. 1989
February 21-23; Burlington, VT. Gen. Tech. Rep. NE-147. USDA, Forest Ser-
CHILDERS, C. C., AND R. J. BESHEAR. 1992. Thrips (Thysanoptera) species associated
with developing citrus flowers in Florida and a key to adult terebrantian fe-
males. J. Entomol. Sci. 27: 393-412.
(Thysanoptera) species, their occurrence and seasonal abundance on develop-
ing buds and flowers of Florida citrus. J. Entomol. Sci. 25: 601-614.
CHILDERS, C. C., S. NAKAHARA, AND R. J. BESHEAR. 1994. Relative abundance of Fran-
kliniella bispinosa and other species of Thysanoptera emerging from soil be-
neath navel orange trees in Florida during spring flowering. J. Entomol. Sci.
29: 318-329.
GILBERT, M. J. 1986. First African record of Scirtothrips dorsalis Hood (Thysanoptera:
Thripidae) a potential pest of citrus and other crops in southern Africa. (Scien-
tific Note). J. Entomol. Soc. South Africa 49(1): 159-170.
GILBERT, M. J. 1990. Relative population levels of citrus thrips, Scirtothrips aurantii
on commercial citrus and adjacent bush. South Africa J. Zool. 25(1): 72-76.
LEWIS, T., L. A. MOUND, S. NAKAHARA, AND C. C. CHILDERS. 1997. Appendix II. Major
crops infested by thrips with main symptoms and predominant injurious spe-
cies. p. 675-709 in T. Lewis [ed.]. Thrips as crop pests. CAB International. Wall-
ingford, United Kingdom.
LIMA, J. E. 0., AND F. S. DAVIES. 1981. Fruit set and drop of Florida navel oranges.
Proc. Florida State Hort. Soc. 94: 11-14.
SAMWAYS, M. J., B. A. TATE, AND E. MURDOCH. 1987. Population levels of adult citrus
thrips, Scirtothrips aurantii Faure (Thysanoptera: Thripidae) relative to sea-
son and fruit-scarring. J. Appl. Entomol. 104: 372-377.

424 Florida Entomologist 82(3) September, 1999

SAS INSTITUTE. 1991. SAS language and procedures: usage 2, version 6, First edition.
SAS Institute, Cary, NC.
SONODA, R. M., AND R. R. PELOSI. 1988. Characteristics of Colletotrichum gloeospori-
oides from lesions on citrus blossoms in the Indian River area of Florida. Proc.
Florida State Hort. Soc. 101: 36-38.
STOUT, R. G. 1962. Estimating citrus production by use of frame count survey. J. Farm
Econ. 44: 1037-1049.
TAKAGI, K. 1981. The injuries on the citrus fruit caused by green leafhoppers, Em-
poasca spp., Tea yellow thrips, Scirtothrips dorsalis Hood, citrus red mite,
Panonychus citri (McGregor) and citrus rust mite, Aculops pelekassi (Keifer).
English summary [In Japanese]. Bull. Fruit Tree Res. Sta. Min. Agric. For. and
Fisheries. Series D (Kuchinotsu). 56(3).
TANIGOSHI, L. K., AND J. Y. NISHIO-WONG. 1982. Citrus thrips: biology, ecology and
control. U.S. Dept. Agric. Tech. Bull. 1668. 17 pp.


Citrus Research and Education Center, 700 Experiment Station Road,
Lake Alfred, FL, 33850


Populations of the brown citrus aphid (BCA), Toxoptera citricida (Kirkaldy), and
associated insects were monitored in citrus groves in western Puerto Rico in 1996 and
1997. Peak populations of the aphid were observed in spring and fall of both years. A
total of 6,737 infested citrus terminals were examined and categorized by growth
stage and BCA colony size. Natural enemies, primarily coccinellids, syrphids,
chrysopids, and aphidiids, were recorded when present in BCA colonies and a list of
species is presented. A complex of large coccinellid species was inferred to have the
greatest potential for impact on BCA populations in Puerto Rico. Predation by syrphid
larvae was another important source of BCA mortality. The parasitoid, Lysiphlebus
testaceipes (Cresson), was abundant in some samples, but usually parasitized only a
small percentage of aphids within colonies. The fungus, Verticillium lecanii (Zimm.),
infected BCA colonies sporadically. The red imported fire ant, Solenopsis invicta Bu-
ren, was often observed harvesting honeydew from BCA colonies and interfering with
natural enemies.


Durante 1996 y 1997 se realize un monitoreo del afido pardo de los citricos (BCA)
Toxoptera citricida (Kirkaldy) e insects asociados en huertos de citricos ubicados en
el occidente de Puerto Rico. En ambos ahos, las poblaciones maximas del afido ocu-
rrieron durante la primavera y el otoho. Se examine un total de 6,737 brotes infesta-
dos, los cuales se clasificaron por etapa de desarrollo y el tamaho de la colonia de BCA.
Se present una lista de enemigos naturales encontrados en las colonies de BCA,
principalmente Coccinellidae, Syrphidae, Chrysopidae y Aphidiidae. Se dedujo que un
complejo de species grandes Coccinellidae tienen el mayor potential para el control

Michaud & Browning: Brown Citrus Aphid

de BCA en Puerto Rico. Otra causa important de mortalidad de BCA fue la predacion
por larvas de Syrphidae. En algunas muestras se encontraron poblaciones altas del
parasitoide Lysiphlebus testaceipes (Cresson), pero 6stos solamente parasitaron pocos
afidos dentro de las colonies. El hongo Verticillium lecanii (Zimm.) se encontr6 infec-
tando colonies de BCA de manera esporadica. Con frecuencia se observe a la hormiga
de fuego roja importada, Solenopsis invicta Buren, cosechando secreciones en colonies
de BCA e interfiriendo con los enemigos naturales.

The discovery of the brown citrus aphid (BCA) in Ft. Lauderdale and Miami, Flor-
ida in 1995 caused widespread concern among citrus growers due to the potential im-
pact of this pest on citrus production (Halbert & Brown 1996). The primary concern
is the efficiency of this aphid in vectoring citrus tristeza virus (CTV) (Meneghini
1948), a serious disease of citrus worldwide (Lee & Rocha-Peha 1992). There are many
strains of CTV that range from benign to virulent, and most are presumed to have en-
tered the country on infected budwood that was subsequently propagated. The strain
commonly referred to as "quick decline" is widespread throughout Florida and in-
duces a hypersensitivity reaction in sweet orange and grapefruit scions on sour or-
ange rootstock, resulting in the death of the tree within two or three years. Although
trees are now being planted on CTV-tolerant or resistant rootstocks, it is estimated
there are as many as 20 million producing trees rooted on sour orange in Florida (Gar-
nsey 1995). Of equal concern are the various "stem-pitting" strains which, although
presently rare, may become more common with the introduction of BCA. Although
these strains do not directly cause tree death, they reduce fruit size and yield. Fur-
thermore, stem-pitting strains are capable of infecting most citrus cultivars indepen-
dent of rootstock. For reasons not yet understood, the BCA is capable of transmitting
CTV, and the severe strains in particular, with far greater efficiency than all other
aphid species found in citrus (Balaraman & Ramakrishnan 1979, Yokomi et al. 1994).
In addition to viral transmission, large infestations of BCA have been implicated in
stunting the growth of young trees and reducing fruit set (Symes 1924, Hall 1930,
Smit 1934, Hely 1968).
Literature on the BCA has recently been summarized (Michaud 1998). In antici-
pation of the need to develop an integrated management program for the BCA in Flor-
ida, we monitored BCA populations and their associated natural enemies in Puerto
Rico where the insect has been present since at least 1992 (Roche-Peha et al. 1995).
Here we summarize our findings of two years observations on the seasonal abundance
of the BCA and its natural enemies in Puerto Rico.


We monitored BCA populations regularly at three sites in western Puerto Rico
throughout 1996 and 1997 and sampled five additional sites intermittently. The
Ensenat and Munez sites were located in the mountainous interior of western Puerto
Rico in the heart of the citrus-growing region. Both sites receive plentiful rainfall
from spring to fall and experience moderate evening temperatures due to their eleva-
tion (300-400 m above sea level). The Ensenat site is a research station of the Univer-
sity of Puerto Rico (Recinto Universitario de Mayagtiez) located at km 10.5, Carretera
124 and bordered on all sides by natural forests. The two groves sampled at this site
both were budwood foundation blocks; one was a 6-10 yr-old grove of Tahiti lime, the
other a 12-20 yr-old block of mixed sweet orange, mostly Valencia.

Florida Entomologist 82(3) September, 1999

The Munez site is a private plantation located at km 1 on Carretera 124 on the out-
skirts of the city of Lares, consisting of ca. 80 hectares of mixed commercial citrus on
the slope of a mountain, interplanted mostly with coffee but with some plantain and ba-
nana at lower elevations. The grove borders semi-riparian land (mostly forest) on two
sides, and urban development on the outskirts of the town of Lares on the other two.
The Canc61l site is a privately owned plantation located 2 km off Carretera 358 at
km 2.5 in San German at the southwestern edge of the citrus growing region, about
8 km from the coast, in somewhat drier climate than the other two sites. The site, at
an elevation of 100-150 m. ASL, was bordered primarily by grazing pastures and
mountain forests and was interplanted with plantain and banana.
Other sites sampled intermittently included Corozal, a research substation of the
University of Puerto Rico located at the intersection of Carretera 647 and Carretera
159 in the district of Corozal; Neuva Era, a private plantation located at km 5 of Car-
retera 124 in the district of Las Marias; Limani, a research substation of the Univer-
sity of Puerto Rico located at km 6.5 of Carretera 525 in the district of Adjuntas; and
La Balear, a private plantation located 2 km off Carretera 135 at km 63.5 near the
town of Castanfer.
We defined an aphid colony as the aggregate of aphids infesting a single citrus ter-
minal. Flushing citrus terminals expand new leaves at the rate of about one per day
and remain suitable for aphid growth and reproduction for a period of 20 to 30 days.
Infested terminals were categorized according to their stage of growth as follows: F =
Feather (most leaves still folded); F/NE = Feather/Newly Expanded (some leaves
folded, others newly expanded); NE = Newly Expanded (most leaves newly ex-
panded); NE/NH = Newly Expanded/Newly Hardened (some leaves newly expanded,
others newly hardened); NH = Newly Hardened (most leaves newly hardened). BCA
colonies were classified into 6 size categories: 1 (1-10 aphids); 2 (11-50); 3 (51-150); 4
(151-500); 5 (501-1000); 6 (> 1000).
Alate adults were also tallied when present in colonies and, during the second sea-
son, we recorded the presence of nymphs with wing buds as an indication of colony
maturation. This enabled us to examine the distribution of immigrant alates by tal-
lying the numbers of alates present on non-maturing colonies, i.e. colonies without
alate nymphs present.
The numbers of natural enemies present at each BCA colony were recorded by life
stage and family, e.g. "1 syrphid egg" "2 coccinellid larvae" etc. Syrphid adults were oc-
casionally observed ovipositing on aphid colonies but their incidence was too low to
permit statistical analysis. Adult coccinellids were frequently collected in the field
and samples of coccinellid, chrysopid, and syrphid larvae were collected periodically,
brought back to the laboratory, and reared through to adult stages for identification
and detection of parasitism. We tallied the number of Lysiphlebus testaceipes (Cres-
son) mummies in BCA colonies, as well as the number of discernibly parasitized
aphids. The latter are distinguishable by their swollen and discolored appearance
when the parasitoid larva is in the later stages of development.
To test for discrimination by predators against parasitized aphids as prey, we ex-
posed mixtures of parasitized and unparasitized BCA to late instar larvae of Pseudo-
dorous clavatus (F.) (n = 8) and Cycloneda sanguine ssp. limbifer (L.) (n = 5) in plastic
petri dishes in the laboratory. A number of ant species were observed tending BCA col-
onies and these were noted when present. In particular, the red imported fire ant, So-
lenopsis invicta Buren, was frequently observed. At certain times and locations, aphid
corpses were observed with fungal hyphae growing out of them and, with some expe-
rience, we were able to distinguish saprophytic fungi consuming deceased aphids
from the entomopathogenic form, Verticillium lecanii (Zimm.).

Michaud & Browning: Brown Citrus Aphid

Statistical Analyses

Site-to-site and year-to-year comparisons were made using a one-way ANOVA de-
sign followed by Fisher's LSD when more than 2 groups were compared (SPSS 1995),
and comparisons between dates within years and sites using ANOVA for repeated
measures (Systat 1989). The distributions of natural enemy life stages across BCA
colonies of different size classes were compared with those expected by chance using
a Chi-Square Goodness-of-Fit Test with 5 df. Similarly, the frequencies with which dif-
ferent insects occurred together on BCA colonies (e.g. predatory larvae and fire ants)
were compared to expected values (calculated from their independent frequencies) us-
ing a Chi-Square Goodness-of-Fit Test with 2 df.


The Ensenat site averaged a higher percentage of shoots infested in 1996 than in
1997, (mean = 43.1% vs. 15.5%, F = 9.360, P < 0.01) and the same trend was evident,
though not significant, at Muhez (mean = 30% vs. 11.4%, F = 2.540, P = 0.13). Cancel
displayed an opposite, though non-significant, trend to higher percentage of shoots in-
fested in 1997 than in 1996 (mean = 17.5% vs. 28.5%,F = 1.980, P = 0.17). Pooling data
from all 3 sites yielded a mean of 31.6% shoots infested in 1996 and 18.7% in 1997 (F
= 5.100, P = 0.03).
Colonies in smaller size classes were invariably more abundant than those in
larger size categories (Table 1), and had a higher frequency of occurrence on terminals
in early, rather than late, developmental stages (Table 2). Many colonies appeared to
be founded by multiple alate foundresses (Fig. 1). The mean number of alate
foundresses per non-maturing colony in 1997 was calculated to be 1.85 + 0.06 (SEM),
almost certainly an underestimate given that a significant proportion of founding
alates would have already died and gone undetected.
We examined 3,667 BCA colonies in 1996 and 3,070 in 1997, for a total of 6,737.
Summary data for the three primary sites in both years are shown in Figs. 2-7, and
for secondary sites, in Table 3. The following differences between primary sites were
significant (Fisher's LSD, P < 0.05) in 1996: Ensenat had fewer coccinellid adults than
either Cancel or Munez (F = 18,459); Cancel had more coccinellid eggs than Ensenat
(F = 4.185) and more coccinellid larvae and syrphid eggs than either Ensenat or
Munez (F = 20.087 and F = 16.308, respectively); Munez had more mummies than
Cancel, which had more than Ensenat (F = 18.329); Cancel had fewer colonies tended
by fire ants than either Ensenat or Munez (F = 18.123) and; Ensenat had more colo-
nies mixed withAphis spiraecola Patch than either Cancel or Munez (F = 19.660). No
other differences between sites in 1996 were significant.
In 1997, the following differences between sites were significant: Cancel had more
coccinellid adults, eggs and larvae than either Ensenat or Munez (F = 11.111, F =
17.100, and F = 20.512, respectively); Cancel had more mummies than Ensenat (F =
8.571) and fewer colonies tended by fire ants than either Ensenat or Munez (F =
19.355); Ensenat had more colonies infected with V lecanii than either Cancel or Munez
(F = 55.996) and; Cancel had more colonies mixed with A. spiraecola than Ensenat,
which had more than Munez (F = 13.026). Our estimate of the percentage of colonies
maturing was not significantly different among sites in 1997 (F = 3.126, P -= 0.058); Can-
cel: Mean = 15.4 + 4.6%; Ensenat: Mean = 7.9 + 2.1%; Munez: Mean = 4.4 2.1%.
There were significant differences between years at all 3 sites. In 1997 the Ensenat
site had significantly more coccinellid adults (F = 14.730, P < 0.001), coccinellid larvae
(F = 12.176, P < 0.001), syrphid eggs (F = 6.061, P < 0.05) and syrphid larvae (F =

Florida Entomologist 82(3) September, 1999


Percent of BCA colonies in various size classes
Total no.
Location samples 1-10 11-50 51-150 151-500 500-1000 >1000

Cancl 2415 31.7 40.4 18.6 5.7 2.0 0.8
Enseiat 1617 29.4 31.2 24.3 10.6 3.6 0.9
Muiez 832 36.0 36.3 18.9 7.5 1.0 0.1


Terminal Stage

Colony Size F F/NE NE NE/NH NH Totals

1-10 554 464 753 160 137 2068
11-50 444 422 942 293 266 2367
51-150 127 237 602 278 239 1483
151-500 16 78 242 142 134 612
500-1000 0 14 61 53 37 165
>1000 0 7 7 19 9 21

Totals 1141 1222 2607 945 822 6737

4.311, P < 0.05) than in 1996, and more colonies attended by L. testaceipes (F = 9.286,
P < 0.01), but fewer colonies mixed withA. spiraecola (F = 26.130, P < 0.001). The Can-
cel site had significantly more coccinellid larvae (F = 16.669, P < 0.001) in 1997 than
in 1996, more colonies attended by L. testaceipes (F = 17.033, P < 0.001), more colonies
mixed withA. spiraecola (F = 9.839, P < 0.01), but fewer syrphid eggs (F = 6.727, P <
0.01). The Munez site also had more coccinellid larvae in 1997 than in 1996 (F =
17.747, P < 0.001) but fewer colonies mixed withA. spiraecola (F = 7.778, P < 0.01).
The following species of natural enemies were observed feeding on the BCA during
the study. Coleoptera, Coccinellidae: Chilochorus cacti (L.), Cladis nitidula (F.), Coelo-
phora inaequalis (F.), Curinus coerulous Mulsant, Coleomegilla innotata (Mulsant),
C. sanguinea ssp. limbifer, Diomus sp., Egius platycephalus Mulsant, Hippodamia
convergens (Guerin), Hyperaspis festiua Mulsant, Olla v-nigrum (Mulsant), Procula
feruuginea (Oliver), Scymnus (Schymnus) floralis (F.); Diptera, Chamaemyiidae: Leu-
copis sp. Diptera, Syrphidae: Allograpta radiata (Bigot), Allograpta exotica (Wied-
mann), Ocyptamus cubanus (Hull), Ocyptamus fuscipennis Say, P. clavatus;
Hymenoptera, Aphidiidae: L. testaceipes; Neuroptera, Chrysopidae: Cereaochrysa sp.
One particular set of observations (March 12, 1997, Limani) revealed an interest-
ing difference in the distribution of coccinellids and syrphids between two adjacent
blocks. The first block was a mixed citrus collection of mature trees surrounded by tall
weeds (n = 76 BCA colonies), whereas the second was a block of recently top-worked

Michaud & Browning: Brown Citrus Aphid 429




0 40-




0 ,I I I f
0 1 2 3 4 5 6 7 8 9 10 11 12

No. of alate foundresses

Fig. 1. Frequency distribution of BCA colonies with different numbers of alate

trees with virtually no canopy structure and very little ground cover (n = 81). The first
block had significantly more coccinellid adults per colony (mean + SE = 0.20 + 0.05)
compared with the second block (0.02 + 0.02) (F = 11.333, P = 0.001) and significantly
more coccinellid larvae (0.29 + 0.06 vs. 0.00 + 0.00, F = 23.543, P < 0.001). On the other
hand, the second block had significantly more syrphid eggs per colony than did the
first block (0.61 + 0.12 vs. 0.01 + 0.01, F = 21.720, P < 0.001), although the numbers
of syrphid larvae were not significantly different (F = 1.156, P = 0.284). Block one also
had significantly more mummies per colony than block two (1.92 + 0.61 vs. 0.02 + 0.02,
F = 10.293, P = 0.002), although the numbers of parasitized aphids was not different
(F = 0.896, P = 0.345).
A total of 821 coccinellid adults were observed on 653 BCA colonies, 969 coccinellid
eggs on 85 colonies, and 1427 larvae on 759 colonies. The egg data represent only the
large coccinellid species (primarily C. sanguine ssp. limbifer and C. inaequalis) which
oviposit in exposed locations. The fact that coccinellid eggs are usually laid in masses
of 5 to 25 accounts for their highly clumped distribution. However, most eggs of these
species appear to be laid outside aphid colonies on older, hardened leaves, which likely

Florida Entomologist 82(3) September, 1999

accounts for their relatively low numbers in this study (we recorded only eggs associ-
ated with BCA colonies). The eggs of smaller coccinellid species (Diomus, Hyperaspis
and Scymnus spp.) were never detected during our observations; eggs of these species
are very small and usually laid in concealed locations. Coccinellid pupae were some-
times observed on the undersides of hardened leaves below aphid colonies, but almost
never within a colony. On several occasions we observed newly closed coccinellid lar-
vae feeding on eggs of their own species in the field. The distribution of coccinellid eggs
and larvae across BCA colonies of different size classes in depicted in Fig. 8.
A total of 1,676 syrphid eggs were counted on 842 BCA colonies, and 533 larvae on
432 colonies. The vast majority (>85%) appeared to be eggs and larvae of P. clavatus,
although 0. fuscipennis comprised up to 25% of specimens in samples of larvae col-
lected in late summer and fall. Syrphid pupae were occasionally recorded on BCA col-
onies, but most mature larvae appeared to leave aphid colonies before pupation. When
syrphid larvae (P. clavatus) were confined together with coccinellid larvae (C. san-
guinea ssp. limbifer) in petri dishes (n = 16) we did not observe any evidence of attacks
by one on the other, even when larval size was asymetric and no aphids were provided
as food. As many as 30% of syrphid larvae in some samples were parasitized by a gre-
garious parasitoid, Syrphophagus nr. aphidivorous (Hymenoptera: Encyrtidae). The
distribution of syrphid eggs and larvae across BCA colonies of different size classes in
depicted in Fig. 9.
The only aphidiid collected in our samples was L. testaceipes. We observed 20,645
parasitized aphids in 1,299 BCA colonies and 8,939 mummies in 892 colonies. We col-
lected 23 samples of mummies from various sites and the emergence of adult wasps
varied from 0 to 33% between samples, with an overall average of 4.2%. The hyper-
parasitoid Pachyneuron nr. siphonophorae (Ashmead) (Hymenoptera: Pteromalidae)
emerged from 2.6% of mummies. We also collected several specimens of another hy-
perparasitoid, Alloxysta sp., that we observed probing aphids on a number of occa-
sions, although this species never emerged from any of our mummy samples. Late
instar larvae of P. clavatus and C. sanguine ssp. limbifer did not appear to discrimi-
nate against parasitized aphids as prey; both species readily consumed parasitized
aphids in our petri dish trials, even with unparasitized aphids present.
The red imported fire ant, S. invicta, was the ant species most frequently observed
tending BCA colonies in Puerto Rico. Other ant species collected at BCA colonies in-
cluded Brachymyrmex obscurion Ford, Monomorium ebeninum Ford, Paratrechina
longicornis (Latrielle), Pheidole fallax, Solenopsis glogularia, and Wasmannia auro-
punctata (Roger). We found that both syrphid and coccinellid larvae occurred on fire
ant-tended BCA colonies at significantly lower frequencies than expected by chance
(Chi-square = 11.4, 2df, P < 0.005 & Chi-square = 35.7, 2df, P < 0.005, respectively).


BCA abundance in Puerto Rico typically peaks twice a year, once in early spring
and again in mid fall, coinciding with peak periods of flush availability during these
seasons. A large proportion of available terminals may be infested during these peri-
ods and a significant number of large colonies are typically present. A third popula-
tion peak may sometimes occur in mid-summer (e.g. Canc61l, 1997, Fig. 3), but
infestation of summer flush is usually low. Mean daily temperatures above 25C have
adverse effects on BCA reproduction and survival (Komazaki 1982, Michaud unpub-
lished). Mean daily temperatures in summer frequently exceed 25C in Puerto Rico
and this may be a key factor limiting population growth during this season. However,
the primary coccinellid species reside in groves throughout the summer, even when

Michaud & Browning: Brown Citrus Aphid

0 (a)



0.6 -

EI L i L

Jan Feb Mar Apr May Jun Jul Aug Sep Oct Nuv Dec

Cancel, 1996

Fig. 2. Summary data for Cancel site, 1996; (a) Percentage of trees with flush (open
bars) and percentage of flushed terminals infested (solid bars); (b) Percentage of BCA
colonies under attack by coccinellids (open bars) and syrphids (solid bars); (c) Mean no
of coccinellid larvae (open bars) and adults (solid bars) per BCA colony; (e) Percentage
of BCA colonies with mummies and or aphids parasitized by L. testaceipes (open bars),
attended by S. invicta (solid bars), and mixed with A. spiraecola (hatched bars). As-
terisks indicate months for which data was not available, all other gaps represent '0'


100 -

100 -
20 -


No 0.8
0.6 -



Florida Entomologist 82(3) September, 1999



Jun Feb Mar Apr May Jun Jul Aug Sep Oct Nov Dec

Canc61, 1997

Fig. 3. Summary data for Cancel site, 1997; (a) Percentage of trees with flush (open
bars) and percentage of flushed terminals infested (solid bars); (b) Percentage of BCA
colonies under attack by coccinellids (open bars) and syrphids (solid bars); (c) Mean no
of coccinellid larvae (open bars) and adults (solid bars) per BCA colony; (e) Percentage
of BCA colonies with mummies and or aphids parasitized by L. testaceipes (open bars),
attended by S. invicta (solid bars), and mixed with A. spiraecola (hatched bars). As-
terisks indicate months for which data was not available, all other gaps represent '0'

Michaud & Browning: Brown Citrus Aphid


1.4 -
N -

0.6 -

* -I rLr n F I -1 I

Jan Feb Mar Apr May Jun Jul Aug Sep Oct Nov Dec

Ensefat, 1996

Fig. 4. Summary data for Ensanat site, 1996; (a) Percentage of trees with flush
(open bars) and percentage of flushed terminals infested (solid bars); (b) Percentage
of BCA colonies under attack by coccinellids (open bars) and syrphids (solid bars); (c)
Mean no of coccinellid larvae (open bars) and adults (solid bars) per BCA colony; (e)
Percentage of BCA colonies with mummies and or aphids parasitized byL. testaceipes
(open bars), attended by S. invicta (solid bars), and mixed withA. spiraecola (hatched
bars). Asterisks indicate months for which data was not available, all other gaps rep-
resent '0' values.

Florida Entomologist 82(3) September, 1999



No. os


l ] Fr-I r- El

Jan Feb Mar Apr May Jun Jul Aug Sep Oct Nov Dec

Ensefiat, 1997

Fig. 5. Summary data for Ensanat site, 1997; (a) Percentage of trees with flush
(open bars) and percentage of flushed terminals infested (solid bars); (b) Percentage
of BCA colonies under attack by coccinellids (open bars) and syrphids (solid bars); (c)
Mean no of coccinellid larvae (open bars) and adults (solid bars) per BCA colony; (e)
Percentage of BCA colonies with mummies and or aphids parasitized byL. testaceipes
(open bars), attended by S. invicta (solid bars), and mixed withA. spiraecola (hatched
bars). Asterisks indicate months for which data was not available, all other gaps rep-
resent '0' values.

Michaud & Browning: Brown Citrus Aphid

100 -



20 r r

I in 6. riE m

r- *

0.2 1 i L h. L

60 -

Jan Feb Mar Apr May Jan Jul Aug Sep Oct Nov Dec

Mufiez, 1996

Fig. 6. Summary data for Mufiez site, 1996; (a) Percentage of trees with flush (open
bars) and percentage of flushed terminals infested (solid bars); (b) Percentage of BCA
colonies under attack by coccinellids (open bars) and syrphids (solid bars); (c) Mean no
of coccinellid larvae (open bars) and adults (solid bars) per BCA colony; (e) Percentage
of BCA colonies with mummies and or aphids parasitized by L. testaceipes (open bars),
attended by S. invicta (solid bars), and mixed with A. spiraecola (hatched bars). As-
terisks indicate months for which data was not available, all other gaps represent '0'


0.2 -



- ----- -------


Florida Entomologist 82(3)

September, 1999

16 A I

Jan Feb Mar Apr May Jun Jul Aug Sep Oct Nov Dec

Mufiez, 1997

Fig. 7. Summary data for Mufiez site, 1997; (a) Percentage of trees with flush (open
bars) and percentage of flushed terminals infested (solid bars); (b) Percentage of BCA
colonies under attack by coccinellids (open bars) and syrphids (solid bars); (c) Mean no
of coccinellid larvae (open bars) and adults (solid bars) per BCA colony; (e) Percentage
of BCA colonies with mummies and or aphids parasitized by L. testaceipes (open bars),
attended by S. invicta (solid bars), and mixed with A. spiraecola (hatched bars). As-
terisks indicate months for which data was not available, all other gaps represent '0'



20 -


Michaud & Browning: Brown Citrus Aphid

40 00 0 0 coC o
d 0 ;d N

0000 0C
4 Clq

^ 0 0 0 0 N;

000 m
N0 2"^ -t -

0 0 10 -

a -os
d a^ a 6 0 06 ao

10 0 0o No N N
*^ C -I CO -N

CO 0 -N 0N O^ (-
0N 0 0; 6 0 0; 0

d~ o 10 d 10 -N -
0o 0^ 0 0 0 0

10 10 dZ dO 0 d

Cl 10 o 10 oo io~

S000 C N 0

ad ld C


T -H ^ 0 TO -^

o a 00 0 a
000 00 -i o o 0

( o 0; 0 0o 6

-N -^ Cl Cld

1- co '- (N -

bo a- ab a a a

tD 0 10 Cl CO 10
10 0 d dO 0 10
Cl Cl '- "^ Cl 10

Cl 10 0 10 W 10
0 COZ NO 10 -^

co 10C1N

d a a d 10 0
-^ CO tD Cl 0; ^O


o -t

10 CM 0

1 0 103

10 0; 0 0*

0 0 0D 0 *


.- Cl
00 0 00

^ CM 01

OZ 1D0 -' -

'* O N CO


0" 01 01 0
0000 o o

0 C






Florida Entomologist 82(3) September, 1999


0 0^ C1j'.1

H- -I1~C

W oz C
o I | -f


Q --

E ~ t 00000 OOO

c .a< d d d d d

5 O~ 0N N 1(2-

E-i S o 1
H c1(2

M ~ ~ ~ C 6 0 0 0 0OO^if

m z '.l

4 c0 0 -t' 0
K Co 0 0

m" S' 5 0 0 0 0d

o || q o

I-. <1 < m 2 zz0

3 Cl1( -a^
ctz^ 0 0

^S t 6c o o o o

St 1(32 ^f N
0< o o+J,
cc C'iO(= 2.
4eT c $4 -i C- ,"
m Q -&

Michaud & Browning: Brown Citrus Aphid

50 --- 50
(a) (b)
40 n=85 40 n=759

30 30-

20 20-

10- 10

0 1 2 3 4 5 6 7 0 1 2 3 4 5 6 7
Colony Size Colony Size

Fig. 8. Distribution of observed (open bars) vs. expected (hatched bars) percentage
of BCA colonies of different size classes with coccinellid (a) eggs and (b) larvae. Aster-
isks indicate significant differences between observed and expected values, P < 0.05.
Chi-square Goodness of Fit test.

BCA density is low, and eliminate many aphid colonies in their early stages. Adult coc-
cinellids were sometimes observed feeding on soft scales or whitefly larvae when very
few aphids were present.
Year to year variation in BCA infestation intensity was evident. The BCA was ap-
parently more abundant in 1996 than in 1997, especially in the Lares district where
the Ensehat site averaged a larger proportion of terminals infested across all sampling
dates. The same trend was evident, though not significant, at the Munez site which is
only 12 km from Ensenat, but was not evident at the Canc6l site which is relatively dis-
tant from the other two. We conclude that regional differences exist with respect to var-
ious ecological parameters that influence BCA populations in particular years.
There were also consistent differences between sites with respect to natural ene-
mies. The Canc6 site had a consistently higher rate of coccinellid attendance at BCA
colonies than either of the other two sites in both years. It is not clear whether this dif-
ference is due to more successful reproduction of coccinellid populations within the
Canc61l grove, or to higher rates of immigration from surrounding habitats, but both
these parameters are likely important determinants of coccinellid impact on BCA

BCA Colony Structure

Small colonies were more often observed on terminals in earlier growth stages,
whereas large colonies were more frequently encountered on mature terminals. This
result is expected if younger terminals are the preferred sites for colonization. A num-
ber of small colonies were observed on relatively mature terminals but many of these
were in decline, i.e. they had been decimated by predation and/or other mortality fac-
tors. Newly expanded terminals can be rare toward the end of a flush cycle and a sig-
nificant number of small colonies may form on older terminals, despite a low
probability of colony maturation prior to shoot hardening.

Florida Entomologist 82(3)

September, 1999

50 50
(a) (b)

40 n =842 40 n =432

30 30 -

20 20

10 10

0 0
0 1 2 3 4 5 6 7 0 1 2 3 4 5 6 7
Colony Size Colony Size

Fig. 9. Distribution of observed (open bars) vs. expected (hatched bars) percentage
of BCA colonies of different size classes with syrphid (a) eggs and (b) larvae. Asterisks
indicate significant differences between observed and expected values, P < 0.05. Chi-
square Goodness of Fit test.

On heavily infested trees, large colonies could be found on shoots in relatively
early developmental stages, probably founded by apterous migrants from adjacent
mature colonies. Apterous adults have a significantly higher age-specific fecundity
and net reproductive rate than their alate counterparts, together with a shorter pre-
reproductive period (Takanashi 1989). Therefore, colonies founded by multiple apter-
ous migrants grow much faster than alate-founded colonies.
The data in Fig. 1 suggest that a large proportion of immigrant alates join colonies
already in progress as opposed to initiating colonies of their own. Recent laboratory
investigations (Michaud 1999) have confirmed that alate BCA have an intrinsic ag-
gregation tendency. It seems likely that multiple foundresses would increase colony
growth rate, although whether this would lead to an increased probability of colony
maturation remains to be shown. We suspect that small colonies are more vulnerable
to elimination by predation events that would have little impact on larger colonies,
leading to individual fitness benefits for foundresses that pool their reproductive ef-
forts to produce faster-growing/larger colonies.


We concur with Frazer & Gilbert (1976) that current methods for counting adult coc-
cinellids in the field greatly underestimate their true numbers. At any given time, only
a proportion of the population may be actively feeding on aphid colonies. This propor-
tion may be even lower following heavy rain or inclement weather when many adults
tend to seek shelter in secluded sites. We are confident, therefore, that our data tend to
underestimate the true numbers of coccinellid adults present in these citrus groves. C.
sanguine ssp. limbifer was the most abundant and ubiquitous coccinellid species in the
study, closely followed by C. inaequalis. C. sanguine is also the most common coccinel-
lid on BCA in the Dominican Republic (Borb6n et al. 1992) and Cuba (Zeleny 1969,
Batista et al. 1995) and one of the prominent species in Venezuela (Morales & Burandt
1985) and Brazil (Lara et al. 1977, Bartoszeck 1980, Santos & Pinto 1981).

Michaud & Browning: Brown Citrus Aphid

It is evident from Fig. 8a that established BCA colonies in the larger size classes
were the preferred oviposition sites of the large coccinellid species and that small,
newly formed colonies were less preferred. However, coccinellid larvae occurred on the
smaller colony size classes at close to expected frequencies (Fig. 8b), indicating that
the larvae search for, and effectively locate, aphid colonies. We sometimes observed
egg masses laid on leaves adjacent to infested terminals, rather than on the termi-
nals/colonies themselves. The tendency for a significant amount of oviposition to occur
outside the actual colonies may account for the relatively low incidence of coccinellid
eggs within colonies. Hemptinne et al. (1992) presented evidence that oviposition by
gravid females ofAdalia bipunctata (L.) was inhibited as a result of exposure to con-
specific larvae or other females. Given the cannibalistic habits of the larvae of many
coccinellids, it may be adaptive for females to place their egg masses near, but not on,
the infested terminals that are likely to attract conspecific larvae. Alternatively, the
survival and/or fecundity of some coccinellid species may be reduced when feeding on
the BCA, as reported by Tao & Shui (1971) and Morales & Burandt (1985). We suc-
cessfully reared larvae of C. sanguine ssp. limbifer and C. inaequalis to adulthood in
the laboratory on an exclusive diet of BCA, but did not assess the fertility of adults.
The difference in distribution of coccinellid adults and larvae observed between
the two adjacent blocks at Limani on March 12, 1997 could be explained by a foraging
preference among adult coccinellids for habitats with substantial ground cover. Dis-
tinct habitat preferences are known to influence the local abundance of many coccinel-
lid species (Honek & Rejmanek 1982, Honek 1985). The generalized escape response
of coccinellid adults is to fall to the ground when a threat is perceived, a behavior
which may be less effective in the absence of ground cover. We observed the assassin
bug Zelus longipes (L.) (Hemiptera: Reduviidae) preying on both adult and larval coc-
cinellids and this species may be an important source of mortality when it becomes
abundant in summer months. However, the most vulnerable coccinellid life stage ap-
pears to be the pupa; we observed high rates of pupal predation but never observed
the predatory agent.


Colonies in the smallest size class had fewer syrphid eggs than would be expected
by chance (Fig. 9a), a result that appears to conflict with accounts of syrphid prefer-
ence for ovipositing on small colonies (e.g. Firempong & Kumar 1975, Kan 1988).
However, syrphid preference for aphid colonies of particular sizes varies among spe-
cies and some are known to prefer larger colonies for oviposition (Dixon 1959, Chan-
dler 1968). There was a tendency for large BCA colonies to receive multiple
ovipositions, which may partially account for the higher than expected proportion of
class 3 and 4 colonies with syrphid eggs (Fig. 9a). Syrphid larvae have lower vagility
than coccinellid larvae and, presumably, a lower probability of locating another aphid
colony should they completely consume their nascent colony. Consequently, the selec-
tion of oviposition sites by adult female syrphids is probably a more important deter-
minant of larval survival than it is for coccinellids. Newly initiated BCA colonies may
be avoided as oviposition sites because their establishment, and their ability to pro-
vide sufficient food for larval development, is uncertain. The largest number of syr-
phid eggs were laid on class 2 and 3 colonies, which presumably have greater survival
probability than class 1 colonies. These data are consistent with those for syrphid lar-
vae which show the largest number appearing on class 3 and 4 colonies (Fig. 9b).
Overall, these data are remarkably similar to those presented by White (1995) for P.
clavatus on BCA colonies of different size classes in Trinidad. Adult syrphids may pre-
fer to forage in open habitats rather than among closely-spaced trees with full cano-

Florida Entomologist 82(3) September, 1999

pies. Spider predation is an important mortality risk for adult syrphids and spider
webs are abundant within and between trees in mature citrus groves. Such a habitat
preference could explain the dramatic difference in numbers of syrphid eggs on colo-
nies in block 1 vs. block two observed on March 12, 1997 at Limani.


Chrysopids are relatively rare in Puerto Rican citrus groves, despite generally low
pesticide use and an abundance of alternate prey species. We collected a total of nine
larvae and three adults of Cereaochrysa sp. The larvae are trash-carriers and one
reared out in the laboratory yielded a parasitoid, Brachycyrtus sp. (Hymenoptera: Ich-


Incomplete development of L. testaceipes inA. spiraecola has been observed in It-
aly (Tremblay & Barbagallo 1983), France (Stary et al. 1988), and Florida (J. P.
Michaud, unpublished), although Costa and Stary (1989) found a strain in Portugal
which did complete development in BCA. The low emergence rate of L. testaceipes
from mummies of BCA in Puerto Rico was very similar to that reported by Yokomi &
Tang (1996) and Carver (1978, 1984) in Australia. On several occasions we tried, and
failed, to initiate laboratory colonies using wasps from samples of mummies with high
percentage emergence. We suspect that local populations of L. testaceipes express ad-
ditive genetic variance for survival in BCA, which would account for the high variabil-
ity in emergence rates among samples and across generations.
Despite the low emergence rate, L. testaceipes females readily oviposit in the BCA,
causing the aphid to act as an 'egg trap' for the parasitoid, as previously noted by
Carver (1984). Considering the paucity of alternate host aphids for L. testaceipes in
citrus (both Toxoptera aurantii Boyer de Fonscolombe and Aphis gossypii Glover are
scarce and A. spiraecola is an unsuitable host), it is doubtful that the parasitoid pop-
ulation is self-maintaining within the citrus ecosystem in Puerto Rico. Nevertheless,
the high floral diversity of the region apparently affords many suitable hosts for
L. testaceipes, and citrus groves receive periodic waves of immigration. We suspect
that the variation in abundance of L. testaceipes in BCA populations, both seasonally
and between sites, largely reflects variation in its emergence rate from alternative
hosts in the vicinity of citrus groves.
Rosenheim et al. (1995) have recently drawn attention to the importance of iden-
tifying intra-guild predation as a potential impediment to biological control programs.
We observed no intra-guild predation under laboratory conditions when C. sanguine
and P. clavatus larvae were placed in petri dishes together, even when larval size was
asymmetric. Consumption of parasitized prey has been previously recorded for coc-
cinellids (Quezada & Debach 1973; Hoelmer et al. 1994, Colfer & Rosenheim 1995)
and syrphids (Kindlmann & Ruzicka 1992). Both the syrphid and coccinellid larvae
we tested readily consumed BCA parasitized by L. testaceipes, and it is therefore
likely that predation of larvae within aphids is a source of mortality for L. testaceipes
in addition to hyperparasitism by Pachyneuron and Alloxysta spp.


The red imported fire ant, S. invicta, is a ubiquitous resident of citrus groves in Pu-
erto Rico and was the most common ant species in attendance at BCA colonies. It has

Michaud & Browning: Brown Citrus Aphid

been reported that ants in general (Shindo 1972), and S. invicta in particular, can have
negative effects on aphid predators (El-Ziady & Kennedy, 1956; Vinson & Scarborough
1989) and parasitoids (Frazer & van den Bosch 1973). The fact that syrphid eggs were
observed on fire ant-tended colonies at close to expected frequencies suggests that syr-
phid adults do not discriminate against fire ant-tended aphid colonies for oviposition.
However, significantly fewer syrphid larvae were observed on fire ant-tended colonies
than would be expected by chance, and the same was true for coccinellid larvae, sug-
gesting a negative impact of fire ants on the beneficial activities of both these important
natural enemies groups. Vinson & Scarborough (1991) showed that fire ants remove L.
testaceipes-parasitized aphids and mummies from Rhopalosiphum maidis (Fitch) colo-
nies growing on sorghum. We observed fire ants carrying mummies away from BCA col-
onies on occasion, although mummies and aphids parasitized by L. testaceipes were
both encountered on fire ant-tended colonies at close-to-expected frequencies.
Vinson & Scarborough (1991) also reported that fire ants carried aphids to unin-
fested leaves and we observed live aphids being transported by fire ants on several oc-
casions. This behavior could potentially lead to amplification of BCA infestations
within trees if fire ants frequently initiate new colonies on uninfested terminals. We
also observed fire ants removing dead aphids from BCA colonies, a behavior which
could conceivably interfere with disease dynamics in the aphid population.

Other Aphididae

The green citrus aphid, A. spiraecola, was frequently observed to form mixed col-
onies with the BCA, even when many uninfested terminals were available for coloni-
zation. The percentage of mixed colonies seemed to peak in April at most sites (see
Ensenat, 1996, Fig, 4; Cancel 1996 & 1997, Figs. 2 & 3; Munfez 1996, Fig. 6; Limani
and La Balear, Table 3) when 50% or more of BCA colonies may be mixed withA. spi-
raecola. This appears to be the period of peak flight activity for A. spiraecola, when
alates are migrating into citrus in large numbers from maturing colonies on other
hosts. Why A. spiraecola alates select terminals already occupied by BCA (or vice
versa) remains to be explored. Both species are attacked by the same guild of natural
enemies and there may be a 'safety in numbers' advantage to the association. A. spi-
raecola could benefit from dilution of predation pressure when in association with
BCA, since the latter is a larger aphid with a higher reproductive rate. Possible ben-
efits for the BCA include the protected microenvironment created by the cupped and
twisted leaves produced by citrus in response to A. spiraecola feeding.

Implications for biological control

Assuming that only alate aphids usually move from tree to tree, only those BCA
colonies which mature and export alates are of economic importance with regard to
the transmission of CTV. We suspect that colony resilience to predation increases with
colony size, simply because larger colonies are more likely to leave survivors. Based
on the high reproductive potential of the BCA (Komazaki 1988), we suspect that only
colonies discovered by natural enemies in relatively early stages have a high proba-
bility of being completely eliminated. Nevertheless, predation appears the most likely
reason that only a relatively small proportion of BCA colonies were observed in ma-
ture stages, i.e. producing migratory alatae.
The BCA is notorious as an aphid species with few effective parasitoids (Carver
1978, Carver & Woolcock 1985, Murakami et al. 1984). In Japan, Lysiphlebiajaponica
is reported to attack the BCA (Takanashi 1990, 1991). Flanders & Fisher (1959) at-

Florida Entomologist 82(3) September, 1999

tribute some control of BCA to parasitism by Trioxys and Aphelinus spp in Kwang-
tung Province, China but relatively little information is available on other potential
parasitoids of BCA in China which is presumably its country of origin. In Puerto Rico,
the relatively low survival of L. testaceipes in the BCA essentially negates any numer-
ical response of this parasitoid within BCA infestations. Furthermore, the tendency of
females to parasitize only a small fraction of the aphids in each colony means that col-
onies are rarely eliminated by parasitism alone.
The pathogenic fungus V lecanii was observed in BCA colonies only sporadically,
but was occasionally a significant mortality factor at the Ensenat site, particularly in
1997. This fungus has been reported to cause significant mortality to BCA popula-
tions in Venezuela (Rond6n et al. 1981) and Argentina (Yasem de Romero 1985) al-
though epizootics are contingent on environmental conditions conducive to infection,
specifically prolonged periods of leaf wetness.
The most important natural enemies of the BCA in Puerto Rico appear to be the
complex of larger coccinellid species which apparently eliminate large numbers of col-
onies in their earliest stages. Syrphids are another important natural enemy group
probably contributing to the elimination of many BCA colonies, and to a reduction in
the number of alates exported from others. The conditions influencing production of
alatae in the BCA have not been studied, but one important factor appears to be
crowding. We suspect that predation within colonies can delay the production of
alates, both through direct reductions in aphid density and by the disturbance and
scattering of aphids within colonies. The potential exists for an additive impact of syr-
phids and coccinellids on BCA populations, in that intra-guild predation does not ap-
pear to occur between the primary species. On the other hand, our observations
indicate that these two predator guilds may express disparate habitat preferences un-
der some circumstances. It is notable that other guilds of aphidophagous predators
such as the Staphylinidae (Coleoptera), Anthocoridae, Miridae and Nabidae (Hemi-
ptera) were not recorded from BCA colonies in this study. However, since our obser-
vations were restricted to daylight hours, nocturnal predators would not have been
observed. Our observations also indicate that red imported fire ants have a measur-
able negative impact on natural enemies when they tend BCA colonies, and may
therefore warrant consideration in integrated management of the BCA.


We thank the personnel of USDA, APHIS, PPQ for their invaluable assistance, in
particular D. Fieselmann (Gulfport, MI), D. Rivera (San Juan, PR), and E. Rodriguez
(Mayagtiez, PR). L. Almodovar, M. Bonet, and V. Torres assisted with data collection.
We are also grateful to G. Evans, R. Gordon, L. Stange, M. Thomas, and H. Weems of
the Florida State Collection of Arthropods for their assistance with specimen identi-
fications. This work was funded by a UFL (IFAS) grant from USDA, APHIS, PPQ.
Florida Agriculture Experiment Station Journal Series No. R-06236.


BALARAMAN, K., AND K. RAMAKRISHNAN. 1979. Transmission studies with strains of
citrus tristeza virus on acid lime. Z. Pflanzenkr. Pflanzenschutz 86: 653-661.
BARTOSZECK, A. B. 1980. The occurrence of Toxoptera citricidus (Homoptera: Aphididae)
and its natural enemies in Imperatriz, Maranhao, Brazil. Dusenia: 12: 9-13.
AMO, R. PEREZ, AND J. L. MORERA. 1995. Tristeza and Toxoptera citricida in
Cuba, incidence and control strategy. p. 197-203 in R. F. Lee, M. Rocha Pena,

Michaud & Browning: Brown Citrus Aphid

C. L. Niblett, F. Ochoa, S. M. Garnsey, R. K. Yokomi, and R. Lastra, [eds.] Pro-
ceedings of the Third International Workshop on Citrus Tristeza Virus and
Brown Citrus Aphid in the Caribbean Basin: Management Strategies. Lake Al-
fred, FL, May 15-18, 1995. University of Florida, Institute of Food and Agricul-
tural Sciences.
BORBON, J. C., A. J. ABUD, P. J. MILLAN, J. ASIATICO, AND N. ABREU. 1992. Presencia
de la tristeza de los citricos y Toxoptera citricidus (Kirkaldy) en la Republica
Dominicana. p. 95-101 in R. Lastra, R. F. Lee, M. Rocha-Pena, C. L. Niblett,
F. Ochoa, S. M. Garnsey, R. K. Yokomi [eds.] Citrus tristeza virus and Toxoptera
citricidus in Central America: Development of management strategies and use
of biotechnology for control. Proceedings of a Workshop, Maracay, Venezuela,
Sept. 14-19, 1992.
CARVER, M. 1978. The black citrus aphids, Toxoptera citricidus (Kirkaldy) and T au-
rantii (Boyer de Fonscolombe) (Homoptera: Aphididae). J. Aust. Ent. Soc. 17:
CARVER, M. 1984. The potential host ranges in Australia of some imported aphid par-
asites (Hymenoptera: Ichneumonoidea: Aphidiidae). Entomophaga 29: 351-360.
CARVER, M., AND L. T. WOOLCOCK. 1985. Interactions betweenAcyrthosiphum kondoi
(Homoptera: Aphidoidea) and Aphelinus asychis (Hymenoptera: Chalcidoidea)
and other parasites and hosts. Entomophaga 30(2): 193-198.
CHANDLER, A. E. F. 1968. The relationship between aphid infestations and oviposition
by aphidophagous Syrphidae (Diptera). Ann. Appl. Biol. 61: 415-423.
COLFER, R. G., AND J. A. ROSENHEIM, 1995. Intraguild predation by coccinellid beetles
on an aphid parasitoid, Lysiphlebus testaceipes. Proc. Beltwide Cotton Conf. pp.
COSTA, A., AND P. STARY, 1989. Lysiphlebus testaceipes, an introduced parasitoid in
Portugal (Hymenoptera: Aphidiidae). Entomophaga 33: 403-412.
DIXON, T.J. 1959. Studies on oviposition behaviour of Syrphidae (Diptera). Trans. R.
Entomol. Soc. Lon. 3: 57-80.
EL-ZAIDY, S., AND J. S. KENNEDY. 1956. Beneficial effects of the common garden ant,
Lasius niger L. on the black bean aphid, Aphis fabae Scopoli. Proc. R. Entomol.
Soc. Lond. Ser. A. Gen. Entomol. 31: 61-65.
FIREMPONG, S., AND R. KUMAR. 1975. Natural enemies of Toxoptera aurantii (Boy.)
(Homoptera: Aphididae) on cocoa in Ghana. Biol. J. Linn. Soc. 7: 261-292.
FLANDERS, S. E., AND T. W. FISHER. 1959. The economic effect of aphidophagous in-
sects on citrus in south China. J. Econ. Entomol. 52: 536-537.
FRAZER, B. D., AND R. VAN DEN BOSCH. 1973. Biological control of the walnut aphid in
California: The inter-relationship of the aphid and its parasite. Environ. Ento-
mol. 2: 561-568.
FRAZER, B. D., AND N. GILBERT. 1976. Coccinellids and aphids: a quantitative study of
the impact of adult ladybirds (Coleoptera: Coccinellidae) preying on field popu-
lations of pea aphids (Homoptera: Aphididae). J. Entomol. Soc. B.C. 73: 34-56.
GARNSEY, S.M. 1995. A brief perspective on tristeza in Florida. Proc. 3rd. Int. Work-
shop on CTV and BCA.
HALBERT, S. E., AND L. G. BROWN. 1996. Toxoptera citricida (Kirkaldy), Brown citrus
aphid-Identification, biology and management strategies. Entomol. Circ. No.
374, Fla. Dept. Agr. & Consumer Services.
HALL, W. J. 1930. Notes on the control of some of the more important insect pests of
Citrus in Rhodesia. Rhodesia Agric. J. 27: 737-747.
HELY, P. C. 1968. The entomology of citrus in New South Wales. Misc. Pub. Aust. En-
tomol. Soc. 1: 1-20.
HEMPTINNE, J. L., A. F. G. DIXON, AND J. COFFIN. 1992. Attack strategy of ladybird
beetles (Coccinellidae): factors shaping their numerical response. Oecologia:
90: 238-245.
HOELMER, K. A., L. S. OSBORNE, AND R. K. YOKOMI. 1994. Interactions of the whitefly
predator Delphastus puclillus (Coleoptera: Coccinellidae) with parasitized sweet
potato whitefly (Homoptera: Aleyrodidae). Environ. Entomol. 23: 136-139.

Florida Entomologist 82(3) September, 1999

KAN, E. 1988. Assessment of aphid colonies by hoverflies. I. Maple aphids and Episyr-
phis balteatus (de Geer) (Dipter: Syrphidae). J. Ethol. 6: 39-48.
KINDLMANN, P., AND Z. RUZICKA. 1992. Possible consequences of a specific interaction
between predators and parasites of aphids. Ecol. Modelling 61: 253-265.
KOMAZAKI, S. 1982. Effects of constant temperature on population growth of Tox-
optera citricidus (Kirkaldy), Aphis citricola van der Groot and Aphis gossypii
Glover (Homoptera: Aphididae) on Citrus. Appl. Entomol. Zool.
KOMAZAKI, S. 1988. Growth and reproduction in the first two summer generations of
two citrus aphids, Aphis citricola Van Der Groot and Toxoptera citricidus
(Kirkaldy) (Homoptera: Aphididae), under different thermal conditions. Appl.
Entomol. Zool. 23(3): 220-227.
LARA, F. M., S. A. DE BORTOLI, AND E. A. OLIVEIRA. 1977. Population fluctuations in
some insects associated with Citrus sp. and their correlation with meteorolog-
ical factors. Cientifica 5: 134-143.
LEE, R. F., AND M. A. ROCHA-PENA. 1992. Citrus Tristeza Virus. p. 226-249 in J. Ku-
mar, H. S. Chaube, U. S. Singh and A. N. Mukhopadhyay [eds.]. Plant Diseases
of International Importance. Prentice Hall, Englewood Cliffs, NJ.
MENEGHINI, M. 1948. Experiencias de transmissao da doenga "tristeza" dos citrus
pelo pulgao preto da laranjeira. 0. Biol6gico 14 (5): 115-118.
MICHAUD, J. P. 1998. A review of the literature on Toxoptera citricida (Kirkaldy) (Ho-
moptera: Aphididae). Fla. Entomol. 81: 37-60.
MICHAUD, J. P. 1999. Aggregation by alatae of Toxoptera citricida (Kirkaldy) (Ho-
moptera: Aphididae). Environ. Entomol. (in press)
MORALES, J., AND C. L. BURANDT, JR. 1985. Interactions between Cycloneda san-
guinea and the brown citrus aphid: Adult feeding and larval mortality. Eviron.
Entomol. 14: 520-522.
MURAKAMI, Y., N. ABE, AND G. W. COSENZA. 1984. Parasitoids of scale insects and
aphids on citrus in the Cerrados region of Brazil (Hymenoptera: Chacidoidea).
Appl. Entomol. Zool. 19(2): 237-244.
QUEZADA, J. R., AND P. DEBACH. 1973. Bioecological and populations studies of the
cottony-cushion scale, Icerya purchase Mask., and its natural enemies, Rodolia
cardenalis Mul. and Cryptochaetum iceryae Will. in southern California. Hil-
garida 41: 631-688.
GARNSEY, AND R. K. YOKOMI. 1995. Citrus tristeza virus and its aphid vector
Toxoptera citricida: Threats to citrus production in the Caribbean and Central
and North America. Plant Dis. 79: 437-443.
RONDON, A., E. ARNAL, AND F. GODOY. 1981. Comportamiento del Verticillium lecanii
(Zimm.) Viegas, pat6geno del afido Toxoptera citricida (Kirk.) en finca citrico-
las de Venezuela. Agron. Trop. 30: 201-212.
traguild predation among biological-control agents: Theory and evidence. Biol.
Con. 5: 303-335.
SANTOS, G. P., AND A. C. Q. PINTO. 1981. Biologia de Cycloneda sanguinea e sua as-
sociagao com pulgao em mudas de mangueira. Pesc. Agropec. Bras. 16(4): 473-
SHINDo, M. 1972. Relation between ants and aphids in a citrus orchard. Proc. Assoc.
P1. Prot. Kyushu 18: 69-71.
SMIT, B. 1934. A study of the most important insect pests that confront the Citrus
grower in the Eastern Province. S. Afr. J. Sci. 31: 439-441.
SPSS". 1995. Statistical Package for the Social Sciences, v. 6.1 for the Power Macin-
tosh". SPSS Inc., Chicago, IL.
STARY, P., J. P. LYON, AND F. LECANT. 1988. Biocontrol of aphids by the introduced Ly-
siphlebus testaceipes (Cress.) (Hymenoptera: Aphidiidae) in Mediterranean
France. J. Appl. Entomol. 105: 74-87.
SYMES, C. B. 1924. Notes on the black citrus aphis. Rhodesia Agric. J. 11: 612-626.
SYSTAT". 1989. SYSTAT v. 5.0 for the Macintosh", SYSTAT Inc., Evanston, IL.

Michaud & Browning: Brown Citrus Aphid

TAKANASHI, M. 1989. The reproductive ability of apterous and alate viviparous mor-
phs of the citrus brown aphid, Toxoptera citricidus (Kirkaldy) (Homoptera:
Aphididae). Jap. J. Appl. Entomol. Zool. 33: 266-269.
TAKANASHI, M. 1990. Development and reproductive ability of Lysiphlebus japonicus
Ashmead (Hymenoptera: Aphidiidae) parasitizing the brown citrus aphid, Tox-
optera citricidus (Kirkaldy) (Homoptera: Aphididae). Jpn. J. Appl. Ent. Zool.
34: 237-243.
TAKANASHI, M. 1991. Percent parasitism by an aphidiid parasitoid, Lysiphlebus
japonicus Ashmead attacking host colonies of different sizes. Proc. Assoc. P1.
Prot. Kyushu 37: 163-166.
TAO, C. C., AND S. C. HUI. 1971. Biological control of citrus, vegetable and tobacco
aphids. Spec. Publ. Taiwan Agric. Res. Inst. 10: 1-110.
TREMBLAY, E., AND S. BARBAGALLO. 1983. Lysiphlebus testaceipes (Cr.) a special case
of ecesis in Italy. p. 65-68 in R. Cavalloro [ed.]. "Aphid Antagonists", Proceed-
ings of a meeting of the EC Experts Group, Portici, Italy, 23-24 Nov. 1982. A. A.
Balkema, Rotterdam.
VINSON, S. B., AND T. A. SCARBOROUGH. 1989. Impact of the imported fire ant on pop-
ulations of cotton aphid (Aphis gossypii) predators. Fla. Entomol. 72: 107-111.
VINSON, S. B., AND T. A. SCARBOROUGH. 1991. Interactions between Solenopsis invicta
(Hymenoptera: Formicidae), Rhopalosiphum maidis (Homoptera: Aphididae),
and the parasitoid Lysiphlebus testaceipes Cresson (Hymenoptera: Aphidi-
idae). Ann. Entomol. Soc. Am. 84: 158-164.
WHITE, G. 1995. Natural enemies of Toxoptera citricida in Trinidad WI. p. 41-43 in R.
F. Lee, M. Rocha Pena, C. L. Niblett, F. Ochoa, S. M. Garnsey, R. K. Yokomi, and
R. Lastra [eds.] Proceedings of the Third International Workshop on Citrus
Tristeza Virus and Brown Citrus Aphid in the Caribbean Basin: Management
Strategies. Lake Alfred, FL, May 15-18, 1995. University of Florida, Institute
of Food and agricultural Sciences.
YASEM, DE ROMERO, M. G. 1985. Verticillium lecanii (Zimm.) Viegas, en afidos de la
provincia de Tucuman, Argentina. CIRPON, Rev. Invest. III: 63-66.
YOKOMI, R. K., AND Y. Q. TANG. 1996. A survey of parasitoids of brown citrus aphid
(Homoptera: Aphididae) in Puerto Rico. Biol. Con. 6: 222-225.
SEY, T. R. GOTTWALD, M. A. ROCHA-PENA, AND C. L. NIBLETT. 1994. Establish-
ment of the brown citrus aphid (Homoptera: Aphididae) in the Caribbean Basin
and transmission of citrus tristeza virus. J. Econ. Entomol. 87: 1078-1085.
ZELENY, J. 1969. A biological and toxicological study of Cycloneda limbifer Casey (Co-
leoptera Coccinellidae). Acta Entomol. Bohemoslav. 66: 333-344.

Florida Entomologist 82(3) September, 1999


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


Cirrospilus floridensis, reared from the citrus leafminer, Phyllocnistis citrella
Staintain, is described and illustrated. Zagrammosoma zebralineatum De Santis and
Pnigalio flavipes (Ashmead) are synonymized with Zagrammosoma multilineatum
(Ashmead) and Pnigalio minio (Walker), respectively.

Key Words: citrus, leafminer, eulophid, biological control, taxonomy


Se describe e ilustra una nueva especie, Cirrospilus floridensis, criada a partir del
minador de los citricos, Phyllocnistis citrella Staintain. Se propone la sinonimia de
Zagrammosoma zebralineatum De Santis con Zagrammosoma multilineatum (Ash-
mead) y de Pnigalio flavipes con (Ashmead) Pnigalio minio (Walker).

The citrus leafminer, Phyllocnistis citrella Staintain, was described from India in
1856. Prior to 1993, this pest was only known from the citrus growing regions of East
Asia, Indonesia, Australia, and East Africa (Heppner, 1993). P. citrella was discovered
in Homestead, Florida in 1993 and quickly spread throughout the citrus growing re-
gions of Florida, Texas and Alabama, the Caribbean, Central America, and Mexico. It
has also been reported throughout the Mediterranean Basin, South Africa and parts
of coastal west Africa (Heppner 1995; Schauffet al. 1998).
Concern over the threat that this species poses to the Florida citrus industry led
to a survey that began in 1994 to determine the native parasitoids that attack citrus
leafminer in Florida (Browning and Pena, 1995). During the course of the study, a new
parasitoid species was discovered and three new synonymies for other parasitoids of
the citrus leafminer were established. One of the synonymies established, Cirrospilus
quadrastriatus (Subba Rao and Ramamani) with Cirrospilus ingenuus Gahan, was
reported by Ujiya and Adachi (1995); Hoy and Nguyen (1997); and Schauff et al.
The following acronyms are used for the institutions where specimens are deposited.
BMNH The Natural History Museum, London, England.
FSCA Florida State Collection of Arthropods, Gainesville, Florida, USA.
UNLP Universidad Nacional de La Plata, La Plata, Argentina.
USNM United States National Museum, Washington, DC, USA

Evans: A New Species of Cirrospilus

Cirrospilus floridensis Evans, new species
(Figs. 1-5)


Female holotype: Coloration (Fig. 1). Fore wing hyaline; body yellow to yellow-or-
ange with the following areas dark brown: central portion of scape, proximal three-
quarters of pedicel, flagellum, compound eyes, base of occiput, 2 diagonal lines on the
pronotum, transverse band on anterior and posterior margins of mesoscutum, no-
taular sutures, apical spot on axillae, axillulae, 2 longitudinal submedial scutellar
grooves, central region of propodeum, apical tarsal segments, valvulae III, base of ovi-
positor, lateral margins of basal tergite and 4-5 transverse bands on metasomal dor-
sum; scutellum of some individuals with a round, dark brown, central spot. Structure.
Body length: 1.7 mm. Head: as wide as mesosoma, malar space 0.6x eye height; OOL
0.4x POL; mandible with 2 large and 4 small teeth; face imbricate. Antennae (Fig. 3)
inserted slightly below lower ocular line. Mesosoma: Pronotum imbricate; midlobe of
mesoscutum length 1.lx width with sharp anterior lateral margins and 6 pairs of pale
setae (Fig. 1), and shallow hexagonal reticulation; axillae advanced and weakly im-
bricate; side lobe rounded with 2 large setae and 5-6 short setae; scutellum surface
somewhat flattened, length 1.2x width with two submedian longitudinal grooves
highlighted with dark pigmentation; placoid sensillae small and round, located about
halfway between anterior (Scl) and posterior (Sc2) pairs of scutellar setae, Scl 0.6x
Sc2 length; dorsellum smooth and semicircular; propodeum short and relatively
smooth except for prominent medial carina and a few short secondary carinae along
the anterior and posterior margins; lateral margin of callus each with approximately
9 long setae; coxae III with shallow imbricate sculpturing, tibial spur of middle leg
1.7x corresponding basitarsus. Fore wing: length 2.4x its maximum width; submar-
ginal vein length 0.4x length of marginal vein with 5 very long and stout setae; mar-
ginal vein with approximately 12 long setae along the anterior margin; stigmal vein
elongate, 2.Ox length of marginal vein; postmarginal vein 0.8x length of stigmal vein
with 3 stout setae along its anterior margin; costal cell with approximately 15 setae;
basal cell with 5-6 setae, speculum broad and closed posteriorly; cubitus slightly
curved; discal hairs dense; marginal fringe short, less than O.1x maximum width of
fore wing. Gaster: length 1.5x width; ovipositor arising at base of tergite II; valvulae
III length 0.2x valvular II length, with numerous elongate setae.


Similar to female in size, structure and coloration, except sensory organ present on
scape (Fig. 2) and transverse bands on gastral tergites II and III broken centrally (Fig. 5).
Morphological variation: Females of C. floridensis vary primarily in the degree of
coloration of the mesosoma. A dark brown, longitudinal line or oval mark is some-
times present on the mesoscutum and scutellum, and the size of the darkened area of
the propodeum may vary usually covering only a small central area but sometimes
covering almost the entire propodeum except for the lateral margins. Specimens of an
unidentified Cirrospilus reared in the Dominican Republic, Bahamas, and Honduras
are nearly identical to C. floridensis in color and structure, but differ in having the ax-
illae completely yellow and the first pair of scutellar setae very short, similar to those
of C. elegantissimus Westwood, a European species. Whether these specimens repre-
sent a distinct species, or a variation of C. floridensis, is not known at this time.

Florida Entomologist 82(3)

September, 1999


/,P A.h'fl -- 'A\

---, 1 *N"/, / "PI
I' '2' '1 j I. N \ .... ; hII

Figs. 1-5. Cirrospilus floridensis-(1,2,4) holotype female; (3,5) male: 1) habitus 2)
antenna 3) antenna 4) gaster venter 5) gaster dorsum. Terms: anelli (A), axilla (Ax),
cerci (Cr), coxa (Cx), dorsellum (Ds), funicle (F), femur (Fm), mesoscutum (Ms), mar-
ginal vein (mv), notaulus (Nt), ovipositor (Ov),pedicel (P), postmarginal vein (pmv),
pronotum (Pn), propodeum (Pp), petiole (Pt), radicle (R), scape (S), scutellum (Sc),
side lobe (Sl), submarginal vein (smv), stigmal vein (sv), tibia (Ti), tarsus (Ta), valvu-
lae III (V3).

Cirrospilus floridensis Evans can be distinguished from other Cirrospilus species
by the dark transverse band along the posterior margin of the midlobe. It is most sim-
ilar to Cirrospilus marilandica (Girault) and can be distinguished by its shorter post-
marginal vein which is less than or subequal to the stigmal vein and by the dark
brown bands along the scutellar grooves. In C. marilandica, the postmarginal vein is
clearly longer than the stigmal vein and the scutellum is entirely yellowish-orange.

Evans: A New Species of Cirrospilus

Boucek (1988) separated the genus Cirrospilus into five groups of species. Cir-
rospilus floridensis is placed in the Cirrospilus s. str. (=Pseudiglyphomyia) group
based upon the propodeum that is at least 1.5x as long as the dorsellum with a dis-
tinct median carina connecting the equally carinate anterior and posterior margins.

Specimens Examined

Holotype female (USNM), U.S.A, Florida, Dade County, Homestead, II 1995,
R. Duncan, reared from Phyllocnistis citrella Staintain (Lepidoptera: Gracillariidae)
on Citrus sp. Paratypes (BMNH,FSCA,USNM): 5 females and 5 males, same collec-
tion data as holotype. Other specimens (FSCA): 4 females, Florida, Dade Co., reared
from Phyllocnistis sp. on mahogany (Swietenia macrophylla King).


This species is named for the state of Florida.

Pnigalio minio (Walker)

Eulophus minio Walker 1847:25; Holotype female, USA: Florida, St. John's Bluff
Pnigalio minio (Walker): Peck 1951: 426.
Elachistus proximus Ashmead 1894: 340.
Pnigalio proximus (Ashmead): Burks 1975: 145 (USNM, examined).
Sympiesis flavipes Ashmead 1886: 133; Holotype female, USA: Florida, Jacksonville,
(USNM Type No. 41330, examined). NEW SYNOMYMY. For complete synon-
ymy of this species see Miller (1970).

Walker (1847) described P. minio from a specimen collected at St. John's Bluff, lo-
cated on the northeast outskirts of Jacksonville, Florida. This is the area from which
the holotype of Sympiesis (=Pnigalio) flavipes Ashmead was collected. I examined the
holotype of Pnigalio flavipes and that of Pnigalio proximus (Ashmead). These species
are very similar in coloration, relative length of antennal segments, and the shape, ve-
nation and ciliation of the fore wing, but differ primarily by the degree of sculpturing
of the mesosoma. In P. flavipes, the mesoscutum, scutellum and dorsellum are
coarsely sculptured, the axillae are alutaceous, the propodeum has a secondary cos-
tula and paramedial carinae are present in the inner quadrant areas. In P. proximus,
the sculpturing of the mesoscutum and scutellum is more shallow, the axillae and
dorsellum are smooth, or nearly so, the propodeum has one costula and the inner
quadrant areas are smooth.
Barrett et al. (1988) reported that P. flavipes individuals varied in color and pro-
podeal sculpturing over the species geographic range. Smaller individuals, particu-
larly the males, tend to have weaker sculpturing, the propodeum may lack the
costula, and the antennal segments are often shorter than those of larger specimens.
Pnigalio proximus individuals are generally smaller (0.9-2.0 mm) than P. flavipes
(1.8-2.7 mm) individuals (Miller, 1970). However, differences in body size may occur
among individuals of the same species reared from different hosts and under different
environmental conditions. I consider the morphological differences that have been
used to distinguish P. minio from P. flavipes to represent intraspecific variation, and
propose that the two species be synonymized. This species, by far, is the most preva-
lent parasitoid reared from the citrus leafminer in Florida.

Florida Entomologist 82(3) September, 1999

Zagrammosoma multilineatum (Ashmead)

Hippocephalus multilineatus Ashmead 1888: 8; Female holotype, USA: Kansas, Riley
Co. (USNM, examined), name preoccupied by Hippocephalus Swainson 1839,
in fishes.
Zagrammosoma zebralineatum De Santis 1983: 9; Holotype female, COLOMBIA,
Pradera, 1982, E. Flores, ex. Leucoptera coffeella. (UNLP, examined) NEW
I examined the holotypes of Zagrammosoma zebralineatum De Santis and
Zagrammosoma multilineatum (Ashmead). The two specimens are nearly identical in
body shape and coloration, but differ in that the gaster ofZ. zebralineatum is slightly
more elongate and has a slightly different banding pattern than that of Z. multilin-
eatum. I consider these differences to represent intraspecific variation, evidenced by
the fact that the gaster of specimens reared from P. citrella from Phyllocnistis citrella
in the Bahamas were similar to Z. multilineatum in the shape (less elongate) but
identical in body coloration to Z. zebralineatum. Similar variation in the coloration of
the gaster of Z. multilineatum has been observed in individuals reared from P. citrella
on citrus in Florida.

Cirrospilus ingenuus Gahan

Cirrospilus ingenuus Gahan 1932: 753, Holotype female, JAVA: Bogor (=Buitenzorg),
Java, ex. Phyllocnistis citrella, A. Vough, (USNM Type No. 43924, examined).
Scotolinx quadristriata Subba Rao and Ramamani 1965:412.
Cirrospilus quadristriatus (Subba Rao and Ramanani).
Cirrospilus ingenuus Gahan: Ujike and Adachi 1995: 96.
The synonymy of Cirrospilus ingenuus Gahan with Cirrospilus quadristriatus
(Subba Rao and Ramamani) was established during the course of this study and con-
firmed by J. LaSalle and Z. Boucek (BMNH). The original description of C. ingenuus
does not mention the four dark, transverse bands on the dorsum of the gaster, which
may have led to some confusion regarding its identity. The synonymy was first re-
ported by Ujike and Adachi (1995), and later by Hoy and Nguyen (1997) and Schauff
et al. (1998).


I thank J. LaSalle and Z. Boucek of the Natural History Museum (BMNH), London
for confirming the synonymy of C. quadrastriata with C. ingenuus; H. Browning,
J. Pena, R. Duncan, M. Pomerinke, P. Stansly, J. Bulloch for the collection of speci-
mens used in the study; I. Jackson, N. Pilcher and H. Davidson for collection and prep-
aration of specimens used in the study; L. De Santis and M. Schauff for the loan of
type specimens Zagrammosoma zebralineatum and Z. multilineatum, respectively;
L. Stange and two anonymous reviewers for reviewing this manuscript; and the
USDA/APHIS/National Biological Control Institute for their financial support. Flor-
ida Agricultural Experiment Station Journal Series No. R-06598.


ASHMEAD, W. H. 1886. Studies on North American Chalcidoidea. American Entomol.
Soc. Trans. 13: 125-135.

Evans: A New Species of Cirrospilus

ASHMEAD, W. H. 1888. Descriptions of some unknown parasitic Hymenoptera in the
collection of the Kansas State Agricultural College, received from Prof. E. A.
Popenoe. Bull. Kansas Agric. Exper. Stat. 3: 1-8.
BARRETT, B. A., J. F. BRUNNER, AND W. J. TURNER. 1988. Variations in color, size, and
thoracic morphology of Pnigalio species (Hymenoptera: Eulophidae) parasitiz-
ing Phyllonorcter elmaeella (Lepidoptera: Gracillariidae) in Utah and Wash-
ington. Ann. Entomol. Soc. America 81(3): 517-521.
BOUCEK, Z. 1988. Australian Chalcidoidea (Hymenoptera). A biosystematic revision
of genera of fourteen families, with a reclassification of species. CAB Interna-
tional, Wallingford, United Kingdom 832 pp.
BROWNING, H. W. AND J. PENA. 1995. Biological control of the citrus leafminer by its
native parasitoids and predators. Citrus Industry, April 1995, pp. 46-48.
BURKS, B. D. 1975. The species of Chalcidoidea described from North America north
of Mexico by Francis Walker (Hymenoptera). Bull. British Mus. (Nat. History),
Entomol. Ser. 32(4): 139-169.
DE SANTIS, L. 1983. Eulofidos (Hymenoptera) de Colombia y Brasil Parasitos de Leu-
coptera coffeella. Rev. Colombiana Entomol. 9: 9-12.
GAHAN, A. 1932. Miscellaneous descriptions and notes on parasitic Hymenoptera.
Ann. Entomol. Soc. America 25: 753.
HEPPNER, J. B. 1993. Citrus leafminer, Phyllocnistis citrella Stainton (Lepidoptera:
Gracillariidae: Phyllocnistinae). Florida Dept. Agric. Cons. Serv., Div. Plant In-
dustry, Entomol. Circ. 359: 1-2.
HEPPNER, J. B. 1995. Citrus leafminer (Lepidoptera: Gracillariidae) on fruit in Flor-
ida. Florida Entomol. 78: 183-186.
HoY, M. AND R. NGUYEN. 1997. Classical biological control of the citrus leafminer
Phyllocnistis citrella Stainton. Trop. Lepid. 8: 1-19.
MILLER, C. D. 1970. The Nearctic species of Pnigalio and Sympiesis (Hymenoptera:
Eulophidae). Mem. Entomol. Soc. Canada, 68: 5-85.
PECK, 0. 1951. In Muesebeck, C. F. W, K. V. Krombein, and H. K. Townes. Hy-
menoptera of America North of Mexico. Synoptic Catalog. Agric. Mono. 2: 410-
SCHAUFF, M. E., L. LASALLE AND G. A. WIJESEKARA. 1998. The genera of chalcid par-
asites (Hymenoptera: Chalcidoidea) of citrus leafminer (Phyllocnistis citrella
Stainton (Lepidoptera: Gracillariidae). J. Nat. Hist.32: 1001-1056.
SUBBA RAO, B. R., AND S. RAMAMANI. 1965. Biology of Cirrospiloideus phyllocnistoides
and description of a new species, Scotolinx quadrastriata as parasites of Phyl-
locnistis citrella. Indian J. Entomol. 27: 408-413.
UJIYE, T., AND I. ADACHI. 1995. Parasitoids of the citrus leafminer, Phyllocnistis cit-
rella Staintain (Lepidoptera: Gracillariidae) in Thailand. Trop. J. App. Ento-
mol. Zool. (Tokyo) 36: 253-255.
WALKER, F. 1847. Characters of undescribed Chalcidites collected in North America
by E. Doubleday, Esq. and now in the British Museum. Ann. Mag. Nat. Hist.
ser. 1, vol 20:23.

Florida Entomologist 82(3) September, 1999


Department of Ecology and Evolutionary Biology, University of Arizona,
Tucson, Arizona, USA 85721


The natural history ofAnidarnes bicolor (Ashmead) Boucek (Hymenoptera: Aga-
onidae: Sycophaginae), a host-specific gallery of the Florida strangling fig (Ficus aurea,
Moraceae), is described.A. bicolor females oviposit through the outside of the globular
fig inflorescence; offspring feed on sterile tissue within galls induced on the inner wall
of fig. Oviposition coincides with entry of the pollinators (Pegoscapus mexicanus; Hy-
menoptera: Agaonidae: Agaoninae) into the fig, and does not interfere with pollina-
tion. Pollinator presence is in fact crucial to A. bicolor success, because unpollinated,
galled figs are aborted by the tree. However, A. bicolor may nevertheless reduce polli-
nator success: maturation of pollinator offspring appears to be negatively affected by
the developingA. bicolor galls. Although the composition of the wasp community as-
sociated with the native Florida figs is relatively well-known, this is one of the first
studies of the natural history of one of the non-pollinator species.

Key Words: gall, wasp, fig, Agaonidae, Ficus aurea, Anidarnes


Se describe la historic natural del agallero Anidarnes bicolor (Ashmead) Boucek
(Hymenoptera: Agaonidae: Sycophaginae), cuyo hospedero especifico es el higo es-
trangulador (Ficus aurea, Moraceae), de Florida. Las hembras deA. bicolor ovipositan
desde afuera dentro de la inflorescencia globular; la progenie se alimenta de tejido es-
teril dentro de las agallas inducidas en la superficie internal del higo. Esta oviposici6n
coincide con la entrada de los polinizadores (Pegoscapus mexicanus; Hymenoptera:
Agaonidae: Agaoninae), y no interfere con la polinizaci6n. La presencia de los polini-
zadores es esencial para el exito deA. bicolor porque higos que no son polinizados son
abortados por el arbol. Sin embargo, es possible que A. bicolor reduzca el 6xito de los
polinizadores: aparentemente el desarrollo de la progenie de los polinizadores es afec-
tado negativamente por el desarrollo de las agallas de A. bicolor. Aunque la composi-
ci6n de la comunidad de avispas asociadas con los arboles de higo nativos de Florida
es relativamente bien conocida, 6ste es uno de los primeros studios sobre la historic
natural de una de las species no polinizadoras.

The mutualism between figs (Ficus spp., Moraceae) and the agaonid wasps that
pollinate them is one of our most spectacular examples of coevolution. Ecological as-
pects of this interaction are now rather well-known (Janzen 1979, Bronstein 1992,
Anstett et al. 1997). Each of the more than 700 fig species can be pollinated by only
one (rarely two) fig wasp species. Pollinators gain access to flowers within the globose
inflorescences, or figs, via a bract-lined entry pore (the ostiole). Once inside, they pol-
linate the flowers, then lay eggs within some of them. Pollinator offspring thus con-
sume many of the developing seeds.

Bronstein:A Galling Fig Wasp in Florida 455

Figs are also exploited by a large community of chalcidoid wasps that develop
within the figs but do not transfer pollen. Little is known about the natural history of
these wasps, their interactions with and impact on the mutualism that they exploit,
and their evolutionary origins (Hamilton 1979, Bronstein 1991, Compton and Hawk-
ins 1992, West and Herre 1994, West et al. 1996, Kerdelhu6 and Rasplus 1996,
Machado et al. 1996). These wasps include gallers, inquilines, cleptoparasites, para-
sitoids of the pollinators, and parasitoids of other non-pollinating wasp species. Up to
32 species have been described in association with a single fig species (J.-Y. Rasplus,
unpublished data). Multiple species often develop side by side in a single fig; in the
New World figs, offspring of non-pollinators can outnumber the pollinator offspring
(Bronstein and Hossaert-McKey 1996). Their sporadic appearance in samples, diverse
feeding niches, often extreme sexual dimorphism, and small size have impeded de-
tailed study. Some genera and many common species remain undescribed.
The interaction between Ficus aurea Nuttall (the Florida strangling fig) and
Pegoscapus mexicanus (Grandi) Wiebes (Agaonidae: Agaoninae) is one of the best un-
derstood fig pollination mutualisms (Frank 1983, 1984, Bronstein and Patel 1992a,
1992b, Bronstein and Hossaert-McKey 1995, 1996, Khadari et al. 1995, Gibernau et
al. 1996, Anstett et al. 1996). The non-pollinating wasp communities associated with
F. aurea and the one other native Florida fig species, a member of the F. citrifolia
P. Miller complex, are also comparatively well-known. Expanding on older work of
Butcher (1964) and Burks (1969), Nadel et al. (1992) identified nine species occupying
F. citrifolia figs and eight occupying F. aurea figs. Five of these wasps were associated
with both fig species, while the others were apparently host-specific. Most of these
species are described under a reworked classification by Boucek (1993), who places
the Florida non-pollinators into the Eurytomidae, Torymidae, and two subfamilies of
the Agaonidae (Otitesellinae and Sycophaginae). The natural history and ecology of
these wasps are unknown, with the exception of one torymid (Physothorax bidentulus
Burks) known to parasitize the galls of a cecidomyiid fly in F. laevigata figs (Roskam
and Nadel 1990).
This paper describes the natural history ofAnidarnes (= Neosycophila) bicolor (Ash-
mead) Boucek (Agaonidae: Sycophaginae), a common non-pollinating wasp associated
exclusively with F. aurea in southern Florida (Nadel et al. 1992). In addition, I examine
the possible impact ofA. bicolor on the fig's interactions with its obligate pollinator.


Ficus aurea is a Caribbean Floristic Region species that reaches its northern limit
in south central Florida (Condit 1969). The study was conducted in a large, natural
F. aurea population located on the campus of The University of Miami in Coral Ga-
bles, Florida. About 50 individuals, ranging in size from 40 to 250 cm d.b.h., are found
at this site growing in hammock remnants along old canals, with the majority in areas
removed from human disturbance. Long-term phenological patterns within this pop-
ulation, as well as their implications for interactions with pollinators and other
wasps, are described by Bronstein and Patel (1992a, 1992b) and Bronstein and Hos-
saert-McKey (1995, 1996).
Anidarnes bicolor is one of the largest nonpollinating wasps found in F. aurea,
reaching about 2.0 mm in length (about 2-3 times larger than the pollinator). Females
and males are both winged and resemble each other closely, but can easily be distin-
guished by the female's ovipositor, about half the length of the gaster. Boucek (1993)
gives a description of this species and provides identification keys; his Figure 41 illus-
trates a female individual.

Florida Entomologist 82(3) September, 1999

The distribution and abundance ofA. bicolor were determined from three sets of
figs. The first set consisted of 748 figs sampled during and soon after pollination (i.e.,
B and early C phase; Galil and Eisikowitch 1968), from trees flowering between Jan-
uary and mid-August 1992. On 15 dates, at about 10-day intervals, 34 trees in the
study population were examined to see if they bore newly pollinated figs. When such
figs were located, a sample of 10 was removed from the tree. In the laboratory, each
fig was split open and examined under the dissecting microscope. A. bicolor galls,
which are already quite obvious at this stage, were counted. We also recorded the
number of dead pollinators within each fig to look for associations between pollination
intensity and A. bicolor galls. Pollinators typically although not invariably become
trapped in the figs they enter, and die within 24 hours (Gibernau et al. 1996, Bron-
stein and Hossaert-McKey 1996). It was noted if each pollinator had died within the
fig cavity or within the ostiole (entry pore), where it might have become trapped either
during entry to or exit from the cavity.
The second set of figs was collected when seeds and wasp offspring were mature (D
phase). On each census date between February and December 1991, 20 trees in the
study population were each examined for the presence of figs in which developing
wasps had reached adulthood and were ready to depart. When detected, a haphazard
sample (usually 5) was removed. In the laboratory, each fig was split open, placed into
a vial, and covered tightly with fine mesh. Each vial was then placed in the dark for 24
hours to allow wasp escape, then filled with a preservative and capped. Seeds and all
wasps were later counted, and wasps were identified to species or morphospecies and
to sex. Wasps still trapped within galls were removed and identified when possible.
A third set of 313 figs was also collected when seeds and wasps were mature. These
figs had been bagged several days prior to receptivity; when receptive, they had polli-
nators but no other insects I.. 1 .1.. :, l..introduced to them (detailed methods are pro-
vided by Khadari et al. 1995 and Anstett et al. 1996). These experiments were
conducted on 13 trees between 1991 and 1995. Wasps were reared from these figs, pre-
served, and later identified, using the methods described above.


General Natural History

Ovipositing A. bicolor were never found within the fig cavity, and presumably
never enter it. Rather, oviposition occurs through the fig wall, as in the other non-pol-
linating wasps associated with F. aurea (unpublished data). Each offspring matures
inside a large, thick-walled gall. When these galls are numerous, they can occupy
nearly the entire volume of the fig. Superficially, these galls appear to be greatly swol-
len female florets similar to, but much larger than, those occupied by offspring of the
pollinators and other non-pollinator species. However, closer inspection reveals them
to be rotund invaginations of the inner wall of the fig. Hence,A. bicolor is apparently
a gallery that is dietarily independent of the seeds or wasps developing simultaneously
in the fig. This type of gall appears to be generally rare among fig wasps, and has not
previously been reported in the New World. It has been observed most commonly in
species of Epichrysomallinae (Agaonidae) associated with African figs (Compton and
van Noort 1992; J.-Y. Rasplus and C. Kerdelhue, unpublished data). The eurytomid
genus Sycophila has also been reared from such galls in Africa, but is apparently a
parasite of the epichrysomallines (J.-Y. Rasplus, unpublished data).
When mature,A. bicolor individuals chew out of their galls and enter the fig cavity.
They depart the fig through the exit tunnel constructed by the males of the pollinator

Bronstein:A Galling Fig Wasp in Florida 457

species. It is not known if or how A. bicolor is able to escape from figs in which all
P. mexicanus males have died prior to maturity (17% of all figs; unpublished data). It
is very likely thatA. bicolor matings occur outside of the fig, as is usual for the subset
of fig nonpollinators in which both males and females are winged (Hamilton 1979).

Phenology ofA. bicolor

Of 748 figs sampled within a week of pollinator arrival, 225 (30.1%) already con-
tained the characteristic galls of A. bicolor. Three lines of evidence suggest that A. bi-
color attack can precede pollination. First, 17.8% of the galled, immature figs sampled
(n = 225) did not contain trapped pollinators. Second, manyA. bicolor galls were al-
ready quite large at the time of sampling, suggesting that they had been developing
for some days. Third, 12.7% of the figs bagged as much as a week prior to pollinator
introduction experiments (n = 313) contained A. bicolor individuals when mature.
This species made up over 80% of the non-pollinator individuals reared in these ex-
periments. It is unlikely that A. bicolor would have regularly penetrated the bagging
material while two significantly smaller and more common species, Idarnes carme
Walker and Heterandrium sp., almost never did. Instead, these observations suggest
thatA. bicolor had attacked these figs prior to bagging, i.e., in A phase, well before the
figs were receptive to pollinators. A number of other fig gallers (Kerdelhue and Ras-
plus 1996), including one epichrysomalline that galls the fig wall (S. G. Compton, un-
published data), are known similarly to oviposit prior to pollination.
AlthoughA. bicolor offspring begin development earlier than the pollinator offspring,
they take as long or somewhat longer to mature. Figs cut open 1-2 days before P. mexi-
canus have completed maturation never contain mature A. bicolor individuals ready to
depart (unpublished data). In fact, in the galled figs examined when P. mexicanus were
departing, 23% ofA. bicolor individuals had not yet left their galls. Males emerged ear-
lier than females: 27% ofA. bicolor females (n = 133) were still within the galls when pol-
linator offspring were mature, compared to only 16.5% of the males (n = 89).
Over 99% of mature galled figs contained seeds and pollinator offspring, implying
that A. bicolor only matures in pollinated figs. While immature figs sometimes con-
tained galls but no trapped pollinators (see above), it is probable that at the time of
sampling, pollinators had either not yet entered these figs or else had entered them
and then departed. Hence, although certain non-pollinator species can induce the tree
to retain rather than abscise unpollinated figs (Bronstein 1991), this does not appear
to be the case for A. bicolor.

Frequency of Attack

Figs attacked byA. bicolor during 1992 (30.1% of sampled figs) contained 5.4 + 4.4
galls (mean + 1 SD, range 1-23, n = 225 figs). Attack was strongly skewed towards few
galls per fig (Fig. 1). The number of galls per fig varied significantly with both tree and
with month of sampling, as well as with the interaction between month and tree (Ta-
ble 1). There were no obvious seasonal patterns of attack in 1992 (Fig. 2), at least up
to mid-August, when winds from Hurricane Andrew blew all figs off all study trees.
Anidarnes bicolor was found once again occupying figs by the second week of January
1993, at which point local pollinator populations had successfully reestablished
(Bronstein and Hossaert-McKey 1995).
Anidarnes bicolor was the third most common non-pollinating wasp in mature F.
aurea figs sampled during 1991, after Idarnes carme (found in 77.8% of the figs) and
a Heterandrium species (39.5% of figs).An average of 4.30 + 3.38 individuals matured

Florida Entomologist 82(3) September, 1999


1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23

Galls per Fig

Fig. 1. Frequency distribution of Anidarnes bicolor galls in 225 immature, newly
galled figs sampled from January to August 1992.

in the 27.1% of sampled figs (n = 52) that containedA. bicolor. The range was large (1-
19 individuals) and skewed towards low numbers, as in the 1991 sample of immature
figs. The number of A. bicolor (which do not mature inside fig florets) was not corre-
lated with the number of florets per fig (Spearman rank correlation, r = -0.205, P =
0.116, n = 52), unlike the numbers of seeds, pollinator offspring, and other non-polli-
nator wasps (Bronstein and Hossaert-McKey 1996, Anstett et al. 1996).
Fifty percent of the mature A. bicolor-infested figs contained either exclusively
males or exclusively females; it was not uncommon to find four or more individuals of
one sex and none of the other. The sex ratio was female-biased, with males making up
40% of all mature individuals (n = 222 wasps).

Impact on the Pollination Mutualism

The presence of the largeA. bicolor galls crowding the cavity apparently did not re-
strict pollinator entry into figs. Galled figs contained no fewer trapped pollinators
than did ungalled ones, although there was a trend in this direction (Mann-Whitney
test, U = 54305, P = 0.082, df = 746). However, the number of pollinator offspring ma-
turing in galled figs decreased as the number of galls increased (Spearman rank cor-
relation, r = -0.353, P = 0.039, n = 32).
There are four possible mechanisms for this negative effect ofA. bicolor on pollina-
tor offspring numbers. First, once pollinators manage to enter the fig, the galls may
impede their access to the flowers. This is unlikely. Pollinators found trapped within
figs were equally likely to have successfully reached the fig cavity in galled as com-
pared to ungalled figs (U = 41712, P = 0.732, df = 746). Furthermore, seed set in galled
figs was not reduced compared to ungalled figs (r = -0.119, P = 0.396, n = 52), implying
that pollinators conducted their usual pollination behaviors once there. Second,

Bronstein:A Galling Fig Wasp in Florida


Sum of Mean
Source df squares square F-Value P-Value

Tree 28 1485.237 53.044 7.494 0.0001
Month 7 310.098 44.300 6.259 0.0001
Tree* month 35 2009.997 57.428 8.113 0.0001
Residual 677 4791.975 7.078

P. mexicanus might pollinate normally in galled figs, but lay fewer eggs there. Third,
they might both pollinate and oviposit normally in galled figs, but the expandingA. bi-
color galls might then crush some of the delicate larvae as they develop. Many fig ova-
ries were in fact obviously flattened by the expansion of these large galls. However,
maturation of other non-pollinator species was not similarly depressed by galling (r. =
0.106, P = 0.600, n = 32), weakening this hypothesis. Finally, it is possible that devel-
opingA. bicolor draw away fig resources that might otherwise be put into P. mexicanus
development. Further studies are needed to evaluate these hypotheses thoroughly.


I I I i I I I I
Jan Feb Mar Apr May Jun Jul Aug

Month Figs Matured (1992)
Fig. 2. Attack of Anidarnes bicolor by month, from 748 figs sampled during 1992
(means + 1 SD shown). Although differences among months were highly significant
(see Table 1), there was no clear seasonality of attack.

FF1 rH


Florida Entomologist 82(3) September, 1999

Hence, despite the fact that A. bicolor depends on P. mexicanus both to assure fig re-
tention to maturity and to create a passage out of the fig, it has a distinctly negative ef-
fect on this species. This would appear to be an unusually clear case of a fig inhabitant
that imposes a cost to the pollination mutualism, despite feeding on sterile tissue rather
than on one of the mutualists (West and Herre 1994, Kerdelhue and Rasplus 1996). Yet,
F. aurea has no apparent adaptations to preventA. bicolor attack. Bronstein (1991) pro-
posed three alternative hypotheses to explain a similar tolerance of non-pollinators in
a Costa Rican fig species. The Commensalism Hypothesis, which states that non-polli-
nators are tolerated by the fig because they inflict no costs, may be relevant in the
present case. AlthoughA. bicolor attack does reduce reproductive success of P. mexica-
nus, the only possible pollen vector of F. aurea, the cost to F aurea itself might still be
negligible. At least in this population, almost every F aurea fig is entered by one or more
pollinators. Seed set is highest when a single pollinator enters (unpublished data), but
in nature, up to 24 can be found crowding into a single fig (Bronstein and Hossaert-
McKey 1996). Hence, if somewhat fewer P. mexicanus departed a tree, that tree's suc-
cess as a pollen donor would not necessarily be depressed. The Coevolutionary Race Hy-
pothesis (Bronstein 1991), which states that non-pollinators evolve adaptations to
exploit figs faster than figs evolve counteradaptations to them, may also be relevant,
however. F. aurea shows no obvious adaptations to avoid attack from any member of its
large community of non-pollinators, some of which are likely to impose even higher fit-
ness costs on it than does A. bicolor. Understanding F. aurea's tolerance of A. bicolor
may well require studies of the broader community of these enigmatic wasps.


Martine Hossaert-McKey, Doyle McKey, Cathy Dull, and Aviva Patel provided in-
valuable assistance with fig collections during 1991 and 1992. I also thank Cecelia
Smith, Terry Brncic and Kristen Iker for help with fig dissections. Hannah Nadel and
Jean-Yves Rasplus identified the wasps. Special thanks to Jean-Yves Rasplus, Carole
Kerdelhue, Steve Compton, Simon van Noort, and George Weiblen for sharing unpub-
lished observations, and to Sandra Perez, Jennifer Weeks, Ted Lee, Margrit McIntosh,
Jean-Yves Rasplus, and Carole Kerdelhue for helpful comments on the manuscript.
This work was supported by grants from the National Science Foundation (BSR/ROW
9007492 and BSR 9407615) and NATO (CRG 921294).


ANSTETT, M.-C., J. L. BRONSTEIN, AND M. HOSSAERT-MCKEY. 1996. Resource alloca-
tion: a conflict in the fig/fig wasp mutualism? J. Evol. Biol. 9: 417-426.
ANSTETT, M.-C., M. HOSSAERT-MCKEY, AND F. KJELLBERG. 1997. Figs and fig pollina-
tors: evolutionary conflicts in a coevolved mutualism. Trends Ecol. Evol. 12: 94-
BOUCEK, Z. 1993. The genera of chalcidoid wasps from Ficus fruit in the New World.
J. Nat. Hist. 27: 173-217.
BRONSTEIN, J. L. 1991. The nonpollinating wasp fauna of Ficus pertusa: exploitation
of a mutualism? Oikos 61: 175-186.
BRONSTEIN, J. L. 1992. Seed predators as mutualists: ecology and evolution of the fig/
pollinator interaction. In Insect-Plant Interactions. Volume IV, edited by
E. Bernays. Boca Raton: CRC Press.
BRONSTEIN, J. L., AND A. PATEL. 1992a. Causes and consequences of within-tree phe-
nological patterns in the Florida strangling fig, Ficus aurea (Moraceae). Am. J.
Bot. 79: 41-48.

Bronstein:A Galling Fig Wasp in Florida 461

BRONSTEIN, J. L., AND A. PATEL. 1992b. Temperature-sensitive development: conse-
quences for local persistence of two subtropical fig wasp species. Am. Midl. Nat.
128: 397-403.
BRONSTEIN, J. L., AND M. HOSSAERT-MCKEY. 1995. Hurricane Andrew and a Florida
fig pollination mutualism: resilience of an obligate mutualism. Biotropica 27:
BRONSTEIN, J. L., AND M. HOSSAERT-MCKEY. 1996. Variation in reproductive success
within a subtropical fig/pollinator mutualism. J. Biogeogr. 23: 433-446.
BURKS, B. D. 1969. Redefinition of two genera of chalcidoids from figs, with new Flor-
ida species (Hymenoptera). Florida Ent. 52: 115-122.
BUTCHER, F. G. 1964. The Florida fig wasp, Secundeisenia mexicana (Ashm.), and
some of its hymenopterous symbionts. Florida Ent. 47: 235-238.
COMPTON, S. G., AND S. VAN NOORT. 1992. Southern African fig wasps (Hymenoptera:
Chalcidoidea): resource utilization and host relationships. Proc. K. Ned. Akad.
Wet. C 95: 423-435.
COMPTON, S. G., AND B. A. HAWKINS. 1992. Determinants of species richness in south-
ern African fig wasp assemblages. Oecologia 91: 68-74.
CONDIT, I. J. 1969. Ficus, The Exotic Species. Berkeley: California Division of Agricul-
tural Sciences.
FRANK, S. A. 1983. Theoretical and empirical studies of sex ratios, mainly in fig wasps.
M.S. thesis, University of Florida, Gainesville.
FRANK, S. A. 1984. The behavior and morphology of the fig wasps Pegoscapus assuetus
and P. jimenezi: descriptions and suggested behavioral characters for phyloge-
netic studies. Psyche 91: 289-308.
GALIL, J., AND D. EISIKOWITCH. 1968. On the pollination ecology of Ficus sycomorus in
East Africa. Ecology 49:259-269.
sequences of protecting flowers in a fig: a one-way trip for pollinators? J. Bio-
geogr. 23: 425-432.
HAMILTON, W. D. 1979. Wingless and fighting males in fig wasps and other insects. In
Sexual Selection and Reproductive Competition in Insects, edited by M. S.
Blum and N. A. Blum. London: Academic Press.
JANZEN, D. H. 1979. How to be a fig. Ann. Rev. Ecol. Syst. 10: 13-51.
KERDELHUE, C., AND J.-Y. RASPLUS. 1996. Non-pollinating Afrotropical fig wasps af-
fect the fig-pollinator mutualism in Ficus within the subgenus Sycomorus. Oi-
kos 75: 3-14.
MCKEY. 1995. When figs wait for pollinators: the length of fig receptivity. Am.
J. Bot. 82: 992-999.
phylogenies of fig pollinating and non-pollinating wasps and the implications
for the origin and evolution of the fig-fig wasp mutualism. J. Biogeogr. 23: 521-
NADEL, H., J. H. FRANK, AND JR. R. J. KNIGHT. 1992. Escapees and accomplices: the
naturalization of exotic Ficus and their associated faunas in Florida. Florida
Ent. 75: 29-38.
ROSKAM, J. C., AND H. NADEL. 1990. Redescription and immature stages of Ficiomyia
perarticulata (Diptera: Cecidomyiidae), a gall midge inhabiting syconia of Fi-
cus citrifolia. Proc. Ent. Soc. Washington 92: 778-792.
WEST, S. A., AND E. A. HERRE. 1994. The ecology of the New World fig-parasitizing
wasps Idarnes and implications for the evolution of the fig-pollinator mutual-
ism. Proc. Roy. Soc. Lond. Ser. B 258: 67-72.
WEST, S. A., E. A. HERRE, D. M. WINDSOR, AND P. R. S. GREEN. 1996. The ecology and
evolution of the New World non-pollinating fig wasp communities. J. Biogeogr.
23: 447-458.

Florida Entomologist 82(3) September, 1999


Department of Zoology, University of Oklahoma, Norman, Oklahoma 73019

Current address: Evergreen State College, Lab I, Olympia, Washington 98505


I investigated the effects of a top predator, Mecistogaster spp. (Odonata: Pseu-
dostigmatidae), on survivorship of the grazer Culex mollis (Diptera: Culicidae) and
decomposition rates of leaf litter in treehole microcosms. In a factorial experiment us-
ing 200 ml cups, less litter mass remained when grazers (51%) and grazers plus pred-
ators (51%) were present, than without grazers (57%). Predators reduced mosquito
survival, but had no indirect effect on litter decomposition rate. Mosquito larvae fa-
cilitated decomposition of litter and may have become food limited.

Key Words: damselflies, detritus, indirect effects, mosquitoes, predation


Se investigaron los efectos del depredador Mecistogaster spp. (Odonata: Pseudos-
tigmatidae) en la supervivencia del pacero Culex mollis (Diptera: Culicidae) y las ta-
sas de descomposici6n de hojarasca en el microcosmos de huecos de arboles. El
experiment se efectu6 en vasos de 200 ml, utilizando un diseno factorial. La cantidad
de hojarasca fue menor (51%) cuando los paceros y los grazer mas depredadores (51%)
estuvieron presents, que cuando no hubo paceros. La presencia de paceros redujo la
sobrevivencia de mosquitos, pero no tuvo un efecto indirecto sobre la tasa de descom-
posici6n de la hojarasca. Las larvas de mosquito facilitaron la descomposici6n de la
hojarasca y probablemente estuvieron limitados de alimento.

Trophic cascades (i.e., indirect effects of predators on productivity at lower trophic
levels) occur in large aquatic systems (e.g., Carpenter et al. 1985, Power 1990) and
some terrestrial systems (e.g., Spiller & Schoener 1994, Moran & Hurd 1998). Fewer
detailed investigations of trophic interactions have been conducted in small aquatic
settings (but see e.g., Addicott 1974, Bradshaw & Holzapfel 1983, Cochran-Stafira &
von Ende 1998, Srivastava & Lawton 1998).
Water-filled treeholes are formed when rain collects in branch crotches and other
wood-lined cavities, and are common aquatic habitats in temperate hardwood and
tropical forests (e.g., Kitching 1971). Unlike large aquatic systems such as lakes and
streams, treeholes are experimentally tractable and can be reproduced as laboratory
microcosms without compromising the scale of the natural system.
Detritus, especially leaf litter, is the base of food webs in most treeholes (Kitching
1971). Leaf litter in treeholes is primarily consumed by bacteria and fungi (Fish &
Carpenter 1982), and mosquito larvae are typically the most common grazers on this

Yanoviak: Tropical Treehole Microcosms

microfauna. In the lowland moist forest of Barro Colorado Island (BCI), Panama, lar-
vae ofAedes, Culex and Haemagogus mosquitoes are the principal prey of top preda-
tors in treeholes, which include odonate larvae, Dendrobates tadpoles, and
Toxorhynchites mosquito larvae (Fincke et al. 1997, Fincke 1999, Yanoviak 1999). Lar-
vae of pseudostigmatid damselflies are the most common top predators in BCI tree-
holes (Fincke 1992a, 1999).
Carpenter (1982) and Fish & Carpenter (1982) showed that consumption of decom-
poser microbes byAedes triseriatus (Say) larvae enhances litter degradation rates in
treehole microcosms. Similar facilitation effects of grazers occur in other systems (e.g.,
Hargrave 1970, Barsdate et al. 1974), and several mechanisms have been suggested:
1) grazers fragment or directly consume leaf material in addition to the decomposer
fauna (e.g., Fenchel 1970, Carpenter 1982); 2) grazing stimulates decomposer metab-
olism (e.g, Lopez et al. 1977); 3) grazing alters the composition of decomposer taxa (e.g,
Harrison 1977, Hanlon & Anderson 1979); and 4) inorganic compounds excreted by
grazers enhance growth of decomposers (e.g., Durbin et al. 1979).
Tropical treeholes support diverse predator assemblages (Kitching 1990, Fincke
1999) which can reduce mosquito survivorship (Fincke et al. 1997, Nannini & Juliano
1998), and influence life history characteristics such as development time and adult
mass (Lounibos et al. 1993, Grill & Juliano 1996). These strong predator effects, in
combination with mosquito effects on litter decomposition, suggest that treehole com-
munities are prone to trophic cascades.
The purpose of this study was to determine the independent and interactive effects
of a top predator and grazers on litter decomposition rates in treehole microcosms on
BCI. I hypothesized that the consumption rate of leaf litter by decomposers would dif-
fer in simple food webs of different length (i.e., in the presence or absence of grazing
mosquitoes, and with and without a predacious odonate). I tested the following pre-
dictions: 1) grazing by larvae of the treehole mosquito Culex mollis Dyar and Knab en-
hances leaf litter degradation; 2) predation by odonates reduces the survivorship (=
successful pupation) and time to pupation of mosquitoes; and 3) the presence of an
odonate indirectly reduces the degradation rate of leaf litter.


Experiments were conducted in the laboratory building of the Smithsonian Trop-
ical Research Institute on BCI during January and February 1998. Treehole micro-
cosms consisted of plastic cups (8.5 cm height x 5 cm diameter) containing 125 ml
filtered rain water. Treehole water was collected and pooled from 5 natural holes on
BCI, and 3 ml was added to each cup as a microfauna inoculum. A strip of balsa wood
(1 cm x 10 cm x 0.15 cm) served as a predator perch site and was included in all treat-
ments. I added 200 + 0.5 mg dried (> 10 d in air, then 48 h at 70C) Platypodium ele-
gans Vogel leaf litter as a nutrient source. I used leaflets of P. elegans because they are
small and are often found in natural treeholes (pers. obs.). Most of the litter in each
cup consisted of whole leaflets (each 15 cm2) collected from recent branch falls, but
every cup also contained leaf fragments (each 1.0 cm2) to standardize the initial lit-
ter mass. I allowed the cup contents to stand for 48 hours prior to the start of the ex-
periment to permit growth of bacteria and fungi. I maintained all cups at 23C and
periodically added rain water to compensate for evaporation.
The experimental design was a 2 x 2 factorial replicated six times, with presence/
absence of predators and grazers as main effects. Treatment 1 contained odonates
(predators) only, Treatment 2 contained mosquito larvae (grazers) only, and Treat-
ment 3 contained both. Control cups lacked grazers and predators. Each cup in Treat-
ments 1 and 3 received one medium size (15-18 mm) larva of the pseudostigmatid

Florida Entomologist 82(3) September, 1999

damselfly Mecistogaster sp. The two species of Mecistogaster on BCI, M. linearis Fab.
and M. ornata Rambur, are difficult to distinguish as larvae and were not differenti-
ated in this experiment. Both species have similar foraging behavior and maximum
growth rates (Fincke 1992b). I obtained odonate larvae from different natural tree-
holes and starved them for > 24 h before the start of the experiment. I biased the pred-
ators used in Treatment 3 toward smaller initial sizes (15-17 mm) because maximal
growth and foraging rates of Mecistogaster tend to occur around that body length
(Fincke 1992b, Fincke et al. 1997).
I used larvae of C. mollis as grazers in the experiment. This species is relatively
common in natural treeholes on BCI (Fincke et al. 1997, Yanoviak 1999), and its egg
rafts are easy to recognize and collect. Treatments 2 and 3 each received 25 neonate
(< 12 h post hatch) C. mollis larvae on experiment days 0, 2,4, and 6 (total = 100 larvae
per cup). Mosquitoes added to a given treatment on a given day were reared from at
least two different field-collected egg rafts (n > 16 rafts). Leaf litter and odonate den-
sities were within the ranges observed for natural treeholes on BCI (pers. obs.). Al-
though total mosquito density was biased toward the maximum found in small (< 500
ml) holes, the gradual addition of neonates over several days approximated mosquito
recruitment patterns in natural and artificial holes augmented with fresh litter. Mos-
quitoes consumed by odonates in Treatment 3 were not replaced with new individuals.
I operationally defined mosquito survivorship as successful transformation to the
pupa stage. I collected pupae from the cups daily and stored them in a drying oven
(70C) until the end of the experiment, when all pupae accumulated within a replica-
tion were weighed to the nearest 0.1 mg. Pupae were not separated by sex. Fresh
weights (_ 0.1 mg) and body lengths (_ 0.25 mm, excluding caudal lamellae) of Mecis-
togaster larvae were recorded at the start and end of the experiment. I blotted the odo-
nate larvae dry with a cotton cloth before all measurements. Leaf litter remaining in
the cups at the end of the experiment was collected with forceps, rinsed in distilled
water, dried (70C, 48 h), and weighed to the nearest 0.1 mg. I terminated the exper-
iment 31 days after the first mosquito input, when all larvae in Treatment 3 (predator
plus grazers) pupated or were consumed.
Proportional data were arcsine-square root transformed (Sokal & Rohlf 1981) and
homogeneity of variances was determined with F-tests prior to analyses. Odonate
fresh weight and body length were highly correlated (N = 12, r = 0.99, P < 0.001), so
only the change in weight was analyzed. One odonate larva in Treatment 3 molted
during the experiment and was excluded from predator size comparisons. Means for
the total number of C. mollis pupae, individual pupal mass (obtained by division), and
days to appearance of the first pupa were compared across grazer and grazer plus
predator treatments by ANOVA with Bonferroni adjustment for multiplicity.


The amount of leaf litter remaining at the end of the experiment was greater in the
absence of mosquitoes, but there was no significant predator effect or grazer*predator
interaction (Fig. 1). The average number of mosquitoes that successfully pupated was
greater in the absence of predators (Table 1), but survival was low in all treatments;
no more than 20 of the 100 mosquitoes in a replicate survived to pupation. Average in-
dividual pupal mass and the mean number of days to appearance of the first pupa did
not differ between the two treatments containing mosquitoes (Table 1).
The foraging behavior of mosquito larvae surviving to the last week of the experi-
ment differed from larvae that successfully pupated in the second week. Older larvae
spent more time actively browsing leaf material and browsed a larger leaf area. These
behavioral shifts were not anticipated and therefore not quantified.

Yanoviak: Tropical Treehole Microcosms

60 Tmt df SS F
Pr 1 1.77 0.47
tM Gr 1 57.2 15.22**
Pr*Gr 1 0.43 0.11
i 55 Error 20 75.2

I 50-

45 ---

Ctrl Pr Gr Gr+Pr

Fig. 1. Mean (+ SE) percent of litter mass remaining in each treatment (Tmt) at
the end of the experiment. Ctrl = control, Pr = predator only, Gr = grazers only, Gr +
Pr = grazers plus predators. ** = P < 0.05.

The mass of odonates in Treatment 1 (predators only), did not differ between the
start and end of the experiment (paired t = 0.074, df = 5, P = 0.94). However, odonate
mass in Treatment 3 (grazers plus predators) increased over the 31 days (paired t =
5.98, df = 4, P = 0.001; Fig. 2).
Treatments containing mosquitoes had comparatively clear water and no appar-
ent microbial biofilm present on the leaf litter. In contrast, the water in the control
cups and predator-only treatment was cloudy (presumed due to abundant bacteria)
throughout the experiment and the leaves were covered with a conspicuous accumu-
lation of decomposers (e.g., fungal strands), especially during the latter three weeks.

The combination of species used in this study did not result in a true 3 trophic level
system due to selective feeding by the odonates (= predator inefficiency; Fretwell 1977,


Grazers Grazers
Variable only + Predator F,o10 P

Days to first pupa 9.17 + 1.72 9.33 + 0.82 0.05 0.83
number of pupae 10.5 + 2.95 3.33 + 1.75 26.19 <0.001
Individual pupa
mass (mg) 0.323 + 0.068 0.251 + 0.097 2.18 0.17

Florida Entomologist 82(3) September, 1999


-M 0.10 -


0.00 -- m

Pr Gr+Pr
Fig. 2. Mean (+ SE) fresh mass of odonates at the start (hatched bars) and end
(open bars) of the experiment in predator only (Pr) and grazer plus predator (Gr + Pr)
treatments. N = 6 for Pr and N = 5 for Pr + Gr. = P < 0.05.

1987). Mecistogaster depressed survivorship of C. mollis, but these odonates (and the
other common pseudostigmatid on BCI, Megaloprepus coerulatus [Drury]) are size-se-
lective predators, preferentially feeding on the largest prey available (Yanoviak 1999).
This size-based refugium allows smaller (i.e., early instar) mosquito larvae to graze on
and limit the growth of decomposers despite the presence of a predator. The lack of a
significant predator*grazer interaction and the similarity in remaining litter biomass
between the two treatments containing mosquitoes support this conclusion.
Although no trophic cascade effect occurred, my results are similar to those of
Power (1990), who found that one of the top predators in her system did not control
the dominant grazers due to the presence of refugia (Power et al. 1992). Other preda-
tors occurring in BCI treeholes, such as larvae of the aeshnid dragonflies Gynacantha
membranalis Karsch and Triacanthagyna dentata Geijskes feed more voraciously and
less selectively than pseudostigmatids (Yanoviak 1999). A duplicate experiment con-
ducted with one of these species may yield a cascade effect.
Results of this study support the prediction that grazing by mosquito larvae facil-
itates the degradation of leaf litter in treehole microcosms. Although direct consump-
tion of leaf material (e.g., Carpenter 1982) is the simplest explanation for this effect,
there was no evidence (e.g., leaf fragmentation or abrasion) that C. mollis larvae con-
sumed a measurable portion of the litter in this experiment. Alternative explanations
for facilitation of litter decay cannot be addressed with data gathered in this study,
but there is circumstantial evidence that grazing by mosquito larvae altered the com-
position of decomposer microbes. Specifically, the minimal fungal development in
treatments containing C. mollis suggests that grazing controlled the growth of fungi
that were abundant in non-grazer treatments.
Because odonates reduce mosquito abundance in artificial and natural tree holes
(Fincke et al. 1997), the negative effect of Mecistogaster on C. mollis survivorship was

Yanoviak: Tropical Treehole Microcosms

expected. Odonates occasionally eat mosquito pupae (Yanoviak 1999), so survivorship
operationally defined as adult emergence rather than successful pupation may have
produced an even stronger result.
I attribute low mosquito survivorship in the grazer-only treatment to nutrient lim-
itation, which commonly affects the abundance and life history characteristics of tree-
hole mosquitoes (e.g., Fish & Carpenter 1982, Carpenter 1983, Hard et al. 1989,
Leonard & Juliano 1995). Evidence for food limitation in this study includes slower
appearance of first pupae (- 9 d compared to about 5 d for C. mollis reared on ad libi-
tum yeast or fish food in the same laboratory), and the relative clarity of the water in
treatments containing mosquitoes. In addition, differences in the foraging behavior of
mosquito larvae surviving to the last week of the experiment compared to those that
successfully pupated in the second week may be a reflection of starvation stress.
Predators can reduce mosquito abundance in treeholes (Fincke et al. 1997) and al-
ter community parameters such as species richness (Yanoviak 1999), but "top-down"
control of energy transfer among lower trophic levels may be relatively weak in this
system. Although preliminary, my results suggest that this is true in the presence of
size-selective predators.


Comments from 0. M. Fincke, K. Gido, V. H. Hutchison, M. E. Kaspari, L. P. Louni-
bos and two anonymous reviewers improved the manuscript. X. L. Longarela assisted
with the Spanish translation. This project was supported by funds from the IIE/Ful-
bright Foundation, the Explorers Club, Sigma Xi, and the Univ. of Oklahoma.


ADDICOTT, J. F. 1974. Predation and prey community structure: an experimental
study of the effect of mosquito larvae on the protozoan communities of pitcher
plants. Ecology 55: 475-492.
BARSDATE, R. J., R. T. PRENTKI, AND T. FENCHEL. 1974. Phosphorus cycle of model ec-
osystems: significance for decomposer food chains and effect of bacterial graz-
ers. Oikos 25: 239-251.
BRADSHAW, W. E., AND C. M. HOLZAPFEL. 1983. Predator-mediated, non-equilibrium
coexistence of tree-hole mosquitoes in southeastern North America. Oecologia
57: 239-256.
CARPENTER, S. R. 1982. Comparisons of equations for decay of leaf litter in tree-hole
ecosystems. Oikos 39: 17-22.
CARPENTER, S. R. 1983. Resource limitation of larval treehole mosquitoes subsisting
on beech detritus. Ecology 64: 219-223.
CARPENTER, S. R., J. F. KITCHELL, AND J. R. HODGSON. 1985. Cascading trophic inter-
actions and lake productivity. Bioscience 35: 634-639.
COCHRAN-STAFIRA, D. L., AND C. N. VON ENDE. 1998. Integrating bacteria into food
webs: studies with Sarracenia purpurea inquilines. Ecology 79: 880-898.
DURBIN, A. G., S. W. NIXON, AND C. A. OVIATT. 1979. Effects of the spawning migra-
tion of the alewife, Alosa pseudoharengus, on freshwater ecosystems. Ecology
60: 8-17.
FENCHEL, T. 1970. Studies on the decomposition of organic detritus derived from the
turtle grass Thalassia testudinum. Limnol. Oceanogr. 15: 14-20.
FINCKE, 0. M. 1992a. Behavioural ecology of the giant damselflies of Barro Colorado
Island, Panama. p. 102-113 in D. Quintero and A. Aiello [eds.]. Insects of Pan-
ama and Mesoamerica. Oxford Univ. Press, Oxford.
FINCKE, 0. M. 1992b. Interspecific competition for tree holes: consequences for mating
systems and coexistence in neotropical damselflies. Am. Nat. 139: 80-101.

Florida Entomologist 82(3) September, 1999

FINCKE, 0. M. 1999. Organization of predator assemblages in Neotropical tree holes:
effects of abiotic factors and priority. Ecol. Entomol. 24: 13-23.
FINCKE, 0. M., S. P. YANOVIAK, AND R. D. HANSCHU. 1997. Predation by odonates de-
presses mosquito abundance in water-filled tree holes in Panama. Oecologia
112: 244-253.
FISH, D., AND S. R. CARPENTER. 1982. Leaf litter and larval mosquito dynamics in
tree-hole ecosystems. Ecology 63: 283-288.
FRETWELL, S. D. 1977. The regulation of plant communities by the food chains exploit-
ing them. Perspect. Biol. Med. 20: 169-185.
FRETWELL, S. D. 1987. Food chain dynamics: the central theory of ecology? Oikos 50:
GRILL, C. P., AND S. A. JULIANO. 1996. Predicting species interactions based on behav-
iour: predation and competition in container-dwelling mosquitoes. J. Anim.
Ecol. 65: 63-76.
HANLON, R. D. G., AND J. M. ANDERSON. 1979. The effects of collembola grazing on mi-
crobial activity in decomposing leaf litter. Oecologia 38: 93-99.
HARD, J. J., W. E. BRADSHAW, AND D. J. MALARKEY. 1989. Resource- and density-de-
pendent development in tree-hole mosquitoes. Oikos 54: 137-144.
HARGRAVE, B. T. 1970. The effect of a deposit-feeding amphipod on the metabolism of
benthic microflora. Limnol. Oceanogr. 15: 21-30.
HARRISON, P. G. 1977. Decomposition of macrophyte detritus in seawater: effects of
grazing by amphipods. Oikos 28: 165-169.
KITCHING, R. L. 1971. An ecological study of water-filled tree-holes and their position
in the woodland ecosystem. J. Anim. Ecol. 40: 281-302.
KITCHING, R. L. 1990. Foodwebs from phytotelmata in Madang, Papua New Guinea.
Entomologist 109: 153-164.
LEONARD, P. M., AND S. A. JULIANO. 1995. Effect of leaf litter and density on fitness
and population performance of the hole mosquito Aedes triseriatus. Ecol. Ento-
mol. 20: 125-136.
LOPEZ, G. R., J. S. LEVINTON, AND L. B. SLOBODKIN. 1977. The effect of grazing by the
detritivore Orchestia grillus on Spartina litter and its associated microbial
community. Oecologia 30: 111-127.
LOUNIBOS, L. P., N. NISHIMURA, AND R. L. ESCHER. 1993. Fitness of a treehole mos-
quito: influences of food type and predation. Oikos 66: 114-118.
MORAN, M. D., AND L. E. HURD. 1998. A trophic cascade in a diverse arthropod com-
munity caused by a generalist arthropod predator. Oecologia 113: 126-132.
NANNINI, M. A., AND S. A. JULIANO. 1998. Effects of the facultative predator Anopheles
barberi on population performance of its preyAedes triseriatus (Diptera: Culi-
cidae). Ann. Entomol. Soc. Am. 91: 33-42.
POWER, M. E. 1990. Effects of fish in river food webs. Science 250: 811-814.
POWER, M. E., J. C. MARKS, AND M. S. PARKER. 1992. Variation in the vulnerability of
prey to different predators: community-level consequences. Ecology 73: 2218-
SOKAL, R. R., AND F. J. ROHLF. 1981. Biometry. W. H. Freeman & Co., N.Y.
SPILLER, D. A., AND T. W. SCHOENER. 1994. Effects of top and intermediate predators
in a terrestrial food web. Ecology 75: 182-196.
SRIVASTAVA, D. S., AND J. H. LAWTON. 1998. Why more productive sites have more
species: experimental test of theory using tree-hole communities. Am. Nat. 152:
YANOVIAK, S. P. 1999. Community ecology of water-filled tree holes in Panama. Ph.D.
dissertation, Univ. of Oklahoma, Norman.

Reinert et al.: Beauveria bassiana us Boxelder


1Texas A&M University Research & Extension Center,
17360 Coit Road, Dallas, TX 75252-6599

2Earthgreen Products, Inc., Dallas, TX 75225

3Biology Department, Labette College, Parsons, KS


Laboratory studies were initiated to assay the efficacy of BioCide-TRFTM (Mycotrol
GH-ES) Beauveria bassiana Vuill. on two nuisance pests of the family Rhopalidae that
frequently take winter protection in and around urban and suburban buildings. Mi-
cro-application treatments (2.5 il droplet) of 6.5 x 106 conidia per insect, applied to the
dorsal prothorax of the red shouldered bug, Jadera haematoloma (Herrich-Schaeffer),
provided 100% control of adults in 8 d with 84% of the dead bugs expressing mycelia
in that time period. Lower dosages did not provide significant mortality at 5 d, but by
8 d, 32-40% control was provided by the 4.0 x 104 and 1.6 x 106 conidia treatments. All
adults of the boxelder bug, Boisea trivittatus (Say), exposed to treated leaves and seeds
with the 1:100 dilution dose of Beauveria were killed within 7 d and 66.7% of the dead
bugs expressed mycelia within 8 d. Significant mortality (83.3, 76.7 and 73.3%) was
also provided by the 1:300, 1:500 and 1:1,000 dilution rates, respectively. This race of
B. bassiana shows promise for control of both of these annual nuisance pests.

Key Words: Boisea trivittatus, Leptocoris, Jadera haematoloma, Sapindus, Acer


Se efectuaron studios de laboratorio para probar la eficacia de Biocide-TRF
(Mycotrol GH-ES) Beauveria bassiana Vuill. en dos molestas plagas de la familiar Rho-
palidae que toman protecci6n en el interior y exterior de edificios urbanos y suburba-
nos. La microaplicacion de una gota de 2.5 il conteniendo 6.5 x 106 conidios/insecto
colocada en el protorax dorsal del insecto de lomo rojo, Jadera haematoloma (Herrich-
Schaeffer) control 100% de los adults en 8 dias. El 84% de los insects muertos mos-
traron desarrollo de micelios. Las dosis de 4.0 y 1.6 x 106 conidios/insecto lograron un
control de 32 a 40% a los 8 dias. Todos los adults del insecto Boisea trivittatus (Say)
expuestos a hojas y semillas tratadas con una diluci6n de 1:100 de Beauveria murie-
ron en 7 dias y 66.7% de los insects muertos mostraron desarrollo de micelios en 8
dias. Las diluciones de 1:300, 1:500, y 1:1000 tambi6n causaron mortalidad significa-
tiva (83.3, 76.7, y 73.3%, respectivamente). Esta raza de B. bassiana muestra poten-
cial para el control de las dos plagas estudiadas.

The family Rhopalidae, or scentless plant bugs, is small but contains several spe-
cies that quite often become serious nuisance pests in the fall as they take winter pro-
tection in and around urban and suburban buildings. The boxelder bug (BEB)
(formerly in the genus Leptocoris Hahn), and the red shouldered bug (RSB), each feed

Florida Entomologist 82(3) September, 1999

on the fruits of their respective hosts throughout their nymphal and adult develop-
ment. Damage by both species usually occurs early in the season as the overwintering
adults feed on the developing buds. This feeding may distort the foliage and cause the
plants to be bushy.
RSB is a household nuisance pest across the range of its primary breeding host,
Chinaberry, Sapindus drummondii Hook & Arn. When Chinaberry trees are present
in the landscape, this pest often reaches high populations on the developing yellow
berries. As the nymphs and adults mature, large numbers are often seen migrating up
and down the tree trunks, across the lawn and landscape, and into houses and other
nearby buildings throughout the summer and early fall. This migration into buildings
takes place both during the breeding period and later in search of overwintering sites.
RSB has a known distribution including AL, AZ, CA, CO, FL, IL, KS, MO, OK, the
West Indies and Mexico to Columbia and Venezuela, but its biology is poorly under-
stood (Henry 1988, Arnold et al. 1992).
BEB is a common household nuisance pest which also leaves its host plant in late
fall to seek overwintering sites in and around buildings. Their annual invasion of
houses on warm days in late fall, winter and spring in search of overwintering sites
is well documented (Arnold et al. 1992, Ebeling 1978, Johnson & Lyon 1992). This in-
sect breeds on the boxelder tree,Acer negundo L., and it also feeds and develops large
populations on other Acer spp. BEB is known from central and eastern Canada
throughout the eastern United States and west to MT, NV and CA (Dreistadt et al.
1994, Henry 1988). Its life history and behavior has been studied by Smith and Shep-
herd (1937) and Tinker (1952). Wheeler (1982) provides an extensive review of the lit-
erature on its economic importance and biology.
Several biocontrol products, containing formulations of Beauveria, have been used
successfully against other families of Hemiptera in other countries around the world
and a review is presented by McCoy et al. (1985). In the Peoples Republic of China, cer-
tain strains of B. bassiana are produced and used for control of the European corn borer,
Ostrinia nubilalis (Hubner). In several eastern European countries, Beauveria is mass
produced for use against the Colorado potato beetle, Leptinotarsa decemlineate (Say).
In the U.S., various strains of B. bassiana have been widely tested against several
primary pests including the Colorado potato beetle (Watt & Lebrun 1984); lygus bugs,
Lygus spp., and several stored products pests (Dunn & Mechalas 1963); chinch bugs,
Blissus leucopterus leucopterus (Say) (Ramoska & Todd 1985); green peach aphid,
Myzus persicae (Sulzer) (Kish et al. 1994); Russian wheat aphid, Diuraphis noxia
Kurdyumov, (Wang & Knudsen 1993); pear psylla, Cacopsylla pyricola (Foerster)
(Puterka et al. 1994) and the red imported fire ant, Solenopsis invicta Buren (Oi et al.
1994). Naturalis" L, a strain of B. bassiana, has been evaluated in cotton and exhib-
ited activity in tests against boll weevil,Anthonomous grandis grandis Boheman; tar-
nished plant bug, Lygus lineolaris (Palisot de Beauvois); cotton fleahopper
Pseudatomoscelis seriatus (Reuter) and silverleaf whitefly, Bemisia argentifolii Bel-
lows & Perring (Knauf 1995).
The purpose of this study was to evaluate the BioCide-TRFT (Mycotrol GH-ES)
strain ofB. bassiana as a low-impact environmentally-sound control measure for RSB
and BEB for use in urban landscapes.


RSB Experiment:
For this study, RSB adults were collected as they migrated up and down the trunk
of several Chinaberry trees on the campus of the Texas A&M University Research and

Reinert et al.: Beauveria bassiana us Boxelder

Extension Center at Dallas, TX. Adults were brushed into 3.79 liter (1 gal) ice cream
containers fitted with 15 cm diam. funnel modified lids to prevent escapes. They were
held for 24 h; only adults that appeared healthy after ld in captivity were used for the
The RSB Experiment was set up on 28 June, 1996. Captive adults were divided
into units of 5 randomly-selected adults in 4-dram shell vials with stoppers. Vials
were randomly assigned to treatments and emptied into 9 cm diam x 15 mm deep
plastic petri dish feeding chambers that were labeled for the respective treatments in
a randomized complete block design with 5 reps. Each chamber was provided with two
7.5 cm diam filter papers moistened with water to provide constant, high humidity
during the study. Additionally, 2 chinaberry fruits were added to each chamber as a
food source during the study.
Four dilutions of BioCide-TRFTM (Mycotrol GH-ES) B. bassiana (2.11 x 10"
conidia/liter) (2.0 x 10" conidia/qt) were compared with chlorpyrifos (as a commercial
standard) and an untreated check (rates listed in Table 1). The adults confined in each
chamber were then anesthetized with CO2 gas and a 2.5 pl droplet of the respective
conidia dilution or chlorpyrifos solution was applied with a syringe to the dorsal pro-
thorax of each of the five individuals. SilwetTM (silicon + polyether copolymer) (0.04%),
a wetting agent, was added to each dilution to insure adherence to the normally hy-
drophobic cuticles of the insects. If a treated bug did not show complete recovery
within 10 min, another adult was treated and replaced it. Treatments were observed
daily for mortality; all dead or moribund insects were held for several days to confirm
the fungus activity. The percent mortality was recorded at 5 and 8 d after treatment
and confirmed as an expression of Beauveria mycelia from the joints of the dead in-
sects. Fungal infection was confirmed from representative cadavers by culturing on
Saubaroud maltose agar (SMA) (Difco, Detroit, MI) petri plates in the lab.

2.11 x 1013 CONIDIA/LITER, DALLAS, TX, JULY 1996 (5 REPS).

5 days' 8 days'

Mortality Mycelia Mortality Mycelia
Treatment Conidia" (%) (%) (%) (%)

BioCide 32% 6.5 x 106 56 a'd 12 acd 100 ac'd 84 ac'd
BioCide 32% 1.6 x 106 4 b 0 b 32 cd 8 b
BioCide 32% 4.0 x 104 16 b 0 b 40 c 4 b
BioCide 32% 1.0 x 104 0 b 0 b 12 d 0 b
6.6% EC 1.2 ge 64 a 0 b 76 b 0 b
Untreated 10 b 0 b 10 d 0 b

a2.5 pl of each concentration was applied to the dorsal prothorax of each adult bug.
Adult mortality and % expression of fungal mycelial growth at days post inoculation.
'Analysis was made on arcsine transformation of the data: untransformed data is presented.
Means in a column not followed by the same letter are significantly different by Waller-Duncan k-ratio t-test
(k= 100) (P= 0.05).
'Rate as 1.2 g al/liter (1 lb al/100 gal).

Florida Entomologist 82(3) September, 1999

BEB Experiment:

Adults for this study were collected from a developing population (ca. 50% adults)
of boxelder bugs on a bigtooth maple,Acer grandidentatum Nutt., that had developed
a heavy seed set, also in the landscape at the TAMU-REC, Dallas, TX. Leaves and
seed were also collected from this host tree to be used as a food source during the
study. A second method of exposure to the Beauveria was also investigated in this
study in the lab. One leaf and 3 seeds were dipped in the appropriate dilutions (Table
2) of BioCideTM (Mycotrol GH-ES)B. bassiana (2.11 x 10" conidia/liter), allowed to air
dry and placed on two water-saturated filter papers (for constant humidity) in each
feeding chamber for exposure to the Beauveria. Silwet was added to the dilutions as
above. On 4 November 1996, 5 randomly-selected adults, that had been held for ld as
above, were introduced into each 9 cm diam x 15 mm deep plastic feeding chambers
labeled for the respective treatment in a randomized complete block design with 6
reps. Both a water + Silwet dipped and an untreated check were included. Treatments
were observed daily for mortality; all dead insects were held for several days to con-
firm the fungal infection and cultured the same as with the RSB.
Data were subjected to analysis by the General Linear Model procedure and treat-
ments were separated by Waller-Duncan k-ratio t-test (k = 100) (P = 0.05) (SAS Insti-
tute 1985). Mortality data were transformed by arcsine before analysis, but
untransformed means are presented here.


RSB Experiment:

Results of the study are summarized in Table 1. All of the adults treated with the
6.5 x 106 conidia per insect dose of Beauveria were killed and 84% of the dead bugs ex-
pressed mycelia within 8 d, and this treatment provided significantly greater mortal-
ity (P < 0.05) than the other treatments, including chlorpyrifos (76%). The lower
dosages did not provide significant mortality at 5 d, but by 8 d, 32-40% control was
provided by the 4.0 x 104 and 1.6 x 106 conidia doses, respectively. The chlorpyrifos
standard, however, provided significantly higher mortality than either of these lower

BEB Experiment:

Results of this study are summarized in Table 2. The 1:100 dilution of Beauveria
applied to leaves and seeds ofAcer grandidentatum killed 100% of the BEB within 7
d and 66.7% of the dead bugs expressed mycelia within 8 d. Significant mortality at 8
d (83.3, 76.7 & 73.3%) was also provided by the 1:300, 1:500 and 1:1,000 dilution rates,
respectively. The 2 lowest dilution rates tested were not significantly better than the
untreated checks. Also, no difference was recorded between the water dipped + Silwet
and the other untreated check. As a general trend, it took longer to kill the BEB, the
more dilute the Beauveria treatment.
The BioCide-TRFTM (Mycotrol GH-ES) strain ofB. bassiana shows promise for con-
trol of each of these annual nuisance pests. Both methods of application yielded a good
separation among the rates tested. For the micro-application directly to the insects,
the separation was between 6.5 x 10' and 1.6 x 106. With treatments applied to the
substrate (air dried), a significant separation occurred between dilutions of 1:100 and
1:300. The primary separation in efficacy occurred between dilutions of 1:1,000 and


5 day"

7 day"

8 days"

9 day"

Dilution Mort (%) Mycelia (%) Mort (%)

Mort (%) Mycelia (%) Mort (%)

BioCide 32% 1:100 73.3 aw" 60.7 a"' 100 ab" 100 ab', 66.7 a"b, 100 ab,'
BioCide 32% 1:300 6.7 b 3.3 b 63.3 b 76.7 b 16.7 b 83.3 b
BioCide 32% 1:500 23.3 b 13.3 b 53.3 bc 56.7 b 33.3 ab 76.7 b
BioCide 32% 1:1,000 3.3 b 0 b 33.3 cd 56.7 b 20.0 b 73.3 b
BioCide 32% 1:10,000 0.0 b 0 b 10.0 d 26.7 c 0 b 26.7 c
BioCide 32% 1:100,000 3.3 b 0 b 6.7 d 6.7 c 0 b 6.7 c
Untreated dipped 13.3 b 0 b 16.7 d 23.3 c 0 b 26.7 c
Untreated 13.3 b 0 b 13.3 d 16.7 c 0 b 23.3 c

aAdult mortality and % expression of fungal mycelial growth at days post inoculation.
bAnalysis was made on arcsine transformation of the data: untransformed data is presented.
Means in a column not followed by the same letter are significantly different by Waller-Duncan k-ratio t-test (k = 100) (P = 0.05).


Florida Entomologist 82(3) September, 1999

1:10,000. These results should serve as a guide in determining field treatment rates.
Field trials in the landscape are planned to confirm these results and determine the
dilution rates required for control under field conditions.


ARNOLD, D., K. PINKSTON, AND B. CARTWRIGHT. 1992. Major horticultural & house-
hold insects of Oklahoma. Oklahoma State Univ., Coop. Ext. Serv. Circ. E-918.
178 pp.
DREISTADT, S. H., J. K. CLARK, AND M. L. FLINT. 1994. Pests of landscape trees and
shrubs, an integrated pest management guide. Univ. California, Div. Agric.
Natural Resources. Pub. 3359. 327 pp.
DUNN, P. H., AND B. J. MECHALAS. 1963. The potential of Beauveria bassiana (Bal-
samo) Vuillemin as a microbial insecticide. J. Insect Pathol. 5: 451.
EBLING, W. 1978. Urban Entomology. Univ. California Press, Berkeley, CA 695 pp.
HENRY, T. J. 1988. Family Rhopalidae Amyot and Serville, 1843 (=Corizidae Douglas
and Scott. 1865). p. 652-664 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, Lieden, NY 958 pp.
JOHNSON, W. T., AND H. H. LYON. 1992. Insects That Feed on Trees and Shrubs. Com-
stock Publ. Assoc., Cornell Univ. Press, Ithaca, NY. 556 p.
KISH, L. P., I. MAJCHROWICZ, AND K. D. BIEVER. 1994. Prevalence of natural fungal
mortality of green peach aphid (Homoptera: Aphididae) on potatoes and non-
solanaceous hosts in Washington and Idaho. Environ. Entomol. 23(5): 1326-
KNAUF, T. A. 1995. NaturalisTM L for control of cotton insects: a review of trials in
Texas and the Mississippi Delta. Addendum. Proc. Beltwide Cotton Conf. 1995:
McCoY, C. W., R. A. SAMSON, AND D. G. BOUCIAS. 1985. Entomogenous fungi, p. 151-
236. in C. M. Ignoffo [ed.]. CRC Handbook of Natural Pesticides, Vol. 5, Micro-
bial Insecticides, Part A, Entomogenous Protozoa and Fungi. CRC Press, Inc.,
Boca Raton, FL 243 pp.
OI, D. H., R. M. PEREIRA, J. L. STIMAC, AND L. A. WOOD. 1994. Field applications of
Beauveria bassiana for control of the red imported fire ant (Hymenoptera: For-
micidae). J. Econ. Entomol. 87(3): 623-630.
PUTERKA, G. J., R. A. HUMBER, AND T. J. POPRAWSKI. 1994. Virulence of fungal patho-
gens (Imperfect fungi: Hyphomycetes) to pear psylla (Homoptera: Psyllidae).
Environ. Entomol. 23(2): 514-520.
RAMOSKA, W. A., AND T. TODD. 1985. Variation in efficacy and viability of Beauveria
bassiana in the chinch bug (Hemiptera: Lygaeidae) as a result of feeding activ-
ity on selected host plants. Environ. Entomol. 14: 146-148.
SMITH, R. C., AND B. L. SHEPHERD. 1937. The life history and control of the boxelder
bug in Kansas. Trans. Kansas Acad. Sci. 40:143-159.
TINKER, M. E. 1952. The seasonal behavior and ecology of the boxelder bug, Leptocoris
triuittatus in Minnesota. Ecology 33: 407-414.
WANG, Z. G., AND G. R. KNUDSEN. 1993. Effect of Beauveria bassiana fungi hypho-
mycetes on fecundity of the Russian wheat aphid Homoptera: Aphididae. Envi-
ron. Entomol. 22(4): 874-878.
WATT, B. A., AND R. A. LEBRUN. 1984. Soil effects of Beauveria bassiana on pupal pop-
ulations of the Colorado potato beetle. Environ. Entomol. 13: 15.
WHEELER, A. G., JR. 1982. Bed bugs and other bugs, p. 319-351. in A. Mallis [ed.].
Handbook of Pest Control. 6 ed. Franzak & Foster Co, Cleveland, OH 1101 pp.

Scientific Notes


1University of Florida, Tropical Research and Education Center, 18905 SW 280 St.,
Homestead FL 33031

2USDA-ARS-Midwest Area National Center for Agriculture, 1815 N University St.,
Peoria IL 61604

Atemoya, Annona squamosa L. x Annona cherimola Miller, and sugar apple, A.
squamosa, have importance for the tropical fruit industry in Florida, but their yields
are unreliable. This is so because fruit set is erratic due to highly variable pollination
and physiological stress. Flowers of atemoyas and sugar apples initially undergo a fe-
male phase during which stigmas are receptive, and later they have a male phase
when the anthers split to shed pollen, but the stigmas are no longer receptive (Gotts-
berger 1970). This prevents autogamy (i.e., fertilization of ovules by pollen from the
same flower) (Nadel and Peha 1994). Most Annonaceae species are cantharophilous
(beetle-pollinated), and a few are sapromyophilus (fly-pollinated) or thrips-pollinated
(Gottsberger 1985). Cross pollination ofAnnona spp. can be carried out by sap beetles
(Nitidulidae) (Reiss 1971, Nagel et al. 1989, George et al. 1992). In south Florida, nit-
idulid pollinators are in the genera Carpophilus (six species) and Haptoncus (one spe-
cies) (Nadel and Peha 1994). Gottsberger (1985) suggested that the Annonaceae and
other primitive plants with cantharophilous pollination had evolved a specialized pol-
lination system by releasing heavy floral volatiles that attract certain beetle species.
For example, atemoya flowers open mid- to late afternoon and allow beetles enter. The
flowers begin to shed pollen around noon of the following day, and drop their petals by
the following noon when the beetles start their departure. The petals emit an odor
reminiscent of ripe, fermenting fruit that attracts the pollinators (Podoler et al. 1984).
Current practices to manage pollinator populations in annona groves have resulted in
inadequate yields (Schroeder 1971, George et al. 1992). Efforts to increase nitidulids
in the flowers with fermenting pineapples as a bait have given mixed results (Galon
et al. 1982, George et al. 1992).
Fruit setting in Annonaceae can be also be improved by hand pollination, but this is
expensive and time consuming (Kahn and Arpaia 1990), or by improving attraction to
the sap beetles by the use of lures (Nadel and Peha 1994). Several pheromones have been
identified in the Carpophilus species complex (Okumura and Savage 1974) and tested in
date groves in California (Bartelt et al. 1992, 1994). The effect of the pheromones is syn-
ergized when they are used in combination with food odors (Bartelt et al. 1992). In this
study we evaluate the effect of sap beetle pheromones and food odor bait as attractants
for beetle pollinators and their effect on fruit set of sugar apple and atemoya.
Synthetic hydrocarbon pheromones used in this study were available from previ-
ous research (Bartelt et al. 1995). The compounds were: (2E, 4E, 6E)-5-ethyl-3-me-
thyl-2,4,6-nonatriene (1), (3E, 5E, 7E)-6-ethyl-4-methyl-3, 5, 7-decatriene (2), (3E, 5E,
7E)-5-ethyl-7-methyl-3,5,7-undecatriene (3), (2E, 4E, 6E, 8E)-3,5,7-trimethyl-2, 4.6,
8-decatetraene (4), (2E, 4E, 6E, 8E)-3, 5, 7-trimethyl-2, 4, 6, 8-undecatetraene (5),
(2E, 4E, 6E, 8E)-7-ethyl-3, 5-dimethyl-2, 4, 6, 8-decatetraene (6), (2E, 4E, 6E, 8E)-7-
ethyl-3, 5-dimethyl-2, 4, 6, 8-undecatetraene (7) and (3E, 5E, 7E, 9E)-6, 8-diethyl-4-
methyl-3, 5, 7, 9-dodecatetraene (8). Rubber septa were prepared with these com-
pounds in the following percentages as baits for five Carpophilus species: C. freeman,
1(96%), 7 (4%); C. mutilatus, 2 (7%), 3 (93%); C. hemipterus, 4 (67%), 5 (21%), 6 (7%),

Florida Entomologist 82(3) September, 1999

7 (5%); C. lugubris, 5 (9%), 7 (91%); and C. dimidiatus 8 (100%). Proportions were cho-
sen so that emissions from septa would be as similar in composition as possible to
those from beetles. For all species except C. dimidiatus, a total of 500 pg of all-E iso-
mers was used in each bait. The pheromone for C. dimidiatus was used at the higher
rate of 1.2 mg per septum because of the lower volatility of compound 8, the phero-
mone for the species. Fermented whole-wheat bread dough (a 4:1:2:0.01 mixture of
whole-wheat flour, sugar, water and yeast) was used as the synergist in all cases (~ 15
ml per trap) (Bartelt et al. 1994).
The studies were conducted from 21 March to 17 May 1995 by placing beetle traps
(Dowd et al. 1992) baited with pheromone septa and bread dough in the perimeters of
a 0.46-ha sugar apple orchard in Homestead, Florida. Traps (n = 12) were hung from
the trees at 1.2 m above the ground and placed at a distance of ca. 28 m apart. Traps
were inspected weekly and trapped species identified.
Effect of Bait Stations on Fruit Set. The study consisted of two experiments, each
having a complete block design. For experiment 1 (27 April, 1995 -7 July 1995), the
study was conducted in the same orchard mentioned above. The treatments included
the pheromones of C. mutilatus, C. freeman, C. lugubris, C. hemipterus and C. dimid-
iatus, each used in combination with the dough as co-attractant. In addition there was
an unbaited control and a treatment containing just the bread dough. Flowers were
recorded per tree and attractants were hung only in those trees with more than 100
tagged flowers. One bait station was hung per tree for 7 days, and then removed. Bait
stations consisted of a wire holding a rubber septum with the pheromone, and a plas-
tic cup with 15 g of bread dough. The baits were covered with a fabric sleeve to prevent
contact by the beetles. Rows of 12 trees per treatment were used. Treatments were set
60 m apart.
For experiment 2 (September 12, 1996-October 11, 1996), the study was conducted
in a 1-ha atemoya orchard in Homestead, Florida. The combinations of pheromones
that gave the best results in experiment 1, plus dough, and the dough alone, were
tested following the same procedure as explained above. From the second to the sev-
enth week after treatment, percent fruit set per tree was calculated every week by di-
viding the number of fruits 1 cm in diameter or higher by the number of flowers
Percentages were transformed to arcsine of the square root of the proportion to
stabilize variance, and these data were analyzed by using Statistical Analysis System
general linear models (SAS 1987) for balanced ANOVA. Means were separated by
Waller-Duncan k-ratio t-tests.
The dominant sap beetle species were C. mutilatus (32.9%), C. freeman (21.7%),
and C. dimidiatus (19.4%), followed by Colopterus sp. (13.6%), Haptoncus luteolus
(7.3%), C. marginelus (1.8%), C. humeralis (1.3%) and C. hemipterus (0.1%) (Table 1).
One of these species, C. mutilatus was the most abundant species collected from ate-
moya flowers in Florida in 1988 followed by H. luteolus. (Nadel and Pena 1994). C. mu-
tilatus responded best to the combination of its own pheromone plus fermenting
whole wheat dough, and catches with this treatment were as high as 92 individuals
in 1 wk in one trap. C. mutilatus responded less strongly to pheromones of other bee-
tles. C. dimidiatus responded to its own pheromone and to the combination of its own
pheromone plus fermenting whole wheat dough. C. hemipterus responded only to the
combination of its own pheromone plus fermenting dough. H. luteolus, for which a
pheromone is not known, responded consistently to the combination of the pheromone
of C. lugubris plus whole wheat dough.
Percent fruit set was higher in 1995 than in 1996 (Tables 2 and 3). In 1995, the
highest percent fruit was obtained for the combination C. dimidiatus plus dough com-

17, 1995.

Beetles per week per trap (mean)
or Pheromone C. mutilatus C. freeman C. dimidiatus C. marginelus C. humeralis C. hemipterus H. luteolus Colopterus sp.

Dough 0.22 b 0.00 b 0.00 b 0.00 b 0.10 a 0.00 a 1.72 ab 0.00 c
Cm phero- .
mone,+ Dough 19.70 a 0.80 b 0.00 b 0.00 b 0.00 a 0.00 a 0.60 ab 0.20 bc
Cf phero-
mone + Dough 1.20 b 23.70 a 0.00 b 0.40 ab 0.20 a 0.00 a 0.20 b 1.20 bc C
Cl, phero-
mone + Dough 1.60 b 0.60 b 0.00 b 0.50 ab 0.40 a 0.00 a 2.10 a 2.80 b
Ch, phero-
mone + Dough 5.11 b 0.55 b 0.00 b 0.70 a 0.10 a 0.10 a 0.89 ab 8.00 a
Cd, 0.40 b 0.00 b 10.50 a 0.00 b 0.30 a 0.00 a 1.00 ab 0.10 bc
pheromone +

'Cm = Carpophilus mutdlatus; Cf = freeman; Cl = C. lugubris; Ch = C. hemipterus; Cd = C. dimidatus.
The square root transformation x'= (x + 0.5 was applied to normalize the means. Original means are shown. Means in the same column followed by the same letter are not significantly
different according to a Waller-Duncan k-ratio t-test (P > 0.05).

Florida Entomologist 82(3) September, 1999

pared to the percent fruit set in untreated trees or that from baited with C. freeman
and C. mutilatus pheromones. Moreover, percent fruit set was low when C. dimidiatus
pheromone was used without food bait. Bartelt (1992) observed that the effect of the
pheromones was synergized by food odors and recommended the use of both attracta-
nts in combination. In 1996, more fruit was observed on those trees baited with either
C. dimidiatus, C. freeman or C. mutilatus pheromones in combination with dough
than on untreated trees or those trees baited with dough alone. In both years fruit re-
tention decreased to 5-6% during the last 5-7 weeks after treatment, when the fruit
reached more than 2 cm in diameter. The lower fruit retention found on weeks 5-7 after
treatment might have been caused by water or fertilizer stress. It is well known that
water stress and lack of nutrients induce fruit abscission (Salisbury and Ross 1978).
Previously, Nagel et al. (1989) demonstrated that fruit set of atemoya caged trees
supplied with nitidulids was significantly higher (21%) than on caged trees without
nitidulids (5.1%). Nitidulids were found to pollinate 10% of sugar apple flowers, and
29% of atemoya flowers in Florida (Nadel and Pena 1994). Our study suggests
strongly that nitidulids increased pollination and percent fruit set in both sugar apple
and atemoya, because significant differences between the treated and non-treated


Pheromone source' % Fruit Set + SE2

Cd 6.87 + 1.00 c
Cd + dough 38.11 + 5.10 a
Cf 11.06 + 2.03 bc
Cf + dough 10.84 + 5.00 bc
Cm 11.14 + 5.21 bc
Cm + dough 10.36 + 6.61 bc
Control (no bait) 4.33 + 4.68 d

Cd = Carpophilus dimidiatus; Cf = C. freeman; Cm = C mutilatus.
Means followed by the same letter are not significantly different according to a Waller-Duncan k-ratio t-test
on arcsine transformed data (p > 0.001, k-ratio = 100). Untransformed means are presented.


Pheromone Source' % Fruit Set + SE2

Cd + dough 10.33 + 3.39 a
Cf + dough 9.42 + 4.37 a
Cm + dough 8.33 + 4.78 a
Dough 4.33 + 3.98 b
Control (no bait) 0.55 1.24 c

SCd = Carpophilus dimidiatus; Cf = C. freeman; Cm = C mutilatus.
Means followed by the same letter are not significantly different according to a Waller-Duncan k-ratio t-test
on arcsine transformed data (p > 0.001, k-ratio = 100). Untransformed means are presented.

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