Ventura et al.: Responses of Diabrotica and Cerotoma 403
RESPONSES OF DIABROTICA SPECIOSA AND
CEROTOMA ARCUATA TINGOMARIANA (COLEOPTERA:
CHRYSOMELIDAE) TO VOLATILE ATTRACTANTS
M. U. VENTURA, M. C. MARTINS AND A. PASINI
Departamento de Agronomia, Universidade Estadual de Londrina.
C.P. 6001, 86051-970, Londrina, PR, Brazil
The relative responses of Diabrotica speciosa (Ger.) and Cerotoma arcuata tingo-
mariana Bechyne (Coleoptera: Chrysomelidae) to traps baited with chemicals were
studied. Volatile substances of Curcubita maxima Duchesne blossoms, other previ-
ously reported volatile attractants for Diabroticites and mixtures of chemicals were
tested in common bean, Phaseolus vulgaris L., and soybean, Glycine max (L.) Mer.,
fields. Traps baited with 1,4-dimethoxybenzene caught 29.4 times more beetles than
solvent controls in fields of soybeans, and 9.4 times more in common bean fields. Traps
baited with VIP (veratrole + indole + phenylacetaldehyde) caught 6.5 times more bee-
tles than solvent controls in soybean and 3.5 times more in common bean plots,
whereas traps baited with TIC (1,2,4-trimethoxybenzene + indole + trans-cinnamal-
dehyde) caught 6.7 times more beetles in soybean and 3.5 times more in common bean
plots. Volatile chemicals used in this study did not attract C. a. tingomariana. In a
dose-response study, captures of D. speciosa increased significantly with increasing
doses of 1,4-dimethoxybenzene.
Key Words: Diabrotica speciosa, Cerotoma arcuata tingomariana, Phaseolus vulgaris,
Glycine max, semiochemical, kairomone
As respostas relatives de Diabrotica speciosa (Ger.) e Cerotoma arcuata tingoma-
riana Bechyne (Coleoptera: Chrysomelidae) para armadilhas com substancias quimi-
cas foram estudadas. Substancias volateis de flores de Curcubita maxima Duchesne,
outros atraentes volateis, previamente reportados para diabroticineos, e misturas de
substancias foram testados em campos de feijao,Phaseolus vulgaris L., e soja, Glycine
max (L.) Mer. Armadilhas com 1,4-dimethoxybenzene capturaram 29,4 vezes mais be-
souros do que a testemunha com solvente em campos de soja, e 9,4 vezes mais em fei-
jao. Armadilhas com VIP (veratrole + indole + phenylacetaldehyde) capturaram 6,5
vezes mais besouros do que a testemunha com solvente em soja e 3,5 vezes mais em
feijao, enquanto que armadilhas com TIC (1,2,4-trimethoxybenzene + indole + trans-
cinnamaldehyde) capturaram 6,7 vezes mais besouros em soja e 3,5 vezes mais em fei-
jao. As substancias nao atrairam C. a. tingomariana. Em estudo de dose-resposta, as
captures de D. speciosa aumentaram significativamente com doses crescentes de 1,4-
Relationships between Diabroticites and plants of the family Cucurbitaceae are
mediated by kairomones. The extremely bitter cucurbitacins are arrestants and feed-
ing stimulants for Diabroticite and Aulacophorite beetles (Luperini tribe) (e.g. Cham-
blis & Jones 1966, Ferguson et al. 1983, Howe et al. 1976). Adult Diabrotica also are
attracted to Cucurbita spp. blossoms by volatile chemicals. These substances play an
404 Florida Entomologist 83(4) December, 2000
important role in cucurbit selection by Diabroticite beetles. Morgan & Crumb (1928)
first reported the attraction of Diabroticites to volatile chemicals when they described
the attraction of D. undecimpunctata howardi (Barber) to cinnamaldeyde and cin-
namyl alcohol baits. Snapp & Swingle (1929) attracted the same species with benzyl
alcohol. D. barberi Smith and Lawrence and D. cristata (Harris) were attracted to eu-
genol, which also attracts the Japanese beetle, Popillia japonica Newman (Ladd et al.
1983, Lampman & Metcalf 1988) and to eugenol analogs (Ladd 1984). The isolation
and identification of several volatile compounds from Cucurbita spp. blossoms
(Andersen & Metcalf 1986, Andersen 1987) contributed to the knowledge of Diabrot-
ica spp. and related genera in their specific responses to chemicals. A series of Cucur-
bita blossom kairomones and closely related compounds (parakairomones) attract
Diabrotica spp. and Acallyma spp. (e.g. Andersen & Metcalf 1986, Lampman & Met-
calf 1987, 1988, Metcalf & Lampman 1989, Lewis et al. 1990, Lance et al. 1992, Deem-
Dickson & Metcalf 1995, Petroski & Hammack 1998).
Diabrotica speciosa (Ger.) and Cerotoma arcuata tingomariana Bechyn6 were cap-
tured in traps baited with cucurbitacins (Roel & Zatarin 1989, Ventura et al. 1996)
upon which these beetles feed compulsively and sequester and store 23,24-dihydrocu-
curbitacin D (Nishida et al. 1986, Nishida & Fukami 1990). This widespread similar-
ity in behavioral response metabolism of cucurbitacins provides strong evidence for
coevolution between these Chrysomelidae and Cucurbitaceae.
Although North American Diabrotica responses to volatile substances have been
studied since the beginning of this century (Morgan & Crumb 1928), no reports are
available on South American pests. We report here the results of field trials testing
the relative attraction of D. speciosa and C. a. tingomariana to volatile kairomones
and mixtures from C. maxima blossoms and some North American Diabrotica spp.
MATERIAL AND METHODS
Field experiments were carried out at the Universidade Estadual de Londrina
School Farm, Londrina (latitude 2319'S, longitude 5112'W), Parana State, Brazil.
Soybean, Glycine max (L.) Mer., cv. Ocepar 14 (sown on December 19, 1996) and com-
mon beans, Phaseolus vulgaris L., cv. Iapar 59 (sown on February 26, 1997; February
10, 1998) fields (0.5-ha plots) were used as testing sites.
In 1997, beetle traps consisted of transparency film (15.2 x 27.9 cm) (3M do Brasil,
Ribeirao Preto, SP, Brazil) painted with yellow gold Suvinil paint 2450-0103 (BASF
S.A., Sao Bernardo do Campo, SP, Brazil) on the interior. Yellow traps previously were
successful in capturing D. speciosa and C. a. tingomariana (Ventura et al. 1996). The
film was clamped into a 15.2-cm tall cylinder and externally coated with the clear in-
sect adhesive, Tangle Trap (Tangle Foot Co., Grand Rapids, MI, USA). In 1998, 750-
ml plastic cups painted with the same paint replaced the transparency film traps.
Dental wicks (40 mm long x 10 mm diameter) soaked with test chemicals were
clamped (transparency film) or glued (on the bottom of the cups) to the traps. Solid
chemicals were prepared as standard wt/vol. solutions in acetone. The baited traps
were placed in the field at 3:00 P.M. and removed after 24 hours.
Traps with 100 pl or 100 mg of each chemical were fixed on a wooden stake above
canopy height in soybeans on March 20, 1997 and 0.25 m height in common beans on
March 22, 1997. Control traps received only acetone. The chemicals tested included
the C. maxima blossom volatile substances 3-hydroxy-3,7,11-trimethyl-1,6,10-dodec-
atriene (nerolidol); benzyl alcohol; 2,3-benzopyrrole (indole); phenylacetaldehyde; 1,2-
dimethoxybenzene (veratrole); 1,2,4-trimethoxybenzene; benzaldehyde; 4-[2,6,6-tri-
Ventura et al.: Responses of Diabrotica and Cerotoma 405
methyl-1-cyclohexen-l-yl]-3-buten-2-one (B-ionone); benzyl acetone; a-ionone (Sigma
Chemical Company, St Louis, MO); 1,4-dimethoxybenzene; 4-methoxyphenethyl alco-
hol; cinnamyl alcohol; trans-cinnamaldehyde (Aldrich Chemical Co., Milwaukee, WI)
and the Diabrotica spp. parakairomone 2-methoxy-4-(-2-propenyl) phenol (eugenol),
and eugenol-related 4-allyl-1,2-methylenedioxybenzene (safrole) (Sigma). SIC (sa-
frole + indole + trans-cinnamaldehyde), TIC (1,2,4-trimethoxybenzene + indole +
trans-cinnamaldehyde) and VIP (veratrole + indole + phenylacetaldehyde) mixtures
were used at a dosage of 100 mg or 100 il of each single chemical per trap. Traps were
returned to the laboratory where the beetles were identified to species and sexed. Sex
ratio of field populations of beetles was also determined from sweep net samples taken
when the traps were removed.
The responses ofD. speciosa and C. a. tingomariana to a range of dosages (1, 3, 10,
30, 100 or 300 mg per trap) of the compound most attractive to D. speciosa were eval-
uated on April 20, 1998. A regression analysis was performed to evaluate the relation-
ship between lure concentration and trap effectiveness.
All experiments were conducted in a four replicate randomized complete block de-
sign. Distance between traps was 5 m within a block, and 10 m between blocks. Anal-
ysis of variance (ANOVA) was performed and Tukey's studentized range test (HSD)
was used to compare individual means (SAS Institute 1989) on volatile chemicals
screening. Data were transformed by log (x + 1) constant to normalize the data and re-
duce heterogeneity of variances. Means and standard errors of means are presented
for untransformed data.
RESULTS AND DISCUSSION
Only traps baited with 1,4-dimethoxybenzene, VIP and TIC mixtures caught sig-
nificantly more D. speciosa than the controls and this was true in both soybeans and
common bean (Tables 1 and 2). 1,4-Dimethoxybenzene was the most attractive com-
pound. Traps baited with the latter compound, VIP and TIC baited traps captured
both males and females. The sex ratio of D. speciosa beetles determined with sweep
net sampling was 1.1 (n = 100) in common beans and 1.0 (n = 100) in soybeans.
The TIC mixture is a strong attractant to North American Diabrotica spp. (Lamp-
man & Metcalf 1987, 1988, Lance et al. 1992) andAcalymma vittatum (F.) (Lewis et
al. 1990). Despite its geographic isolation from the North American inhabitants, Di-
abrotica spp. andA. vittatum, D. speciosa was also attracted to the simplified blend of
C. maxima blossoms (Tables 1 and 2).
The Diabrotica genus has been grouped in three taxonomic units; two of which
contain pest species (Branson & Krysan 1981). The fucata species group in which D.
speciosa is included is multivoltine, polyphagous and overwinters as adults in regions
where frost seldom occurs. The virgifera species group is univoltine, oligophagous, has
an egg diapause and overwinters in soil at temperatures below zero (Branson & Kry-
san 1981, Krysan et al. 1989). D. speciosa shows responses similar to D. u. howardi, a
species also belonging to the fucata group, in its responses to VIP and 1,4-dimethox-
ybenzene. D. u. howardi was attracted to other benzenoid compounds (Lampman et
al. 1987). In contrast, VIP was reported to be largely non-attractive to the virgifera
group species (Lampman & Metcalf 1987).
D. speciosa was not attracted to single-component lures that are known to attract
other species of Diabrotica in either the virgifera or fucata groups [i.e. benzyl acetone,
benzaldehyde, cinnamyl alcohol, eugenol, indole, B-ionone, phenylacetaldehyde, cin-
namaldehyde, and veratrole (Andersen & Metcalf 1986, Lampman et al. 1987, Lamp-
man & Metcalf 1987, 1988, Metcalf & Lampman 1989, Lewis et al. 1990)] but
Florida Entomologist 83(4)
TABLE 1. MEAN NUMBER (--SE) OF ADULTS AND THE SEX RATIO (FEMALES PER MALE) OF
D. SPECIOSA AND C. A. TINGOMARIANA CAUGHT PER YELLOW TRANSPARENCY
FILM STICKY TRAPS IN A SOYBEAN CROP AFTER 24 H (MARCH 20, 1997).
Treatment2 D. speciosa C. a. tingomariana
Benzyl acetone 5.0 + 1.0bc (1.5) 4.2 + 2.4ab (1.4)
Benzyl alcohol 8.7 + 3.5bc (1.3) 6.0 + 1.7ab (1.0)
Benzaldehyde 8.2 + 3.2bc (0.9) 7.2 + 0.9ab (0.9)
Cinnamyl alcohol 6.5 + 1.lbc (1.0) 5.0 + 3.3ab (0.7)
1,4-Dimethoxybenzene 108.7 + 25.6 a (1.2) 5.7 + 1.2ab (1.1)
Eugenol 4.7 + 1.lbc (0.9) 5.7 + 1.5ab (0.9)
Indole 10.2 + 3.9bc (1.1) 5.5 + 1.Oab (1.0)
a-ionone 2.5 1.5c (1.0) 4.5 1.Oab (0.8)
B-ionone 7.5 + 2.4bc (1.3) 7.7 + 2.3ab (1.3)
4-Methoxyphenethyl alcohol 4.5 + 2.2bc (1.2) 3.5 + 1.Oab (1.8)
Nerolidol 20.7 + 9.4bc (1.1) 13.2 + 3.6a (0.9)
Phenylacetaldehyde 11.2 + 1.lbc (1.2) 3.7 + 1.3ab (0.7)
Safrole 5.0 + 1.2bc (1.0) 4.5 + 1.Oab (0.8)
Trans-cinnamaldehyde 3.7 + 1.8c (2.0) 6.2 + 2.4ab (1.5)
1,2,4-Trimethoxybenzene 10.0 + 4.8bc (1.7) 4.0 + 1.3ab (1.3)
Veratrole 7.7 + 1.9bc (1.1) 5.2 + 2.8ab (0.9)
SIC3 12.5 + 1.4bc (1.5) 2.2 + 1.3b (1.2)
TIC4 24.7 + 2.2b (1.2) 4.0 + 1.5ab (1.7)
VIP5 24.2 + 5.4b (1.1) 3.0 + 1.5ab (1.0)
Control 3.7 + 0.9c (1.5) 6.2 + 1.7ab (0.7)
'Means in the same column with different letter are significantly different based on Tukey's studentized range
test (P < 0.05), n= 4.
Single and mixed chemicals are dosed at 100 mg or 100 1l of each compound per trap.
Safrole + indole + trans-cinnamaldehyde.
1,2,4-Trimethoxybenzene + indole + trans-cinnamaldehyde.
Veratrole + indole + trans-cinnamaldehyde.
exhibited its own species-specific pattern of response. 1,4-Dimethoxybenzene is the
major floral volatile component in Curcubita maxima Duchesne cv. True Hubbard and
the 4th major one in cv. Blue Hubbard (Andersen 1987). C. maxima blossoms attract
adults of Diabrotica species (Fischer et al. 1984, Andersen & Metcalf 1987). However,
despite the great proportion of 1,4-dimethoxybenzene in the C. maxima floral odor, no
records of its attractiveness to Luperini beetles have been reported. Metcalf & Metcalf
(1992) reviewed the attractiveness of volatile chemicals from blossoms to Diabroticite
beetles and attributed a >100-mg threshold of response by D. barberi, D. cristata, D.
u. howardii and D. v. virgifera to 1,4-dimethoxybenzene.
D. speciosa apparently is attracted to methoxylated or methylene-bridged analogs,
with or without an allyl or propenyl moiety, possibly with or without a free phenolic
group. Further investigations concerning these structure-activity aspects might be
The response by D. speciosa appeared to vary in magnitude (only one comparison)
depending on the host (differences in soybeans versus common beans). The common
bean is recognized as a very attractive crop to this beetle, mainly early in its pheno-
logical cycle (Ventura et al. 1996). Similarly, D. virgifera virgifera LeConte was re-
Ventura et al.: Responses of Diabrotica and Cerotoma 407
TABLE 2. MEAN NUMBER (--SE) OF ADULTS AND THE SEX RATIO (FEMALES PER MALE) OF
D. SPECIOSA AND C. A. TINGOMARIANA CAUGHT PER YELLOW TRANSPARENCY
FILM STICKY TRAPS IN COMMON BEAN CROP AFTER 24 H (MARCH 22, 1997).
Treatment2 D. speciosa C. a. tingomariana
Benzyl acetone 8.5 + 3.0c (1.4) 0.7 + 0.2a(2.0)
Benzyl alcohol 17.2 + 3.5bc (1.3) 2.0 + 1.4a(1.0)
Benzaldehyde 10.7 + 4.2bc (1.1) 1.2 + 0.2a(1.5)
Cinnamyl alcohol 15.0 + 7.4bc (1.0) 2.7 + 0.5a (0.8)
1,4-Dimethoxybenzene 77.0 + 32.Oa (1.0) 1.0 + 1.0a(3.0)
Eugenol 11.7 + 2.8bc (0.7) 1.5 + 0.3a(0.5)
Indole 16.0 + 4.9bc (1.1) 1.0 + 0.6a(0.3)
a-ionone 6.7 + 2.4c (0.8) 0.7 + 0.2a(0.5)
B-ionone 8.5 + 3.3c (1.0) 1.5 + 0.9a(2.0)
4-Methoxyphenethyl alcohol 16.7 + 3.9bc (0.8) 3.7 + 2.1a(0.5)
Nerolidol 18.5 _+ 2.8bc (1.0) 4.5 _+ 2.9a(1.2)
Phenylacetaldehyde 15.5 + 5.9bc (0.9) 0.7 + 0.5a(0.5)
Safrole 9.5 + 3.3bc (1.0) 4.0 + 1.3a(1.5)
Trans-cinnamaldehyde 9.0 + 1.1c (0.9) 2.2 + 1.3a(1.2)
1,2,4-Trimethoxybenzene 11.2 + 4.1bc (0.7) 1.5 + 0.9a(2.0)
Veratrole 14.7 + 3.6bc (1.0) 3.7 + 1.9a(1.1)
SIC3 17.5 + 6.8bc (0.7) 0.5 + 0.3a(0.0)
TIC4 28.5 + 13.5b (0.7) 4.0 + 2.0a.(06)
VIP5 29.0 + 13.Ob (1.3) 3.7 + 1.9a (0.5)
Control 8.2 + 1.7c (1.1) 2.7 + 1.5a (1.2)
'Means in the same column with different letter are significantly different based on Tukey's studentized range
test (P < 0.05), n= 4.
Single and mixed chemicals are dosed at 100 mg or 100 1l of each compound per trap.
Safrole + indole + trans-cinnamaldehyde.
1,2,4-Trimethoxybenzene + indole + trans-cinnamaldehyde.
Veratrole + indole + trans-cinnamaldehyde.
corded as responding differently to volatile attractants according to the host plant
phenology (Andersen & Metcalf 1987, Lampman et al. 1987).
D. speciosa is a polyphagous beetle associated with numerous plant species and
plant parts, but principally leaves and flowers (Lima 1952, Krysan 1986). Further in-
vestigation of the attraction and composition of volatile chemicals in flowers of species
more frequented by D. speciosa, especially the wild South American Cucurbitaceae,
may reveal more chemicals involved in insect-host interactions. The response of this
pest to 1,4-dimethoxybenzene indicates that cucurbitacin-baited traps could be im-
proved by adding this volatile chemical. This would be useful for crops in which D. spe-
ciosa is a rootworm pest, such as corn, Zea mays L.; wheat, Triticum aestivum L.; and
potato, Solanum tuberosum L.; in which growers are not able to easily and quickly as-
sess economic thresholds.
There were no significant differences between captures of C. a. tingomariana in traps
baited with single test compounds or mixture of compounds and the controls (Tables 1
and 2). The sex ratio of C. a. tingomariana beetle samples with a sweep net was 1.1 (n =
100) when collected in soybean and 1.0 (n = 100) in common bean. Although C. a. tingo-
mariana feeding is strongly stimulated by cucurbitacins (Nishida et al. 1986, Nishida &
Florida Entomologist 83(4)
Fukami 1990, Ventura et al. 1996), no records of this species infesting flowers of cucurbits
exist. While the polyphagous D. speciosa is a pollen feeder that responds to volatile sub-
stances in Cucurbita spp., as well as a pest of corn, beans and cucurbits (among other
hosts) (Krysan 1986), the oligophagous C. a. tingomariana is associated with beans (lar-
vae and adults) and Cucurbitaceae (adults). It has been suggested that Luperine species,
including Diabroticite and Aulacophorite, primarily coevolved with Cucurbitaceae, and
their preference for other hosts is recent (Metcalf & Lampman 1989). These beetles retain
an attraction to volatile substances in Cucurbita spp. blossoms and a feeding stimulation
by cucurbitacins. It is possible that squash-bean-corn plantings in the pre-Columbian
New World influenced host plant range (Metcalf & Lampman 1989). However the non-
pollen feeder C. a. tingomariana might have a more recent relationship with wild Cucur-
bitaceae because the lack of response of C.a. tingomariana to attractants indicates the as-
sociation with cucurbitacins is not as strong as with Diabrotica species. Feeding by
another bean leaf beetle of the Cerotoma genus, C. trifurcata Forster, is deterred by cur-
cubitacins (Metcalf et al. 1980). C. a. tingomariana must have expanded its host range to
tolerate the bitter cucurbitacins. This species sequesters 23,24-dihydrocucurbitacin D, as
D. speciosa does, after which it gains bitterness in body tissue, and strongly deters feeding
by predators (Nishida & Fukami 1990). The lack of attractiveness of volatile compounds
to C. a. tingomarina is a limitation in the improvement of lures to be used in beans and
cucurbits, crops in which D. speciosa and C. a. tingomarina are simultaneous pests.
o Cerotoma arcuata tingomariana
3 10 30 100 300
1,4-dimethoxybenzene dose pertrap (mg)
Fig. 1. Relationship between mean number of adults caught -+SE per yellow plastic
cup sticky traps (n = 4) of D. speciosa and C. a. tingomariana after 24 h (on 20 April
1998) and dosage of 1,4-dimethoxybenzene. The linear regression equations were y =
0.936242 + 47.084 log x (r2 = 0.74, P < 0.0001, n = 6) for D. speciosa and y = 3.930574
- 0.2467 log x (r2 = 0.0136, P < 0.5871, n = 6) for C. a. tingomariana.
Ventura et al.: Responses of Diabrotica and Cerotoma 409
Captures of D. speciosa in traps increased significantly with rising doses of 1,4-
dimethoxybenzene (Fig. 1) in a dose-dependent manner. The dose-dependent response
pattern would be especially advantageous to a mass trapping concept (Hoffmann et
al. 1996). D. speciosa is a very important pest in many species of vegetables and fruits
cultivated in small field areas or in greenhouses in Latin America. In such crops, traps
could be used baited with 1,4-dimethoxybenzene to reduce beetle populations.
We thank A. R. Panizzi (EMBRAPA-Soja, Londrina, PR.), C. C. Niva and Jocelyn
G. Millar (Department of Entomology, University of California, Riverside) for criti-
cally reviewing the manuscript, Edio Vizoni (Department of Applied Mathematics,
Universidade Estadual de Londrina, Londrina, PR) for his assistance with equation
fitting and statistics, R. L. Metcalf (University of Illinois, Urbana, IL) for chemical in-
formation and David C. Tramontina (Laboratory of Entomology, Universidade Estad-
ual de Londrina, Londrina, PR) for logistical help.
ANDERSEN, J. F., 1987. Composition of the floral odor of Cucurbita maxima Duchesne
(Cucurbitaceae). J. Agric. Food Chem. 35: 60-62.
ANDERSEN, J. F., AND R. L. METCALF. 1987. Factors influencing distribution of Di-
abrotica spp. in blossoms of cultivated Curcubita spp. J. Chem. Ecol. 13: 681-
ANDERSEN, J. F., AND R. L. METCALF. 1986. Identification of a volatile attractant for
Diabrotica andAcalymma spp. from blossoms of Cucurbita maxima Duchesne.
J. Chem. Ecol. 12: 687-699.
BRANSON, T. F., AND J. L. KRYSAN. 1981. Feeding and oviposition behavior and life cy-
cle strategies of Diabrotica: An evolutionary view with implications for pest
management. Environ. Entomol. 10: 826-831.
CHAMBLIS, 0. L., AND C. M. JONES. 1966. Cucurbitacins: specific insect attractants in
Cucurbitaceae. Science 153: 1392-1393.
DEEM-DICKSON, L., AND R. L. METCALF. 1995. Attractants for adult corn rootworm
monitoring and control, pp. 93-100. In: Proceedings, Illinois Agricultural Pesti-
cide Conference, 3-6 Jan 1995, U. Illinois, Urbana, II.
FERGUSON, J. E., E. R. METCALF, R. L. METCALF, AND A. M. RHODES. 1983. Influence
of cucurbitacin content in cotyledons of Cucurbitaceae cultivars upon feeding
behavior of Diabroticina beetles (Coleoptera: Chrysomelidae). J. Econ. Ento-
mol. 76: 47-51.
FISCHER, J. R., T. F. BRANSON, AND G. R. SUTTER. 1984. Use of common squash culti-
vars, Cucurbita spp. for mass collection of corn rootworm beetles, Diabrotica
spp. (Coleoptera: Chrysomelidae). J. Kansas Entomol. Soc. 57: 409-412.
HOFFMANN, M. P., J. J. KIRKWYLAND, R. F. SMITH, AND R. F. LONG. 1996. Field tests
with kairomone-baited traps for cucumber beetles and corn rootworms in cu-
curbits. Environ. Entomol. 25: 1173-1181.
HOWE, W. L., J. R. SANBORN, AND A. M. RHODES. 1976. Western corn rootworm adults
and spotted cucumber beetle associations with Cucurbita and cucurbitacins.
Environ. Entomol. 5: 1043-1048.
KRYSAN, J. L. 1986. Introduction: Biology, distribution and identification of pest Di-
abrotica, pp. 1-23. In: A. B. Krysan, and T. A. Miller [eds.] Methods for Study of
Pest Diabrotica. Springer-Verlag N.Y. 260 pp.
KRYSAN, J. L., I. C. MCDONALD, AND J. H. TUMLINSON. 1989. Phenogram based on al-
lozymes and its relationship to classical biosystematics and pheromone struc-
ture among eleven (Coleoptera: Chrysomelidae). Ann. Entomol. Soc. America
410 Florida Entomologist 83(4) December, 2000
LADD, T. L. JR. 1984. Eugenol-related attractants for the northern corn rootworm (Co-
leoptera: Chrysomelidae). J. Econ. Entomol. 77: 339-341.
LADD, T. L. JR., B. R. STINNER, AND H. R. KRUEGER. 1983. Influence of color and height
of eugenol-baited sticky traps on attractiveness to northern corn rootworm bee-
tles (Coleoptera: Chrysomelidae). J. Econ. Entomol. 77: 652-654.
LAMPMAN, R. L., AND R. L. METCALF. 1987. Multicomponent kairomonal lures for
southern corn rootworms (Coleoptera: Chrysomelidae Diabrotica spp.). J. Econ.
Entomol. 80: 1137-1142.
LAMPMAN, R. L., AND R. L. METCALF. 1988. The comparative response of Diabrotica
species (Coleoptera: Chrysomelidae) to attractants. Environ. Entomol. 17: 644-
LAMPMAN, R. L., R. L. METCALF, AND J. F. ANDERSEN. 1987. Semiochemical attracta-
nts of Diabrotica undecimpunctata howardi Barber, southern corn rootworm
and Diabrotica virgifera virgifera Le Conte, the western corn rootworm (Co-
leoptera: Chrysomelidae). J. Chem. Ecol. 13: 959-975.
LANCE, D. R., W. SCHOLTZ, J. W.STEWART, AND J. K.FERGEN. 1992. Non-pheromonal
attractants for mexican corn rootworm beetles, Diabrotica virgifera zea (Co-
leoptera: Chrysomelidae). J. Kansas Entomol. Soc. 65: 10-15.
LEWIS, P. A., R. L. LAMPMAN, AND R. L. METCALF. 1990. Kairomonal attractants for
Acalymma vittatum (Coleoptera: Chrysomelidae). Environ. Entomol. 19:8-19.
LIMA, A. C. 1952. Insetos do Brasil, 7 Tomo, Coleoptera, 290 cap. Escola Nacional de
Agriculture (S6rie didatica 9), Rio de Janeiro, 371 pp.
METCALF, R. L., AND R. L. LAMPMAN. 1989. Chemical ecology of and Curcubitaceae.
Experientia 45: 240-247.
METCALF, R. L., AND E. R. METCALF. 1992. Plant Kairomones in Insect Ecology and
Control. Chapman and Hall, New York, 168 pp.
METCALF, R. L., R. A. METCALF, AND A. M. RHODES. 1980. Cucurbitacins as
kairomones for Diabroticite beetles. Proc. Nat. Acad. Sci. USA 77: 3769-3772.
MORGAN, A. C., AND S. E. CRUMB. 1928. Notes on the chemotropic responses of certain
insects. J. Econ. Entomol. 21: 913-920.
NISHIDA, R., AND H. FUKAMI. 1990. Sequestration of distasteful compounds by some
pharmacophagous insects. J. Chem. Ecol. 16: 151-164.
NISHIDA, R., H.FUKAMI, Y. TANAKA, B. P. MAGALHAES, M. YOKOYAMA, AND A. BLU-
MENSCHEIN. 1986. Isolation of feeding stimulants of Brazilian leaf beetles (Di-
abrotica speciosa and Cerotoma arcuata) from the root of Ceratosanthes
hilariana. Agric. Biol. Chem. 50: 2831-2836.
PETROSKI, R. J., AND L. HAMMACK. 1998. Structure activity relationships of phenyl
alkyl alcohols, phenyl alkyl amines, and cinnamyl alcohol derivates as attrac-
tants for adult corn rootworm (Coleoptera: Chrysomelidae: Diabrotica spp.).
Environ. Entomol. 27: 688-694.
ROEL, A. R., AND ZATARIN. 1989. Eficiencia de iscas a base de ab6bora d'agua, Lage-
naria vulgaris (Cucurbitaceae) tratadas com inseticidas, na atratividade a Di-
abrotica speciosa (Germar, 1824) (Coleoptera: Chrysomelidae). An. Soc.
Entomol. Brasil 18: 213-219.
SAS INSTITUTE.1989. SAS/STAT. User's Guide, Version 6, 4th. ed. SAS Institute, Cary,
SNAPP, 0. I., AND H S. SWINGLE. 1929. Preliminary report on attractants for peach in-
sects. J. Econ. Entomol. 22: 98-101.
VENTURA, M. U., M. ITO, AND R. MONTALVAN. 1996. An attractive trap to capture Di-
abrotica speciosa (Ger.) and Cerotoma arcuata tingomariana Bechyne. An. Soc.
Entomol. Brasil 25: 529-535.
Mankin et al.; Enhancement of Female Caribfly Response 411
EXPOSURE TO MALE PHEROMONES ENHANCES
ANASTREPHA SUSPENSE (DIPTERA: TEPHRITIDAE)
FEMALE RESPONSE TO MALE CALLING SONG
RICHARD W. MANKIN1, ERIK PETERSSON2, NANCY D. EPSKY',
ROBERT R. HEATH3 AND JOHN SIVINSKI'
'United States Department of Agriculture, Agricultural Research Service,
Center for Medical, Agricultural, and Veterinary Entomology,
PO Box 14565, Gainesville, Florida 32604
'Department of Animal Ecology, Evolutionary Biology Centre, Uppsala University,
Norrbyvagen 18 D, S-752 36 Uppsala, Sweden
'United States Department of Agriculture, Agricultural Research Service, Subtropical
Horticultural Research Station, 13601 Old Cutler Rd, Miami, FL 33158
Anastrepha suspense (Loew) females are attracted to traps baited with male pher-
omone and/or broadcast calling song, but a high variability in female responsiveness
has hindered attempts to use such attractants in practical trapping systems. Prior ex-
perience is one factor that may contribute to variability in female responses. To inves-
tigate this possibility, female responses to male calling song were compared after 38-
40-h prior exposure to different combinations of live males, live females, synthetic
pheromone components, and broadcast song. The broadcast song, obtained from a sex-
ually successful male, contained a series of wing-fanning pulse trains averaging 0.31-
s in duration, separated by 0.36-s quiet intervals. Within the pulse trains, the fre-
quency rose quickly from -125 to 148 Hz and then slowly declined to -120 Hz. The
proportions of females responding to the broadcast calling song were greatest when
females were first exposed to live males or pheromone components. These proportions
significantly exceeded 0% difference between the proportions under silent and broad-
casting speakers. The proportion of females that responded after prior exposure to
broadcast song alone was significantly higher than the proportion that responded af-
ter no prior exposure to sexual stimuli, but it did not significantly exceed 0%.
Key Words: calling song, pheromone, Anastrepha suspense, attraction
Hembras deAnastrepha suspense Loew son atraidas a trampas con senuelo de fe-
romonas masculinas y/o transmisi6n de canci6n de llamado, pero una alta variabili-
dad de reacci6n en la hembra ha impedido intentos para desarrollar sistemas
practices de trampas. Un factor que puede contribuir a la variabilidad de reacci6n es
experiencia previa a senales sexuales. Para investigar este efecto, respuestas de hem-
bras a canci6n de llamado fueron comparadas despues de 38 a 40-h de pre-exposici6n
a diferentes combinaciones de machos, hembras, feromona masculina, y transmisi6n
de canci6n. El estimulo de canci6n fue estandarizado por copias concatenadas de una
grabaci6n de 7.2-s de un macho sexualmente exitoso, y contenia trenes de pulsos de
abanicos de ala promediando 0.31-s de duraci6n, separados por intervalos de silencio
de 0.36-s. Dentro de los trenes de pulso, la frecuencia subi6 rdpidamente de -125 a
148 Hz y despues declinaron lentamente a -120 Hz. Las proporciones de hembras que
respondieron a las transmisiones de canci6n de llamada fueron las mayores cuando
las hembras fueron pre-expuestas a machos o feromona masculina. Estas proporcio-
412 Florida Entomologist 83(4) December, 2000
nes significativamente excedieron niveles nulos. Las proporciones de hembras que
respondieron despues de pre-exposici6n solo a transmisi6n de canci6n fueron signifi-
cativamente mayores que las proporciones de hembras respondientes que no habian
tenido pre-exposici6n al estimulo sexual, pero en ningun grupo las proporciones exce-
dieron niveles nulos.
Sexual courtship in the Tephritid pest,Anastrepha suspense (Loew) ('caribfly'), in-
volves male-produced pheromonal (Nation 1972), acoustic (Webb et al. 1976), and vi-
sual signals (Burk 1981) that are attractive to females and consequently have
potential uses in trapping and monitoring programs. These signals are usually pre-
sented to females by males that aggregate in leks on larval host plants in late after-
noon (Burk 1983, Norrbom & Kim 1988). Males compete for single-leaf territories on
which they emit pheromone, produce a calling song generated by repeated bursts
(trains) of wing-fanning pulses, and semaphore with patterned wings (e.g., Sivinski &
Burk 1989, Aluja et al. 2000). When a female lands on his leaf, he approaches and
mounts her if she permits, producing an intense precopulatory song while attempting
to engage her genitalia (Sivinski et al. 1984). The precopulatory song is continuous
rather than pulsed, and is more intense and higher in frequency than the calling song
(Webb et al. 1984). As in most systems where courtship occurs in aggregations (Alex-
ander et al. 1997), males vary considerably in their mating success.
Bioassays designed to identify the signal characteristics that distinguish sexually
successful males from nonmaters have yielded ambiguous results. For example, ini-
tial laboratory studies comparing female responses to different combinations of pher-
omone and calling song suggested that calling song combined with pheromone was
more attractive than pheromone alone (Webb 1973, Chambers 1975). In field-cage
studies, however, statistically significant numbers of females were captured in traps
baited with live males, pheromone alone, and broadcast calling song alone, but not in
traps with a combination of broadcast song and pheromone (Webb et al. 1983).
Additional ambiguities have appeared in studies comparing responses to songs
that contained systematic differences in acoustic parameters. Sexually successful car-
ibfly males produce songs generated from precisely featured bursts of wing fanning
pulses (example in Fig. 1). The bursts have a mean pulse-train duration (PTD) of
-0.32 s and are separated by -0.34-s pulse train intervals (PTI) (Sivinski & Webb
1986). The mean inter-pulse interval (IPI) is ~7.14 ms, corresponding to a frequency
of ~140 Hz (Sivinski & Webb 1986). Changes in the magnitudes of one or more of these
parameters often result in reduced responsiveness of females. For example, Sivinski
et al. (1984) found that female caribflies failed to respond to songs recorded from con-
specific males with pulse trains of typical frequency and duration (~140-Hz pulses,
0.273-s PTD) but atypically long PTI (1.115s). Female caribflies were also unrespon-
sive to male precopulatory song and to song produced by a male Queensland fruit fly,
Bactrocera neohumeralis Hardy. The Queensland fruit fly produces song with the cor-
rect PTI but a higher frequency (-379 Hz) and shorter PTD (-0.12 s) (Sivinski et al.
1984). Such results suggest that the songs most attractive to female caribflies are
those with low frequencies, long PTDs, and short PTIs. Indeed, Burk & Webb (1983)
had reported that females mate preferentially with larger males, and that larger
males produced lower frequency songs with shorter PTIs than smaller males. Other
studies, however, did not always find that larger males produce song with shorter
PTIs (Webb et al. 1984).
Mankin et al.; Enhancement of Female Caribfly Response 413
#1 2 3 4 5
0.00 025 0.50 0.75 1.00
Fig. 1. Example of 2 pulse trains in a recording of male caribfly calling song, with
inset showing 6 individual pulses at the start of 1st train. The horizontal axis shows
time in seconds. The vertical axis shows the microphone signal on a relative scale (see
text for amplification details): PTD, pulse train duration; PTI, pulse train interval;
IPI, interpulse interval; horizontal axis shows time (s); vertical axis shows micro-
phone signal in relative scale.
Some of this ambiguity may have resulted from a lack of understanding of poten-
tially important features of caribfly pulse trains. Historically, calling songs were de-
scribed simply by their mean frequencies (e.g., Webb et al. 1983, Sivinski & Webb
1986). However, Webb et al. (1987) and others observed later that the frequency is not
constant within pulse trains, but decreases toward the end of each train. The magni-
tudes of these frequency changes have not been characterized and their effects on fe-
male response are unknown.
Additional ambiguity may be explained by variability in the levels of female re-
sponsiveness (cf Searcy & Andersson, 1986). Prior experience with sexual signals is
one potential contributor to this variability. Caribfly males are known to change their
calling patterns in the presence of females and other males (Sivinski & Webb 1986).
However, the effect of prior experience on female caribfly behavior has not been inves-
The objective of this study was to investigate the effect of prior exposure to sexual
stimuli on female caribfly responsiveness to broadcast song. We conducted a series of
bioassays during which we exposed virgin females to different combinations of pher-
omone and/or male calling song prior to experiments, and then assessed their re-
sponses to play-back of precisely characterized song recorded from a sexually
414 Florida Entomologist 83(4) December, 2000
MATERIALS AND METHODS
Caribflies used in this study were obtained as pupae from the Florida Department
of Agriculture, Division of Plant Industry in Gainesville, Florida. After eclosion, adult
flies were given water and a 3:1 mixture of refined cane sugar and hydrolyzed
brewer's yeast. They were maintained in a laboratory with a photoperiod of 12:12 h
(L:D) at room temperature and ambient humidity. Adult flies were sorted by sex 3-4
days after eclosion, placed in cubic screen cages (30-cm per side), and the females
thereafter were kept in a separate "female room". Because most males do not signal
sexually until after 5-7 d (Sivinski 1994), the sorted females were unlikely to have
been exposed to significant amounts of male pheromones.
Sexually mature females (10-18 d old) were used for all experiments. They had no
exposure to adult live males from time of sorting until use in an experiment unless
Male Calling Song
The acoustic signals were generated by concatenating multiple copies of a 7.2-s
segment of song from a sexually successful male (Fig. 2) onto an endless loop tape.
Webb et al. (1983) used this same song segment to produce a continuous-loop record-
ing that successfully captured female caribflies when it was broadcast from a trap in
a field cage. The first two pulse trains of the signal are shown on smaller time scales
in Figure 1.
Webb et al. (1976, 1983) measured temporal patterns and the mean frequencies
and Sound Pressure Levels (SPLs) ofA. suspense calling song bursts, but frequency
patterns and SPLs within bursts could not be measured using the technology then
available. For this report, we analyzed the dynamic features of the song bursts using
a Bruel and Kjaer (B & K) model 4145 microphone, a model 2639 preamplifier, and
a model 2610 measuring amplifier (Mankin 1994). Signals were amplified 20-40 dB
(where dB = 20 Logl(V, /VT,), and V,, V, are the amplifier output and input volt-
age levels, respectively) and digitized at 25 kHz using a 12-bit MetraByte (Keithley/
MetraByte Inc., Taunton, MA) DAS-16G A/D converter installed in a Pentium 350-
mHz microcomputer. The digitized signals were analyzed with custom-written soft-
ware (Mankin 1994, Mankin et al. 1996a). The customized software located the peak
of each wingbeat pulse, marked its time within the recording, and measured the in-
ter-pulse interval (IPI, Fig. 1). A pulse train was identified as a series of unbroken
pulses separated by an IPI of no more than 20 ms (approximately 3 typical pulses).
A custom-written subroutine noted the order of each pulse within the train and cal-
culated the instantaneous frequency (1/IPI). Another subroutine noted the begin-
ning and end of each pulse train for calculations of pulse train duration (PTD) and
The mean frequency within the pulse train was calculated as an average for each
IPI measurement in the train, based on its order number from the beginning of the
train (e.g. IPI #'s 1-5 in Fig. 1). The mean frequency at the 5th IPI for example, was
the average of all values of 1/IPI between the 5th and 6th pulse.
Sound Pressure Levels were calibrated as in Mankin et al. (1996b), and the
speaker output was adjusted to produce 55 dB SPL (relative to 20 (Pa) mean signal
level at a distance of 12 mm. This is the level used in previously successful trapping
Mankin et al.; Enhancement of Female Caribfly Response 415
1.8 3.6 5.4
Fig. 2. Oscillograph of 7.2-s repeated section of male caribfly calling song used in
female attraction bioassay. Horizontal and vertical axes, same as in Figure 1.
studies (Webb et al. 1983). Such sounds are audible to humans over distances of 2-4
m in the laboratory, but the range of detectability by female caribflies has not been
Bioassay Arenas and Response Measurements
Separate exposure treatments were conducted in a laboratory and a wind tunnel.
The effect of prior exposure to live males was tested by moving some of the caged vir-
gin females into a 1.8 x 1.8 x 2.5 m "male room" containing several hundred caged
males of all ages. These females were adjacent to, but physically isolated from males,
and were exposed to sight, sound, and pheromone from males. After 38-40 h exposure
in the "male room", 20 females were moved to a (20 x 20 x 20 cm) screen cage in a sep-
arate room for the acoustic bioassay.
In the acoustic bioassays, 2 pairs of monaural headphones (Realistic or Archer,
both including the foam pads) were placed on top of the cage, facing down. One pair
was silent and the other was connected to a recorder (Realistic CTR-62 or CTR-66)
playing the endless-loop tape. The signals were broadcast during the peak of the daily
courtship-signaling period, 8.5 h after the start of the 12-h photophase. The females
standing directly under each speaker were counted every five minutes, a total of 13
times for each 1-h replicate. The locations of silent and broadcasting speakers were al-
ternated at least once during each trial.
Attraction to male song was measured as the mean of the 13 measurements of the
difference between the proportions of females under the broadcasting and the silent
speakers. The responses of females exposed in the "male room" prior to assays were
416 Florida Entomologist 83(4) December, 2000
compared with the responses of females kept continuously in the "female room". Thir-
teen replications were done in the "female room" tests and seven replications in the
"male room" tests.
In the second experiment, females were exposed in a wind tunnel (Heath et al.
1993) to different combinations of male courtship signals for 38-40 h before the acous-
tic bioassay. The combinations were filtered air in a "clean tunnel", air from the "male
room", "live males" (10 males in a cage, 1.1 m upwind), "pheromone" (exposure to pu-
tative synthetic pheromone), or "sound only" (calling song broadcast continuously
from speakers on top of the cage). The putative synthetic pheromone was composed of
~5% ocimene; ~2% nonenols ((Z)-3-nonen-l-ol and (Z,Z)-3,6-nonadien-1-ol);~10% sus-
pensolide ((E,E)-4,8-dimethyl-3,8-decadien-10-olide); -5% E,E-a-farnesene; -35% B-
bisabolene; ~10% anastrephin (trans-hexahydro-trans-4,7a-dimethyl-4-vinyl-2-(3H)-
benzofuranone); and ~33% epi-anastrephin (trans-hexahydro-cis-4,7a-dimethyl-4-vi-
nyl-2-(3H)-benzofuranone), which approximated the ratio of components that are re-
leased under natural light condition in late afternoon (Heath et al. 1993). Synthetic
components were formulated in glass capillaries (ocimene and nonenols) and on rub-
ber septa (remaining components) using protocols reported previously (Weatherston
et al. 1985a, 1985b, and Heath et al. 1986, respectively). The putative synthetic pher-
omone was formulated to release ~900 ng per h or the equivalent of release from 10
male caribflies in late afternoon (Heath et al. 1993).
The cages used for the wind tunnel treatments had solid sides (14.2 x 30 cm) and
bottoms (13.3 x 30 cm) with single-screen covered circular openings (9 cm diameter)
cut into the front and back pieces (13.3 x 14.2 cm). Two screen-covered circular open-
ings were cut into the top for placement of the speakers. The foam pads were removed
to prevent oviposition on the speakers in the "sound only" experiment. Attractiveness
of broadcast songs after the wind tunnel treatment was measured as in the previous
experiment. This experiment had 10 replications.
Behavioral Response Analysis
Nonparametric statistical analyses were used due to nonnormal frequency distri-
butions of the raw data. Dunn's multiple comparison (based on Kruskal-Wallis rank
sums) (Hollander & Wolfe, 1973) was used to compare the treatments in the second
Female Acoustic Attraction Response
Male caribfly calling song elicited a detectable attraction response when females
were in the presence of synthetic pheromone or had been exposed to males. In the first
experiment, females that had been kept in the male-room responded at significant
levels to calling song, but the responses of females that had been kept in the female
room were not significantly different from zero (Fig. 3). The two treatments differed
in the proportion of females under broadcasting speakers (z = 2.774, p < 0.005, Mann-
Whitney U-test). In the second experiment, females that had been kept in the male
room, exposed to live males, or exposed to synthetic pheromone in the wind tunnel
were attracted to male calling song (Fig. 4). Females that had been kept in the clean
tunnel or exposed only to male calling song showed no attraction toward broadcast
Mankin et al.; Enhancement of Female Caribfly Response 417
0.02 I a
Female cage location
Fig. 3. Comparison of responses to calling song by females kept only with other fe-
males or kept in a room with caged males. The vertical axis shows the difference in the
proportion of females at the broadcasting and the silent speakers, with positive values
indicating a greater proportion of females observed under the broadcasting speakers.
Medians are indicated by the larger circles and vertical lines indicate the 0-25th and
75-100th percentiles. Smaller dots (right) show the measured proportion differences.
Filled median circles indicate that the median differs significantly from zero accord-
ing to the sign test (p < 0.016). The numbers underneath list the number of cages (rep-
licates) in each treatment.
songs. The "clean tunnel" and "male room" tests in the second experiment were essen-
tially equivalent to the "female room" and the "male room" tests in the first experi-
ment. The results of the two experiments were comparable.
Dynamic Features of Male Calling Song
The 7.2-s segment of male calling song broadcast in these recordings (Figs. 1 and
2) consisted of 11 pulse trains with a mean frequency + Standard Error (SE) of 140.38
+ 3.86 Hz. The pulse trains had a mean duration (PTD) of 0.31 + 0.02 s, separated by
silent intervals (PTI) of 0.36 + 0.05 s. The frequency varied within the pulse train. On
average, the train began with pulses of ~125 Hz (Fig. 5). The frequency increased rap-
idly to a maximum of ~148 Hz by the 5th IPI (~36 ms into the train), and then grad-
ually declined to below 120 Hz by the 30th IPI (-225 ms into the train).
-- --- -----------
418 Florida Entomologist 83(4) December, 2000
r 0.04 -
J 0.02- c go c
S1 b 0;* b
a 0.00 _- _--- __ "
0 -0.02 I t
J -0.04 -
CLEANTUNNEL LIVEMALES MALEROOM PHEROMONE SOUNDONLY
Female cage location
Fig. 4. Comparison of female responses to male calling song after previous expo-
sure in a wind tunnel to clean air, caged males, putative synthetic male pheromone,
or calling song. Positive values indicate that proportionally more females were ob-
served under the broadcasting speakers. Medians are indicated by the larger circles
and vertical lines indicate the 0-25th and 75-100th percentiles. Smaller dots (right)
show the measured proportion differences. Filled median circles indicate that the me-
dian differed significantly from zero according to the sign test. Treatments with the
same letter had medians that were not significantly different according to Dunn's
multiple comparison (experimentwise error rate = 0.09).
The result that females exposed to male pheromone before a bioassay were more
responsive to calling song than unexposed females is consistent with at least two al-
ternative hypotheses. One is that exposure to pheromone is necessary to trigger a re-
sponse to calling song. A second is that experience with the male courtship repertoire
may increase the responsiveness to subsequent courtship signals. The first hypothesis
is plausible because, in nature, (cf. Sivinski & Burk, 1989) females usually smell the
males before they hear or see them. However, Sivinski et al. (1984) found that virgin
females with no experience of male pheromone increased their levels of activity when
they heard broadcast male calling song. In this study, the responses of females that
had been exposed only to broadcast song were significantly greater than the responses
of females that had never been exposed to any sexual stimuli. Such results suggest
that acoustic attraction can occur without exposure to pheromone, but the attraction
is weaker than that to pheromone.
The results here and those of Sivinski et al. (1984) are most consistent with the
second hypothesis, that prior experience of the female with the male courtship reper-
toire increases female responses to newly encountered courtship stimuli. This effect
Mankin et al.; Enhancement of Female Caribfly Response 419
1 1 0 i . ..i -
0 18 36
No. in Burst
Fig. 5. Temporal pattern of wing-beat frequency within calling song pulse trains of
a sexually successful caribfly male. The vertical axis indicates the wing-beat fre-
quency (1/IPI) in s1'. The horizontal axis indicates the number of pulses from the be-
ginning of the pulse train, counted as in the inset in Figure 1.
might be similar to one observed in wind tunnel experiments with female parasitoid
attraction to host larvae (e.g. Drost et al. 1986; Eller et al. 1988; Turlings et al. 1989).
If the second hypothesis is correct, however, the nonsignificant result for females that
had been exposed to calling song but not pheromone suggests that exposure to pher-
omone has a greater effect on subsequent responsiveness.
Because the female responses to the broadcast calling song in these bioassays were
too low for practical applications, the main benefit of this study is some insight for im-
provements in the design of fruit fly acoustic bioassays. First, female responsiveness
to acoustic signals can be increased by exposure to pheromone, either before or during
the acoustic testing. Second, the broadcasting of acoustic stimuli by speakers may
present an inadequate stimulus to female fruit flies. The adequacy of acoustic stimuli
for attraction has been a problem in many other insect bioassays as well (Searcy &
The response in this bioassay was low despite the use of song generated by a sex-
ually successful male. The mean frequency is lower than the 149-Hz average of sexu-
ally successful males in Webb et al (1984), but a lower frequency may correlate with
larger size and improved mating propensity (Burk & Webb 1983). The PTI and PTD
are comparable to measurements from other sexually successful males (Webb et al.
1984). Consequently, it is not likely that the signal pattern itself was deficient but
420 Florida Entomologist 83(4) December, 2000
some other stimulus feature. One potential contributor to the low female responsive-
ness was that the sound was produced by a speaker rather than a vibrating wing. The
speaker generates a signal of larger spatial extent and lower air velocity than the vi-
brating wing. In addition, the vibrating wing has visual components that have not yet
been demonstrated as affecting caribfly female mate choice (Aluja et al. 2000) but
have been shown to affect mate choice in other Diptera (e.g., Lunau 1992). The con-
tinued development of new acoustic signal analysis and signal presentation capabili-
ties may improve our future ability to elicit an attraction response of female fruit flies
to synthetic courtship stimuli.
We thank Everett Foreman for preparing endless-loop tapes and calibrating the
speakers, and Barbara Dueben for formulating the synthetic pheromone. We also
thank Christer Lofstedt, Tim Forrest, Bo Svensson, James Nation, Phil Taylor, and
Amos Mizrach for valuable points on earlier drafts of the manuscript. This study was
performed during E. Petersson's post-doctoral stay at USDA-ARS, Gainesville, Flor-
ida, on the courtesy of the Swedish Natural Science Research Council.
ALEXANDER, R. D., D. C., MARSHALL, AND J. R. COOLEY. 1997. Evolutionary perspec-
tives on insect mating, pp. 4-31, in Mating systems in insects and arachnids, J.
Choe & B. Crespl, [eds.]. Cambridge Univ. Press, Cambridge.
ALUJA, M., J. PINERO, I. JACOME, F. DIAZ-FLEISCHER, AND J. SIVINSKI. 2000. Behavior
of flies in the genus Anastrepha (Trypetinae: Toxotrypanini), pp. 375-406, in
Fruit Flies (Tephritidae): phylogeny and evolution of behavior, M. Aluja and A.
L. Norrbom [eds], CRC Press, Boca Raton, FL.
BURK, T. 1983. Behavioral ecology of mating in the Caribbean fruit fly Anastrepha
suspense (Loew). Florida Entomol. 66: 330-344.
BURK, T., AND J. C. WEBB. 1983. Effect of male size on calling propensity, song param-
eters, and mating success in Caribbean fruit flies (Anastrepha suspense
(Loew)). Ann. Entomol. Soc. Am. 76: 678-682.
CHAMBERS, D. L. 1975. Quality in mass-produced insects: Definition and evaluation,
pp. 19-32, in Controlling fruit flies by the sterile-insect technique. Interna-
tional Atomic Energy Agency, Vienna, 175 pp.
DROST, Y. C., W. J. LEWIS, P. 0. ZANEN, AND M. A. KELLER. 1986. Beneficial arthropod
behavior mediated by airborne semiochemicals. I. Flight behavior and influ-
ence of preflight handling of Microplitis croceipes (Cresson). J. Chem. Ecol. 12:
ELLER, F. J., J. H. TUMLINSON, AND W. J. LEWIS. 1988. Beneficial arthropod behavior
mediated by airborne semiochemicals: Source of volatiles mediating the host-
location flight behavior ofMicroplitis croceipes (Cresson) (Hymenoptera: Bra-
conidae), a parasitoid of Heliothis zea (Boddie) (Lepidoptera: Noctuidae). Envi-
ron. Entomol. 17: 745-753.
NORRBOM, A. L., AND K. C. KIM. 1988. A list of the reported host plants of the species
ofAnastrepha (Diptera: Tephritidae). USDA-APHIS-PPQ, AHPIS 81-52.
HEATH, R. R., A. MANUKIAN, N. D. EPSKY, J. SIVINSKI, C. 0. CALKINS, AND P. J.
LANDOLT. 1993. A bioassay system for collecting volatiles while simultaneously
attracting tephritid fruit flies. J. Chem. Ecol. 19: 2395-2410.
HEATH, R. R., P. E. A. TEAL, J. H. TUMLINSON, AND L. J. MENGELKOCH. 1986. Predic-
tion of release ratios of multicomponent pheromones from rubber septa. J.
Chem. Ecol. 12: 2133-2143.
HOLLANDER, M., AND D. A. WOLFE. 1973. Nonparametric statistical methods. John
Wiley & Sons, New York.
Mankin et al.; Enhancement of Female Caribfly Response 421
LUNAU, K. 1992. Mating behavior in the long-legged fly Poecilobothrus nobilitatus L.
(Diptera, Dolichopodidae): courtship behaviour, male signalling and mating
success. Zool. Beitr. 34: 465-479.
MANKIN, R. W. 1994. Acoustical detection ofAedes taeniorhynchus swarms and emer-
gence exoduses in remote salt marshes. J. Am. Mosq. Cont. Assoc. 10: 302-308
MANKIN, R. W., A. MALAVASI, AND C. AQUINO. 1996a. Acoustical comparisons of call-
ing songs from Anastrepha species in Brazil, pp. 37-42. In Fruit Fly Pests: A
World Assessment of Their Biology and Management, B. A. McPheron and G. J.
Steck, [eds]. St. Lucie Press, Delray Beach, FL.
MANKIN, R. W., D. SHUMAN, AND J. A. COFFELT. 1996b. Noise shielding of acoustic de-
vices for insect detection. J. Econ. Entomol. 89: 1301-1308.
NATION, J. L. 1972. Courtship behavior and evidence for a sex attractant in the male
Caribbean fruit fly, Anastrepha suspense. Ann. Entomol. Soc. Am. 65: 1364-
SEARCY, W. A., AND M. ANDERSSON. 1986. Sexual selection and the evolution of song.
Ann. Rev. Ecol. Syst. 17: 507-533.
SIVINSKI, J. 1994. Longevity in the Caribbean fruit fly: effects of sex, strain and sexual
experiences. Florida Entomol. 76: 635-644.
SIVINSKI, J., AND T. BURK. 1989. Reproductive and mating behaviour, pp. 343-351 in
A. S. Robinson and G. Hooper, eds., Fruit flies, their biology, natural enemies
and control Elsevier Sci. Publ., Amsterdam.
SIVINSKI, J., AND J. C. WEBB. 1986. Changes in a Caribbean fruit fly acoustic signal
with social situation (Diptera: Tephritidae). Ann. Entomol. Soc. Am. 79: 146-
SIVINSKI, J., T. BURK, AND J. C. WEBB. 1984. Acoustic courtship signals in the Carib-
bean fruit fly,Anastrepha suspense (Loew). Anim. Behav. 32: 1011-1016.
SIVINSKI, J., N. D. EPSKY, AND R. R. HEATH. 1994. Pheromone deposition on leaf ter-
ritories by male Caribbean fruit flies, Anastrepha suspense (Loew). J. Insect
Behav. 7: 43-52.
TURLINGS, T. C. J., J. H. TUMLINSON, W. J. LEWIS, AND L. E. M. VET. 1989. Beneficial
arthropod behavior mediated by airborne semiochemicals. VIII. Learning of
host-related odors induced by a brief contact experience with host by-products
in Cotesia marginiventris (Cresson); a generalist larval parasitoid. J. Insect Be-
hav. 2: 217-225.
WEATHERSTON, I. D. MILLER, AND L. DOHSE. 1985a. Capillaries as controlled release
devices for insect pheromones and other volatile substances-a reevaluation:
Part I. Kinetics and development of predictive model for glass capillaries. J.
Chem. Ecol. 11: 953-965.
WEATHERSTON, I., D. MILLER, AND L. DOHSE. 1985b. Capillaries as controlled release
devices for insect pheromones and other volatile substances a reevaluation.
Part II. Predicting release rate from celcon and teflon capillaries. J. Chem. Ecol.
WEBB, J. C. 1973. Analysis and identification of specialized sounds possibly used by
the Caribbean fruit fly Anastrepha suspense for communication purposes.
Ph.D. Dissertation, University of Tennessee, Knoxville.
WEBB, J. C., T. BURK, AND J. SIVINSKI. 1983. Attraction of female Caribbean fruit flies,
Anastrepha suspense (Diptera: Tephritidae), to the presence of male and male-
produced stimuli in field cages. Ann. Entomol. Soc. Anm 76: 996-998.
WEBB, J. C., J. L. SHARP, D. L. CHAMBERS, J. J. McDOW, AND J. C. BENNER. 1976. The
analysis and identification of sounds produced by the male Caribbean fruit fly,
Anastrepha suspense (Loew). Ann. Entomol. Soc. Am. 69: 415-420.
WEBB, J. C., J. SIVINSKI, AND C. LITZKOW. 1984. Acoustical behavior and sexual suc-
cess in the Caribbean fruit fly, Anastrepha suspense (Loew) (Diptera: Tephriti-
dae). Environ. Entomol. 13: 650-656.
WEBB, J. C., J. SIVINSKI, AND B. J. SMITTLE. 1987. Acoustical courtship and sexual
success in irradiated Caribbean fruit flies (Anastrepha suspense) (Diptera: Te-
phritidae). Florida Entomol. 70: 104-109.
422 Florida Entomologist 83(4) December, 2000
PARASITOID-HOST MATCHING BETWEEN THE LITTLE
DECAPITATING FLY PSEUDACTEON CURVATUS
FROM LAS FLORES, ARGENTINA AND THE
BLACK FIRE ANT SOLENOPSIS RICHTERI
SANFORD D. PORTER1 AND JUAN A. BRIANO2
1USDA-ARS, Center for Medical, Agricultural and Veterinary Entomology
PO Box 14565, Gainesville, FL 32604
2USDA-ARS, South American Biological Control Laboratory, Bolivar 1559 (1686)
Hurlingham, Buenos Aires Province, Argentina
Matching biotypes of potential biocontrol agents to target host populations can
greatly improve the effectiveness of control. This study was designed to determine if
the fly Pseudacteon curvatus Borgmeier from Las Flores, Buenos Aires Province, Ar-
gentina prefers its natural host, the black fire ant, Solenopsis richteri Forel. We found
that P. curvatus strongly preferred S. richteri from Argentina, S. richteri from the
United States, and hybrid (S. richteri x S. invicta) fire ants from the United States
when each was tested against S. invicta from the United States. The time to pupation
of developing parasitoids was 10% and 21% longer in hybrid and red fire ants than in
black fire ants. Parasitism rates, however, were not significantly different among
these ant hosts in no-choice parasitism tests.
Key Words: Diptera, Phoridae, Hymenoptera, Formicidae, Solenopsis invicta, host
preference, hybrid fire ants, biocontrol, biotype
Hacer corresponder a biotipos de agents potenciales de control biol6gico con po-
blaciones huespedes puede, en gran media, mejorar la eficiencia en el control. Este
studio fue disenado para determinar si la mosca Pseudacteon curvatus Borgmeier de
Las Flores, Provincia de Buenos Aires, Argentina, prefiere a su huesped natural, la
"hormiga brava" negra, Solenopsis richteri Forel. Encontramos que P. curvatus prefi-
ri6 marcadamente a S. richteri de Argentina, a S. richteri de los EE.UU. y a la forma
hibrida (S. richteri x S. invicta) de los EE.UU. cuando fue comparada con S. invicta de
los EE.UU. El period hasta pupaci6n de los parasitoides en desarrollo fue 10% y 21%
mas largo en la forma hibrida y en la "hormiga brava" roja que en la "hormiga brava"
negra. Las proporciones de parasitismo, sin embargo, no fueron significativamente di-
ferentes entire los huespedes en pruebas de parasitismo de no-elecci6n.
Two exotic fire ants have become established in the United States, the black im-
ported fire ant, Solenopsis richteri Forel, and the red imported fire ant, Solenopsis in-
victa Buren (Trager 1991). The black imported fire ant is found in northern
Mississippi and Alabama where it occupies about 30,000 km2(Shoemaker et al. 1994).
The red imported fire ant is found in 11 southeastern states from North Carolina
through Texas and occupies about 1,100,000 km2 (Callcott & Collins 1996). Between
these two species is a broad band of hybridization from the Mississippi River to At-
lanta, GA (Shoemaker et al. 1994). Hybrid fire ants occupy about 130,000 km2. De-
Porter & Briano: Fire Ant Decapitating Fly
spite this broad zone of hybridization, red and black imported fire ants are still
considered separate species because they apparently do not hybridize in South Amer-
ica (Ross & Trager 1990).
Pseudacteon curvatus Borgmeier is a small decapitating fly from South America
that parasitizes Solenopsis fire ant workers (Porter 1998). This species was released
in several states in the spring of 2000 as a potential biocontrol agent for red and black
imported fire ants in the United States (unpublished data). In South America, P. cur-
vatus occurs over a very wide range from Sao Paulo, Brazil westward into Mato
Grosso do Sul, Brazil and southward to Buenos Aires Province, Argentina (Borgmeier
1925; SDP-unpublished data). Over this range, P. curvatus is known to parasitize at
least three species of South American fire ants: Solenopsis saevissima (F. Smith), S.
invicta, and S. richteri (SDP-unpublished data).
Host-specificity tests in the United States (Porter 2000) demonstrated that P. cur-
vatus flies from Argentina strongly prefer red imported fire ants over the native fire
ants Solenopsis geminata (Fab.) and Solenopsis xyloni (McCook). This preference is
not surprising because P. curvatus is not a natural parasite of either S. geminata or
S. xyloni. Forced laboratory rearing tests showed that S. geminata and S. xyloni are
both very poor hosts for P. curvatus (Porter 2000).
Matching parasitoid biotypes to target host populations can greatly improve the
success of biocontrol programs (Van Driesche and Bellows 1996, p. 149). P. curvatus
flies from Buenos Aires Province, Argentina normally parasitize the black fire ant S.
richteri. The objective of this study was to determine if P. curvatus flies from Buenos
Aires Province are better adapted to S. richteri (their normal host) than the red fire
ant, S. invicta (a host in other parts of South America). Results of this study will help
us decide where P. curvatus should be released in the United States.
MATERIALS AND METHODS
P. curvatus flies used in this study were originally collected from El Toro Ranch
southeast of Las Flores, Buenos Aires Province, Argentina in March 1997 (Porter
2000). A few flies from the same location were added to the lab colony several times
up to December 1998.
To examine P. curvatus preferences for S. richteri, S. invicta, and hybrid fire ants,
3-hour old and 1-day old flies were introduced into white plastic trays (42 x 28 x 15 cm)
with screened vents and tight-fitting glass lids (described in detail by Porter 2000). We
used both 3-hour and 1-day old flies to produce an age mixture similar to what might
occur in the field. In the bottom of each tray, were two parallel chambers (7 x 30 x 5
cm, 1 x w x h) for two kinds of ants. Ants were contained in the two bottom chambers
by coating the sides with Fluon CICI, Wilmington, DE).
A small opaque inverted cup (4 cm diameter) was placed on the bottom of each of
the two small parallel chambers. These cups were moved back and forth from one end
of a chamber to the other with a long aspirator arm (Porter and Alonso 1999) each
time most of the ants had crawled under a cup to hide. This procedure kept the ants
in both sides trailing continuously from one end of a bottom chamber to the other so
the flies always had an opportunity to attack workers of either type of ant.
We used 7 colonies of S. richteri from Las Flores, Buenos Aires Province, Argen-
tina, 9 colonies of S. richteri from northeastern Mississippi (Tupelo 4 colonies,
Booneville 3, Corinth 1, Mayhew 1), and 7 colonies of hybrid fire ants from around
Starkville, MS (USDA Lab 4, Mayhew 3). The identities ofS. richteri and hybrid fire
ants from Mississippi were confirmed by gas chromatography (Vander Meer et al.
1985). For each trial S. richteri and hybrid fire ants were paired with similar-sized S.
invicta workers from Gainesville, FL. Different colonies of each kind of ant were used
424 Florida Entomologist 83(4) December, 2000
for each trial to assure that results were not due to differences in the attractiveness
of individual colonies. Tests with S. richteri fire ants from Argentina were conducted
in January 1999. Tests with hybrid fire ants and S. richteri from the United States
were conducted in June 1999. Tests for all three kinds of ants were run 1-3 weeks after
colonies were collected in the field. Voucher specimens have been deposited in the
Florida Museum of Arthropods, Gainesville, Florida, USA.
Each test run lasted about 3 h and used 14-18 female flies with an equivalent num-
ber of males. Test ants contained 0.25 g of workers (~400) and 0.5 g of brood. The trays
were inspected every 10 min and the number of female flies hovering in attack mode
over each species of ant was recorded by visual count. Females considered in attack
mode hovered 3-10 mm above the ants and oriented to their movements. Males of this
species are not attracted to the ants.
To determine if P. curvatus flies were equally successful in parasitizing black, hy-
brid, and red fire ants, we conducted a series of no-choice parasitism tests. The trays
used in these tests contained a single solid bottom covered with moistened plaster as
described by Porter (2000). Timer motors were used to automatically raise an inverted
cup in one end of each tray while lowering a cup at the other end of each tray. This
caused the test ants to continuously trail back and forth between the two cups. Timer
motors were set to run for 8 h a day (10:00 to 18:00 h).
We conducted 6 trials each with: S. richteri from northeastern Mississippi (Corinth
- 1 colony, Booneville -2, Tupelo -3), hybrid fire ants from Starkville, MS (3 colonies)
and Mayhew, MS (3 colonies), and S. invicta from Gainesville, FL (6 colonies). All col-
onies were collected in June 1999 and used 1-2 weeks after collection. Tests contained
0.5 g of workers (~800) and 1.0 g of brood. Different colonies were used for each test
replicate. We used mostly the same colonies for the no-choice parasitism tests as we
did for the paired preference tests.
Fifteen to sixteen female flies and an equivalent number of males were added to all
no-choice trials on day 1. Tests lasted 2 days. P. curvatus adults usually only live a day
or two in the attack trays; consequently, most of the flies were dead by the end of the
trials. Inactive flies usually live several days longer in the lab. Longevity in the field
is unknown, but it is likely to be intermediate between inactive flies and flies in the
attack trays. At the end of each trial, worker ants were transferred into small boxes
(20 x 2 x 5 cm) with tight-fitting vented lids. Ants were fed fresh sugar water every 3-
4 days. We inspected the head capsules of dead workers for fly pupae every 1-2 days
for a period of 30 days.
When given a choice in paired tests, about 70% of the P. curvatus females preferred
to attack black fire ants or hybrid fire ants over red fire ants (Fig. 1). We found highly
significant differences in the number of attacking flies for each of the following pairs
using paired t-tests: S. invicta versus S. richteri from Argentina (t = 4.95, d.f. = 6, P =
0.0026), S. invicta versus S. richteri from Mississippi (t = 3.48, d.f. -= 8, P -= 0.0083) and
S. invicta versus hybrid fire ants from Mississippi (t = 7.11, d.f. = 6, P = 0.0004). How-
ever, no significant differences were found between the number of flies preferring S.
richteri from Argentina, S. richteri from Mississippi, or hybrid workers (S. richteri x
S. invicta) from Mississippi (all tested against S. invicta from Florida; ANOVA, F =
0.81, d.f. = 2,20, P = 0.46).
Once a fly began attacking workers in the choice tests, the average number of ovi-
position strikes per 15 seconds was respectively 0.67 + 0.05, 0.87 + 0.13, 1.15 + 0.14,
and 1.57 + 0.26 for S. richteri workers from Argentina, S. invicta from Florida, S.rich-
teri from the United States, and hybrid fire ants. The attack rate for hybrid fire ants
Porter & Briano: Fire Ant Decapitating Fly
60 9 -
S. richteri S. invicta S. richteri S. invicta Hybrid S. invicta
(Arg.) (U.S.) (U.S.) (U.S.) (U.S.) (U.S.)
Fig. 1. The percent of female Pseudacteon curvatus decapitating flies from Las
Flores Argentina preferring to attack S. richteri from Argentina, S. richteri from Mis-
sissippi, or hybrid (S. richteri x S. invicta) fire ants from Mississippi, each in paired
tests with S. invicta from Gainesville, Florida. The number of trials is indicated below
each pair of bars. Error bars indicate standard errors of the mean.
was significantly higher (Fisher's PLSD, P < 0.004; 1-way ANOVA) than rates for ei-
ther S. invicta or S. richteri from Argentina. Other pairwise comparisons were not sta-
tistically significant. The biological basis and importance of this pattern is not clear.
While P. curvatus strongly preferred black and hybrid fire ants when given a
choice, significant differences were not found in the number of pupae produced in the
no-choice parasitism tests (Table 1, ANOVA, F = 0.42, df = 2,15, P = 0.66). However,
the mean time to pupation varied significantly among hosts (Table 1, ANOVA, F =
12.6, df = 2,13, P = 0.0009, data were log transformed to equalize variance, two colo-
nies were deleted [S. invicta 1, S. richteri 1] because they each produced less than
40 pupae). The development time to pupation was 21% longer in S. invicta than in S.
richteri and 10% longer in hybrid fire ants than S. richteri (Table 1). The mean vari-
ability of pupation time (as measured by SD) was also significantly larger for flies de-
veloping in S. invicta and hybrid fire ants than in S. richteri (Table 1;ANOVA, F = 5.7,
df = 2,13 P= 0.017).
P. curvatus from Las Flores, Argentina appears to have evolved a specialized rela-
tionship with S. richteri, its natural host. Specifically, these flies demonstrated a
strong preference for S. richteri and hybrid fire ants over S. invicta (Fig. 1). The fact
that hybrid workers were apparently as attractive as S. richteri workers suggests that
that hybrid workers were apparently as attractive as S. richteri workers suggests that
426 Florida Entomologist 83(4) December, 2000
TABLE 1. PARASITISM RATES OF THE FLY PSEUDACTEON CURVATUS ATTACKING DIFFER-
ENT KINDS OF FIRE ANTS AND MEAN DEVELOPMENTAL TIMES OF FLY PUPAE IN
Fire Ant Species (U.S.)1
S. richteri Hybrid S. invicta
Pupae Produced/Female Fly
(number + SE) 8.3 + 1.7 a 9.4 + 1.0 a 7.6 + 1.5 a
Mean Development Time
(egg to pupae, days + SE) 12.9 + 0.2 a 14.2 + 0.3 b 15.6 + 0.5 c
Mean Standard Deviation
in Devel. Time (days + SE) 2.2 + 0.3 a 3.5 + 0.5 b 4.2 + 0.4 b
'Means within a row with different letters were significantly different (Fisher's PLSD, P s 0.05).
the source of this attraction is a qualitative trait that is not diminished in the hybrid.
However, a one-on-one comparison would be necessary to determine whether the flies
prefer black and hybrid fire ants equally. Preferences for specific hosts are likely
based on chemical cues (Porter 1998a, b). Which cues these might be are unknown,
but black, red, and hybrid fire ants exhibit distinctive differences in their cuticular
hydrocarbons, venom alkaloids, and pheromones (Vander Meer et al. 1985, Obin &
Vander Meer 1989, Vander Meer & Lofgren 1989). It is notable that a strong prefer-
ence for S. richteri was maintained, even after flies had been cultured for 1-2 years
(about 8-16 generations) in the lab using exclusively S. invicta workers as hosts. Re-
tention of a strong preference for S. richteri over this period demonstrates that this
preference was not quickly obscured either by behavioral experience or genetic adap-
S. richteri populations in the United States are much more likely to have origi-
nated from Argentine or Uruguayan port areas rather than landlocked Las Flores
(170 km south of Buenos Aires). The fact that the percent preference for S. richteri
workers from Las Flores, Argentina and the preference for S. richteri workers from
northeastern Mississippi were quite similar suggests that host preferences are prima-
rily species-level rather than population-level differences. Head-to-head comparisons
of fire ant workers from a variety of locations would, of course, be necessary to fully
evaluate the extent and nature of parasitoid preferences for different host ant popu-
We found that fly developmental rates increased significantly from S. richteri to
hybrids to S. invicta (Table 1). This relationship is what would be expected by non-
dominant genetic hybridization.
In contrast with preferences and developmental rates, rates of parasitism in no-
choice laboratory tests were not clearly associated with the type of ant tested (Table
1); perhaps additional replicates would eventually show a modest effect, but this is
not certain. Previous tests showed that P. curvatus females do not do well at parasit-
izing the two most common native fire ants in the United States (Porter 2000).
The practical implications of this study are that P. curvatus flies from Las Flores,
Argentina may do best if they are released onto imported black or hybrid fire ant pop-
ulations in Alabama, Georgia, Mississippi, and Tennessee. Similarly, P. curvatus bio-
types collected from regions where they normally parasitize the red fire ant S. invicta
may be more effective in regions of the United States where this species predomi-
nates. Matching specific P. curvatus biotypes to their normal host would be especially
Porter & Briano: Fire Ant Decapitating Fly
important if the host preferences that we observed in the lab are associated with the
fly's ability to locate potential hosts at distances of several meters or more when vi-
sual abilities are likely to be ineffective. The actual importance of matching P. curva-
tus biotypes to their normal host populations will be evaluated during field releases
of this parasitoid that are currently in progress in Florida, Alabama, and Tennessee.
Robert Vander Meer (USDA-ARS, Gainesville, FL) identified hybrid and black fire
ants from Mississippi using gas chromatography. Lloyd Davis (USDA-ARS, Gaines-
ville, FL) set up and ran many of the preference tests. Cynthia Vann, Barbara May-
field, Laura Collins, Damali Kelly, and David Almquist (USDA-ARS, Gainesville, FL)
ably assisted with various aspects of this study. Lloyd Davis, Lloyd Morrison (USDA-
ARS, Gainesville, FL), and Kathy Flanders (Auburn Univ., AL) read the manuscript
and provided a number of valuable suggestions.
BORGMEIER, T. 1925. Novos subsidies para o conhecimento da familiar Phoridae
(Dipt.). Arch. Mus. Nac. Rio de Janeiro 25: 85-281.
CALLCOTT, A.-M. A., AND H. L. COLLINS. 1996. Invasion and range expansion of red
imported fire ant (Hymenoptera: Formicidae) in North America from 1918-
1995. Florida Entomol. 79: 240-251.
OBIN, M. S., AND R. K. VANDER MEER. 1989. Between- and within-species recognition
among imported fire ants and their hybrids (Hymenoptera: Formicidae): Appli-
cation to hybrid zone dynamics. Ann. Entomol. Soc. Amer. 82: 649-652.
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. 2000. Host specificity and risk assessment of releasing the decapitating
fly, Pseudacteon curvatus, as a classical biocontrol agent for imported fire ants.
Biol. Control. 19: 35-47.
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-
Ross, K. G., AND J. C. TRAGER. 1990. Systematics and population genetics of fire ants
(Solenopsis saevissima complex) from Argentina. Evolution 44: 2113-2134.
SHOEMAKER, D. D., K. G. ROSS, AND M. L. ARNOLD. 1994. Development of RAPD
markers in two introduced fire ants, Solenopsis invicta and Solenopsis richteri,
and their application to the study of a hybrid zone. Mol. Ecol. 3: 531-539.
TRAGER, J. C. 1991. A revision of the fire ants, Solenopsis geminata group (Hy-
menoptera: Formicidae: Myrmicinae). J. New York Entomol. Soc. 99: 141-198.
VAN DRIESCHE, R. G., AND T. S. BELLOWS, JR. 1996. Biological Control. Chapman &
Hall, New York.
VANDER MEER, R. K., C. S. LOFGREN, AND F. M. ALVAREZ. 1985. Biochemical evidence
for hybridization in fire ants. Florida Entomol. 68: 501-506.
VANDER MEER, R. K., AND C. S. LOFGREN. 1989. Biochemical and behavioral evidence
for hybridization between fire ants, Solenopsis invicta and Solenopsis richteri
(Hymenoptera: Formicidae). J. Chem. Ecol. 15: 1757-1765.
428 Florida Entomologist 83(4) December, 2000
PARASITISM OF BEMISIA ARGENTIFOLII ON COLLARD
WITH REDUCED OR NORMAL LEAF WAX
HEATHER J. McAUSLANE1, ALVIN M. SIMMONS2
AND D. MICHAEL JACKSON2
1Department of Entomology & Nematology, University of Florida,
Gainesville, FL 32611-0620
2USDA-ARS, U.S. Vegetable Laboratory, 2875 Savannah Highway,
Charleston, SC 29414-5334
Collard, Brassica oleracea var. acephala L., cultivars with reduced leaf wax (i.e.,
glossy phenotypes) possess ovipositional antixenotic resistance to the silverleaf white-
fly, Bemisia argentifolii Bellows & Perring (Homoptera: Aleyrodidae). We investigated
parasitism by 2 parasitoids of B. argentifolii reared on 2 phenotypes of the collard cul-
tivar 'Green Glaze', differing in amount of leaf wax. When Eretmocerus sp. (Hy-
menoptera: Aphelinidae) parasitoids were given a choice between parasitizing whitefly
nymphs on glossy and normal-wax collard, there were no significant differences in the
number of parasitized nymphs on the 2 plant phenotypes. However, 4.5 times more En-
carsia pergandiella Howard (Hymenoptera: Aphelinidae) emerged from whiteflies on
glossy than on normal-wax plants. In a no-choice test, the number of Eretmocerus sp.
emerging on glossy and normal-wax plants did not differ significantly. In a similar no-
choice test, more than twice as manyE. pergandiella emerged from whiteflies on glossy
collard than on normal-wax collard. Time to 50% emergence for whiteflies and both
species of parasitoids did not differ on the 2 collard types in any of the no-choice tests.
We conclude that management of B. argentifolii populations can be improved on col-
lard, and probably other B. oleracea vegetables, through the use of reduced leaf wax
cultivars that have antixenotic resistance to B. argentifolii and have no detrimental ef-
fects, possibly even beneficial effects, on important whitefly natural enemies.
Key Words: Brassica oleracea, plant resistance, Eretmocerus, Encarsia pergandiella,
parasitoid, leaf wax, tritrophic interactions
Cultivos de acelga, Brassica oleracea var. acephala L., con reducci6n de cera foliar
(por ejemplo, fenotipos glaseados) poseen resistencia antixenotica oviposicional a la
mosquita blanca, Bemisia argentifolii Bellows & Perring (Homoptera: Aleyrodidae).
Investigamos el parasitismo por 2 parasitoides de B. argentifolii criados en 2 fenotipos
del cultivo de acelga "Glaseado Verde", con diferencia en la cantidad de cera foliar.
Cuando species parasitoides de Eretmocerus (Hymenoptera: Aphelinidae) fueron da-
das opci6n entire parasitar ninfas de mosquita blanca sobre acelga glaseada o con cera
normal, no hubieron diferencias significativas en el numero de ninfas parasitadas de
los 2 fenotipos de plants. Sin embargo, emergieron 4,5 veces mas Encarsia pergan-
diella Howard (Hymenoptera: Aphelinidae) de mosquitas blancas en plants glasea-
das que en plants con cera normal. En una prueba sin opci6n, el numero de species
de Eretmocerus emergiendo en plants glaseadas y con cera normal no difiri6 signifi-
cativamente. En una prueba similar sin opci6n, mas de 2 veces E. pergandiella emer-
gieron de mosquitas blancas en acelga glaseada que en acelga de cera normal. El
tiempo para 50% de surgimiento de mosquitas blancas y las dos species de parasitoi-
des no difiri6 en los 2 tipos de acelga en cualquiera de las pruebas sin opci6n. Conclui-
mos que la administraci6n de populaciones de B. argentifolii puede ser mejorada en
acelga, y probablemente otros vegetables de B. oleracea, a trav6s del uso de cultivos con
McAuslane et al.: Leaf Wax Effects on Bemisia Parasitoids 429
cera foliar reducida que tienen resistencia antixenotica a B. argentifolii, y que no tie-
nen efectos perjudiciales, posiblemente hasta efectos beneficiosos, a importantes ene-
migos naturales de la mosquita blanca.
The silverleaf whitefly, Bemisia argentifolii Bellows & Perring, also known as the
"B" strain of the sweetpotato whitefly, Bemisia tabaci (Gennadius), is a serious pest of
vegetable, ornamental, and agronomic crops throughout tropical and, increasingly,
temperate regions of the world. Cruciferous vegetables, such as collard, Brassica oler-
acea var. acephala, and broccoli, B. oleracea var. italica, are important overwintering
hosts for B. argentifolii in the southern United States (Simmons & Elsey 1995) and are
sources of infestation for spring and summer crops (Coudriet et al. 1985, Simmons &
Elsey 1995). Reduction of whitefly populations in Brassica vegetables is desirable, not
only to reduce the need for insecticides and prevent economic loss in these vegetables,
but also to reduce spring whitefly populations available to infest newly planted crops.
Host plant resistance to B. argentifolii has been investigated in many crops. Leaf char-
acteristics, such as trichome abundance and orientation (McCreight & Kishaba 1991,
Kishaba et al. 1992, Wilson et al. 1993, Heinz & Zalom 1995, Lambert et al. 1995, McAus-
lane et al. 1995, McAuslane 1996), presence of glandular exudates (Liedl et al. 1995), and
vascular bundle density (Cohen et al. 1996) and depth within the leaf (Chu et al. 1998,
1999), have been implicated in resistance to B. argentifolii. Recently, antixenotic resis-
tance to whiteflies has been demonstrated in collard and broccoli genotypes that have re-
duced leaf wax (Farnham & Elsey 1995, Jackson et al. 2000). These genotypes have a
glossy or shiny appearance due to their smaller wax load. Whiteflies preferred to oviposit
on normal-wax genotypes; however, if offered no choice of oviposition host, whiteflies ovi-
posited similar numbers of eggs and their progeny developed and survived equally well on
glossy and normal-wax plants (Elsey & Farnham 1994, Jackson et al. 2000).
Bemisia argentifolii can suffer much mortality by natural enemies in many crops that
are not sprayed extensively with broad-spectrum insecticides, such as peanut (McAus-
lane et al. 1993), organic vegetables (Stansly et al. 1997), and collard (Simmons & Jackson
2000). It is well known that plant characteristics can affect the behavior and physiology
of the predators and parasitoids at the third trophic level (Price et al. 1980). For example,
leaf hairs on cucumber (van Lenteren et al. 1995) and tomato (van Roermund & van Len-
teren 1995) interfere with locomotion and parasitization efficiency of Encarsia formosa
Gahan on Trialeurodes vaporariorum (Westwood). Hairs on soybean reduced parasitism
of B. argentifolii by Encarsia and Eretmocerus species (McAuslane et al. 1995). Eigen-
brode and colleagues (Eigenbrode et al. 1995, 1996, 1999) have demonstrated that several
generalist predators control diamondback moth, Plutella xylostella L., more effectively on
glossy cabbage cultivars and that this is due to more efficient locomotion and prey location
behaviors on glossy than on normal-wax genotypes. Little is known, however, about the
potential influence of leaf epidermal waxes on parasitoids of B. argentifolii.
The purpose of this research was to determine the potential effect of collard leaf wax
on parasitoids of B. argentifolii. We selected one Eretmocerus species, a thelytokous un-
described species from Hong Kong (McAuslane & Nyugen 1996), and one Encarsia spe-
cies, Encarsia pergandiella Howard, a common species native to the New World
(Polaszek et al. 1992). We chose one species of each genus because of the different ovipo-
sition habits of the genera. Encarsia species oviposit through the nymphal host exoskel-
eton whereas Eretmocerus species insert the ovipositor between the whitefly nymph and
the leaf surface. In this study, we measured parasitism by these parasitoids when pre-
sented whiteflies on normal-wax or glossy collard in no-choice and choice situations.
430 Florida Entomologist 83(4) December, 2000
MATERIALS AND METHODS
Plants and Insects
Seeds of 'Green Glaze' collard were obtained from M. W. Farnham (U.S. Vegetable
Laboratory, Charleston, SC). This cultivar segregates in a 3:1 ratio for individual
plants with either glossy (i.e., reduced foliar waxbloom) or normal-wax appearance
(Jackson et al. 2000). Seeds were sown in a greenhouse in a soil-less medium (Metro-
mix 200, Grace Sierra, Milpitas, CA). When seedlings could be distinguished as either
glossy or normal-wax, they were transplanted into 12-cm-diameter pots filled with a
1:1 mixture of Metromix 200 and Metromix 500, and were fertilized with approxi-
mately 5 g of a slow-release fertilizer (14-14-14, N-P-K, Osmocote, Scotts-Sierra,
Marysville, OH). Plants were used for experiments 5 to 9 weeks post-germination.
Whiteflies, B. argentifolii, used in experiments withEretmocerus sp. were obtained
from a colony reared on cotton, Gossypium hirsutum L., 'DPL 90', and collard, 'Georgia
Southern', in a climate-controlled room (28C, 14:10 [L:D] photoperiod, 30-50% RH).
The thelytokous Eretmocerus sp. has been maintained on B. argentifolii on hibiscus,
Hibiscus rosa-sinensis L., since it was introduced into the United States from Hong
Kong in October 1992 (McAuslane & Nguyen 1996). Rearing conditions for Eret-
mocerus sp. were the same as those for B. argentifolii. Bemisia argentifolii used in ex-
periments with E. pergandiella were from a colony maintained in a greenhouse on
several vegetable species. The original feral adults were collected from a field of
sweetpotato in Charleston Co., SC (Simmons 1994); feral adults from sweetpotato
were added to the colony annually. An endemic population of E. pergandiella was
maintained on B. argentifolii on several species of vegetables in a greenhouse. The
parasitoids were collected at the same time as the whiteflies. The colony was occasion-
ally supplied with cotton wicks soaked in 10% honey water.
Parasitism by Eretmocerus sp. in a no-choice test
Experiments with Eretmocerus sp. were conducted in an indoor climate-controlled
room (28C day/24C night, 14:10 [L:D)], 30-50% RH) illuminated with high output
110-W cool- white fluorescent lights. Fifteen glossy and 15 normal-wax 6-week-old col-
lard plants bearing 6 to 9 leaves were placed individually in plastic cylindrical cages
(15 cm diameter x 30 cm high) with lids and 2 side openings screened with fine plastic
mesh (94 x 94 mesh). The 2 oldest leaves were removed from each plant and then each
cage was infested with 30 pairs of whiteflies. Whiteflies were removed after 72 hours.
We assumed that whitefly oviposition on the 2 collard types was equal because the
number of eggs laid on normal-wax and glossy collard is equal in no-choice situations
(Elsey & Farnham 1994); however, we did not count whitefly eggs. Five female Eret-
mocerus were added to each cage 10 days later when whiteflies had developed to the
first or second instar. Parasitoids were removed 24 hours later. When emergence be-
gan, newly-emerged whiteflies and parasitoids were aspirated from the plants and
their exuvia were counted and removed from the leaves with a pin each day. This was
continued until no further emergence was noted.
Parasitism by Eretmocerus sp. when presented a choice between glossy and normal-
Foraging behavior and plant preference was indirectly studied by allowing parasi-
toids a choice of glossy or normal-wax collard plants on which to forage for whitefly
nymphs. Plants were 6 weeks old bearing 5 to 6 leaves. Four plants of the same phe-
notype were placed in a screened cage (70-mesh organdy fabric bag supported on a 60
McAuslane et al.: Leaf Wax Effects on Bemisia Parasitoids 431
cm x 60 cm x 60 cm plastic PVC-pipe frame) and were infested with 150 pairs of white-
flies. Whiteflies were removed 24 hours later. As in the previous experiment, we did
not count whitefly eggs but assumed that there were similar numbers on glossy and
normal-wax plants. The infested plants were then rearranged randomly among cages
so that each screened cage contained 2 whitefly-infested glossy and 2 infested normal-
wax plants. Ten cages were set up in this manner. Twelve female parasitoids were re-
leased into the center of each cage 11 days after adult whiteflies were removed when
first and second instars were present. Parasitoids were not removed. Fifteen days
later, leaves were cut from the plants and examined under a microscope. Whitefly ex-
uvia and parasitized whitefly nymphs were counted on upper and lower surfaces of all
leaves. Emergence of whiteflies and number of parasitized whiteflies were calculated
on a per cage basis (= sum of whitefly exuvia or parasitized nymphs on 2 plants of the
Parasitism by E. pergandiella in a no-choice test
Collard seeds were germinated in a greenhouse and then grown in an indoor tem-
perature-controlled room under fluorescent lighting (40-W cool white and 40-W Vita-
lite Duro-test Power-Twist). Upon reaching the 4-5 leaf stage, the plants were
placed in an open greenhouse colony of B. argentifolii. Since whiteflies had a choice of
ovipositing on normal-wax or glossy collard plants during the infestation procedure,
normal-wax plants were exposed to whiteflies for 2 hours and glossy plants for 1 hour
longer to compensate for reduced oviposition on the glossy collard. Exposure times
were based on data obtained from field experiments (Jackson et al. 2000) and prelim-
inary greenhouse studies (unpublished data). The plants were then moved from the
colony and adult whiteflies were removed first by the air flow from an electrical fan
and then with an aspirator. Two plants of each collard type were placed in a Plexiglass
cage (45 cm wide x 45 cm long x 46 cm high) below fluorescent lamps (as described
above) in a temperature-controlled room (14:10 L:D photoperiod supplying ca. 452 lux
at plant height). Four cages per trial were set up. Temperature within the cages was
26-27C. After the whiteflies developed to the second to third nymphal instar, all
leaves below a single tagged target leaf (3-4 from bottom) were detached, as were any
leaves younger than the targeted leaf that contained whitefly nymphs. Forty E. per-
gandiella (unsexed) were released into each cage. The parasitoids were retrieved with
an aspirator after 24 hours. Upper and lower surfaces of the tagged leaves were
checked daily for whitefly or parasitoid emergence. Any adults and exuvia observed
were removed daily, and exuvial counts were recorded. This was continued until no
further emergence was noted. The experiment was repeated to obtain a second trial.
Emergence of whiteflies and parasitoids was calculated on a per cage basis (= sum of
emergence on 2 plants of the same genotype).
Parasitism by E. pergandiella when presented a choice between glossy and normal-
Collard plants at the 5- to 6-leaf stage were infested during a 12-14 hour exposure
to whiteflies in a greenhouse. The adult whiteflies were removed using an aspirator
and then the plants were transferred to a temperature-controlled room. One glossy
and one normal-wax plant were placed in a Plexiglass cage (45 cm wide x 45 cm long
x 46 cm high) and 8 replicate cages were set up. Forty parasitoids (unsexed) were
added to each cage when whiteflies had developed to the second or third nymphal in-
star. After 3 weeks, exuvia from which either a whitefly or a parasitoid emerged were
432 Florida Entomologist 83(4) December, 2000
We compared whitefly emergence, parasitoid emergence, and number of parasit-
ized nymphs between glossy and normal-wax collard using analysis of variance
(PROC GLM; SAS Institute 1997). Because only a small percentage (<5%) of imma-
ture whiteflies develop on the top leaf surface of collard (Simmons 1994) and leaf sur-
face does not affect development (Simmons 1999), emergence data were pooled
between leaf surfaces. Data for emerged whiteflies and parasitoids or parasitized
nymphs were log (x + 0.1)-transformed, when necessary, to correct for variance in-
creasing with the mean. Means shown in tables are untransformed. We compared de-
velopmental time of whiteflies and parasitoids between the 2 collard types by
estimating time to 50% emergence on each plant using linear regression (PROC REG;
SAS Institute 1997) and performing analysis of variance (PROC GLM) on estimated
times to 50% emergence. Significant least square means were separated by the prob-
ability of a significant difference at = 0.05 (PROC GLM).
Parasitism by Eretmocerus sp.
In the no-choice test, one plant which was initially classified as a normal-wax col-
lard was in fact glossy and another normal-wax collard was destroyed during the ex-
periment leaving 13 replicate normal-wax collard plants and 16 glossy collard
replicates. Whiteflies emerged over a 15-day period beginning 21 days after the plants
were infested. The time to 50% emergence of whiteflies did not differ on the 2 collard
wax types (F, 12) = 0.11; P = 0.75) and averaged 4.4 + 0.8 days (mean + se), with the
first day of whitefly emergence being day 1. Eretmocerus parasitoids emerged over an
11-day period, beginning on day 13 of whitefly emergence. The time to 50% emergence
of parasitoids did not differ between the 2 collard types (F,1 12, = 0.41; P = 0.53) and av-
eraged 4.4 + 0.3 days. The number of parasitoids emerging was not influenced by col-
lard type (F(, 1,= 1.48; P = 0.25) nor was whitefly emergence (F ,12)= 1.69; P = 0.22)
In the choice test, neither whitefly emergence (F),,9) = 0.24; P = 0.63) nor the num-
ber of parasitized nymphs (F,,,, = 0.92; P = 0.37) differed significantly between nor-
mal-wax and glossy collard (Table 1).
Parasitism by E. pergandiella
In the no-choice test, whiteflies emerged over a 13-day period in trial 1 and an 11-
day period in trial 2. Time to 50% emergence was not influenced by collard type (F,
= 0.04; P = 0.86), but there was a significant effect of trial (F1, = 13.06; P = 0.009).
Time to 50% emergence of whiteflies was 4.7 + 0.4 days in trial 1 and 2.2 + 0.7 days
after first whitefly emergence in trial 2. Encarsia pergandiella emerged over a 12-day
period in trial 1 and an 8-day period in trial 2, but there was no significant effect of
collard type (F,,4) = 0.31; P = 0.61) or trial (F,4) = 0.06;P = 0.82) on time to 50% emer-
gence, which averaged 3.8 + 0.5 days after the first parasitoid emerged.
Whitefly emergence was influenced by trial (F )1, = 33.91; P = 0.001), collard type
(F(, ,= 12.00; P = 0.013), and the interaction of trial x collard type (F,,) = 14.05; P =
0.0095). Significantly more whiteflies emerged from glossy collard in trial 1 than from
normal-wax collard (Table 2). Emergence did not differ in trial 2. Emergence of E. per-
gandiella was significantly influenced by trial (F,1 ,12= 5.39; P = 0.039) and collard type
McAuslane et al.: Leaf Wax Effects on Bemisia Parasitoids 433
TABLE 1. NUMBER OF EMERGED ERETMOCERUS SP. PARASITOIDS AND WHITEFLIES ON
GLOSSY AND NORMAL-WAX 'GREEN GLAZE' COLLARD IN NO-CHOICE AND CHOICE
TESTS (MEANS + SE, RANGE).
Wax type No. Eretmocerus emerged No. whiteflies emerged
Glossy 36.0 + 6.3 134.8 + 24.0
(0- 74) (19- 439)
Normal-wax 24.2 + 3.8 168.2 + 29.5
(5 47) (28 408)
Glossy 88.5 + 12.3 940 + 211
(36 160) (382 2644)
Normal-wax 97.9 + 11.4 937 + 206
(45 166) (415 1565)
(F(1,12)= 11.38; P = 0.006), but there was no trial x collard type interaction (F(1,12)= 0.15;
P = 0.71). More than twice as many parasitoids emerged from glossy plants than from
normal-wax plants (an average of 35.5 + 4.8 vs. 14.6 + 5.0 per plant, respectively) (Ta-
In the choice test, there was no significant difference in whitefly emergence on the
2 collard types (F,,,7, = 0.01; P = 0.93) (Table 2). However, 4 times as many E. pergan-
diella emerged from glossy collard than from normal-wax collard (Fi ,, = 23.87; P =
The glossy leaf-wax trait in Brassica vegetables has been associated with resis-
tance to several important insect pests such as the cabbage aphid, Brevicoryne bras-
sicae (L.), the imported cabbageworm,Artogeia rapae (L.), P. xylostella (Eigenbrode &
Shelton 1990, Stoner 1990, Eigenbrode et al. 1991), and B. argentifolii (Elsey & Farn-
ham 1994, Farnham & Elsey 1995, Jackson et al. 2000). In the case of P. xylostella, re-
duced pest populations on glossy plants were due partly to the direct physical and
allelochemical effects on the insect of leaf wax components (Eigenbrode et al. 1991),
and partly to enhanced predation by natural enemies (Eigenbrode et al. 1995). Eigen-
brode et al. (1995) stated that the importance of predation should be evaluated during
development of glossy Brassica for resistance to insects.
Our study indicates that parasitism of whitefly nymphs by E. pergandiella is en-
hanced on glossy phenotype 'Green Glaze' collard compared with a normal-wax phe-
notype. These phenotypes are isogenic except for the single gene mutation causing
glossiness, hence, we would not expect any nutritional effects of the plant acting
through the host on the parasitoid. We did not find the same increase in whitefly par-
asitism by Eretmocerus sp. on glossy collard. However, and of more importance to reg-
ulation of whitefly populations, we saw no decrease in parasitism by Eretmocerus sp.
on glossy collard.
We had expected, given the nature of the oviposition behavior of these 2 parasitoid
species, that Eretmocerus sp. would be more affected, either negatively or positively, by
Florida Entomologist 83(4)
TABLE 2. NUMBER OF EMERGED E. PERGANDIELLA PARASITOIDS AND WHITEFLIES ON
GLOSSY AND NORMAL-WAX 'GREEN GLAZE' COLLARD IN NO-CHOICE AND CHOICE
TESTS (MEANS SE, RANGE).
No. E. pergandiella emerged No. whiteflies emerged
19.8 + 4.4b
'Means within a column and within a test followed by different letters
nificant difference of least squares means, a = 0.05).
Stiffer significantly (probability ofa sig-
collard leaf wax because Eretmocerus sp. females must locate a suitable gap between
the whitefly nymph and the leaf surface through which to insert their ovipositor. If the
marginal wax laid down by whitefly nymphs adheres differently on normal-wax and
glossy plants, we might expect the ability of Eretmocerus sp. to insert its ovipositor to
be different on the 2 plant types. On the other hand, E. pergandiella females oviposit
through the dorsum of their host and the adhesion of the host to the leaf should not in-
fluence oviposition success. The eggs of Eretmocerus species, in general, lie underneath
the host on the leaf surface for several days before the first instar parasitoid ecloses
and chews into the whitefly nymph (Foltyn & Gerling 1985, McAuslane & Nyugen
1996). The egg is presumably in contact with leaf waxes and allelochemicals and could
be influenced by physical and chemical characteristics of this wax layer. Encarsia per-
gandiella immature stages are never in direct contact with the leaf surface. In our
study, contrary to our expectations, we found that parasitism of whitefly by E. pergan-
diella was in fact improved on the glossy collard while parasitism success of Eret-
mocerus was unchanged. Reasons other than those proposed above must account for
the different parasitism success of these two species on glossy and normal-wax collard.
Large-bodied generalist predators of diamondback moth, such as adults of Orius
insidiosus (Say) and Hippodamia convergens Guerin-Meneville, and larval Chrysop-
erla carnea (Stephens) are more mobile on glossy cabbage genotypes (Eigenbrode et
al. 1996) and consequently locate P. xylostella larvae better on glossy cabbage than on
normal-wax cabbage (Eigenbrode et al. 1995). Their greater mobility is due to the fact
McAuslane et al.: Leaf Wax Effects on Bemisia Parasitoids 435
that they spend less time scrambling (i.e., slipping while walking), falling off the
plant, and grooming off wax particles that had accumulated on their tarsi on glossy
cabbage than on normal-wax cabbage (Eigenbrode et al. 1996). Much of the reason
that predators can move more efficiently on glossy genotypes is because they can gen-
erate much greater adhesive force on glossy leaves than on normal-wax cabbage
leaves (Eigenbrode et al. 1999). Aphelinid parasitoids fly from plant to plant when for-
aging for whitefly hosts and often fly within the plant, from leaf to leaf. Parasitoids do,
however, walk extensively on the leaf searching for host patches. It is not known
whether these very small-bodied hymenopterans suffer the same reduced traction on
normal-wax plants as do larger predators. This aspect of parasitoid behavior needs to
be studied more carefully.
Other possible reasons for the difference in parasitism of Eretmocerus and E. per-
gandiella are potential differences in whitefly nymph distribution on leaves. If white-
fly oviposition behavior differs on glossy and normal-wax collard, leading to different
dispersion of nymphs, this may differently affect the foraging behavior and parasiti-
zation success of these 2 parasitoid species. Finally, although we tried to perform ex-
periments under similar environmental conditions, Eretmocerus and E. pergandiella
were studied in different laboratories. It is known that water saturation and other en-
vironmental conditions can alter the composition and amount of leaf waxes (reviewed
in Eigenbrode & Espelie 1995). This may have affected parasitoid behavior and/or
Host plant resistance and biological control have long been considered the corner-
stones of pest management strategies. While generally thought to be compatible and
additive in nature (Bergman & Tingey 1979), these tactics have not always been so.
We have demonstrated that the nonpreference ovipositional resistance (antixenosis)
in glossy collards to B. argentifolii is fully compatible with biological control by species
in the 2 most important genera of whitefly parasitoids. Parasitism of whiteflies is at
least as high, if not higher in the case ofE. pergandiella, on a glossy collard phenotype
compared with that on a normal-wax phenotype. Use of glossy collard in a manage-
ment program that includes natural enemies should lead to smaller whitefly infesta-
tions in collard, leading to a reduced need for insecticide application and reduced
populations infesting spring plantings of other crops.
The authors thank D. Boyd, J. Squitier, and Y. Cardoza (University of Florida,
Gainesville, FL), and J. Bhamoo, B. Davis, J. Day, L. Cauthen, and J. Cook (USDA ARS,
Charleston) for assistance with experiments and rearing of insects and plants. Thanks
also to G. S. McCutcheon (Clemson University, Charleston, SC) and T.-X. Liu (Texas
A&M University, Weslaco, TX) for providing valuable comments on an earlier version
of this manuscript. This article is from a cooperation between the University of Florida
and the U.S. Vegetable Laboratory, USDA, Agricultural Research Service. Funding was
provided by Experiment Station Project FLA-ENY-03194 and the U.S. Vegetable Lab-
oratory. It is published with the approval of both agencies as Florida Agricultural Ex-
periment Station Journal Series No. R-07369. Mention of a proprietary product does
not constitute an endorsement or recommendation by the USDA for its use.
BERGMAN, J. M., AND W. M. TINGEY. 1979. Aspects of interaction between plant gen-
otypes and biological control. Bull. Entomol. Soc. Am. 25: 275-279.
436 Florida Entomologist 83(4) December, 2000
CHU, C.-C., A. C. COHEN, E. T. NATWICK, G. S. SIMMONS, AND T. J. HENNEBERRY. 1999.
Bemisia tabaci (Homoptera: Aleyrodidae) biotype B colonization and leaf mor-
phology relationships in upland cotton cultivars. Austral. J. Entomol. 38: 127-
CHU, C.-c., E. T. NATWICK, A. C. COHEN, G. S. SIMMONS, D. E. BRUSHWOOD, AND T. J.
HENNEBERRY. 1998. Bemisia argentifolii colonization, vascular bundle depth
relationships, lint yield, and nymphal parasitism in selected Deltapine cotton
cultivars. Southwest. Entomol. 23: 293-299.
COHEN, A. C., T. J. HENNEBERRY, AND C.-C. CHU. 1996. Geometric relationships be-
tween whitefly feeding behavior and vascular bundle arrangements. Entomol.
Exp. Appl. 78: 135-142.
COUDRIET, D. L., N. PRABHAKER, A. N. KISHABA, AND D. E. MEYERDIRK. 1985. Varia-
tion in developmental rate on different hosts and overwintering of the sweet-
potato whitefly, Bemisia tabaci (Homoptera: Aleyrodidae). Environ.
EIGENBRODE, S. D., T. CASTAGNOLA, M.-B. ROUX, AND L. STELJES. 1996. Mobility of
three generalist predators is greater on cabbage with glossy leaf wax than on
cabbage with a wax bloom. Entomol. Exp. Appl. 81: 335-343.
EIGENBRODE, S. D., AND K. E. ESPELIE. 1995. Effects of plant epicuticular lipids on in-
sect herbivores. Annu. Rev. Entomol. 40: 171-194.
EIGENBRODE, S. D., N. N. KABALO, AND K. A. STONER. 1999. Predation, behavior, and
attachment by Chrysoperla plorabunda larvae on Brassica oleracea with differ-
ent surface waxblooms. Entomol. Exp. Appl. 90: 225-235.
EIGENBRODE, S. D., S. MOODIE, AND T. CASTAGNOLA. 1995. Predators mediate host
plant resistance to a phytophagous pest in cabbage with glossy leaf wax. Ento-
mol. Exp. Appl. 77: 335-342.
EIGENBRODE, S. D., AND A. M. SHELTON. 1990. Behavior of neonate diamondback
moth larvae (Lepidoptera: Plutellidae) on glossy-leafed resistant Brassica ole-
racea L. Environ. Entomol. 19: 1566-1571.
EIGENBRODE, S. D., K. A. STONER, A. M. SHELTON, AND W. C. KAIN. 1991. Character-
istics of glossy leaf waxes associated with resistance to diamondback moth (Lepi-
doptera: Plutellidae) in Brassica oleracea. J. Econ. Entomol. 84: 1609-1618.
ELSEY, K. D., AND M. W. FARNHAM. 1994. Response of Brassica oleracea L. to Bemisia
tabaci (Gennadius). HortScience 29: 814-817.
FARNHAM, M. W., AND K. D. ELSEY. 1995. Recognition of Brassica oleracea L. resis-
tance against the silverleaf whitefly. HortScience 30: 343-347.
FOLTYN, S., AND D. GERLING. 1985. The parasitoids of the aleyrodidBemisia tabaci in
Israel: development, host preference and discrimination of the aphelinid wasp
Eretmocerus mundus. Entomol. Exp. Appl. 38: 255-260.
HEINZ, K. M., AND F. G. ZALOM. 1995. Variation in trichome-based resistance to Be-
misia argentifolii (Homoptera: Aleyrodidae) oviposition on tomato. J. Econ. En-
tomol. 88: 1494-1502.
JACKSON, D. M., M. W. FARNHAM, A. M. SIMMONS, AND W. A. VAN GIESSEN. 2000.
Evaluation of collards for resistance to whiteflies (Homoptera: Aleyrodidae). J.
Econ. Entomol. (in press).
KISHABA, A. N., S. CASTLE, J. D. MCCREIGHT, AND P. R. DESJARDINS. 1992. Resistance
of white-flowered gourd to sweetpotato whitefly. HortScience 27: 1217-1221.
LAMBERT, A. L., R. M. MCPHERSON, AND K. E. ESPELIE. 1995. Soybean host plant re-
sistance mechanisms that alter abundance of whiteflies (Homoptera: Aley-
rodidae). Environ. Entomol. 24: 1381-1386.
LIEDL, B. E., D. M. LAWSON, K. K. WHITE, J. A. SHAPIRO, D. E. COHEN, W. G. CARSON,
J. T. TRUMBLE, AND M. A. MUTSCHLER. 1995. Acylsugars of wild tomato Lyco-
persicon pennellii alters settling and reduces oviposition ofBemisia argentifolii
(Homoptera: Aleyrodidae). J. Econ. Entomol. 88: 742-748.
MCAUSLANE, H. J. 1996. Influence of leaf pubescence on ovipositional preference of
Bemisia argentifolii (Homoptera: Aleyrodidae) on soybean. Environ. Entomol.
McAuslane et al.: Leaf Wax Effects on Bemisia Parasitoids 437
McAUSLANE, H. J., F. A. JOHNSON, D. A. KNAUFT, AND D. L. COLVIN. 1993. Seasonal
abundance and within-plant distribution of parasitoids of sweetpotato whitefly,
Bemisia tabaci (Homoptera: Aleyrodidae) in peanuts. Environ. Entomol. 22:
McAUSLANE, H. J., F. A. JOHNSON, D. L. COLVIN, AND B. SOJACK. 1995. Influence of
foliar pubescence on incidence and parasitism of Bemisia argentifolii (Ho-
moptera: Aleyrodidae) on soybean and peanut. Environ. Entomol. 24: 1135-
MCAUSLANE, H. J., AND R. NGUYEN. 1996. Reproductive behavior and biology of a th-
elytokous species of Eretmocerus (Hymenoptera: Aleyrodidae) parasitizing Be-
misia argentifolii. Ann. Entomol. Soc. Am. 89: 686-693.
MCCREIGHT, J. D., AND A. N. KISHABA. 1991. Reaction of cucurbit species to squash
leaf curl virus and sweetpotato whitefly. J. Am. Soc. Hort. Sci. 116: 137-141.
POLASZEK, A., G. A. EVANS, AND F. D. BENNETT. 1992. Encarsia parasitoids of Bemisia
tabaci (Hymenoptera: Aphelinidae, Homoptera: Aleyrodidae): a preliminary
guide to identification. Bull. Entomol. Res. 82: 375-392.
PRICE, P. W., C. E. BOUTON, P. GROSS, B. A. MCPHERON, J. N. THOMPSON, AND A. E.
WEIS. 1980. Interactions among three trophic levels: Influence of plants on in-
teractions between insect herbivores and natural enemies. Annu. Rev. Ecol.
Syst. 11: 41-65.
SAS INSTITUTE. 1997. SAS/STAT Software: Changes and Enhancements through
Release 6.12. Cary, N.C.
SIMMONS, A. M. 1994. Oviposition on vegetables by Bemisia tabaci (Homoptera: Aley-
rodidae): Temporal and leaf surface factors. Environ. Entomol. 23: 381-389.
SIMMONS, A. M. 1999. Nymphal survival and movement of crawlers of Bemisia argen-
tifolii (Homoptera: Aleyrodidae) on leaf surfaces of selected vegetables. Envi-
ron. Entomol. 28: 212-216.
SIMMONS, A. M., AND K. D. ELSEY. 1995. Overwintering and cold tolerance of Bemisia
argentifolii (Homoptera: Aleyrodidae) in coastal South Carolina. J. Entomol.
Sci. 30: 497-506.
SIMMONS, A. M., AND D. M. JACKSON. 2000. Evaluation of foliar-applied insecticides
on abundance of parasitoids of Bemisia argentifolii (Homoptera: Aleyrodidae)
in vegetables. J. Entomol. Sci. 35: 1-8.
STANSLY, P. A., D. J. SCHUSTER, AND T.-X. LIU. 1997. Apparent parasitism of Bemisia
argentifolii (Homoptera: Aleyrodidae) by Aphelinidae (Hymenoptera) on vege-
table crops and associated weeds in south Florida. Biol. Control 9: 49-57.
STONER, K. A. 1990. Glossy leaf wax and plant resistance to insects in Brassica oler-
acea under natural infestation. Environ. Entomol. 19: 730-739.
VAN LENTEREN, J. C., Z. H. LI, J. W. KAMERMAN, AND R. Xu. 1995. The parasite-host
relationship between Encarsia formosa (Hym., Aphelinidae) and Trialeurodes
vaporariorum (Hom., Aleyrodidae) XXVI. Leaf hairs reduce the capacity of En-
carsia to control greenhouse whitefly on cucumber. J. Appl. Entomol. 119: 553-
VAN ROERMUND, H. J. W., AND J. C. VAN LENTEREN. 1995. Residence times of the
whitefly parasitoid Encarsia formosa Gahan (Hym., Aphelinidae) on tomato
leaflets. J. Appl. Entomol. 119: 465-471.
WILSON, F. D., H. M. FLINT, B. R. STAPP, AND N. J. PARKS. 1993. Evaluation of culti-
vars, germplasm lines, and species of Gossypium for resistance to biotype "B" of
sweetpotato whitefly (Homoptera: Aleyrodidae). J. Econ. Entomol. 86: 1857-
438 Florida Entomologist 83(4) December, 2000
RESIDUAL EFFICACY OF BLATTICIDES APPLIED TO
SURFACES CONTAMINATED WITH GERMAN COCKROACH
(DICTYOPTERA: BLATTELLIDAE) FECES
CHARLES A. STRONG, STEVEN M. VALLEYS, PHILIP G. KOEHLER1
AND RICHARD J. BRENNER
Center for Medical, Agricultural and Veterinary Entomology, USDA-ARS,
1600 SW 23rd Drive, Gainesville, Florida 32608
1Department of Entomology and Nematology, University of Florida,
Gainesville, Florida 32611
Crack-and-crevice treatments were simulated in the presence and absence of Ger-
man cockroach [Blattella germanica (L.)] feces to evaluate its effect on insecticide ef-
ficacy toward the German cockroach. The LTs, of German cockroaches exposed to 0.39
ig of cypermethrin/cm2 (Demon EC formulation) on glass Mason jars was 26 min. The
LTs, increased 2.5- and 4.5-fold when Demon EC was mixed with 123 and 184 mg of
cockroach feces, respectively. The presence of German cockroach feces increased the
LT,0 2.5-fold in Dursban EC (chlorpyrifos) and 1.2-fold in Baygon EC (propoxur). Lon-
gevity experiments with 3 X-cyhalothrin formulations in the presence of German cock-
roach feces resulted in significant decreases in insecticide efficacy. Feces reduced the
performance of Commodore WP (X-cyhalothrin) by 12.5, 35, 55, and 97.5% on days 0,
10, 20, and 30, respectively. Initial reductions in efficacy were observed for the De-
mand CS (X-cyhalothrin) and Karate (X-cyhalothrin) formulations when in the pres-
ence of German cockroach feces.
Key Words: Blattella germanica, feces, insecticide efficacy, formulation
Tratamientos de grietas y hendiduras fueron simulados en la presencia y ausencia
de heces fecales de la cucaracha alemana [Blattella germanica (L.)] para evaluar su
efecto en eficacia del insecticide hacia la cucaracha alemana. El LT0, de cucarachas
alemanas expuestas a 0,39 ig de cypermethrin/cm2 (formulaci6n Demon EC) enjarras
de vidrio Mas6n fue 26 minutes. El LT0, increments 2,5 y 4,5 veces cuando se mezcl6
Demon EC con 123 y 184 mg de heces fecales de cucaracha, respectivamente. La pre-
sencia de heces fecales de la cucaracha alemana incremento el LT0, 2,5 veces en Dur-
sban EC (chlorpyrifos) y 1,2 veces en Baygon EC (propoxur). Experimentos de
longevidad con 3 formulaciones X-cyhalothrin en presencia de heces fecales de cuca-
racha alemana result en disminuciones significativas en eficacia del insecticide. Las
heces fecales redujeron el desempeno de Commodore WP (X-cyhalothrin) por 12,5, 35,
55, y 97,5% en los dias 0, 10, 20, y 30, respectivamente. Reducciones iniciales en efi-
cacia fueron observadas en formulaciones de Demand CS (X-cyhalothrin) y Karate (k-
cyhalothrin) cuando en presencia de heces fecales de la cucaracha alemana.
Recent public concern over the environmental and health impacts of pesticides has
discouraged the traditional use of broad spray treatments for control of German cock-
roach, Blattella germanica (L.). Although residual insecticide treatments are still
Strong et al.: Feces Effect On Insecticide Efficacy
used against German cockroaches, the method of application is typically confined to
suspected cockroach aggregation sites or harborages. This method of insecticide ap-
plication is commonly known as crack-and-crevice treatment. Crack-and-crevice in-
secticide placement reduces the amount of toxicant required for control (Bennett et al.
1988) and minimizes insecticide exposure risks to humans and pets.
German cockroaches are gregarious (Roth & Willis 1960) often spending up to 75%
of their lifetime at rest in harborages (Cornwell 1968). Within these harborages, feces
accumulate and may affect the efficacy of crack-and-crevice insecticide applications.
Surface type (Cornwell 1972), organic matter (Niemczyk & Krueger 1987, Kamm &
Montgomery 1990), and oils (Newton & Coombes 1990) influence the performance of
various residual insecticide formulations.
We simulated crack-and-crevice treatments in the presence and absence of Ger-
man cockroach feces for the purpose of evaluating the effect of feces on insecticide ef-
MATERIALS AND METHODS
Cockroaches and Feces
Adult males (1-2 wk old) of the insecticide susceptible Orlando strain of German
cockroach (Koehler & Patterson 1986) were used for all bioassays. The cockroaches
were reared as described by Koehler et al. (1994).
German cockroach feces were collected from rearing containers of final instar cock-
roaches. Feces were separated from cast skins and other debris with a steel sieve
(0.71-mm2 openings) followed by fine mesh steel sieve (0.48-mm2 openings) for use in
Emulsifiable concentrate (EC) formulations of cypermethrin (Demon EC, 25.3%
[Al]; Zeneca, Wilmington, DE.), chlorpyrifos (Dursban 2E, 24.1% [Al]; Dow Agro-
Sciences, Indianapolis, IN), X-cyhalothrin (Karate, 10% [Al] Zeneca), and propoxur
(Baygon 1.5 EC, 14.7% [Al], Bayer, Kansas City, MO) were used in efficacy experi-
ments. Additionally, the wettable powder (WP) and microencapsulated formulations
of X-cyhalothrin (Commodore WP 10% [Al] and Demand CS 9.7% [Al], respectively)
were included in the study. All formulations were diluted in water to form an emulsion
or suspension and pipetted into jars for tests.
Time-mortality relationships (Cochran 1997) were first established for formulated
cypermethrin (Demon EC) in the presence of increasing quantities of cockroach feces.
Cockroach feces (0, 0.24, 0.48, or 0.72 mg per cm2) were added to Mason jars (473 ml,
surface area = 256 cm2; Ball Corp., Muncie, IN). The Demon EC formulation of cyper-
methrin was prepared at a rate of 0.1 mg[AI]/mlby adding 40 P1 of the EC to 99.96 ml
of water. One ml of this solution was pipetted into each jar. An additional 3 ml of water
also was added to each jar to facilitate even coating with the feces-insecticide mixture.
The jars were placed on a roller on their sides and rotated continuously with a gentle
stream of compressed air directed into each jar. After the jars were dry (typically 4 h)
the upper 2 cm of the interior was coated with a petroleum jelly:mineral oil mixture
440 Florida Entomologist 83(4) December, 2000
(3:2) to prevent cockroaches from escaping. Ten adult male cockroaches of the Orlando
strain were placed into each jar individually with featherweight forceps. The cock-
roaches were not anesthetized with CO2. The jars were placed into an environmental
chamber and held at 24.9 + 0.4C and 69.6 + 3.9% RH. Control jars were treated with
water and feces only. A repeated measures method was used to assess lethal time val-
ues. Cockroaches were monitored every 5 min and the number dead recorded until
80% of the cockroaches in the jar were killed. Cockroaches were considered dead or
moribund if unable to right themselves within 15 sec after being flipped onto their
dorsum. Treatment mortality was corrected for control mortality with Abbott's for-
mula (1925). The entire experiment was replicated 3 times. Mortality was analyzed
with the Probit procedure (SAS Institute 1988). Significant differences were deter-
mined by nonoverlap of 95% confidence intervals.
The effect of feces on lethal time was compared among EC formulations of chlorpy-
rifos, cypermethrin, and propoxur. Dursban 2E, Baygon 1.5 EC and Demon EC were
prepared in water at 3, 11, and 0.1 mg[AI]/ml, respectively. Mason jars were treated
with 0.72 mg of feces/cm2 and 1 ml of formulated insecticide as described previously.
Additional water (3 ml) was added and the jars were rolled until dry. Cockroaches
were added to each jar, mortality recorded, and data analyzed as described previously.
The experiment was repeated three times. Data were analyzed by the Probit proce-
dure with mortality as dependent variable (SAS 1988).
The last part of the study was to determine the effect of feces on the residual per-
formance of 3 different formulations of X-cyhalothrin. The wettable powder (Commo-
dore WP), capsulated suspension (Demand CS) and emulsifiable concentrate (Karate)
formulations were diluted in water at 0.1 mg[AI]/ml and 1 ml was applied to the inner
walls of the jars. Each insecticide formulation was added to 2 jars, one containing 0.72
mg of feces/cm2 and one not treated with feces. Ten adult male cockroaches were
placed in each jar. After a 1-min exposure to the treated glass surface the cockroaches
were anesthetized with CO2 (15 sec), placed in an untreated glass Petri dish, and sub-
sequently placed in another untreated plastic Petri dish (100 by 15 mm). Any feces de-
posited with the insects was returned to its respective jar. Cockroaches were held in
an environmental chamber at 24.9 + 0.4C and 69.6 + 3.9% RH. Mortality was as-
sessed 24 h later. Jars were stored in the dark at room temperature (approximately
24C). The bioassay was repeated at 10, 20, and 30 d after treatment by using the
same jars. The experiment was replicated 5 times. Control jars were treated with wa-
ter and feces. Mortality in the presence and absence of feces was compared by Stu-
dent's t-test for each time period and formulation.
The LTs, of Orlando cockroaches exposed to glass Mason jars treated with 0.39
pg[AI]/cm2 of Demon EC was 26 min (Table 1). This value was not significantly differ-
ent from jars treated with Demon EC and 0.24 mg/cm2 of German cockroach feces.
However, the LTs, increased 2.5- and 4.5-fold when Demon EC was mixed with 0.48
and 0.72 mg of feces/cm2, respectively. No mortality was observed in control jars de-
void of insecticide but containing feces.
Among the 3 insecticide classes used in this study, cockroach feces had the greatest
impact on the efficacy of cypermethrin (Table 2). The inhibition ratio (IR = LT0, feces
contaminated jar/LT0, clean jar) for cypermethrin was 4.4-fold. The presence of Ger-
man cockroach feces increased the time to mortality 2.5-fold in Dursban 2E (chlorpy-
rifos) and had a small, yet statistically significant (based on nonoverlap of 95%
confidence intervals), effect on the toxicity of Baygon 1.5 EC (propoxur).
Strong et al.: Feces Effect On Insecticide Efficacy
TABLE 1. THE LTs0 VALUES (MIN) FOR ADULT MALE SUSCEPTIBLE COCKROACHES EX-
POSED TO A DEMON EC (CYPERMETHRIN) TREATED MASON JAR (0.39 G/CM2) IN
THE PRESENCE OF INCREASING QUANTITIES OF GERMAN COCKROACH FECES.
Feces (mg/cm2) Obs Slope SE LT,, (95% CI) Z2 IRb
0 420 1.67 + 0.20 25.55 (21.22-32.16) 1.98 1.00
0.24 400 1.15 + 0.27 20.63 (10.24-28.23) 8.65 0.81
0.48 400 1.31 + 0.27 63.77 (50.79-92.24) 9.96 2.50
0.72 810 1.50 + 0.22 115.59 (98.18-166.65) 1.09 4.52
"Total number of observations recorded until 80% mortality was achieved (repeated measure).
IR = inhibition ratio (LT,, contaminated surface/LT0 clean surface).
Residual activity of various X-cyhalothrin formulations in the presence and ab-
sence of German cockroach feces is illustrated in Figure 1. Mortality was reduced sig-
nificantly among all formulations in the presence of feces. Feces reduced the
performance of the WP by 12.5, 35, 55, and 97.5% on days 0, 10, 20, and 30, respec-
tively (Fig. 1A). Initial reductions in efficacy were severe for the Demand CS and Ka-
rate formulations (Fig. 1B, C). Although Demand CS was effective on clean glass
throughout the study, feces nearly eliminated its ability to kill cockroaches. Mortality
on the Karate-treated surface declined sharply in the absence of feces, and no signif-
icant differences were observed between treatments with and without feces at 10, 20,
and 30 d.
Cracks and crevices in furniture, kitchen equipment, wall voids, and elsewhere in
structures are primary harborage sites for the German cockroach (Cornwell 1968).
These harborages often become heavily littered with cockroach feces (Stejskal 1997).
Treatment of these areas with residual insecticides is a recommended method for
cockroach control in food-handling establishments (Rust 1986, Bennett et al. 1988).
Unfortunately, based on our data, fecal deposits found in these areas may signifi-
cantly reduce the efficacy of some insecticides.
Decreased insecticide efficacy in the presence of German cockroach feces was an-
ticipated based on results of previous reports. For example, organic matter in soil re-
duce insecticide toxicity by acting as an adsorbent (Hamaker & Thompson 1972).
Similarly, activated carbon has been used to protect grass seed from herbicides (Lee
1973). German cockroach control failures in kitchens also have been associated with
insecticide affinity to cooking oils (Ree 1980, Schal 1988, Rust & Reierson 1988). In
addition to reducing insecticide efficacy by adsorption, microbial degradation also
may effect insecticide efficacy. Feces may contain microbes that could have the capac-
ity to metabolize the insecticide. However, preliminary experiments with autoclaved
feces resulted in comparable decreases in insecticide efficacy indicating that microbial
degradation was not likely to be the mechanism responsible (C. Strong, unpublished
Among the EC formulations of 3 insecticide classes evaluated in this study, cock-
roach feces were most detrimental to the performance of cypermethrin, Demon EC
(Table 2). Cypermethrin-treated jars containing feces (0.72 mg/cm2) required 4.5-fold
more time to kill cockroaches than insecticide in clean jars. Chlorpyrifos, Dursban 2E,
toxicity also was significantly reduced (2.5-fold) by the presence of feces. The toxicity
TABLE 2. THE LTso VALUES (MIN) FOR ADULT MALE SUSCEPTIBLE COCKROACHES EXPOSED TO AN INSECTICIDE FORMULATION IN THE PRESENCE OR
ABSENCE OF GERMAN COCKROACH FECES (0.72 MG/CM2).
Treatment' Surface Obsb Slope + SE LT5 (95% CI) 2 IR
Cypermethrin clean 420 1.67 + 0.20 25.55 (21.22-32.16) 1.98 1.00
Cypermethrin feces 1080 1.75 + 0.11 113.35 (105.03-123.28) 14.54 4.44
Chlorpyrifos clean 600 15.24 + 1.08 47.12 (46.28-47.98) 3.86 1.00
Chlorpyrifos feces 810 13.22 0.83 117.21 (114.82-119.48) 2.65 2.49
Propoxur clean 500 11.00 + 0.83 23.30 (22.63-23.97) 5.23 1.00
Propoxur feces 500 10.21 1.48 28.69 (26.38-31.30) 7.67 1.23
"cypermethrin = 0.39 g[AI]/cm', chlorpyrifos = 11.7 pg[AI]/cm', propoxur = 43 pg [AI]/cm'.
'Total number of observations recorded until 80% mortality was achieved (repeated measure).
'IR = inhibition ratio (LT,, contaminated surface/LT,, clean surface).
Strong et al.: Feces Effect On Insecticide Efficacy
0 10 20 30
Days after treatment
Fig. 1. Residual efficacy of 3 X-cyhalothrin-cyhalothrin formulations (A, wettable
powder; B, capsulated suspension; C, emulsifiable concentrate) applied to glass Ma-
son jars in the presence (0) and absence (@) of German cockroach feces.
444 Florida Entomologist 83(4) December, 2000
of propoxur, Baygon 1.5 EC, was least affected by the presence of feces. The effective-
ness of another carbamate, bendiocarb, was unaffected by soil carbon content (Kamm
& Montgomery 1990).
Dramatic losses in X-cyhalothrin efficacy were observed in the longevity experi-
ments. For example, although the WP and CS formulations caused 95 to 100% mor-
tality through 30 d in the absence of feces, their efficacy was reduced to nearly 0%
mortality in the presence of feces at day 30.
Cockroach feces in the home may present problems other than control failures.
Human consumption of food products contaminated with cockroach feces may lead to
digestive disorders (Mullins & Cochran 1973) and feces often contain bacterial and
fungal pathogens (Koehler et al. 1990). Additionally, cockroach feces contain potent
allergens responsible for asthma and related respiratory disorders (Brenner 1995).
Therefore, cleaning cockroach feces from structures before insecticide treatment pro-
vides a 2-fold benefit-reduction of disease potential and improved insecticide effi-
We thank R. S. Patterson, J. Hogsette, (USDA-ARS, Center for Medical, Agricul-
tural and Veterinary Entomology, Gainesville, Florida) and S. J. Yu (University of
Florida, Department of Entomology and Nematology) for critical reviews of the manu-
script. We thank Priamo Mena and Deanna Harrison for technical assistance. This is
Florida Agricultural Experiment Station Journal Series No. R-06821.
A commonly used insecticide application method known as crack and crevice treat-
ment, was simulated in the laboratory to evaluate the effect of cockroach excrement
found in these areas on insecticide efficacy. The presence of cockroach excrement sig-
nificantly decreased the efficacy of several emulsifiable concentrate insecticides (chlo-
rpyrifos, cypermethrin, and propoxur). The efficacy of other insecticide formulations
(Wettable Powder, and Capsulated Suspension) were also decreased in the presence of
cockroach excrement. These results help to explain control failures when this method
of insecticide application is employed. Recommendations for improving the method
ABBOTT, W. S. 1925. A method for computing the effectiveness of an insecticide. Jour-
nal of Economic Entomology 18: 265-267.
BENNETT, G. W., J. M. OWENS, AND R. M. CORRIGAN. 1988. Truman's scientific guide
to pest control operations. pp. 127-145. Edgell Communications, Duluth, MN.
BRENNER, R. J. 1995. Economics and medical importance of German cockroaches, pp.
86-91. In M. K. Rust. J. M. Owens and D. A. Reierson [eds.], Understanding and
controlling the German cockroach. Oxford University Press, New York.
COCHRAN, D. G. 1997. Misuse of the tarsal-contact method for detecting insecticide in
the German cockroach (Dictyoptera: Blattellidae). Journal of Economic Ento-
mology 90: 1441-1444.
CORNWELL, P. B. 1972. The cockroach. Insecticides and cockroach control. pp. 151-152.
Associated Business Programmes LTD. London.
CORNWELL, P. B. 1968. The cockroach. A laboratory insect and industrial pest. pp. 249-
277. Hutchinson & CO LTD. London.
Strong et al.: Feces Effect On Insecticide Efficacy
HAMAKER, J. W., AND J. M. THOMPSON. 1972. Adsorption, pp. 49-143. In C. H. Goring
and J. W. Hamaker [eds.], Organic chemicals in the soil environment, vol. I.
Marcel Dekker, N.Y.
KAMM, J. A., AND M. L. MONTGOMERY. 1990. Reduction of insecticide activity by car-
bon residue produced by burning grass seed fields after harvest. Journal of Eco-
nomic Entomology 83: 55-58.
KOEHLER P. G., AND R. S. PATTERSON. 1986. A comparison of insecticide susceptibility
in seven nonresistant strains of the German cockroach, Blattella germanica
(Dictyoptera: Blattellidae). Journal of Medical Entomology 23: 298-299.
KOEHLER P. G., R. S. PATTERSON, AND R. J. BRENNER. 1990. Cockroaches, In A. Mallis,
[ed.], Handbook of pest control. Franzak and Foster Co., Cleveland, OH.
KOEHLER P. G., C. A. STRONG, AND R. S. PATTERSON. 1994. Rearing improvements for
the German cockroach (Dictyoptera: Blattellidae). Journal of Medical Entomol-
ogy 31: 704-710.
LEE, W. 0. 1973. Clean grass seed crops established with activated carbon bands and
herbicides. Weed Science 21: 537-541.
MULLINS, D. E., AND D. G. COCHRAN. 1973. Nitrogenous excretory materials from the
American cockroach. Journal of Insect Physiology 19: 1007-1018.
NEWTON, J., AND D. S. COOMBES. 1990. The efficacy against Blattella germanica of in-
secticides applied to greasy and clean surfaces of ceramic tile, stainless steel or
painted hardboard. International Pest Control 32 (2): 69-99.
NIEMCZYK, H. D., AND H. R. KRUEGER. 1987. Persistence and mobility of isazonos in
turfgrass thatch and soil. J. Econ. Entomol. 80: 950-952.
REE, 0. W. 1980. A laboratory evaluation of the effects of conditions and materials
found in food handling establishments on insecticides used for the control of
Blattella germanica (L). M.S. thesis, Oklahoma State University, Stillwater.
ROTH L. M., AND E. R. WILLIS. 1960. The biotic association of cockroaches. Smithso-
nian Institution Miscellaneous Collection 141: 1-470.
RUST, M. K. 1986. Managing household pests, pp. 335-368. In G. W. Bennett & J. M.
Owens [eds.], Advances in urban pest management. Van Nostrand Reonhold,
RUST, M. K., AND D. A. REIERSON. 1988. Performance of pyrethroids against German
cockroaches Blattella germanica (L.) (Dictyoptera: Blattellidae). Bulletin of the
Society of Vector Ecology 13: 343-349.
SAS INSTITUTE. 1988. SAS/STAT Guide for personal computers. SAS Institute, Cary,
SCHAL, C. 1988. Relation among efficacy of insecticides, resistance levels, and sanita-
tion in the control of the German cockroach (Dictyoptera: Blattellidae). Journal
of Economic Entomology 81: 536-544.
STEJSKAL, V. 1997. Distribution of faeces of the German cockroach, Blattella german-
ica, in a new refuge. Entomologia Experimentalis et Applicata 84: 201-205.
446 Florida Entomologist 83(4) December, 2000
SAMPLING OF DIAPHORINA
CITRI (HOMOPTERA: PSYLLIDAE) ON
ORANGE JESSAMINE IN SOUTHERN FLORIDA
JAMES H. TSAI', JIN-JUN WANG1'2 AND YING-HONG LIU1'2
1Fort Lauderdale Research and Education Center, IFAS, University of Florida
3205 College Avenue, Fort Lauderdale, FL 33314-7799
2Department of Plant Protection, Southwest Agricultural University,
Chongqing 400716, P. R. China
Dispersion indices and related statistics of Asian citrus psyllid, Diaphorina citri
Kuwayama, on orange jessamine [Murraya paniculata (L.) Jack] shoots in southern
Florida from 1998 to 1999 were determined with 235 data sets and used to develop
sampling plans. Three regression models, Taylor's power law, Iwao's patchiness re-
gression, and k = c +dm [k = m2 / (S2- m)] (where k is the parameter for the negative
binomial distribution) were used to analyze the data. Taylor's power law (a = 0.3407
+ 0.03, b = 1.2971 + 0.03, 7r = 0.88) fit the data better than Iwao's model (a = -0.3217
+ 0.12, B = 1.6979 + 0.06, r- = 0.76). Taylor's b and Iwao's B were both significantly >
1, indicating that D. citri populations were aggregated. Iwao's a was significantly < 0,
indicating that the basic distribution component ofD. citri was the individual insect.
The slope d (0.7489 + 0.48) was indistinguishable from 0, indicating the existence of
a common k (estimated as 1.2741). The incidence (P,, proportion infested) and mean
density (m) relationship was developed by negative binomial distribution (NBD) basis
and Nachman's model [ln (m) = 0.2277 + 1.2444 In (-In (Po)] (where Po = Proportion of
uninfested sampling units in a sample). The NBD was appropriate for studying D.
citri distribution based on comparison of NBD basis and Nachman's models. The re-
lationship to determine sample sizes for fixed levels of precision and fixed-precision-
level stop lines for sequential sampling was also developed.
Key Words: Asian citrus psyllid, Taylor's power Law, Iwao's patchiness regression,
negative bionomial distribution
Indices de dispersion y estadisticas relacionadas con el psila de citrico Asidtico,
Diaphorina citri Kuwayama, en brotes de Muralla paniculata (L.) Jack en el Sur de
Florida entire 1998 y 1999 fueron determinados con 235 conjuntos de datos y usados
para el desarrollo de planes de muestreo. Tres models de regresi6n, la ley de poder
Taylor, la regresi6n Iwao, y k = c + dm [k = m2 / (S2 m)] (donde k es el parametro para
la distribuci6n binomial negative) fueron usados para analizar los datos. La ley de po-
der Taylor (a = 0,3407 + 0,12, b = 1,2971 + 0,03, r2 = 0,88) encaja los datos mejor que
el modelo de Iwao (a = -0,3217 + 0,12,/3 = 1,6979 + 0,06, r2 = 0,76). La b de Taylor y el
/f de Iwao fueron ambos significativamente > 1, indicando que poblaciones de D. citri
fueron agregadas. El a fue significativamente < 0, indicando que el component basico
de distribuci6n de D. citri era el insecto individual. La pendiente d (0,7489 + 0,48) fue
indistinguible de 0, indicando la existencia de una k en comun (estimada a 1,2741). La
relaci6n entire incidencia (P,, proporci6n infestada) y densidad promedio (m) fue desa-
rrollada a base de la distribuci6n negative binomial (NBD) y el modelo de Nachman
[ln (m) = 0,2277 + 1,2444 In (-In (P,)] (donde P, = proporci6n de unidades de muestreo
Tsai et al.: Sampling of D. citri on orange jessamine
sin infestar en una muestra). El NBD fue apropiado para estudiar la distribuci6n de
D. citri basado en la comparaci6n entire la NBD y el modelo Nachman. La relacion
para determinar los tamafios de muestreo para niveles fijos de precision y sus lines
que contienen este parametro para muestreo secuencial tambien fueron desarrolla-
Citrus is one of the most important economic crops in the U.S. with about 500,000
ha in citrus orchards mostly in California, Florida, Texas, and Arizona. In Florida
alone, citrus encompasses 389,857 planted hectares with a total of 107 million trees
in the 33 citrus producing counties. The annual earning on citrus is estimated at
$1.1 billion (Tsai 1998). Citrus greening disease or Huanglungbin is the most seri-
ous disease of citrus in the world (Aubert et al. 1996, Tsai et al. 1988). The Asian cit-
rus psyllid, Diaphorina citri Kuwayama, is the most efficient vector of citrus
greening bacterium, Liberobacter asiaticum Jagoueix, Bove & Garnier, throughout
Asia and the Far East (Catling 1970, Pande 1971, Tsai et al. 1988). The combined
presence of a psyllid vector and a greening agent has been the limiting factor in cit-
rus production in these areas (Ke et al. 1988, Tsai et al. 1988). On June 3, 1998 the
Asian citrus psyllid was first found in southern Florida, with the subsequent discov-
ery of D. citri in Broward, Palm Beach, Martin, Dade, St. Lucie, Hendry, and Collier
Counties in a 3-month period (Halbert et al. 1998). Given high reproductive poten-
tial of this vector during favorable conditions of weather and food availability (J. H.
T., unpublished data), this pest is expected to spread throughout citrus producing
area in Florida in 2-3 years. It poses a serious threat to other citrus producing states
in the future. Based on our observations, this pest is most abundant on orange jes-
samine, Murraya paniculata (L.) Jack (J. H. T., unpublished data), which is widely
planted as hedges in the urban landscape in southern Florida. It could serve as an
alternate host for maintaining psyllid populations when young citrus shoots are not
Data on dispersion of pest populations is an important aspect of population biology
because it is a result of the interaction between individuals of the species and their
habitat (Sevacherian & Stern 1972). Knowledge of this dispersion allows a better un-
derstanding of the relationship between an insect and its environment and provides
basic information for interpreting spatial dynamics, designing efficient sampling pro-
grams for population estimation, and pest management (Harcourt 1961, Iwao 1970,
Sevacherian & Stern 1972, Taylor 1984), and the development of population models
(Croft & Hoyt 1983). Methods that are commonly used to describe dispersion of ar-
thropod populations have been summarized by Southwood (1978). Several estimates
based on sample mean (m) and variance (S2) are used as indices for aggregation (Lloyd
1967) and the dispersion parameter k for the negative binomial distribution (South-
wood 1978). Moreover, these indices are often convertible from one to another. Sam-
pling plans based on these descriptions of dispersion (Kuno 1969, Green 1970) reduce
sampling effort and minimize variation of sampling precision (Hutchison et al. 1988,
Kuno 1991, Trumble et al. 1989). Little is known about the dispersion of D. citri be-
cause of its new pest status in USA. To fill this void, we gathered data on the disper-
sion of D. citri adults on orange jessamine in southern Florida from 1998 to 1999.
From this information, two incidence-density relationships, the optimal sample sizes
for estimating density, and the sequential sampling plans suitable for intensive pop-
ulation research and pest surveys were developed.
448 Florida Entomologist 83(4) December, 2000
MATERIALS AND METHODS
Sampling of Citrus Psyllid Population
A field survey for sampling populations of citrus psyllids was conducted from Oc-
tober 1998 to May 1999 in ten orange jessamine fields in Broward County, Florida.
The plants were not sprayed with insecticides during the course of the study.
For the purpose of sampling, the field in each location was divided into five areas
(10 x 2 m). At weekly intervals, one shoots (about 6-10 cm long) was selected at ran-
dom from each square meter area by throwing a pointed object. Thus a total of 20
shoots were selected from each of the 5 areas on each sampling date. Numbers of cit-
rus psyllid adults per shoot were counted and recorded.
The mean density (m) per shoot and variance (S2) were calculated for shoots in
each field per sampling date and related to each other using Taylor's power law (Tay-
lor 1961, 1971, Taylor et al. 1978) and Iwao's patchiness regression (Lloyd 1967, Iwao
1968, Iwao & Kuno 1971).
Taylor's power law states that the variance (S2) of a population is proportional to
a fractional power of the arithmetic mean (m): S2= am'b. To estimate a and b, the values
of In (S2) were regressed against those of ln(fn) using the model
ln(S2) ln(a)+ bln(m) (1)
where the parameter a is a scaling factor related to sample size (Southwood 1978), the
slope b is an index of aggregation which indicates a uniform, random and aggregated
dispersion when b < 1, b =1, b >1, respectively.
Iwao's patchiness regression method quantifies the relationship between Lloyd's
(1967) mean crowding index (m*) and mean (m) by:
m* = a+ p (2)
where m* was determined as [m + (S2 / m-1)] (Lloyd 1967). The intercept (a) is the in-
dex of basic contagion and the slope (B) is the density contagiousness coefficient inter-
preted in the same manner as b of Taylor's regression.
Estimation of Incidence
The relationship between the proportion of samples with one or more animals (the
incidence, P,) and the density of animals (mn) per sample unit was developed using the
following two methods.
One was developed by assuming that a negative binomial distribution (NBD) with
variable k would describe the distribution of psyllids on the shoots. This assumption
was later tested. The NBD-based relationship was chosen because of the close rela-
tionship between NBD and Taylor's power law (Binns 1986). Estimated S2 was de-
scribed as a function of mn (Taylor 1961). With this relationship, k of the NBD can be
Tsai et al.: Sampling of D. citri on orange jessamine
calculated as [m2 / (am' mn)]. The incidence is then one minus the zero term of the
NBD (Wilson & Room 1983, Nyrop et al. 1989):)
P1 1/[(1 + i/k)k] (3)
Another P, and nr relationship was developed using the model proposed by Nach-
man (1984). Because this model does not use a theoretical probability distribution as
a basis, it was fit to the data to check the assumed applicability of the NBD (Nyrop et
al. 1989). In Nachman's model, the proportion of sample units with no animal (Po) is
related to the mean density as:
PO exp(-6mi) (4)
where 6 and y are parameters of the model. To fit the model to the data, the data were
calculated based on the sampling dates and locations. The model is linearized with the
mean density regressed on P, (Nyrop et al. 1989) as:
ln(m) A+Bln[- ln(Po)] (5)
To test the applicability of the negative binomial distribution, the proportion of
sample units with no psyllids (P,) was calculated for different means using the two in-
cidence and mean relationships and compared by chi square test.
Estimation of Common k for the NBD
The estimates of the dispersion parameter k for the NBD, computed as m2 / (S2 -
mn), were linearly regressed on nm,
k = c+dm (6)
to test for the existence of a common k (k) for each of the data sets (Southwood 1978,
Feng & Nowierski 1992). A d value significantly > 0 indicates the dependence of k on
mean density. The variance and mean within each area where the variance exceeded
the mean were used to estimate k, for a negative binomial distribution (Fleischer et
al. 1991). Estimates of k, were made using Elliot's (1977) techniques, which estimates
k, by regressing y'= (S2 m) on x'= (nz2- S2 I n), and k, was defined by k, = 1/slope. An
index for spatial aggregation of arthropod populations, I/k, which is equal to m* / mn
-1 (Southwood 1978) and is the same as Cassie's index C (Cassie 1962), was also em-
ployed to evaluate the dispersion patterns.
The general linear model procedure (GLM) of SAS (SAS Institute 1988) was used
to estimate the linear regression. Student's t tests were used to determine if the slopes
(b) of the regression lines were significantly > 1.0 (equations 1, 2) or 0 (equation 6)
(Sokal & Rohlf 1981).
We determined the sample sizes for fixed levels of precision by substituting Tay-
lor's variance-mean relationship into the usual expression for the standard error of
the mean and rearranging:
n = am -2/D2 (7)
450 Florida Entomologist 83(4) December, 2000
where n is the sample size and D is the required level of precision expressed as a pro-
portion of the mean, and a and b are the coefficients from Taylor's power law (Pena &
Duncan 1992, Walker & Allsopp 1993). We used two values of D, 0.10 and 0.25; the lat-
ter allows detection of doubling or halving of sample means (Southwood 1978),
whereas the former would be useful in detecting smaller changes in ecological studies
(Walker & Allsopp 1993).
The number of samples after which sampling can be terminated (T ) for a constant
precision, D, of the mean [D = (S / n)2 / m], was determined using the equation de-
rived by Green (1970):
logT, log(D2/a) + (b 1) logn (8)
where T is the stop line for sample size n, a and b are from Taylor's power law, and D
is defined as above.
RESULTS AND DISCUSSION
The complete data set consisted of 235 psyllid samples from ten locations covering
the period of October 1998 through March 1999. The mean density ofD. citri in sam-
ples ranged from 0.1 to 8.5 adults per shoot. The highest number of psyllid adults on
a shoot was 43. These 235 psyllid data sets were used for dispersion analysis.
The results of Iwao's regression of m* on m and Taylor's power law analysis are
listed in Table 1. Iwao's patchiness regression described well the relationship be-
tween mean crowding and density for D. citri (Table 1, Fig. 1). The constant a indi-
cates the tendency to crowding (+ ve) or repulsion (- ve) defined as the 'Index of Basic
Contagion' by Iwao (1970). For D. citri, the value of a was < 0 (t = -2.546, df = 233, P
< 0.05), indicating that for adults the basic component of the population is a single in-
dividual. Estimate of B, the density contagiousness coefficient, was significantly > 1 (t
=11.27, df = 233, P < 0.001).
Taylor's power law provided a highly significant relationship between variance (S2)
and mean density (Table 1, Fig. 2). Taylor's intercept, In (a), was significantly > 0 (t =
11.57, df = 233, P < 0.001; Table 1). Estimate of b was significantly > 1 (t =9.26, df =
233, P < 0.001).
TABLE 1. PARAMETER ESTIMATIONS OF FOUR LINEAR MODELS FOR D. CITRI ON ORANGE
Model Slope + SEM' Intercept + SEM2 N r2
Taylor's 1.2971 + 0.03*** 0.3407 + 0.03*** 235 0.8753
Iwao's 1.6979 + 0.06*** -0.3217 + 0.12* 235 0.7637
Nachman's 1.2444 + 0.06*** 0.2277 + 0.04*** 47 0.9189
Equation 6 0.7489 + 0.48 1.4216 + 0.99 235 0.0102
'Table entries are significant at level of P < 0.05 (*) or 0.0001 (***) for H,:n = 1 for Taylor's and Iwao's model
(t= [slope-1]/SE,,,o, df= N-1), and H,: a = 0 for Nachman's model and equation 6 (t= Slope/SE, oo, df= N-1).
'Table entries are significant at level of P < 0.05 (*) or 0.0001 (***) for H,: n = 0.
Tsai et al.: Sampling of D. citri on orange jessamine
2 4 6 8
Mean No. Psyllids Per Shoot (m)
Fig. 1. Regression analysis of Iwao's mean crowding index (m*) on mean density
(m) for D. citri populations on orange jessamine (m* = -0.3217 + 1.6979m, dots are es-
timated values based on observations).
Taylor's power law generally fit the data better, yielding higher value of r (0.8753)
than Iwao's model (0.7637). The aggregation indices (slopes, b and B) of Taylor's
power law and Iwao's patchiness regression were all significantly > 1 (P < 0.05), in-
dicating an aggregated dispersion distribution of D. citri on orange jessamine (Table
1). The causes of aggregation could be attributed to either active aggregation on the
part of this psyllid (such as behavior and reproductive biology), or to some heteroge-
niety of the environment (such as microclimate, preferred part of plant, and natural
enemies) (Southwood 1978). The similar observations were also reported by Van den
Berg et al. (1991) for the citrus psylla (Trioza erytreae Del Guercio), and Tret'yakov
(1984) for the apple psylla (Cacopsylla mali Schmidtb). Van den Berg et al. (1991) re-
ported that higher numbers of citrus psylla adults were apparently related to the
prevailing wind direction. Catling (1970) stated the population fluctuations of psyl-
lids were closely correlated with flushing rhythm and flush quality because eggs are
laid exclusively on young flush points and nymphs develop on immature leaves. Al-
though no data on D. citri is currently available for direct comparison, however, the
observed values for B and b are similar to those for many moderately aggregated in-
sects (Taylor 1961, 1971). Comparing to other aggregated Homopterans, the values
of B and b for D. citri were lower than those for the citricola scale Coccus pseudomag-
Florida Entomologist 83(4)
-4 -3 -2 -1 0 1 2 3
Fig. 2. Regression analysis of Taylor's power law (ln (S2) = In (0.3407) + 1.2971
ln(m)) for D. citri populations on orange jessamine (dots are estimated values based
noliarum (Kuwana) on citrus (Trumble et al. 1995), the wooly apple aphid Erisoma
lanigerum (Hausmann) on apple trees (Asante et al. 1993), and the Russian wheat
aphid Diuraphis noxia (Mordvilko) in wheat field (Feng & Nowierski 1992), but was
higher than those for the pea aphid Acyrthosiphon pisum (Harris) in alfalfa field
(Hutchison et al. 1988).
Estimation of Incidence
Nachman's model gave an excellent fit to the relationship between the proportion
of shoots without psyllids (Po) and mean density (m) ofD. citri (Table 1, Fig. 3). Using
the parameter estimates (Table 1), the proportion of shoots with or without psyllids
can be estimated from mean density with equation 5. For example, samples with
mean densities of 0.5 and 2 psyllids per shoot correspond to -38 and -77% infested
shoots, respectively. This model could be used for grove managers who wish to develop
the decision-making plans when economic threshold of D. citri becomes available in
The values of P, for various means calculated according to Nachman's model and
the NBD model are presented in Figure 4. Values of P, calculated with the NBD model
Tsai et al.: Sampling of D. citri on orange jessamine 4
-2 -1 0 1 2
Fig. 3. Regression analysis of Nachman's model (in (m) = 0.2277 + 1.2444 In(-
ln(Po)) for D. citri populations on orange jessamine (dots are estimated values based
were greater than those calculated by using Nachman's model at lower population
density. On the contrary, at higher population density level, the values of Po become
smaller. Generally, this deviation was small (<0.01) and the chi square test indicated
that the two models were similar and interchangeable (P > 0.05).
Estimation of Common k for the NBD
Figure 5 gives an overall picture on the relationship between k and the mean num-
ber of psyllids from the 177 samples where the variance exceeded the mean. Regres-
sion of k on the mean density per shoot using all data was not significant (F = 2.388,
7- = 0.0102, P = 0.1236). Moreover, the slope of regression (d) was not significantly
greater than 0 (t = 1.545, df= 233, P = 0.1236). Independence of k with the mean den-
sity suggests the existence of common k for the NBD of the psyllid populations. The
estimates of a common k was 1.2741 using Elliot's (1977) method.
Southwood (1978) states that changes in the density of an insect often lead to
changes in the distribution. However, we did not detect apparent density-dependent
distribution changes in the psyllid population (Table 1, Fig. 5). Similar results were
reported by Feng & Nowierski (1992) for the summer populations of Russian wheat
Florida Entomologist 83(4)
0 1 2 3 4 5 6
Mean No. Psy[llids Per Shoot (m)
Fig. 4. Proportion of shoots without D. citri predicted from the density per shoot
using NDB (dotted line) and Nachman's models (solid line).
aphid, Diuraphis noxia (Mordvilko), on winter wheat. The values of aggregation index
(1/k) <, =, and > 0 represent regularity, randomness, and aggregation of populations
in spatial patterns, respectively (Cassie 1962, Southwood 1978). In our study, 177
sample observations were > 0. This indicates that the populations of D. citri generally
were highly aggregated on orange jessamine. This further supports the results ob-
tained using Taylor's power law and Iwao's models. It should be noted that the values
of 1/ k of the remaining 58 observations (m < 0.5) were < 0 which suggested a spatial
distribution of regularity; which is consistent with the estimate of a (< 0) by Iwao's re-
gression (Table 1).
The relationships between mean psyllid density and required sample size for fixed
precision levels of 10 and 25% are shown in Figure 6. The stop line of the fixed-preci-
sion-level of 25% of the mean for sequential sampling is presented in Figure 7. The
stop line of the fixed-precision-level of 10% of the mean was not presented because of
the requirement for extremely large samples from the field. Based on computer sim-
ulation, the performance of the sequential sampling procedures improved with in-
Tsai et al.: Sampling of D. citri on orange jessamine 4
/ 5 *
0 2 4 6 8 10
Mean No. Psyllids Per Shoot (m)
Fig. 5. Scatter plots of k for the NBD over mean (m) for D. citri populations on or-
creasing psyllid density. Also because the variance-mean regression provided a good
description of the data (Table 1), regression variability had only a minor effect at very
low mean density.
Figures 6 and 7 show that these sampling plans required quite large sample sizes
to obtain relatively precise density estimates. Although such precision in density es-
timates may be required for research purpose, it will probably not be of practical use
in commercial citrus production.
A person sampling D. citri could use Figure 7 by plotting T7 (accumulated adults)
and n (number of shoots sampled) after each sample was taken. When the plot falls
above the line, sampling is stopped and the mean density (m) is within 25% of m = T
/ n. This sequential-count plan permits researchers and pest managers to describe
the mean density more accurately than before. It may lead to a better determination
of the economic threshold in the future.
Sampling small arthropods is operationally difficult and often time consuming.
In this paper, we have developed a sequential sampling plan based on counts of
psyllids which will be useful to anyone requiring accurate decisions based on mean
numbers. As a way of easing this burden presence-absence, (binomial) sampling
has been used in place of complete counts for estimating or classifying densities of
Florida Entomologist 83(4)
5 10 15 20
Psyllids per Sample
Fig. 6. Relationship between required sample size and mean density for achieving
a fixed precision level of 10 and 25% for D. citri populations on orange jessamine.
these organisms (Nyrop et al. 1989). Binomial sampling is appealing because it is
often easier to determine whether one or more animals reside on a sample unit
than it is to make a complete count. It is usually faster and therefore less costly on
a per-sample-unit basis. In addition, there are organisms, such as psyllids, for
which binomial sampling is the only feasible field-sampling method. Sequential
sampling plans can result in saving up to 75% of the time compared with fixed sam-
ple size procedures having comparable error rates (Harcourt 1966). When sampling
is used for decision making, it often suffices to classify a population density as op-
posed to obtaining an estimate. Many sequential sampling programs are based on
this premise. However, due to the new pest status and vector ability of D. citri, fur-
ther research on the biology, ecology, disease transmission and integrated manage-
ment are needed.
Appreciation is extended to Jon Allen, Jorge Pena and Bill Ruesink for reviewing
this manuscript and providing insightful comments. This is Florida Agricultural Ex-
periment Station Journal Series No. R-07603.
Tsai et al.: Sampling of D. citri on orange jessamine
5 10 15
No. of Samples Taken
Fig. 7. Sequential count plan for D. citri populations on orange jessamine, showing
the stop line at a fixed precision level of 25%.
ASANTE, S. K., W. DANTHANARAYANA, AND S. C. CAIRNS. 1993. Spatial and temporal
distribution patterns of Eriosoma lanigerum (Homoptera: Aphididae) on Ap-
ple. Environ. Entomol. 22: 1060-1065.
AUBERT, B., M. GRISONI, M. VILLEMIN, AND G. ROSSOLIN. 1996. A case study of huan-
glongbing (greening) control in Reunion, pp. 276-278. In: J. V. Da Graca, P.
Moreno and R. K. Yokimi [eds.] Proc. 13th Conf. Int. Org. Citrus Virologists
(IOCV). University of California, Riverside. 420 pp.
BINNS, M. R. 1986. Behavioral dynamics and the negative binomial distribution. Oi-
kos 47: 315-318.
CASSIE, R. M. 1962. Frequency distribution models in the ecology of plankton and
other organisms. J. Anim. Ecol. 31: 65-92.
CATLING, H. D. 1970. Distribution of psyllid vectors of citrus greening disease with
notes on the biology and bionomics of Diphorina citri. FAO Plant Prot. Bull. 18:
CROFT, B. A., AND S. C. HOYT. 1983. Integrated management of insect pests of pome
and stone fruits. Wiley, New York 454 pp.
ELLIOT, J. M. 1977. Some methods for the statistical analysis of samples of benthic in-
vertebrates, 2nd ed. Freshwater Biological Association, Cambridge, U.K.
458 Florida Entomologist 83(4) December, 2000
FENG, M. G., AND R. M. NOWIERSKI. 1992. Variation in spatial patterns of the Russian
wheat aphid (Homoptera: Aphididae) among small grains in the northwestern
United States. Environ. Entomol. 21: 1029-1034.
FLEISCHER, S. J., F. W. RAVLIN, AND R. C. REARDON. 1991. Implementation of sequen-
tial sampling plans for gypsy moth (Lepidoptera: Lymantriidae) egg masses in
eastern hardwood forests. J. Econ. Entomol. 84: 1100-1107.
GREEN, R. H. 1970. On fixed precision level sequential sampling. Res. Popul. Ecol. 12:
HALBERT, S., L. BROWN, AND W. DIXON. 1998. Asian citrus psyllid update. Fla. Dep.
Agric. Consum. Serv., Div. Plant Industry, 18 November 1998.
HARCOURT, D. G. 1961. Design of a sampling plan for studies on the population dy-
namics of the diamond-back moth, Plutella maculipennis (Curt.) (Lepidoptera:
Plutellidae). Canadian Entomol. 93: 820-831.
HARCOURT, D. G.1966. Sequential sampling for the imported cabbage worm,Pieris ra-
pae (L.). Canadian Entomol. 98: 741-746.
HUTCHISON, W. D., D. B. HOGG, M. ASHRAF POSWAL, R. C. BERBERET, AND G. W. CU-
PERUS. 1988. Implications of the stochhastic nature of Kuno's and Green's
fixed-precision stop lines: sampling plans for the pea aphid (Homoptera: Aphi-
didae) in alfalfa as an example. J. Econ. Entomol. 81: 749-758.
IWAO, S. 1968. A new regression method for analyzing the aggregation pattern of an-
imal populations. Res. Popul. Ecol. 10: 1-20.
IWAO, S. 1970. Problems in spatial distribution in animal population ecology, pp. 117-
149. In: G. P. Patil [ed.] Random counts in models and structures, 1. Pennsyl-
vania State University Press, University Park. 268 pp.
IWAO, S., AND E. KUNO. 1971. An approach to the analysis of aggregation patterns in
biological populations. pp. 461-513. In: G. P. Patil, E. C. Pielou and W. E. Wa-
ters [eds.] Statistical ecology, vol. 1. Pennsylvania State University Press, Uni-
versity Park. 582 pp.
KE, S., K. B. LI, C. KE, AND J. H. TSAI. 1988. Transmission of the Huanglungbin agent
from citrus to periwinkle by dodder. Pp. 258-264. In: S. M. Garnsey and L. W.
Timmer [eds.] Proc. 10th Conf. Int'l Org. Citrus Virol. 420 pp.
KUNO, E. 1969. A new method of sequential sampling to obtain the population esti-
mates with a fixed level of precision. Res. Popul. Eocl. 5: 31-43.
KUNO, E. 1991. Sampling and analysis of insect populations. Annu. Rev. Entomol. 36:
LLOYD, M. 1967. Mean crowding. J. Anim. Ecol. 36: 1-30.
NACHMAN, G. 1984. Estimates of mean population density and spatial distribution of
Tetranychus urticae (Acari: Tetranychidae) and Phytoseiulus persimilis (Aa-
rina: Phytoseiidae) based on the proportion of empty sampling units. J. Appl.
Ecol. 21: 903-913.
NYROP, J. P., A. M. AGNELLO, J. KOVACH, AND W. H. REISSIG. 1989. Binomial sequen-
tial classification sampling plans for European red mite (Acari: Tetranychidae)
with special reference to performance criteria. J. Econ. Entomol. 82: 482-490.
PANDE, Y. D. 1971. Biology of citrus psyllid, Diaphorina citri Kum. (Hemiptera: Psyl-
lidae). Israel J. Entomol. 5: 307-311.
PENA, J. E., AND R. DUNCAN. 1992. Sampling methods for Prodiplosis longifila
(Diptera: Cecidomyiidae) in limes. Environ. Entomol. 21: 996-1001.
SAS INSTITUTE. 1988. SAS user's guide: statistics. SAS Institute, Cary, NC. 1028 pp.
SEVACHERIAN, V., AND V. M. STERN. 1972. Spatial distribution patterns of Lygus bugs
in California cotton fields. Environ. Entomol. 1: 695-704.
SOKAL, R. R., AND F. J. ROHLF. 1981. Biometry, 2nd ed. Freeman, San Francisco, CA.
SOUTHWOOD, T. R. E. 1978. Ecological methods with particular reference to the study
of insect populations, 2nd ed. Chapman and Hall, London. 524 pp.
TAYLOR, L. R. 1961. Aggregation, variance and the mean. Nature 189: 732-735.
TAYLOR, L. R. 1971. Aggregation as a species characteristic, pp. 357-377. In: G. P. Pa-
til, E. C. Pielou and W. E. Waters [eds.] Statistical ecology, vol. 1. Pennsylvania
State University Press, University Park. 582 pp.
Davidson et al.: Whitefly Feeding System
TAYLOR, L. R. 1984. Assessing and interpreting the spatial distribution of insect pop-
ulation. Annu. Rev. Entomol. 29: 321-358.
TAYLOR, L. R., I. P. WOIWOD, AND J. N. PERRY. 1978. The density-dependence of spa-
tial behavior and the variety of randomness. J. Anim. Ecol. 47: 383-406.
TRET'YAKOV, N. N. 1984. The distribution of the apple weevil and the apple sucker in
orchards. Zashch. Rast. 9:42-43.
TRUMBLE, J. T., M. J. BREWER, A. M. SHELTON, AND J. P. NYROP. 1989. Transportabil-
ity of fixed-precision level sampling plans. Res. Popul. Ecol. 31: 325-342.
TRUMBLE, J. T., E. E. GRAFTON-CARDWELL, AND M. J. BREWER. 1995. Spatial disper-
sion and binomial sequential sampling for citricola scale (Homoptera: Coc-
cidae) on citrus. J. Econ. Entomol. 88: 897-902.
TSAI, J. H. 1998. Development, survivorship, and reproduction of Toxoptera citricida
(Kirkaldy) (Homoptera: Aphididae) on eight host plants. Environ. Entomol. 27:
TSAI, J. H., Z. Y. CHEN, C. Y. SHEN, AND K. X. JIN. 1988. Mycoplasmas and fastidious
vascular prokaryotes associated with tree diseases in China, pp. 69-240. In: C.
Hiruki [ed.] Tree mycoplasmas and mycoplasma disease. The University of Al-
berta Press. 245pp.
VAN DEN BERG, V. E. DEACON, AND P. J. STEENEKAMP. 1991. Dispersal within and be-
tween citrus orchards and native hosts, and nymphal mortality of citrus psylla,
Trioza erytreae (Hemiptera: Triozidae). Agric. Ecosystems Environ. 35: 297-
WALKER, P. W., AND P. G. ALLSOPP. 1993. Sampling distribution and sequential sam-
pling plans for Eumargarodes laingi and Promargarodes spp. (Hemiptera:
Margarodidae) in Australian sugarcane. Environ. Entomol. 22: 10-15.
WILSON, L. T., AND P. M. ROOM. 1983. Clumping patterns of fruit and arthropods in
cotton, with implications for binomial sampling. Environ. Entomol. 12: 50-54.
Davidson et al.: Whitefly Feeding System
IMPROVED ARTIFICIAL FEEDING SYSTEM FOR
REARING THE WHITEFLY BEMISIA ARGENTIFOLII
ELIZABETH W. DAVIDSON1, MICHELLE L. FAY', JACQUELYN BLACKMER2
AND MARK LAVINE3
'Department of Biology, Arizona State University, Tempe, AZ 85287-1501
2USDA/ARS Western Cotton Research Laboratory, Phoenix, AZ 85040
'Department of Entomology, University of Wisconsin, Madison, WI 53706
The artificial rearing system for Bemisia argentifolii Bellows and Perring has been
improved by the selection of an autoclavable, reusable membrane, reduction of the su-
crose concentration, choice of yeast extract, use of an antifungal agent in the petri
dish chamber, choice of surface sterilization agent, egg storage, and maintenance of
high humidity in the chamber during the entire nymphal development. We can now
produce small numbers of adult whiteflies of both sexes on these chambers, confirm-
460 Florida Entomologist 83(4) December, 2000
ing the utility of these improvements. We also reared two Bemisia tabaci A-strain to
adults and several Trialeurodes vaporariorum nymphs to the 3rd and 4th instar on
the improved feeding system.
Key Words: artificial diet, Bemisia tabaci A-strain, Trialeurodes vaporariorum
El sistema de crfa artificial para Bemisia argentifolii Bellows y Perring ha sido me-
jorado tras la selecci6n de una membrana reutilizable y autoclaveable, reducci6n en
concentraci6n de sacarosa, selecci6n de extract de levadura, uso de un agent anti-
hongo en la camara de placa petri, selecci6n de agent de esterilizaci6n de superficie,
almacenaje de huevos, y mantenimiento de alta humedad en la camara durante el
complete desarrollo ninfaceo. Ahora podemos producer numerous pequenos de moscas
blancas adults de ambos sexos en estas camaras, confirmando la utilidad de estas
mejorias. Tambien criamos dos Bemisia tabaci linaje-A con adults de varias ninfas
Trialeurode vaporariorum al 3ro y 4to instar bajo el nuevo sistema de alimentaci6n.
We previously reported an artificial system for rearing the silverleaf whitefly, Be-
misia argentifolii Bellows and Perring (= B. tabaci B-biotype), to the 3rd instar (Jan-
covich et al. 1997). This system used a polycarbonate chamber, a Parafilm
membrane, and a filter-sterilized diet consisting of 30% sucrose and 5% yeast extract
in distilled water. Although this system has proven useful for gut function studies
(E. W. D. & R. Rosell, unpublished) and for assays of potential ingested toxins (Jan-
covich et al. 1997; E. W. D., unpublished), it requires very careful, aseptic techniques
and rarely permits development beyond the 4th instar. Assays using this system often
failed due to fungal contamination. Our goals in improving the feeder system and diet
were to increase egg hatch, to reduce fungal contamination, and to bring a high pro-
portion of B. argentifolii nymphs to 3rd or 4th instar within 14 days, in order to use
these nymphs as hosts for parasitic wasps (Davidson & Jones 1999). We also wished
to investigate the use of the system for the culture of other whitefly species. We report
here improvements in this system that accomplish some of these goals and also enable
successful development of a proportion of B. argentifolii nymphs to adults.
MATERIALS AND METHODS
Several commercially available, autoclavable filtration membranes were tested to
replace the Parafilm@ membrane, including MSI@ TefSep (Micron Separations-Os-
monics Inc., Westboro, MA), Durapore SVLP (Millipore Inc., Bedford, MA) and Mil-
lipore TCTP membranes with pore sizes of 0.2-10 im. TefSep PTFE autoclavable
filtration membranes, 1.0 im, 45 mm diameter (MSI-Osmonics) were found to be most
acceptable and were used in all further experiments.
The feeder assembly, which consists of a bottom chamber, membrane, and upper
plate held together by binder clips (Jancovich et al. 1997), was fully assembled and au-
toclaved (120C for 20 min) before filling with the diet solution. Stainless steel 20 mm
electrophoresis binder clips ("Father Time Clips", Research Products International,
Mt. Prospect, IL) were used.
Davidson et al.: Whitefly Feeding System
To inhibit fungal contamination of feeders, the interior of sterile glass or dispos-
able plastic petri dishes was rinsed with a 0.1% solution of miconazole (Sigma, St.
Louis, MO) in 95% ethanol which was then permitted to evaporate, leaving a residue
of antifungal agent. High humidity was maintained by adding a damp filter paper tri-
angle to each petri dish and by placing a sterile slide, held in place with a bent hair-
clip, over the diet chamber after eggs were deposited. Sterile, filled feeders were held
individually in sterile petri dishes. Assembly took place under a laminar-flow hood
with a germicidal ultraviolet lamp, and all equipment was exposed to the ultraviolet
light for approximately 30 min before assembly.
Egg Harvest and Treatment
Bemisia argentifolii eggs were harvested from cotton, collards or melon. Leaves
were chosen that contained primarily darkened eggs, that were close to hatching.
Leaves were dipped sequentially into a detergent solution, distilled water, a 10%
household chlorine bleach solution (final concentration 0.5% sodium hypochlorite) for
2-3 min, to loosen the egg pedicel, and distilled water. In some experiments, after the
10% bleach step, leaves were dipped in a 3% solution of sodium thiosulfate to remove
residual chlorine, followed by a rinse with distilled water. Eggs were removed using
a WaterPik dental device, filtered through 3 layers of organdy cloth (6-8 fibers/mm)
and collected on coffee filters. Eggs were then transferred to sterile 50-ml plastic cen-
trifuge tubes and surface sterilized using 70% ethanol (approximately 1 min) followed
by either a 10% bleach solution or a 10% Roccal II solution (final concentration 1%
alkyl dimethyl butyl ammonium chloride, Sterling Drug) (2-3 min). In some experi-
ments, a rinse of 3% sodium thiosulfate solution followed the bleach step, to neutral-
ize residual chlorine. As Roccal is no longer commercially available, the product that
has replaced Roccal, Lysol IC (final concentration 1.1% alkyl dimethyl butyl ammo-
nium chloride, 0.12% didecyl dimethyl ammonium chloride), was also tested, as well
as 3% hydrogen peroxide. Eggs were rinsed in sterile distilled water and pipetted onto
feeder membranes, then excess water was removed. Assembled feeders were held in
sealed plastic boxes with an open water container, on the laboratory bench at 24-25C
and a photoperiod of 10:14 (L:D).
Bemisia tabaci Gennadius (A-strain) eggs were obtained from a colony maintained
at the University of California, Riverside. Trialeurodes vaporariorum Westwood eggs
were obtained from the USDA-ARS Biological Control of Insects Research Unit,
Weslaco, TX. Eggs were harvested from leaves and surface sterilized using Roccal as
Storage of Eggs
One cohort of B. argentifolii eggs washed from melon leaves was divided into 4 lots
and held in water at 4C for 0, 1, 4, or 7 days. The eggs were then surface sterilized
using Roccal and placed on feeders.
Yeast extract lots manufactured by Difco (Detroit, MI) and BBL (Becton Dickinson
Co., Cockeysville, MD) were compared at 5% concentration. Sucrose concentration
was compared at 30% and 15%. Dietary pH was adjusted to pH 5 to 8. The antifungal
agents methyl paraben and potassium sorbate (Sigma) were added to the standard
diet at concentrations shown to inhibit growth in whitefly diet of fungi isolated from
462 Florida Entomologist 83(4) December, 2000
Eggs were counted at day 1 after setup and nymphs were counted by instar at days
5 and 14 (+/- 1 day). Feeders were then held until day 28 to observe adult emergence.
Egg sterilization procedures and modifications to the diet were evaluated based upon
hatch percentage and percentage of nymphs that had achieved 3rd or 4th instar (in-
cluding the "red eye" stage) by day 14, based upon total nymphs at day 5. All diet mod-
ifications were evaluated in comparison to cohorts reared on standard diet (15% or
30% sucrose, as noted, plus 5% yeast extract). Means and standard deviations were
calculated using Excel 97 SR-2 (Microsoft), and ANOVA followed by Tukey's separa-
tion of means was performed using Systat version 8.0. Means were compared within
but not between experiments.
RESULTS AND DISCUSSION
The greatest improvement in the rearing technique has resulted from the adoption of
an autoclavable, commercially available membrane to replace Parafilm. TefSep filter
membranes are composed of Teflon (PTFE), are very thin (175 pm), and are hydropho-
bic. Higher hatch percentages were obtained on PTFE filter membranes than on Parafilm
(data not shown). The nymphs attached and fed readily in the oval spaces between the
plastic screen fibers that support the membrane. These feeding sites are equally spaced
across the membrane surface and occur in parallel rows, which facilitated counting of
eggs and nymphs. In preliminary trials, nymphs were unable to feed on membranes with
pore sizes of 0.2 or 0.5 im. The requirement for pores above 0.5 pm is probably related to
the cross-section diameter of the stylet bundle. In adult B. argentifolii, the stylet bundle
is approximately 0.3 im at the tip and 1.8 im in cross section closer to the head (Rosell
et al. 1995). These results suggest that the nymphs insert stylets through the pores in the
filter membrane, but do not puncture the membrane itself. Membranes with pore sizes of
2 im or larger were unacceptable due to leakage. The MSI TefSep 1 im membranes are
now used in all experiments, and the ability to autoclave the entire feeder system has
been a major improvement in reducing microbial contamination. These membranes are,
however, significantly more expensive than Parafilm (about $2.00 each).
Ten percent bleach was previously used both to loosen egg pedicels and to surface-
sterilize eggs washed from leaves (Jancovich et al. 1997). Rinses of leaves with sodium
thiosulfate 'N:, S 1-1 to neutralize chlorine led to decreased egg hatch but did not af-
fect development (Table 1A). Rinses of bleach-treated eggs with sodium thiosulfate
did not markedly improve hatch or development to 3rd or 4th instars (Table 1A). Sub-
stitution of the antibacterial-antifungal agent, Roccal, for bleach during egg surface-
sterilization resulted in a higher percentage of nymphs that reached the 3rd or 4th in-
star by day 14 (Table 1B). Unfortunately, Roccal is no longer manufactured. The prod-
uct that is sold as a replacement, Lysol IC, contains didecyl dimethyl ammonium
chloride in addition to the active ingredient in Roccal, alkyl dimethyl butyl ammo-
nium chloride. Lysol IC reduced egg hatch and development to 3rd or 4th instars
when compared with Roccal (Table 1C). Hydrogen peroxide, shown to be useful in sur-
face sterilization of leafhopper eggs (Wayandande & Fletcher 1998), led to clumping
of eggs on the membrane and reduced egg hatch, but development of nymphs that be-
came established was equivalent to that in Roccal and Clorox treatments (Table 1C).
High humidity appears to be essential to development of B. argentifolii nymphs on
the artificial diet system. Placement of sterile slides over diet chambers to maintain
high humidity within the chambers significantly improved development of nymphs to
3rd or 4th instars (Table ID). Leaf osmotic potential was similarly found to influence
hatch and survival of greenhouse whitefly eggs Castaine6 & Sav6 1993). However, all
water must be removed from the eggs after deposition on the membrane, as even a
small amount of liquid water inhibits egg ( hatch.
Davidson et al.: Whitefly Feeding System
TABLE 1. EFFECTS OF EGG STERILIZATION PROCEDURES, AND ADDITION OF STERILE
SLIDES TO GROWTH CHAMBERS ON HATCH PERCENTAGE AND NYMPHAL DEVEL-
OPMENT. PERCENTAGE OF NYMPHS ACHIEVING 3RD OR 4TH INSTAR BY DAY 14
WAS BASED UPON TOTAL NYMPHS AT DAY 5. STANDARD DIET (5% DIFCO YEAST
EXTRACT IN 15% SUCROSE, PH 7.0), 24-25C (UNLESS OTHERWISE NOTED),
PHOTOPERIOD 10:14 (L:D). MEANS ARE COMPARED WITHIN EACH EXPERIMENT.
MEANS FOLLOWED BY THE SAME LETTER ARE NOT SIGNIFICANTLY DIFFERENT
(ANOVA, a = 0.05).
Mean % egg hatch, Mean % 3-4 instars, Total nymphs,
Treatment day 5 (+/- S.D.) day 14 (+/- S.D.) day 5 (total feeders)
A. sodium thiosulfate neutralization of chlorine
Na2S20, rinse 51.0 (2.6) a 27.6 (6.5) a 1349 (7)
No rinse 61.8 (5.3) b 26.4 (3.9) a 1084 (7)
B. egg sterilization
Clorox 61.0 (8.1) a 36.5 (8.1) a 771 (6)
Clorox + Na2S20, 56.9 (2.5) a 39.5 (5.2) a,b 633 (6)
Roccal 59.5 (5.2) a 49.0 (7.5) b 596 (6)
C. egg sterilization
Hydrogen peroxide 24.3 (5.0) a 34.1 (3.3) a 333 (6)
Lysol IC 42.4 (5.7) b 27.1 (3.0) b 575 (6)
Roccal 54.7 (4.1) c 34.2 (4.8) a 924 (6)
D. addition of sterile slides to chambers
Slides 54.2 (4.5) a 42.3 (5.6) a 1175 (7)
No slides 55.6 (5.1) a 33.6 (2.8) b 1327 (7)
Eggs washed from leaves, but not surface-sterilized, can be stored in water at 4C
for at least one day with no reduction in egg hatch (Fig. 1A) or development to 3rd or
4th instars (Fig. 1B). Leaves bearing eggs can also be stored at least one day under re-
frigeration (data not shown). These procedures facilitate setup of experiments, as egg
harvest can be done at least one day in advance.
Difco and BBL yeast extract produced similar results (Table 2A, B). One lot of
yeast extract from Sigma did not produce any 3rd or 4th instar nymphs by day 14
(data not shown). Difco yeast extract that had been stored at room temperature for
more than 3 years was significantly less effective in maintaining whitefly growth than
fresh yeast extract (Table 2A, B). Yeast extract is now purchased in small quantities
and stored in a dessicator at room temperature or at -20oC.
Increasing the concentration of yeast extract in the diet from 5% to 7.5% resulted
in lower egg hatch, but did not affect development to 3rd or 4th instars (Table 2B). The
diet used in our original study (Jancovich et al. 1997) contained 30% sucrose, based
upon analysis of phloem sap. Reduction of sucrose concentration to 15%, however, had
slight or no adverse effect on nymphal development and resulted in greater egg hatch
(Table 2C, D). Egg hatch was greater on TefSep membranes than on Parafilm mem-
branes (data not shown). These results taken together suggest that water from the
diet may evaporate slowly through the porous membranes, contributing to higher hu-
midity around the eggs.
464 Florida Entomologist 83(4) December, 2000
60 A 60- B
Sa .o a
S240 0 -
20 2 7 0 1 4 7
Fig. 1. Effect of storage of eggs in water for 0, 1, 4 or 7 days at 4C, on A) hatch rate
and B) development. Percentage of nymphs achieving 3rd or 4th instar at day 14 was
based upon total nymphs at day 5. Standard diet (5% Difco yeast extract in 15% su-
crose, pH 7.0), 24-25C, photoperiod 10:14 (L:D). Bars with the same letter are not sig-
nificantly different (ANOVA, a= 0.05).
Diet pH did not significantly affect egg hatch. Developmental response was variable,
but pH 6.5 and 8.0 diets supported the highest percentage of nymphs to the 3rd and 4th
instar (Table 2E). A pH of 5.0 failed to support development to the 3rd or 4th instar (data
not shown). Salvucci et al. (1997) found that in adult whiteflies the optimum pH for in-
gestion of a 20% sucrose diet was between 6.5 and 7.5, in a tested range of 4.5 to 8.5.
The antifungal agent methyl paraben (p-hydroxybenzoic acid methyl ester), +/- po-
tassium sorbate, added to the diet at concentrations effective in inhibiting fungi, was
strongly inhibitory to both hatching and nymphal development (Table 2F). These
agents were apparently toxic to eggs and may have acted as antibiotics against the
symbiotic microorganisms that are necessary for whitefly development (Costa et al.
1997). The reduction of egg hatch due to addition of these compounds to the diet also
implies that dietary components other than water may make contact with the eggs
through the membranes. Eggs did not hatch when miconazole was used as an egg
rinse (data not shown). However, miconazole residue in the petri dishes holding the
feeders was beneficial in inhibiting fungal development in the petri dish chambers.
It is difficult to compare survivorship on artificial diet with that observed on
plants, since predation, parasitism, disease, plant quality and weather conditions are
not factors in mortality of artificially reared nymphs. On plants, survival from eggs to
adults can range from approximately 40% to over 80% (Horowitz et al. 1984, Powell
& Bellows 1992, Wagner 1995). Although production of adult whiteflies on the artifi-
cial diet was not the goal of this study, we observed emergence of adult whiteflies of
both sexes by 28 days equivalent to 0.5% to 2% of the total nymphs counted at day 5.
The highest percentages of B. argentifolii adults were produced when Roccal was used
to sterilize eggs, slides were added to chambers, and sucrose was reduced to 15%.
Davidson et al.: Whitefly Feeding System
cc cc cc c
<1 c w
00 o LC LC L0--
M-- cc Ll
10 L L0 0
CD t- t^
K ,K ,K ,
&) &)~ &
CM M 0
M 0 CM
C r-! rq
ccM M C
,q It cc
c~ c~ c~
~' 0 ~
cM C'] L- .
co C' C'] Ci
MD CD C)^
CO M cq i-
C'] -! 0 0
Florida Entomologist 83(4)
H 'm P
'- Q0m 1
o1 >-; -
C Q 1-] ~
(Z 0 1'
CO CO C)
CCi C] 'D
c~ -~ -~
0 ~ C]
in ~ 0
in ,~ ,~
~, ,~ ,~
in in in
,~ ,~ ,~
in in in
C) C) C)
~- ~- ~-
Davidson et al.: Whitefly Feeding System
TABLE 3. DEVELOPMENT OF BEMISIA TABACI A-STRAIN AND TRIALEURODES VAPORARI-
ORUM ON THE ARTIFICIAL FEEDING SYSTEM. PERCENTAGE OF NYMPHS ACHIEV-
ING 3RD OR 4TH INSTAR AT DAY 14 WAS BASED UPON TOTAL NYMPHS AT DAY 5.
STANDARD DIET (5% DIFCO YEAST EXTRACT, 15% SUCROSE, PH 7.0), 24-25oC,
PHOTOPERIOD 10:14 (L:D).
Mean % egg hatch, Mean % 3-4 instars, Total nymphs,
day 5 (+/- S.D.) day 14 (+/- S.D.) day 5 (total feeders)
Bemisia tabaci A-strain
19.8 (1.7) 26.0 (7.7) 271 (6)
Trialeurodes uaporariorum, experiment 1
22.3 (9.8) 4.7 (2.1) 84 (3)
Trialeurodes uaporariorum, experiment 2
31.2 (13.8) 3.7 (6.4) 78 (3)
Twenty-six percent of Bemisia tabaci (A-strain) developed to 3rd or 4th instar
within 14 days and two adult whiteflies emerged after 28 days on standard diet (5%
Difco yeast extract, 15% sucrose) (Table 3). Although percent hatch and development
for the A-strain whitefly was lower than that normally observed with B. argentifolii,
these results suggest that the feeder system is adequate for the development of B.
tabaci A-strain whiteflies, and may prove useful in investigating the physiological dif-
ferences between these closely-related species. The greenhouse whitefly, T vaporari-
orum, hatched and began to feed on the artificial diet system, but only a small number
developed to the 3rd or 4th instar by 14 days (Table 3). Nonetheless, development of
even a few T vaporariorum nymphs on the Bemisia artificial diet suggests that this
diet could provide the basis for a diet for the greenhouse whitefly. Short-term bioassay
of ingested compounds against greenhouse whitefly nymphs is possible using this as-
say system, similar to the plant tissue culture system used by Mele et al. (1992).
We are grateful to C. LaVesque, T. Perring and W. Jones for providing eggs of B.
tabaci A-strain and T vaporariorum, and to L. Lee for technical assistance. This re-
search was supported by USDA CSREES 9702182 and a Cooperative Agreement with
the USDA-ARS Biological Control of Insects Research Unit, Weslaco, TX.
CASTAlNE, C., and R. SAVE. 1993. Leaf osmotic potential decrease: a possible cause of
mortality of greenhouse whitefly eggs. Entomol. Exper. Appl. 69: 1-4.
COSTA, H. S., T. J. HENNEBERRY, AND N. C. TOSCANO. 1997. Effects of antibacterial
materials on Bemisia argentifolii oviposition, growth, survival, and sex ratio. J.
Econ. Entomol. 90: 333-339.
DAVIDSON, E. W., AND W. JONES, 1999. Successful rearing of parasitoid wasps on Be-
misia argentifolii cultured on artificial diet. Silverleaf Whitefly, National Re-
search, Action and Technology Transfer Plan, 1997-2001, Second Annual
Review of the Second 5-Yr Plan. USDA, ARS 1999-01, p. 69.
468 Florida Entomologist 83(4) December, 2000
HOROWITZ, A. R., H. PODOLER, AND D. GERLING. 1984. Life table analysis of the to-
bacco whitefly Bemisia tabaci in cotton fields in Israel. Acta Oecol. Applic. 5:
JANCOVICH, J. K., E. W. DAVIDSON, M. LAVINE, AND D. L. HENDRIX. 1997. Feeding
chamber and diet for culture of nymphal silverleaf whitefly, Bemisia argenti-
folii. J. Econ. Entomol. 90: 628-633.
MELE, E., J. MESSEGUER, R. GABARRA, J. TOMAS, J. COLL, AND F. CAMPS. 1992. In
vitro bioassay for the effects of Ajuga reptans phytoecdysteroids on Trialeu-
rodes vaporariorum larval development. Entomol. Exp. Appl. 62: 163-168.
POWELL, D. A. AND T. S. BELLOWS, JR. 1992. Preimaginal development and survival
of Bemisia tabaci on cotton and cucumber. Environ. Entomol. 21: 359-363.
ROSELL, R.C., J. E. LICHTY, AND J. K. BROWN. 1995. Ultrastructure of the mouthparts
of adult sweetpotato whitefly, Bemisia tabaci Gennadius (Homoptera: Aley-
rodidae). Int. J. Insect Morphol. & Embryol. 24: 297-306.
SALVUCCI, M. E., G. R. WOLF, AND D. L. HENDRIX. 1997. Effect of sucrose concentra-
tion on carbohydrate metabolism in Bemisia argentifolii: Biochemical mecha-
nism and physiological role for trehalulose synthesis in the silverleaf whitefly.
J. Insect Physiol. 43: 457-464.
WAGNER, T. L. 1995. Temperature-dependent development, mortality, and adult size
of sweetpotato whitefly Biotype B on cotton. Environ. Entomol. 24: 1179-1188.
WAYANDANDE, A. C. AND J. FLETCHER. 1998. Development and use of an established
cell line of the leafhopper Circulifer tenellus to characterize Spiroplasma citri-
vector interactions. J. Invertebr. Pathol. 72: 126-131.
Florida Entomologist 83(4)
COMPARING EFFECTS OF THREE ACARICIDES ON
VARROA JACOBSONI (ACARI: VARROIDAE) AND APIS
MELLIFERA (HYMENOPTERA: APIDAE) USING
TWO APPLICATION TECHNIQUES
GERARDO PEREZ SANTIAGO1, GABRIEL OTERO-COLINA2, DAVID MOTA SANCHEZ2,
MARTHA ELVA RAMIREZ GUZMAN3 AND REMY VANDAME4
'Instituto Politecnico Nacional, CIIDIR Unidad Durango. Zarco 106,
34890 Vicente Guerrero, Dgo., MEXICO.
'Colegio de Postgraduados, Instituto de Fitosanidad, 56230 Montecillo,
'Colegio de Postgraduados, Instituto de Socioeconomia, Estadistica e Informatica,
56230 Montecillo, Mex. MEXICO.
4Colegio de Postgraduados, IFIT-Campo C6rdoba, and El Colegio de la Frontera Sur.
Apdo. postal 36, 30700 Tapachula, Chis. MEXICO.
Two bioassays were administered to determine the dose-lethality response of Var-
roa jacobsoni Oudemans and the honey bee, Apis mellifera L., to amitraz, flumethrin
and fluvalinate. The first bioassay method was spraying by means of the Potter-
Bourgerjon's tower. The results are expressed in mean lethal concentrations (LC,0).
Perez-Santiago et al.: Acaricides on varroa and honey bees 469
The second method was topical application by means of microsyringe and manual ap-
plicator. The results are expressed in mean lethal doses (LDs). Both LCs, and LD0, val-
ues were considerably higher in honey bees than in varroa mites, showing that a wide
margin of safety exists between effective doses against mites and harmful doses for
honey bees. Both methods gave similar confidence intervals; they showed a compara-
ble sensitivity to changes in dose or concentration of pesticides.
Key Words: amitraz, bioassays, flumethrin, fluvalinate, honey bees, susceptibility,
toxicity, varroa mites
Se probaron dos m6todos de bioensayos toxicol6gicos para determinar la respuesta
dosis-letalidad de amitraz, flumetrina y fluvalinato sobre Varroajacobsoni Oudemans
yApis mellifera L. El primero fue aspersi6n por medio de la torre de Potter-Burgerjon;
sus resultados se expresan en concentraciones letales medias (CL,,). El segundo fue
aplicaci6n t6pica por medio de microjeringa y aplicador manual; sus resultados se ex-
presan en dosis letales medias (DL,,). Las DL0, y CLs, de todos los products fueron
considerablemente mas altas en abejas que en acaros, lo cual muestra que existe un
amplio margen de seguridad entire dosis que son lo suficientemente t6xicas sobre los
acaros, sin llegar a ser peligrosas para las abejas. Ambos m6todos de bioensayo dieron
intervalos de confianza comparable y presentaron similar sensibilidad en la res-
puesta a los cambios de dosis y concentraci6n aplicados.
Beekeepers in many parts of the world face severe problems because of recent intro-
ductions of a parasitic mite, Varroa jacobsoni Oudemans (Acari: Varroidae), known as
varroa. Originally from tropical Asia and found on the Indian honey bee, Apis cerana
Fabricius, this mite has shifted to its new host A. mellifera L. Owing to human activi-
ties, it has infested most of honey bee colonies around the world, causing severe losses.
Many control measures have been developed for varroa. Most include the use of
chemicals. However, chemical control has the disadvantages of variable efficacy, in-
creased costs, contamination of hives and hive products and the risk of target pest resis-
tance. Varroa resistance to fluvalinate was documented by the first time in Italy
(Lodesani et al. 1995) and soon in several European countries (Londzin & Sledzinski
1996, Moosbeckhofer & Trouiller 1996, Bruneau et al. 1997, Vandame et al. 1995). Elzen
et al. (1999), by application of discriminating doses of fluvalinate, found indications that
varroa mites from Florida and California were developing resistance to this acaricide.
Development of acaricide resistance by varroa is of concern. Chemical control nec-
essarily involves contact of pesticides with bees and hives. When resistance occurs,
doses should not be increased because of the risk of harming or killing bee hosts and in-
creasing contamination in the hive environment and hive products (Lodesani et al.
1992). Toxicological bioassays can track changes in pesticide susceptibility of a popula-
tion, by detecting changes in the calculated mean lethal concentrations or doses (LCs,
or LD50, respectively), compared to a maximum reference susceptibility or baseline
(Georghiou 1963). Early detection of pesticide resistance is mandatory for developing
a long-term strategy of chemical control, based on replacing ineffective pesticides. Bio-
assay methods must be sensitive to dose variations and easily repeatable, to allow com-
parison of lethal values (Lagunes-Tejeda & Villanueva-Jimenez 1994).
Topical application bioassays have been conducted on varroa by various research-
ers. Ritter & Roth (1986) determined mite susceptibility to Folbex VA (bromopropi-
470 Florida Entomologist 83(4) December, 2000
late) and K79 (clorodimeformidrochloride); they found a positive correlation between
lethal doses and number of previous treatments, suggesting early manifestations of
resistance. Also by topical application, Abed & Ducos de Lahitte (1993) estimated
LDs0's of amitraz and coumaphos.
A spraying method of application for toxicological bioassays has been proposed by
Colin et al. (1994), who used the Potter-Burgerjon's tower in testing lethality and be-
havioral effects of pesticides on varroa mites. This device sprays doses onto an area,
simulating a field application. In this method, data are expressed in lethal concentra-
tions (LC,0) of the material surrounding the specimen; the exact quantity of pesticide
contacting the specimen is unknown. Units are mg L1', parts per million (ppm), g cm2 or
Study of varroa populations established in Mexico may provide useful information
to other parts of the world. According to Otero-Colina & Santillan-Galicia (1996), these
mites were first detected in Veracruz state in the Mexican Gulf Coast lowlands in
1992, although they probably were already present there since about 1989. Before
their discovery and at least three years afterwards, they were seldom chemically
treated. Thus, they have been almost free of selection pressure by pesticides for at least
six years and supposedly show maximal levels of susceptibility to most acaricides.
The present study had the following objectives: a) to estimate LC,0 and LD50 on V
jacobsoni and A. mellifera to the acaricides amitraz, flumethrin and fluvalinate, and
b) to compare two toxicological bioassay methods for determining susceptibility to
these pesticides of varroa mites and honey bees.
MATERIALS AND METHODS
All varroa specimens were obtained from a commercial apiary that had received a
single treatment of fluvalinate (Apistan, Novartis) one year before. Adult female mites
were collected manually, from CO2 anesthetized worker bees or by uncapping parasitized
worker pupae. Mites were kept at 25C and 50% R. H. and put on pupae until they were
used in the tests, up to 4 hours later. Worker bees were collected from combs of healthy
(non-parasitized or with low levels of infestation) European colonies (Apis mellifera li-
gustica Spinola). In order to avoid recently emerged nurse bees and to use bees of similar
age, collections were made from combs without open brood (Felton et al. 1986). Adult
bees were transported to the laboratory and used in bioassays about two hours later.
All acaricides were used in commercial formulations; they were amitraz (Taktic,
12.5%, liquid, Hoechst), flumethrin (Bayticol, 3%, emulsifying concentrate, Bayer)
and fluvalinate (Fluvalin, emulsifying concentrate 25%, Isihara). Commercial for-
mulations were preferred as they are easily available and because they are currently
in use against varroa in many countries (Arculeo et al. 1989, Benitez-Reynoso 1998,
Cardenal Galvan et al. 1989). In the spraying method, the solvent was water; in top-
ical application, the solvent was acetone.
For each pesticide, preliminary bioassays were conducted to obtain maximal dose
causing 0% mortality and minimal dose causing 100% mortality. Then, logarithmic
intermediate doses were applied to obtain the LC,0 and LDs0. Four to seven interme-
diate doses plus extreme values were used in each replication. All dilutions were pre-
pared immediately before being used.
Bioassays on Mites
When the spraying method was carried out, a Potter-Burgerjon's tower was cali-
brated for applying 1.7 mg cm-2 (s.d. = 0.14) of acaricide solutions, by spraying 15 mL
solution at a pressure of 0.703 kg cm-2, then waiting one minute for sedimentation of
Perez-Santiago et al.:Acaricides on varroa and honey bees 471
droplets. A solid cone nozzle (Cat. 1/4J-SS+SU1A-SS, Spraying Systems) was used.
Groups of 14 varroa females were placed in a 14 cm diameter Petri dish containing a
floor of filter paper; each group was treated by an acaricide, then transferred to an-
other Petri dish (5 cm diameter), and incubated at 32 + 2C, 70 + 10% RH. To feed the
mites, two or three worker pupae one to three days old were placed in each Petri dish.
Pupal age was determined by their light yellow thorax, according to Jay (1953). Mor-
tality data were taken 24 hours after the treatment.
For topical application, groups of 14 varroa females were stuck ventral side up on
a microscope slide with Scotch double sided tape; 0.1 mL of pesticide solution was
then applied ventrally to each mite using a microsyringe and a microapplicator. This
contrasted with the method proposed by Ritter & Roth (1986), who applied 0.2 mL so-
lution. The slides were placed in an incubator at 25 + 2oC and 70 + 10% R.H. (instead
of 16C and 98% R.H., by the same authors). Mortality was recorded 24 h later. A spec-
imen was considered dead when it did not respond to tactile stimuli. All tests com-
prised four replications per dose on different days; a solvent-only control was included.
Bioassays on Honey Bees
To compare results, the same bioassay methods were used on bees, with some dif-
ferences owing to size, flying ability and nutritional requirements as given below. All
bees were anesthetized with a stream of CO; in the spraying test, groups of 30 work-
ers were confined in a galvanized iron cage (15 x 20 x 25 cm, 4 mm mesh) with a filter
paper floor, then sprayed in the Potter-Burgerjonas tower. In the topical application
test, every bee in a group of 30 received 1 mL solution dorsally on the thorax. In both
tests, after exposure to chemicals, the groups of bees replicationss) were confined in 1
L plastic cages; they were supplied with solid food (candy) and water, and incubated
at 25 + 2C and 70 + 10% R.H. Every bioassay had three replications.
Analysis of Results
Percentages of mortality were corrected byAbbotes (1925) formula when mortality
was found in the control; when mortality of one bioassay exceeded 10% of bees and
15% of mites, the results were discarded. Values of LCs, and LDs, and their confidence
intervals were estimated by Probit analysis. Relative toxicity of all products was esti-
mated in varroa and in bees, by dividing experimental lethal values by the most toxic
value. Toxicity of all products was also compared on mites vs. bees, by dividing LC,0
and LD0, values.
Results of aspersion and topical methods are expressed in different units and
their values are not comparable. However, an attempt was made to compare these
methods taking the width of confidence intervals as a measure of precision and slopes
as a measure of sensitivity, the last by means of Student's t-test (Dittrich 1962). Ease
of bioassay methods was also considered.
Susceptibility of varroa
Spraying. LC0, and confidence intervals are shown in Table 1. Previous studies of
fluvalinate LC0, levels on varroa were conducted using a residual application tech-
nique, but these results are not comparable with those of the current work, because
different bioassay methods were used. Milani (1995) placed varroa specimens on flu-
472 Florida Entomologist 83(4) December, 2000
TABLE 1. LC5o, LDso AND RANGE (CONFIDENCE INTERVALS, 95%) OF FIVE ACARICIDES
AGAINST VARROA JACOBSONI, BY SPRAYING AND TOPICAL APPLICATION, RE-
LC,, mg Confidence HV1/LV2 LD5o pg Confidence HV/LV
Acaricide L-1 int. (LV-HV) (LC;o) mite int. (LV-HV) (LD;o)
Amitraz 0.23 0.14-0.37 2.68 1.7 1.21-2.39 1.98
Flumethrin 875.083 201-6554 32.61 0.46 0.36-0.59 1.62
Fluvalinate 0.19 0.13 -0.29 2.31 15.42 9.91-24.94 2.52
valinate-impregnated paraffin and determined a LC5o of 20 mg L1 for a susceptible
population from Udine, whereas a resistant population from Lombardy (both in Italy)
showed a LC,5 higher than 200 mg L1. Vandame et al. (1995), using fluvalinate-
sprayed surfaces, estimated a LC5o of 0.21 mg per mL of sedimented solution in sam-
ples from Brignoles, while specimens from Draguignan (both in France) had a LC5o of
2.67 mg mL'1, indicating a twelve fold resistance factor.
According to the above statements, Mexican varroa populations are considered to
have maximum levels of susceptibility to most acaricides, owing to their isolation
from chemically-selected strains. Thus, LC5o values obtained in this study are pro-
posed as baselines for testing acaricides.
Topical application. Table 1 shows LD5, against varroa. Abed & Ducos de Lahitte
(1993) estimated an amitraz LD5o of 2.16 pg per mite, with a confidence interval of
1.46-3.2 pg. These values are close and overlap values obtained in the current work;
this fact suggests comparable levels of susceptibility in both mite populations. Base-
line data expressed as LD,5 are proposed now as they appear in Table 1.
Susceptibility ofApis mellifera
Spraying. Results are shown in Table 2. There are no published data for direct com-
parison with our results, since most research on bee toxicology used oral and contact
bioassay methods (Oomen 1986). According to a pesticide classification of Felton et al.
(1986) of toxicity to honey bees, flumethrin and fluvalinate belong to Group 1, highly
toxic pesticides, with LD,5 < 1 pg/bee. Amitraz belongs to Group 2, moderately toxic,
with LD5U 1-10 pg/bee.
TABLE 2. LC5o, LD5u AND RANGE (CONFIDENCE INTERVALS, 95%) OF FIVE ACARICIDES
AGAINST APIS MELLIFERA, BY SPRAYING AND TOPICAL APPLICATION, RESPEC-
LC5U Confidence HV1/LV2 LDo5 Confidence HV/LV
Acaricide pg L1 int. (LV-HV) (LC5o) pg/bee int. (LV-HV) (LD5o)
Amitraz 1636 983.79-2825 2.32 2.55 1.57-4.32 2.75
Flumethrin 46.87 21.15-95.61 4.52 0.05 0.03-0.09 3.26
Fluvalinate 1601 1429-1803 1.26 0.97 0.57-1.66 2.91
Perez-Santiago et al.:Acaricides on varroa and honey bees 473
Topical application. LDso and confidence intervals appear in Table 2; previous data
were obtained by Oomen (1986) for amitraz: LD5, > 16 LD5o ig/bee, and by Bornek
(1989) for fluvalinate: LD5, = 4.66 pg/bee, using Mavrik; LD5, = 9.12 pg/bee, using
Klartan. Amitraz and fluvalinate LD5o values estimated herein are lower than those
obtained by both authors; however, data cannot be accurately compared because of
different experimental conditions and analytical methods.
Tables 3 shows relative toxicity values for all acaricides used on varroa mites and
honey bees. Consistently, flumethrin was the most toxic product, while fluvalinate
and amitraz showed a lesser similar toxicity.
The rates of bee LCo, or LDo divided by mite LCo, or LD, are presented in Table 4.
These data show that all products have acaricidal, rather than insecticidal action; dif-
ferent toxicity ranges from 500 fold to more than one million fold. This indicates a
wide safety margin between lethal levels against mites and toxic levels for honey bees.
Comparison of bioassay methods
As a measure of sensitivity, slopes resulting from spraying and topical application
were analyzed. In most cases they attained the quality criteria proposed by Ibarra &
Federici (1987) for toxicological bioassays. Table 5 shows a comparison of slopes for
spraying vs. topical application (Student t test, a = 0.05). Significantly higher slopes
for spraying method occurred only in amitraz and fluvalinate applied to honey bees,
representing their greater sensitivity to spraying.
DISCUSSION AND CONCLUSIONS
Precision, as estimated by means of the confidence intervals, is shown in Tables 1
and 2. Although in several cases the quotient HV/LV exceeded the value of 2 (proposed
by Ibarra & Federeci 1987, as the highest permissible limit), sufficiently accurate es-
TABLE 3. RELATIVE TOXICITY OF ACARICIDES ON VARROA JACOBSONI AND APIS MEL-
on varroa on bees
Acaricide Spraying Topical Spraying Topical
Flumethrin 1 1 1 1
Amitraz 262.83 3.7 34.9 51
Fluvalinate 217.12 33.52 34.16 19.4
474 Florida Entomologist 83(4) December, 2000
TABLE 4. COMPARATIVE LCo AND LDso FOR ACARICIDES USED AGAINST VARROA JACOB-
SONI AND APIS MELLIFERA.
LCo..... LC LD o ..... LE,
Acaricide varroa varroa
Amitraz 7113.04 1.5*10'
Flumethrin 5360.81 1*105
Fluvalinate 8426.31 6.3*104
timates of LC50 and LD5s were obtained in both aspersion and topical methods. An im-
portant exception is the large confidence interval shown by spraying of flumethrin on
varroa; no explanation for this fact can be given.
Samples included mixed specimens obtained from adult bees and uncapped pupae,
which constituted a potential source of variation (Milani & Della Vedova 1996), and no
attempt was made to detect differences in susceptibility between such origins. How-
ever, obtaining female mites from a single source was a difficult task, and confidence
intervals may reflect this possible variation.
The spraying method has the advantage of treating all insects or mites at the same
time; sticking individual mites to a slide as well as topical application to honey bees
and mites are very laborious procedures. In addition, by using the Potter-Burgerjon's
tower, the amount of applied droplets could be narrowly regulated. Thus spraying
proved to be more practical for testing on varroa mites, regardless of the need to reg-
ularly calibrate the spraying nozzle.
Although both application methods can be useful, spraying showed a more sensi-
tive response of honey bees and it is easier in both species. So we consider it the best
Fluvalinate has been widely used and, as expected, mites have developed resis-
tance to it in many localities. Reproduction of whole bioassays as well as use of their
estimated LCgo or LDgo as a discriminant screen will aid to decide its eventual replac-
ing in a local or regional basis. Like fluvalinate, flumethrin is a pyrethroid. Thus, a
risk exists of cross-resistance, as shown by Milani (1995). Its useful life is expected to
be shorter and so early detection of resistance is important. Since amitraz is not
chemically related to the pyrethroids, if an efficient and environmentally acceptable
acaricide containing amitraz is available to beekeepers, it could be an option for alter-
nating with pyrethroid treatments.
TABLE 5. COMPARISON OF SLOPES (8) FOR THE DOSE-LETHALITY RELATIONSHIP OF
SPRAYING (S) VS. TOPICAL (T) TESTS CONDUCTED ON V. JACOBSONI AND A.
MELLIFERA, BY MEANS OF A STUDENT T TEST (A = 0.25). Ho: BS = BT.
On Varroa On bees
Acaricide Bs Results of t test
1> Ho rejected, bs > bt.
= H, accepted.
Perez-Santiago et al.:Acaricides on varroa and honey bees 475
ABBOTT, W. S. 1925. A method of computing the effectiveness of an insecticide. J. Econ.
Entomol. 18: 265-267.
ABED, T., AND J. DUCos DE LAHITTE. 1993. De-'terminatios de la DL,0 de l'amitraze et
du coumaphos sur Varroa jacobsoni Oud. au moyen des acaricides Antivarroa
(Schering) et Perizin (Bayer). Apidologie 24(2): 121-128.
ARCULEO, P., C. GALLO, AND P. GENDUSO. 1989. Comparative tests with fluvalinate,
coumaphos, amitraz and semiochemicals (Ecomone) against Varroa jacobsoni
(Oud.). In: R. CAVALLORO (ed.). Present status of varroatosis in Europe and
progress in the varroa mite control. Proceedings of a Meeting of the EC Ex-
perts' Group/Udine, Italy. ECSC-EEC-EAEC, Luxembourg: 311-314.
BENITEZ-REYNOSO, R. 1998. Control de Varroa jacobsoni (Acari: Varroidae) con flu-
valinato en insertos preparados artesanalmente. Tesis professional, Univer-
sidad Aut6noma Chapingo, 72 pp.
BORNECK, R. 1989. Fluvalinate, an interesting molecule against Varroa jacobsoni. In:
R. CAVALLORO (ed.). Present status of varroatosis in Europe and progress in
the Varroa mite control. Proceedings of a Meeting of the EC. Expert's Groupe/
Udine Italy ECSC-EEC-EAEC, Luxembourg: 223-229.
BRUNEAU, E., F. JACOBS, AND J. TROUILLER 1997. Re-'sultats de la champagne de-2detection
de la resistance de varroa aux pyrethrinoides en Belgique 1997. Belgique-Api-
cole 61(10): 294-297.
CARDENAL-GALVAN, J. A., A. GOMEZ-PAJUELO, AND E. C. LOPEZ-SEPiLVEDA GARCIA.
1989. Using fluvalinate inserts in Varroa control. In: R. CAVALLORO (ed.).
Present status of varroatosis in Europe and progress in the varroa mite control.
Proceedings of a meeting of the EC Experts' Group/ Udine, Italy, 28-30 Novem-
ber 1988. ECSC-EEC-EAEC, Luxembourg: 339-342.
COLIN, M. E., F. CIAVARELLA, G. OTERO-COLINA, AND L. P. BELZUNCES. 1994. A
method for characterizing the biological activity of essential oils against Varroa
jacobsoni. In: A. MATHESON (ed.). New perspectives on Varroa. Intl. Bee Re-
search Assoc. 109-114.
DITTRICH, V. 1962. A comparative study of toxicological test methods on a population
of the two spotted spider mite (Tetranychus telarius). J. Econ. Entomol. 55(5):
ELZEN, P. J., F. A. EISCHEN, J. R. BAXTER, G. W. ELZEN, AND W. T. WILSON. 1999. De-
tection of resistance in US Varroa jacobsoni Oud. (Mesostigmata: Varroidae) to
the acaricide fluvalinate. Apidologie 30(1): 13-17.
FELTON, J. C., P. A. OOMEN, AND J. H. STEVENSON. 1986. Toxicity and hazard of pes-
ticides to honey bees: harmonization of test methods. Bee World. 67: 114-124.
GEORGHIOU, G. P. 1963. Comparative susceptibility to insecticides of two green
peach aphid populations, collected 16 years apart. J. Econ. Entomol. 56(5):
IBARRA, J. E., AND B. A. FEDERICI. 1987. An alternative bioassay employing neonate
larvae for determining the toxicity of suspended particles to mosquitoes. J.
Amer. Mosquito Control Assoc. 3(2): 187-192.
JAY, C. 1953. Colour changes in the honey bee pupae. Bee World 34: 48-51.
LAGUNES-TEJEDA, A., AND J. A. VILLANUEVA-JIMENEZ. 1994. Toxicologia y manejo de
insecticides. Colegio de Postgraduados. M6xico. 264 pp.
LODESANI, M., A. PELLACANI, S. BERGOMI, E. CARPANA, T. RABITTI, AND P LASAGNI.
1992. Residue determination for some products used against Varroa infestation
in bees. Apidologie 23(3): 257-272.
LODESANI, M., M. COLOMBO, AND M. SPREAFICO. 1995. Ineffectiveness of Apistan
treatment against the mite Varroa jacobsoni Oud. in several districts of Lom-
bardy (Italy). Apidologie 26: 67-72.
LONDZIN, W., AND B. SLEDZINSKI. 1996. Opornosc roztocza Varroa jacobsoni na srodki
warroabojcze zawierajace tau-fluwalinat. Medycyna-Weterynaryjna 52(8): 526-
476 Florida Entomologist 83(4) December, 2000
MILANI, N. 1995. The resistance ofVarroajacobsoni Oud. to pyrethroids: a laboratory
bioassay. Apidologie 26: 415-429.
MILANI, N., AND G. DELLA VEDOVA. 1996. Determination of the LC,0 in the mite Var-
roajacobsoni of the active substances in Perizin and Cekafix. Apidologie 27:
MOOSBECKHOFER, R., AND J. TROUILLER. 1996. Apistanresistente Varroamilben in Os-
terreich entdeckt! Bienenvater 117(10): 372-373.
OOMEN, P. A. 1986. A sequential scheme for evaluating the hazard of pesticides to
bees,Apis mellifera. Med. Fac. Landbouww. Rijksuniv. Gent. 51/3b.: 1205-1213.
OTERO-COLINA, G., AND M. T. SANTILLAN-GALICIA. 1996. Dispersal of Varroa jacob-
soni (Gamasina: Varroidae) in Mexico. In: R. Mitchell, D. J. Horn, G.R.
Needham and W. C. Welbourn (eds.). Acarology IX, Proceedings. The Ohio Bio-
logical Survey, Columbus Ohio: 137-138.
RITTER, W., AND H. ROTH. 1986. Experiments with mite resistance to varroacidal sub-
stances in the laboratory. In: R. CAVALLORO (ed.). European research on var-
roatosis control. Proceedings of a Meeting of the EC Expert's Group.; Bad
VANDAME, R., M. E. COLIN, L. P. BELZUNCES, AND P. JOURDAN. 1995. Resistance de
varroa au fluvalinate. Les Carnets du Cari. 46: 5-11.
Florida Entomologist 83(4)
EVALUATION OF "TRED-NOTTM DEERFLY PATCHES"
AGAINST HOST-SEEKING DEER FLIES
(DIPTERA:TABANIDAE) IN NORTH FLORIDA
JAMES E. CILEK
John A. Mulrennan, Sr. Public Health
Entomology Research & Education Center, Florida
A & M University, 4000 Frankford Avenue, Panama City, FL 32405
"TRED-NOTT DEERFLY PATCHES" (6.4 x 14.2 cm adhesive strips) affixed to the
back and front of nylon mesh solid black and solid white "baseball" caps were evalu-
ated for their ability to trap host-seeking Chrysops celatus Pechuman, C. vittatus
Weidemann, and Diachlorus ferrugatus (F.). Trials were conducted in a commercial
pine bottomland forest habitat in northwestern Florida during peak seasonal abun-
dance of these species. No D. ferrugatus were captured on patches but approximately
26% of host seeking Chrysops (regardless of patch location, cap color or fly species)
were captured compared with a standard aerial sweep net method. Significantly more
deer flies were captured on patches affixed to the back of the cap compared with
patches placed on the front. No statistical difference (>0.05) existed in number of flies
trapped on patches when cap colors (white versus black) were compared.
Key Words: Chrysops celatus, Chrysops vittatus, Diachlorus ferrugatus, personal pro-
Parches marca "TRED-NOTR"" para la capture de moscas Chrysops celatus Pechu-
man (tiras adhesivas de 6,4 x 14,2 cm) pegadas al frente y al dorso de mallas de nylon
Cilek: Deer fly patches in north Florida
en gorras de "beisbol" de colors s61lidos blanco y negro fueron estudiadas para deter-
minar su habilidad de atrapar a los busca-hospedantes Chrysops celatus Pechuman,
C. Vittatus Weidemann, y Diachlorus ferrugatus (F.). Se realizaron pruebas en un ha-
bitat de bosque de pino commercial al noroeste de Florida durante un period estacional
de alta abundancia para estas species. No fueron capturados D. ferrugatus en par-
ches pero aproximadamente 26% de Chrysops busca-hospedantes (sin importar lugar,
color de la gorra o especie de mosca) fueron capturados en comparaci6n a un m6todo
estandar de rastrear con red area. Mas moscas significativamente fueron capturadas
en parches pegados al dorso de la gorra comparado con parches colocados al frente. No
existi6 diferencia estadistica (>0,05) en el numero de moscas capturadas en parches
cuando se compararon los colors de las gorras (blanco versus negro).
Host-seeking deer flies can often become a serious nuisance and large pestiferous
populations can certainly discourage enjoyment of outdoor activities. Repellents ap-
plied to exposed skin have not been very effective against these pests especially for ex-
tended periods of time (Anderson 1985). Insecticides applied to, or impregnated in,
clothing have been reported to provide some repellency in field situations (Schreck et
al. 1978 and Carlson 1996). Recently, adhesive patches (affixed to headwear) have
been advertised in various retail/outdoor recreational supply catalogs as an effective
way to "stop biting deer flies". This author is unaware of any published studies, con-
ducted under Florida conditions, where such products have been evaluated. As a re-
sult, a field study to evaluate a commercially available adhesive patch against three
species of host-seeking deer flies was conducted late spring, 1998.
MATERIALS AND METHODS
This study was conducted from May 27 through June 8, 1998 in Walton and Bay
Counties, Florida, at a time when Chrysops vittatus Weidemann, C. celatus Pechuman
and Diachlorus ferrugatus (F.) were at seasonal peak abundance as documented by
Jones & Anthony (1968), Cilek and Schreiber (1996, 1999) and Cilek et al. (1994), re-
spectively "TRED-NOTT DEERFLY PATCHES" (6.4 x 14.2 cm double-sided adhe-
sive-coated fabric patches Detex, Leroy, Michigan) were used in all evaluations.
Although package directions indicated that one strip be affixed to the back of a hat,
or cap, comparisons were made with a strip placed in front and back to determine if
location affected patch trapping effectiveness. Patches were affixed to "baseball" caps
made of nylon mesh with solid foam fronts (Cobra Caps, Bangladesh). Solid-colored
white and black hats were compared to determine if color influenced fly collections.
Controls consisted of similar mesh caps with patches affixed in same locations as
treatments but covered with a non-adhesive backing strip (used by the manufacturer
to prevent adhesive strips from adhering to the packaging material). Adhesive
patches were used once per test.
Evaluations were conducted in three geographically separate but similar habitats
(at least 50 km apart) where only one fly species occurred. Each habitat consisted of
abandoned commercial pine bottomland forests that contained a mixture of slash pine
(Pinus elliotii Ex. Chapm.) magnolia (Magnolia grandiflora L.), and live oak (Quercus
virginiana Mill.). Two non-continuous linear transects, each 30-m, were staked out in
each location. One person walked the length of each transect back and forth (i.e. 60 m)
and total number of deer flies attached to adhesive strip(s) recorded at the end of that
478 Florida Entomologist 83(4) December, 2000
transect. After this, aerial net (32 cm diam) samples were then conducted by the same
person in the same area. Sampling consisted of continually swinging the net in figure
"8" sweeps that started at ankles and ended above the head while walking each
transect (Cilek and Schreiber 1996). This method (herein referred to as a standard)
was used as a "best estimate" to quantitatively compare abundance of host-seeking
flies in the immediate vicinity of the sampler (i.e. control) with those captured on
patches (i.e., treatment). Net collected flies were counted at the end of each transect
and released. Treatments and standards were replicated twice per habitat per species
on ten different dates.
Data were transformed using /vx+1 prior to analysis and subjected to ANOVA
(PROC GLM, SAS Institute 1990). A Student-Newman-Keuls test was used to deter-
mine differences (<0.05) in overall mean fly abundance relative to patch (treatment)
vs standard (control) collections, patch location (front vs back), hat color (white vs
black), and Chrysops species (Sokal and Rohlf 1981). These data sets did not include
D. ferrugatus as none were captured on adhesive patches. All mean data in tables are
Overall, significantly fewer host-seeking Chrysops (F = 343.07, df 1, 159; P <
0.0001) were collected from adhesive patches compared with the standard (Table 1).
Adhesive patches captured approximately 26% of the fly population netted by the
standard. Moreover, about 21% (17 out of 80) of the patches captured no flies at all.
Patches affixed to the back of caps captured significantly more flies than those af-
fixed to the front (F = 193.03 df 1, 159; P < 0.0001) (Table 1). No significant difference
was observed in number of deer flies caught on white hats versus black hats (F = 0.81;
df 1, 159; P = 0.37).
Overall, significantly more C. vittatus (9.9 + 0.9) were collected from the standard
and patches compared with similar collections for C. celatus (5.4 + 0.5) (F = 6 8.98, df
1, 159; P < 0.0001). This difference was probably related to location/habitat, and/or
relative population size, rather than species preference. As stated earlier, no D. ferru-
gatus were trapped on adhesive patches regardless of cap color or patch location al-
though they were collected in the aerial net sampling standard (mean 11.4 + 0.5).
The effectiveness of TRED-NOTT DEERFLY PATCHES to trap deer flies was in-
fluenced by a fly's host-seeking behavior. Chrysops preferred the upper regions, espe-
cially the head, and were readily trapped on the patches. D. ferrugatus was not
captured because this species primarily visited the lower extremities when trying to
obtain a blood-meal (Fairchild and Weems 1973, McKeever and French 1997).
Adhesive patches did capture both Chrysops species. Collection differences rela-
tive to patch location (i.e. front vs back) were interesting. Attraction of host-seeking
deer flies to a person walking is well known (Bram 1978) but the orientation to the
back of a human's head may result from a "downwind" plume of expired carbon diox-
ide. Carbon dioxide has long been recognized as a strong attractant for host-seeking
Tabanidae (Bram 1978).
Cilek: Deer fly patches in north Florida
TABLE 1. MEAN (--SE) COMPARISON OF HOST-SEEKING CHRYSOPS CELATUS AND C. VIT-
TATUM (POOLED DATA) TRAPPED BY TRED-NOT DEERFLY PATCHES"M,
Treatment n Mean flies --SE
I. Overall abundance
adhesive patch 80 3.2 + 0.4a
standard 80 12.1 + 0.7b
II. Patch location
front 80 3.0 + 0.4b
back 80 0.3 + 0.1a
III. Cap color
white 80 1.5 + 0.3a
black 80 1.8 + 0.3a
Paired means within rows (I, II, and III) followed by different letters are significantly different (P < 0.05; SNK)
after /X+1 transformation of means), untransformed means are shown in table.
There appeared to be no cap color preference relative to number of Chrysops
trapped on patches, although, it has been well documented that tabanids are gener-
ally attracted to dark objects (Bram 1978). Because the study area bordered well-
known tabanid developmental habitats (i.e. bottomland swamps/marshes), color pref-
erence may have not been an important factor for short-range host seeking, when ex-
pired carbon dioxide (signalling a potential blood host) was present.
In conclusion, TRED-NOT"T DEERFLY PATCHES captured some of the Chrysops
attracted to a person's head but did not completely trap all these pests visiting this
body region. However, the amount of personal annoyance perceived from host-seeking
deerflies is often relative. Therefore any device or method, including the one evalu-
ated here, that removed or reduced host-seeking Chrysops (either percieved or actual)
may be of general benefit to those seeking relief from such outdoor pests.
ANDERSON, J. F. 1985. The control of horse flies and deer flies (Diptera: Tabanidae).
Myia 3: 547-598.
BRAM, R. A. (ed.) 1978. Surveillance and collection of arthropods of veterinary impor-
tance. USDA. Animal Plant Health Insp. Serv. Agric. Handbook 518.
CARLSON, D. A. 1996. Insect repellents. Pp. 283-297. In: Pest management in the sub-
tropics. Integrated pest management-a Florida perspective. Intercept Ltd.,
CILEK, J. E., AND E. T. SCHREIBER. 1996. Diel host-seeking activity of Chrysops celatus
(Diptera: Tabanidae) in northwestern Florida. Florida Entomol. 79: 520-525.
CILEK, J. E., AND E. T. SCHREIBER. 1999. Diel host-seeking activity of adult Diachlorus
ferrugatus (F.) (Diptera: Tabanidae) in northwestern Florida. J. Entomol. Sci.
34: 462- 466.
CILEK, J. E., W. ANSELL, AND G. MEDRANO. 1994. Seasonal abundance of Diachlorus
ferrugatus (Diptera: Tabanidae) in north Florida. J. Florida Mosq. Control As-
soc. 65: 45- 48.
FAIRCHILD, G. B., AND H. V. WEEMS. 1973. Diachlorus ferrugatus (Fabricius), a fierce
biting fly (Diptera: Tabanidae). Florida Dept. Agric. Consum. Serv., Div. Plant
Industry, Entomol. Circ. 139.
480 Florida Entomologist 83(4) December, 2000
JONES, C. M., AND D. W. ANTHONY. 1964. The Tabanidae (Diptera) of Florida. USDA,
Agric. Res. Serv. Tech. Bull. 1295.
McKEEVER, S., AND F. E. FRENCH. 1997. Fascinating beautiful blood feeders. Amer.
Entomol. 43: 217-226.
SAS INSTITUTE. 1990. SAS user's guide: statistics, version 6 ed. SAS Institute, Cary,
SCHRECK, C. E., K. POSEY, AND D. SMITH. 1978. Durability of permethrin as a poten-
tial clothing treatment to protect against blood-feeding arthropods. J. Econ.
Entomol. 71: 397-400.
SOKAL, R. R., AND F. J. ROHFL. 1981. Biometry (2 ed.), W. H. Freeman and Co., San
480 Florida Entomologist 83(4) December, 2000
FRUIT FLIES (DIPTERA: TEPHRITIDAE) INFESTING FRUITS
OF THE GENUS PSIDIUM (MYRTACEAE) AND THEIR
ALTITUDINAL DISTRIBUTION IN WESTERN VENEZUELA
KAMTA P. KATIYAR', JESUS CAMACHO MOLINA2 AND RICARDO MATHEUS1
1Unidad T6cnica Fitosanitaria (UTF), Facultad de Agronomia,
La Universidad del Zulia, Apdo. 526, Maracaibo, Estado Zulia 4002, Venezuela
2Museo de Artr6podos (MALUZ), Facultad de Agronomia,
Apdo. 526, Maracaibo, Estado Zulia 4002, Venezuela
A survey of fruit flies infesting Psidium fruits was conducted in western Venezuela
from June 1992 through December 1995. Of 201 fruit samples collected from 139 lo-
calities at altitudes between sea level and 2,000 m, four species of Psidium plants
were found in the western region of Venezuela. These were P. guajava L. (10-1930 m),
P. guineense Sw. (100-1950 m), P. caudatum McVaugh (1800-1950 m) and P. friedrich-
sthalianum (Berg) Niedenzu (35-1700 m). Four tephritid fly species were reared:
Anastrepha striata Schiner, A. fraterculus (Wiedemann), A. obliqua (Macquart), and
Ceratitis capitata (Wiedemann). All four fruit fly species emerged from P. guajava.
A. striata was the most common on P. guajava, P. guineense and P. friedrichst-
halianum, with an infestation range of 96.1%-97.0%. P. caudatum was more fre-
quently infested by A. fraterculus (94.5% adults emergence); the plant's distribution
was restricted to highlands. Observations on the altitudinal distribution ofA. striata
on P. guajava showed that the highest infestation (253.9 adults/kg fruits) occurred at
about 1,000 m. The infestation rate of P. guajava byA. fraterculus andA. obliqua var-
ied with elevation. In low elevation areas (0-1,200 m),A. obliqua was found more fre-
quently thanA. fraterculus, whereasA. fraterculus was found more frequently thanA.
obliqua in high altitude areas (1,201-2,000 m). C. capitata was erratically encoun-
tered in this study.
Key Words:Anastrepha, Ceratitis capitata, guava, Psidium spp., altitudinal distribution
Desde junio de 1992 a diciembre de 1995 se estudiaron las moscas de las frutas
(Diptera: Tephritidae) que infestan plants del g6nero Psidium en el occidente de Ve-
Katiyar et al.: Psidium spp. fruit flies in Venezuela
nezuela. Se recolectaron un total de 201 muestras de frutas en 139 localidades com-
prendidas desde el nivel del mar hasta 2,000 m de altitude. Se encontraron cuatro
species de plants del g6nero Psidium: P. guajava L. (10-1930 m), P. guineense SW.
(100-1950 m), P. caudatum Mc Vaugh (1800-1950) y P. friedrichsthalianum (Berg)
Niedenzu (35-1700 m). Se lograron criar cuatro species de moscas de la Familia Te-
phritidae: Anastrepha striata Schiner, A. fraterculus (Weidemann), A. obliqua (Mac-
quart) y Ceratitis capitata (Weidemann)]. De P. guajava emergieron las 4 species de
moscas de las frutas encontradas en el present estudio.A. striata result ser la mosca
mas comun en P. guayaba, P. guineense y P. friedrichsthalianum encontrandose infes-
taciones comprendidas entire 96.1%-97.0%. P. caudatum fue encontrada como la
plant hospedera preferida porA. fraterculus con un 94.5%. Ademas, su distribuci6n
esta restringida a tierras altas. La distribuci6n altitudinal de A. striata muestra que
la mayor infestaci6n en frutos de P. guajava ocurre alrededor de los 1,000 m de altitude
(253.9 adultos/Kg de frutas). La infestaci6n relative deA. fraterculus y A. obliqua en
P. guavaja varia con la altitude. En tierras bajas (0-1,200 m), A. obliqua se encontr6
como la especie predominante sobreA. fraterculus. En cambio, en tierras altas (1,201-
2,000 m), A. fraterculus fue la especie predominante sobreA. obliqua. La distribuci6n
geografica y altitudinal de C. capitata fue muy erratica.
The Genus Anastrepha is endemic to the Americas and is restricted to tropical and
subtropical environments. Its range extends from the southernmost part of the
United States (Rio Grande Valley of Texas and southern Florida) to South America,
with the exception of the southern parts of Argentina and Chile. Fruit flies of the ge-
nus Anastrepha compose one of the largest and most economically important insect
groups in the tropics and subtropics due to their damage to cultivated fruits. This
group comprises more than 190 identified species but hosts are known for less than
half (Norrbom & Kim 1988).
In Venezuela there are four economically importantAnastrepha species: the South
American fruit fly, A. fraterculus (Wiedemann), the West Indian fruit fly, A. obliqua
(Macquart), the guava fruit fly, A. striata Schiner, and the zapote fruit fly, A. serpen-
In the 1980's, cultivation of guava, Psidium guajava, in the northern region of Zu-
lia State was expanded. By 1992, in the lake Maracaibo plain, about 4,000 ha of guava
orchards were in production (Araujo et al. 1997). Several species of fruit flies of the
family Tephritidae, especially A. striata and C. capitata, are very important from a
quarantine point of view when fruit export is the objective.
P. guajava is found from sea level to 1,930 m in commercial orchards, backyards of
houses, roadsides, pasture lands, and forests throughout western Venezuela. The
work described in this paper was done to obtain basic information about the altitudi-
nal distribution of different fruit flies infesting cultivated or wild fruits belonging to
the family Myrtaceae, genus Psidium, which can be used in the economic manage-
ment of these fruit flies.
MATERIALS AND METHODS
From June 1992 to December 1995, 201 Psidium spp. fruit samples were collected
whenever available from sea level to 2,000 m elevation in 139 localities in the western
Venezuelan states of Falc6n, M6rida, Tachira, Trujillo, and Zulia comprising an area
of 117,700 sq-km. The Northern and Southern borders of the study area are delimited
482 Florida Entomologist 83(4) December, 2000
with latitudes 1145'N and 732'N respectively, while Eastern and Western borders
are delimited with longitudes 6830'W and 7240'W respectively.
Fruit samples were collected from four species of Psidium: P. guajava L. (common
guava) from sea level to 1,930 m, P. guineense Sw. (mountain guava) from 100-1,950 m,
P. caudatum McVaugh (jumangue) from 1,800-2,000 m, and P. friedrichsthalianum
[Berg] Ndz. (cas or sour guava) from 35-1,700 m altitude. The total number of fruits
collected from each Psidium spp. host plant comprised 7,015, 255, 3,816 and 59 fruits
from P. guajava, P. guineense, P. caudatum and P. friedrichsthalianum respectively.
Following the technique described by Katiyar et al. (1995), fruit samples were in-
cubated and processed in the laboratory. Mature fruits were picked from sample
trees as well as from the ground and were placed in open top wooden boxes (30 x 20 x
10 cm). A sheet of plastic screening (about 4 mm/mesh) had been fitted about 2 cm
from the bottom of each box. The wooden boxes containing fruit samples were placed
in plastic rearing containers (35 x 24 x 13 cm). The tops of the rearing containers
were fitted with a fine-screened window (15 x 8 cm) for aeration. Each container had
a layer of moist sawdust about 2 cm deep at the bottom as a pupation media for the
larvae. The fruit samples were taken to the laboratory in this manner. In the labora-
tory, samples were removed from rearing containers. The fruits were counted,
weighed, and put back in the containers. Every 2-3 days the sawdust was sieved, and
recovered larvae and pupae were placed in 500 cc plastic cups containing a thin layer
(2-3 cm) of moist saw dust. Each container with pupae was placed inside an adult
emergence cage to recover fruit flies and parasitoid adults. The emerged adults were
preserved in 70% ethyl alcohol. Rearing was carried out in the laboratory at 26 + 3C
and 60 + 10% RH.
The climatic condition of the study area is characterized by a rainy season from
April to November, followed by a dry period from December to March. The rainfall in
Western Venezuela varies widely from one place to another. This variation can be ob-
served in lowland areas (0-1,000 m) as well as in highland areas (1,001-2,000 m). Dur-
ing rainy season the temperature is slightly higher compared with the dry season.
Table 1 presents climatic data (temperature, rainfall and RH) from 17 meteorological
stations located between 5 and 2,200 m elevation in the study area.
Samples of fruit fly adults were identified by A. L. Norrbom, Systematic Entomol-
ogy Laboratory, USDA, PSI, ARS, Washington, D.C.
RESULTS AND DISCUSSION
A total of 11,145 fruits (261.5 kg) were collected from four Psidium species. A total
of 30,530 pupae and 20,970 adults of both sexes belonging to four species of tephritids
were reared from these fruits.
Table 2 summarizes the relative abundance of each tephritid species infesting the
four Psidium host plants (based on adult emergence/kg fruits). P. guajava was in-
fested by all four tephritid species found, P. guineense by three species (A. striata,A.
fraterculus andA. obliqua), P. caudatum by two species (A. fraterculus andA. striata)
and P. friedrichsthalianum by two species (A. striata and A. obliqua). The results
show that A. striata was the most common fruit fly pest of Psidium species in the
western region of Venezuela. Based on total adult emergence from all four Psidium
spp., the proportion of each species consisted of A. striata (73.9%), A. fraterculus
(24.2%) andA. obliqua (1.9%).
The results also show that A. striata was the most common fruit fly in three Psid-
ium spp. guajavaa, guineense, and friedrichsthalianum) with an infestation range of
96.1%-97.0% (based on number of adults emerged/kg fruit). Whereas A. fraterculus
Katiyar et al.: Psidium spp. fruit flies in Venezuela
C LC LC x (Z CM
m tm t- D D X x
CO CO D C
L- cD It 0 0 TIt c0q-
m D cc i L-- 0^ LC
C- CO ] Cj; CO t'C
m] 0 m L '-- Ctm
C6 CM C CO 7 1-4
,I Cc CC 0 CM L- m
CM CM CM CM CM CM CM
0C (D LC CM m L-- L--
L C- m CC L-
CC CM CM CM
(Z [ -
CO 0-I 0 T- 0 0 0
C M CM Cc CM cc
0 00 0^ 0^ CC
t C] T- C] 0 C] I0
^DC] CO C] C0 C] IC
i-i M c m
cc t- z -f
CM 0 i-
CcO CO cc L-
0 L-- cc CM T-I CM It CM 0 cc
cc cc CM CM T-I 0 L-- 00 Co ^
m C tm tL- mC
L --I m C L-- m
em cc c
L-- -I cq D mO m It L-- cc cc
CC((Z O O T-~I C [-CCCt
0 (CLC L T-- ClCC
m- ccI L-- D m~ CO D tCO CO
6- [ -- L6 [ 6 C6 I4 6 0C
~cC^ O (ZCCC' CZICCZ
T- -I -I t-1 1-1 1-1 1- 1-1 1-
,I cci LD-- m O It -1 0 0 o c-1
'^CO t[- ICICCCT- O D -
4^ CM CM T-4 T-4 7i 06 06 L6 L6
CM CMCM C CM C- -ITICCI -
cc To LC L-- L--l m CM CM
t0 t" m
o -I CM
T--l T- 0
0 0 CM
;--I CM o 0 o Zo Zo DMq
]O O O T- M
00-10 000 000
CO LCO m' D OC O cc
C)2 CO 0 C
-2 1 t as ia ^ a r
cc r^^r r U U V Cd
NVN NH HHOC ?'O'O ^'OC
C, C C
cc F4 U .U E .c
m-~ -a~ 2 i *-A2 S
- | ? rz | 1^
ca >u60 i maim
T-l 0^ T-
484 Florida Entomologist 83(4) December, 2000
TABLE 2. RATE OF FRUIT INFESTATION OF PSIDIUM SPECIES BY ANASTREPHA SPP. IN
THE WESTERN REGION OF VENEZUELA, 1992-1995.
Host fruit A. striata A. fraterculus A. obliqua Totals
P. guayaba' 110.6 2.4 1.3 114.3
P. guineense2 91.2 0.4 2.4 94.0
P. caudatum3 0.8 13.7 0.0 14.5
P. friedrichsthalianum4 29.8 0.0 1.2 31.0
'Indicates (n = 185 samples and 249,2 Kg).
'Indicates (n = 5 samples and 3.1 Kg).
'Indicates (n = 7 samples and 6.9 Kg).
'Indicates (n = 4 samples and 2.3 Kg).
was the predominant species in P. caudatum (the emerged adults were 94.5%A. frater-
culus and 5.5% A. striata). P. caudatum is a native wild Psidium species found at rel-
atively high elevation (1,500-2,000 m) and has a small fruit, mean weight of 1.8 g.
In general A. striata was the dominant fruit fly in Psidium species except P. cau-
datum in the western part of Venezuela. A. striata is also reported as the major pest
of P. guajava in several other Latin American countries. In Costa Rica, 97.8% of fruit
samples were reported infested byA. striata (Jiron & Hedstr6m 1988), and in Ecuador
A. striata emerged from 70.8% of fruit samples examined (Hedstr6m 1987).
Figure 1A shows the relative intensity of infestation byA. striata in P. guajava at
different altitudes (8 strata). Results demonstrate that the most prevalent distribu-
tion for A. striata infestations occur between 500 to 1,500 m and the highest infesta-
tions occur at about 1,000 m.
Figure 1B shows the altitudinal distribution ofA. fraterculus andA. obliqua. The
results show that at low altitudes (0-1,200 m),A. obliqua was more prevalent thanA.
fraterculus, whereas at higher altitudes (1,201-2,000 m), A. fraterculus was more
prevalent thanA. obliqua. Similar results, which indicate thatA. obliqua prefers low-
land zones andA. fraterculus prefers higher elevation areas, have been found in other
studies (Celedonio-Hurtado et al. 1995, Eskafi & Cunningham 1987, Hedstr6m 1987).
C. capitata was reared only from seven fruit samples of P. guajava collected be-
tween 50-1650 m. The distribution of these fruit samples were four (Zulia state) at 0-
250 m and one each from 250-500 m (Trujillo state), 1250-1500 m (Tachira state), and
1500-1750 m (M6rida state) elevation ranges. The altitudinal distribution and pres-
ence of this fruit fly was erratic and the infestation rate was very low (1.4 adults/Kg).
We thank Alan L. Norrbom (Systematic Entomology Laboratory, USDA, PSI, ARS,
Washington, D.C.), Daniel S. Moreno (Subtropical Agricultural Research Laboratory,
Crop Quality and Fruit Insects Research, USDA, ARS, Weslaco, TX), and Andrew F.
Neild, 116 Crosslet Wale Greenwich, London SE10 8DL for their comments and sug-
gestions in the review of earlier draft of the manuscript. We also thank 3 anonymous
reviewers for critically reviewing and improving this manuscript. This study was
funded by Consejo de Desarrollo Cientifico y Humanistico de La Universidad del Zulia
(CONDES), through project 1912-96 "Programa Museo de Artr6podos".
Katiyar et al.: Psidium spp. fruit flies in Venezuela
250 500 750 1000 1250 1500 1750 2000
250 500 750 1000 1250 1500 1750 2000
Fig. 1. Infestation of Psidium guajava by three tephritid species in western Vene-
zuela, 1992-1995. (A) byA. striata. (B) byA. obliqua andA. fraterculus.
ARAUJO, F., S. QUINTERO, J. SALAS, J. VILLALOBOS, AND A. CASANOVA. 1997. Crec-
imiento y acumulaci6n de nutrients del fruto de Guayaba (Psidium guajava
L.) del tipo "Criolla Roja" en la planicie de Maracaibo. Rev. Fac. Agron. (LUZ)
486 Florida Entomologist 83(4) December, 2000
CELEDONIO-HURTADO, H., M. ALUJA, AND P. LIEDO. 1995. Adult population fluctua-
tion ofAnastrepha species (Diptera: Tephritidae) in tropical orchard habitats of
Chiapas, Mexico. Environ. Entomol. 24: 861-869.
ESKAFI, F. M., AND R. T. CUNNINGHAM. 1987. Host plants of fruit flies (Diptera: Te-
phritidae) of economic importance in Guatemala. Florida Entomol. 70: 116-123.
HEDSTROM, I. 1987. Fruit flies (Diptera: Tephritidae) infesting common guava (Psid-
ium guajava L.) M;.N ... '-. in Ecuador. Rev. Biol. Trop. 35: 373-374.
JIRON, L. F., AND I. HEDSTROM. 1988. Occurrence of fruit flies of the genera Anas-
trepha and Ceratitis (Diptera: Tephritidae), and their host plant availability in
Costa Rica. Florida Entomol. 71: 62-73.
KATIYAR, K. P., J. CAMACHO, F. GERAUD, AND R. MATHEUS. 1995. Parasitoides hy-
men6pteros de moscas de las frutas (Diptera: Tephritidae) en la region occiden-
tal de Venezuela. Rev. Fac. Agro. (LUZ) 12: 303-312.
NORRBOM, A. L., AND K. C. KIM. 1988. A list of the reported host plants of the species
ofAnastrepha (Diptera: Tephritidae). USDA-APHIS-PPQ Bull. 81-52, 114 pp.
FIRST RECORD OF CUTEREBRA SP. (DIPTERA:
CUTEREBRIDAE) INFECTION IN OTOTYLOMYS PHYLLOTIS
P. MANRIQUE-SAIDE', S. HERNANDEZ-BETANCOURT1 AND M. T. QUINTERO2
1Universidad Autonoma de Yucatan, Facultad de Medicina Veterinaria y Zootecnia
Department of Zoology, PO Box: 4-116 Itzimna. Merida, Yucatan, Mexico
E-mail: email@example.com; firstname.lastname@example.org
2Universidad Nacional Autonoma de Mexico, Facultad de Medicina
Veterinaria y Zootecnia, Cd. Universitaria, D.F. C.P. 4510 Mexico D.F.
Maggots of the fly genus Cuterebra (hot flies) are cutaneous tissue parasites of wild
species of rodents and lagomorphs. The latter are their natural hosts, although other
accidental or aberrant hosts as cats, dogs, deer, cattle, mules, skunks and even hu-
mans are known to harbour Cuterebra larvae (Sabrosky 1986). The distribution of the
72 known species is restricted to the New World; 36 have neartic distribution, 36 neo-
tropical and four overlap their geographic distribution (Catts 1982). Most of these spe-
cies appear to be host specific to the rodent species infected or to a group of closely
The genus Cuterebra has been reported in six genera of rodents (Microtus,
Neotoma, Peromyscus, Sciurus, Tamias and Thomomys) (Catts 1982, Sabrosky 1986).
The former three are murids with distribution in the Americas (Wilson & Reeder
1993). From January of 1997 through May of 1998, a population study of wild rats was
carried in a dry tropical forest in the locality of Hobonil (2000'58"N, 89001'13"W) in the
Mexican state of Yucatan (Southeast Mexico). Monthly trapping samples were carried
during five days in an area of 1200 m2. Rats were live-trapped using Sherman traps.
The rats were examined for bot infection. Detection of infection was done visually
and by palpation and infection signs and reports ranged from cyst location to larvae
extraction. From an overall sample of 427 rats of Heteromysgaumeri (269 individuals/
62.99% of overall population), Ototylomys phyllotis (122/28.57%), Oryzomys melanotis
(27/6.32%), Peromyscus yucatanicus (7/1.64%) and Sigmodon hispidus (2/0.46%), 16
rats were founded to be infected with larvae of Cuterebra sp. All the Cuterebra larvae
or infection signs were founded in Ototylomys phyllotis. This is the first record of Cute-
rebra infection in the genus Ototylomys, the big-eared climbing rats.
Nine Cuterebra sp larvae were collected (one of second instar and eight of third in-
star). The larvae were found in cutaneous cysts close to the genitals and anus in
twelve rats; close to the forelegs in two rats, and in both once. In four of the reported
rats, the detection was merely done by palpation and observation of the cyst.
The incidence of infection in the population of Ototylomys phyllotis was 13.11%.
Thirteen of the infected rats were adult males (86.67%) and three adult females
(13.33%). Excepting for two rats (male and female), that had two larvae, all had only
one each. From the rats where myiasis was detected for the first time, eight were re-
captured without showing any re-infection signs and one was re-infected.
The first infected rat was observed in February of 1997, five during October and
September, four in January of 1998, and six during February and March (Fig. 1).
Five larvae were cultured and three adult female flies emerged, two larvae were
preserved in alcohol 75%. Specimens could only been identified to genus and are de-
posited at the Coleccion Entomologica Regional, Universidad Autonoma de Yucatan.
Although the specimens were identified as Cuterebra according to Sabrosky (1986),
488 Florida Entomologist 83(4) December, 2000
Sn 3(13.63) 4(19.04)
S1(9.09) 2t PC
Z 11 i i i
C3 I I I I I [ Ri l l
Jan Feb Mar Apr May Jul Aug Sep Oct Nov Dec Jan Feb Mar May
Fig. 1. Incidence of Cuterebra infection in a population of Ototylomys phyllotis at
Hobonil, Yucatan, Mexico. The white area indicates the abundance of the population
of the rats and the black area indicates the incidence (percent of hosts infected).
more taxonomical work is needed. The internal arrangement of Cuterebra stills under
discussion because some authors (Sabrosky, 1986) have decided to recognize at least
two genera within the genus (the Neartic Cuterebra and the Neotropical Metacutere-
bra), while others have synonymized most of the genus-group names under Cuterebra
(Chillcott, 1965; Guimaraes, 1967).
We would like to thank Hugo Delfin for his comments to this manuscript.
This work is the first report of infection caused by Cuterebra on rodents in the ge-
nus Ototylomys. In a wild population of Ototylomys phyllotis, 16 of 122 mice were in-
fected. Nearly 87% of infected individuals were adult males, and displayed cysts
around the genital organs and anus, and on the arms. Of those animals showing first-
time miasis, eight recaptured individuals did not show signs of re-infection and one
individual was re-infected.
CATTS, E. P. 1982. Biology of New World bot flies: Cuterebridae. Ann. Rev. Entomol.
CHILLCOTT, J. F. 1965. Family Cuterebridae, pp. 1109-1110 in A. Stone et al. A catalog
of the Diptera of America north of Mexico, Agriculture Handbook 276.
GUIMARAES, J. H. 1967. Family Cuterebridae, Fascicle 105: 1-11 in N. Papavero (ed.),
A catalogue of the Diptera of the Americas South of the United States. Museo
de Zoologia, Sao Paulo.
SABROSKY, C. W. 1986. North American Species of Cuterebra, the Rabbit and Rodent
Bot Flies (Diptera: Cuterebridae). Entomol. Soc. Amer. Thomas Say Foundation
Monograph. Maryland. p. 240.
WILSON, D. E., AND D. M. REEDER. 1993. Mammal species of the world. A taxonomic
and geographical reference. Smithsonian Institution Press. Washington. p. 1206.
GUT CONTENT ANALYSIS OF THREE SPECIES OF SAC
SPIDERS BY ELECTROPHORESIS
DIVINA M. AMALIN', JORGE E. PENA1 AND ROBERT MCSORLEY2
1University of Florida Institute of Food and Agricultural Sciences, Tropical Research
and Education Center, 18905 SW 280 St., Homestead, FL 33031
2University of Florida Department of Entomology and Nematology,
Gainesville, Florida 32601
Predation by spiders under field conditions is inherently difficult to study; how-
ever, the existence of specific predation wounds produced by three species of sac spi-
ders on citrus leafminer (CLM), Phyllocnistis citrella Stainton, could make it possible
to quantify their predation. One drawback in this system is that one of the predation
marks made by the sac spiders on CLM larvae is somewhat similar to the feeding
mark made by Pnigalio minio (Walker), an eulophid ectoparasitoid of CLM. A way to
overcome this predation assessment problem is through analysis of predator gut con-
tents. Preliminary observations on the analysis of the prey remnants inside the gut of
the spiders are presented and discussed. The main purpose of this study is to develop
a method to detect prey remnants in the gut of spiders, which will lead to better as-
sessment of the efficiency of spiders on CLM control in nurseries and orchards.
The prey remnants inside the gut of the three species of sac spiders, Chiracanth-
ium inclusum Hentz, Hibana velox (Becker), and Trachelas volutus (Gertsch), were
detected by polyacrylamide gel electrophoresis (PAGE). This method is based on the
detection of prey enzymes in homogenates of the predator after PAGE and staining for
esterase activity (Van Der Geest & Overmeer 1985, Murray & Solomon 1978, Solomon
et al. 1985). Esterase was selected as the indicator protein because its detection em-
ploys an enzymic reaction with substrate yielding a stain with a high extinction coef-
ficient as shown from the previous study on analysing diets of invertebrate predators
by electrophoresis (Murray & Solomon 1978) which allow detection of very small
quantities of enzyme by staining for extended periods.
Spiderlings of C. inclusum, H. velox, and T volutus were obtained from laboratory
cultures. Spiders fed with an artificial diet were individually reared in laboratory
glass vials (15 mm diameter x 60 mm long). The artificial diet consisted of a mixture
of soybean liquid, homogenized milk, and egg yolk (Amalin et al. 1999). Spiders fed
with CLM larvae were reared individually in plastic petri dishes (10 cm diameter x 1
cm high). Samples of CLM larvae were gathered from field collections. Homogenates
were obtained from fourth-nymphal spiderlings fed for 2 days with a total of 5 second
larval instars of CLM, spiders fed with artificial diet, and second larval instars of
CLM. Spiders were placed in a Perspex plate and squashed individually with a glass
rod in 5-10 pl of maceration fluid (1X TBE buffer [0.09 M Tris-borate + 0.002 M EDTA]
with 0.2% Triton X-100 and 10% sucrose). A similar maceration procedure was used
for CLM larvae except that the numbers of larvae varied from 1, 2, 5, to 10 CLM in dif-
ferent homogenate samples. This range of larval densities was used in order to deter-
mine the difference in the intensity of esterase bands with varying numbers of CLM
larvae. For each sample, 20 pl homogenate was dispensed with the aid of a loading tip
to the sample holders that were positioned on top of the gel.
Electrophoresis was carried out as described by Murray & Solomon (1978) and by
Solomon et al. (1985) with some modifications. Polyacrylamide slab gels with a total
gradient concentration of 5-28% and a cross-link gradient of 2.5-6.2% (Margolis &
490 Florida Entomologist 83(4) December, 2000
Wrigley 1975) were prepared between two glass plates using a gel gradient maker. A
1X TBE buffer pH 8.3, to which 0.2% w/v Triton X-100 was added, was used as the gel
buffer. The running buffer was also 1X TBE without Triton X-100. The samples were
run to endpoint for 20 h at 200 volts.
After electrophoresis, gels were incubated in a medium containing 30 mM 1-naph-
thyl acetate and 0.2% Fast Blue RR Salt in 0.2 M phosphate buffer, pH 6.0, in order
to stain proteins with esterase activity. The gel separated from the glass plate was
submerged in the medium and kept inside a dark container and shaken on a shaker
(Gio Gyrotory) at a speed of 28 RPM for 24 hours or until the bands appeared. Con-
clusions about the identity of prey remnants inside the gut of predators were drawn
by visual comparison of esterase patterns of the artificially fed spiders and CLM-fed
In many cases with other predators, the specific esterase activity patterns allow
the identification of prey remnants inside the predator gut (Van Der Geest & Over-
meer 1985). After electrophoresis and staining for esterase activity, the CLM larvae
and prepupae showed only one esterase band (Figs. 1 and 2). There was a difference
in the intensity of the esterase bands on the different numbers of CLM included in
each macerated sample. The intensity increased as the number of individuals per
sample increased (Figs. 1 and 2). This difference in intensity could possibly be used to
quantify the number of prey consumed by the predator. No esterase was obtained
from the C. inclusum and T volutus fed with artificial medium (Fig. 1, lanes 3 to 5);
however, H. velox that fed on the artificial medium gave one esterase band with a
higher molecular weight than the CLM esterase (Fig.2, lanes 2 to 4). The esterase ob-
tained from C. inclusum fed with CLM larvae in the laboratory was somewhat similar
to the esterase of CLM (Fig. 1, lane 2). Similarly, a single esterase band was obtained
from H. velox that fed on CLM, and this band also appears to be identical to the CLM
esterase (Fig.2, lane 9). The result of this experiment is similar to that of Murray &
Solomon (1978) on the single esterase pattern of Panonychus ulmi (Koch), the Euro-
pean red spider mite, and differs from the results obtained by Dicke & DeJong (1988)
who obtained several esterases also on P. ulmi. During the current test, the CLM es-
terase appears to have been expressed only in spiders that fed on CLM except for T
volutus. No esterase activity was obtained from T volutus that fed on CLM larvae.
This could be attributed to the retention time of the CLM esterase in the gut of T vo-
lutus. Probably, the lifetime of CLM esterase in the gut of T volutus is shorter com-
pared to that of C. inclusum and H. velox. It is then worthwhile investigating the
retention time of CLM esterase in the gut of the three species of sac spiders.
Fig. 1. Esterase patterns of CLM, C. inclusum, and T volutus. The identity of the
different lanes are as follows: 1-standard, 2-two C. inclusum fed with CLM in the
laboratory, 3 to 5 -two C. inclusum fed with artificial diet, 6-ten CLM prepupae, 7-
five CLM prepupae, 8 to 9-two T volutus fed with CLM in the laboratory, 10-two T
volutus fed with artificial diet, 11-ten CLM larvae, 12-five CLM larvae, 13-blank,
14-two CLM larvae.
- 2 3 4 S 6 7 8 9 10 11 12 13 14
Fig. 2. Esterase patterns of CLM and H. velox. The identity of the different lanes
are as follows: 1-Standard, 2 to 4-two H. velox fed with artificial diet, 5-blank, 6-
ten CLM prepupae, 7-five CLM prepupae, 8-blank, 9-two H. velox fed with CLM
in the laboratory, 10-ten CLM larvae, 11-five CLM larvae, 12-two CLM larvae,
13-one CLM larva, 14-blank.
More electrophoresis runs should be done including field-collected individuals.
However, using field-collected spiders, a difficulty may arise in the identification of
prey because spiders in the field are generally feeding on multiple prey species. More-
over, not all esterases of a prey may be found in the gut of the spider. Certain esterases
may be localized in tissues that are not ingested, complicating the identification of the
prey for field-collected specimens. Therefore, it is necessary to co-electrophorese all
suspected prey species with the spiders on the same gel. The use of other enzymes (i.e.
fumarate hydratase, glucose-6 dehydrogenase, hexokinase, and other enzymes) must
also be considered in order to look for a more stable enzyme that will be used for gut
We would like to thank Drs. Paulene Lawrence and David Moraga for allowing us
to use their laboratory equipment. We are also grateful to Drs. Denise Johanawicz,
Jesusa Legaspi, and Scott Morrical for reviewing the manuscript. Florida Agricul-
tural Experiment Station Journal Series No. R-07139.
Gut content analysis using polyacrylamide gel electrophoresis (PAGE) was per-
formed on three sac spider species. Results from the electrophoresis showed that H.
velox fed on artificial medium gave one esterase band; whereas, no esterase was ob-
tained from the C. inclusum and T volutus fed on artificial medium. The esterase ob-
tained from C. inclusum and H. velox fed with citrus leafminer (CLM) larvae in the
laboratory seems to be similar to the CLM esterase. No esterase activity was obtained
from T volutus that fed on CLM larvae. The preliminary result of the gut content
analysis using PAGE showed the potential of this method in studying the predatory
spider CLM relationship in lime orchards.
AMALIN, D. M., J. REISKIND, R. MCSORLEY, AND J. PENA. 1999. Survival of the hunting
spider, Hibana velox (Becker), raised on different artificial diets. J. Arachnol.
DICKE M., AND M. DEJONG. 1988. Prey preference of the phytoseeid mite Typhlodro-
mus pyri 2. Electrophoresis Diet Analysis. Exp. Appl. Acar. 4: 15-25.
492 Florida Entomologist 83(4) December, 2000
MARGOLIS J., AND C. W. WRIGLEY. 1975. Improvement of pore gradient electrophoresis
by increasing the degree of crosslinking at high acrylamide concentrations. J.
Chromatogr. 106: 204-209.
MURRAY, R. A., AND M. G. SOLOMON. 1978. A rapid technique for analyzing diets of in-
vertebrate predators by electrophoresis. Ann. Appl. Biol. 90: 7-10.
SOLOMON, M. G., R. A. MURRAY, AND L. P. S. VAN DER GEEST. 1985. Analysis of prey
by means of electrophoresis. In: W. Helle and M. Sabelis (eds.), Spider Mites
and Their Control. Elsevier Science Publishers, Amsterdam.
VAN DER GEEST, L. P. S., AND W. P. J. OVERMEER. 1985. Experiences with polyacryla-
mide gradient gel electrophoresis for the detection of gut contents of phytoseeid
mites. Meded. Fac. Landbouww. Rijksuniv. Gent. 50: 469-471
492 Florida Entomologist 83(4) December, 2000
A SURVEY OF PARASITOIDS OF TRIALEURODES
VAPORARIORUM AND BEMISIA TABACI (HOMOPTERA:
ALEYRODIDAE) IN EASTERN GUATEMALA
HUGH ADAM SMITH', GREGORY A. EVANS2 AND ROBERT MCSORLEY1
1Department of Entomology and Nematology University of Florida,
Gainesville, FL 32611
2Division of Plant Industry Gainesville, FL 32614
Trialeurodes vaporariorum (Westwood), the greenhouse whitefly, and Bemisia
tabaci (Gennadius), the sweetpotato whitefly, are serious economic pests of agro-
nomic, horticultural, and ornamental crops throughout warm regions of the world
(Byrne et al. 1990, Brown 1994). Both species also affect glasshouse production of
plants in temperate regions (Byrne et al. 1990). In the tropics, T. vaporariorum is
more common above elevations of 500 m, and B. tabaci tends to be the predominant
species below 500 m (Caballero 1994).
Whitefly nymphs are sessile and susceptible to parasitism (Gerling 1990). Trialeu-
rodes vaporariorum has been successfully managed in glasshouse systems with parasi-
toids (primarily Encarsia formosa Gahan, Hymenoptera: Aphelinidae) (Vet et al. 1980).
Efforts to reduce populations of B. tabaci with both introduced and native natural ene-
mies are ongoing (Roltsch & Pickett 1995, Hoelmer 1996, Goolsby & Ciomperlik 1999).
There is very little information available on whitefly parasitoids from Guatemala.
A preliminary survey was carried out during April-May 1998 in eastern Guatemala to
determine which whitefly parasitoid species were present. The survey was carried out
at the end of the dry season, when whitefly populations, and presumably populations
of whitefly parasitoids, are at their highest levels. Parasitized whitefly nymphs were
collected from three areas: the Salama Valley (approx. 1000 meters above sea level
[masl]), Sanarate (approx. 850 masl), and the Motagua Valley (230-340 masl). Prelim-
inary observations indicated that T vaporariorum is the predominant whitefly spe-
cies on horticultural crops in the Salama Valley and in the Sanarate area, and B.
tabaci is the predominant species in the Motagua Valley.
Material was collected in the Salama Valley from the Instituto de Ciencia y Tec-
nologia Agrfcolas (ICTA) field station in San Jer6nimo (15003'40"N, 90 15'00"W) and
at the farms of Rene Santos and Margarito Cordova. In the Sanarate area, material
was collected from Finca Monte Grande (1447'02"N, 9012'15"W), Finca El Comun,
and the farm of Francisco del Cid. In the Motagua Valley, material was collected from
Usumatlan (1456'45"N, 90W), San Augustin, and the banks of the Rio Hato where
it crosses beneath the Atlantic highway. San Jer6nimo is about 30 km north of San-
arate. Usumatlan is about 50 km southeast of San Jer6nimo and about 50 km north-
east of Sanarate.
Parasitized whitefly nymphs were collected from common bean (Phaseolus vulgaris
L.), cucumber (Cucumis sativus L.), guisquil (Sechium edule Schwartz., a cucurbit),
squash (Cucurbita pepo L.), tomato (Lycopersicon esculentum Mill.) and watermelon
(Citrullus vulgaris Schrad.). Cucumber and tomato were the only plant species col-
lected from each of the three general areas. Plants were examined in the field, and
leaves which appeared to have high numbers of late-instar and parasitized nymphs
were placed in unwaxed cylindrical 0.95-liter cardboard cartons (Fonda Group Inc.,
Union, NJ, USA) for parasitoid emergence. After 4 wk, dead parasitoid adults were re-
moved from the containers and placed on cotton wool in gel capsules. These were then
mailed to the Division of Plant Industry and Consumer Services in Gainesville, FL, for
identification. The dried host plant material was placed in plastic bags and mailed to
Dr. Andrew Jensen, formerly of the United States Department of Agriculture in Belts-
ville, MD, who identified the whitefly species from nymphs on the dried leaves.
Trialeurodes vaporariorum was the only whitefly species found in material col-
lected from the Salama Valley and Sanarate, and B. tabaci was the only whitefly spe-
cies identified from material collected in the Motagua Valley (Table 1). The parasitoid
species recovered consisted of Encarsia pergandiella Howard, Eretmocerus sp. (Hy-
menoptera: Aphelinidae), and Signophora aleyrodis Ashmead (Hymenoptera: Signa-
phoridae), a hyperparasitoid (Table 1). In the Salama Valley, all but one of the 1150
parasitoid adults collected were E. pergandiella. One Eretmocerus sp. was collected
from cucumber in that area. Encarsia pergandiella predominated in the material from
Sanarate, although Eretmocerus sp. was present in higher numbers than in the
Salama Valley (Table 1). The ratio ofE. pergandiella to Eretmocerus from the Sanarate
area was 158 to 8 (20:1). Eretmocerus was present in higher numbers than E. pergan-
diella in material collected from the Motagua Valley. The ratio of E. pergandiella to
Eretmocerus was 409 to 555 (1:1.4). One S. aleyrodis female was recovered from cu-
cumber in the Motagua Valley. It is unclear from this study if the shift in parasitoid ra-
tio was due to changes in elevation, changes in whitefly host, or a combination of both.
There are apparently two distinct color forms of E. pergandiella. The light form is
entirely yellowish in color and the dark form has dark brown areas on the mesoscu-
tum, axillae, and gaster. Light and dark individuals were collected from T vaporari-
orum on cucumber, guisquil, squash, and tomato in the Salama Valley, and from B.
tabaci on cucumber and tomato in the Motagua Valley.
The significance of these dark and light forms is unclear. Laudonia and Viggiani
(1993) found that Encarsia partenopea Masi produced light color individuals at tem-
peratures around 30 C, and darker individuals at 15 C. In Guatemala, populations of
both the dark and light form emerged from the same whitefly species collected on the
same plant at the same time. This suggests that the observed color variation inE. per-
gandiella females is not induced by either differences in host, host plant, relative hu-
midity, or temperature. Perhaps closer examination of these morphologically similar
forms occurring sympatrically may reveal the existence of two distinct species.
Collections of whitefly parasitoids from tomato in the Salama Valley from Nov.-
Dec. 1998 indicate that parasitoid diversity may be greater in the rainy season than
in the dry season (Smith 1999). The current study indicates that there are important
regional differences in the species composition of whitefly parasitoids as well.
494 Florida Entomologist 83(4) December, 2000
zt -zt ; -
tCOtC ;Ot' kOt. "0t. t. "0" k"
m b L--O O C -1 m 1-1 m m C
2.0 0 0 0 2
a a ,aB B
_^_ O.k iC .k Ck .kiCO. kO.^
Primary parasitoids collected from T vaporariorum and B. tabaci on a variety of
horticultural crops in eastern Guatemala consisted of Encarsia pergandiella and
This research was supported by the Academy for Educational Development, the
Institute de Ciencias y Tecnologia Agricolas, and by the Institute of Food and Agricul-
tural Sciences at the University of Florida. The authors wish to thank Ing. Agr. Mario
Morales of ICTA and Lic. Margarita Palmieri of the Universidad del Valle for their as-
sistance, and Dr. Andrew Jensen for whitefly species identifications. This is Florida
Agricultural Experiment Station Series No. R-07560.
BROWN, J. K. 1994. Current status of Bemisia tabaci as a plant pest and virus vector
in agroecosystems worldwide. FAO Plant Prot. Bull. 42: 3-32.
BYRNE, D. N., T. S. BELLOWS, JR., AND M. P. PARRELLA. 1990. Whiteflies in agricul-
tural systems. Pp. 227-261 in D. Gerling, (ed.), Whiteflies: their bionomics, pest
status and management. Intercept Ltd, Andover, Hants, UK.
CABALLERO, R. 1994. Clave de campo para inmaduros de moscas blancas de Cen-
troamerica. Escuela Agricola Panamericana, Zamorano, Honduras.
GERLING, D. 1990. Natural enemies of whiteflies: predators and parasitoids. pp 147-
185 in D. Gerling (ed.) Whiteflies: their bionomics, pest status and manage-
ment. Intercept Ltd., Andover, Hants, UK.
GOOLSBY, J. A., AND M. A. CIOMPERLIK. 1999. Development of parasitoid inoculated
seedling transplants for augmentative biological control of silverleaf whitefly.
Fl. Entomol. 82: 532-545.
HOELMER, K. A. 1996. Whitefly parasitoids: can they control field populations of Be-
misia? Pp. 451-476 in D. Gerling and R. Mayer (eds.) Bemisia 1995: taxonomy,
biology, damage, control, and management. Intercept, Andover, Hants, UK.
LAUDONIA, S., AND G. VIGGIANI. 1993. Effetto della temperature sulla colorazione de-
gli adulti di Encarsia partenopea Masi (Hymenoptera: Aphelinidae). Bollettino
del Laboratorio di Entomologia Agraria "Filippo Silvestri" Portici. 50: 141-146.
ROLTSCH, W., AND C. PICKETT. 1995. Silverleaf whitefly natural enemy refuges in the
Imperial Valley. P. 135 in T. J. Henneberry, N. C. Toscano, R. M. Faust, and J. R.
Coppedge (eds.) Silverleaf whitefly: 1995 supplement to the 5-year national re-
search and action plan. Agriculture Research Service No. 125, U.S. Department
of Agriculture, Washington, DC.
SMITH, H. A. 1999. Intercropping and whitefly management. Ph. D. dissertation. In-
stitute of Food and Agricultural Sciences, University of Florida. Gainesville,
VET, L. E. M., J. C. VAN LENTEREN, AND J. WOETS. 1980. The parasite-host relation-
ship between Encarsia formosa and Trialeurodes vaporariorum. IX. A review of
the biological control of the greenhouse whitefly with suggestions for future re-
search. Z. Ang. Ent. 90: 26-51.
POLYMORPHIC LARVAL RETREATS IN THE NET-SPINNING
CADDISFLY MACROSTEMUM CAROLINA (TRICHOPTERA:
HYDROPSYCHIDAE): FORM AND PUTATIVE FUNCTION
GORDON R. PLAGUE AND J VAUN MCARTHUR
Savannah River Ecology Laboratory, University of Georgia,
PO Drawer E, Aiken, SC 29802
Larval net-spinning caddisflies of the genus Macrostemum Kolenati (Trichoptera:
Hydropsychidae) construct their catchnets within protective retreats. This genus is
composed of 88 species and is distributed worldwide (Morse 1999). However, the re-
treat architecture has only been described for three North American species (Wallace
& Sherberger 1974, 1975, Wallace 1975) and one South American species (Sattler
1963). The described retreats share a general, rather elaborate architecture (Fig. la)
with the following characteristics: (i) two chambers, one housing the catchnet and one
housing the insect, and (ii) the water entrance hole is at the end of a silken or sand
grain tube that leads into the two-chamber area. Here we report an alternate retreat
design constructed by some Macrostemum carolina (Banks) individuals in the Savan-
nah River, Georgia and South Carolina.
Macrostemum carolina is widely distributed throughout the southeastern United
States, and has been recorded west to Texas (Moulton & Stewart 1997) and north to
New York (Ross 1944). In coastal plain streams with shifting sand streambeds, M.
carolina primarily inhabits submerged snags (i.e. fallen trees or branches), gouging
the base of their retreats out of the wood and covering the top of the structure with
silk. In their original description, Wallace & Sherberger (1974) noted that some M.
carolina individuals in the Apalachicola River construct a second, slightly different
retreat than the one described above. This alternate retreat lacks a silken tube and
simply has the entrance hole open into the chamber area (Fig. ib) (some Macroste-
mum zebratum (Hagen) individuals construct a similar, alternate retreat (see Wallace
1975)). In the Savannah River, M. carolina individuals construct the two retreats de-
scribed above as well as a third type with yet a different entrance hole configuration.
The entrance hole of this third retreat also lacks a silken tube and instead has a ~ 180
silken backstop, with the other -180 essentially flush with the top of the retreat (Fig.
ic). These backstops vary in size, from 3-8 mm in height, though some of this variation
is positively correlated with instar (G. R. P., personal observation). Macrostemum
carolina is common in the Savannah River (Cudney & Wallace 1980), and each retreat
morph is regularly encountered. Individuals of a single morph are often clustered on
snags, although the "flush" phenotype is generally the most prevalent (G. R. P., per-
Although these three retreat morphs are discrete behaviors (though see below),
the individuals in the Savannah River represent a single, panmictic population
(Plague et al., in press). Therefore, retreat construction in M. carolina is either: (i)
phenotypically plastic, with environmental cues influencing retreat design (e.g.
Emlen 1994), (ii) genetically polymorphic, with alternative alleles at a retreat gene (or
genes) controlling the design (e.g. Hori 1993), or (iii) partially heritable, i.e., a combi-
nation of genetic and plastic control (e.g. Roff 1986). Whichever is the case, natural se-
lection probably plays a role in maintaining the alternative phenotypes (Hartl &
Clark 1997, Futuyma 1998). The adaptive value of each design is likely related to the
maintenance of adequate water flow through the retreat, and specifically the net.
Therefore, each morph may be adapted to a particular microhabitat on the snag. For