ECOLOGY AND POPULATION BIoLoCY
Natural History of a Gall-Inducing Weevil Collabismus clitellae
(Coleoptera: Curculionidae) and Some Effects on its Host Plant
Solanum lycocarpum ( Solanaceae ) in Southeastern Brazil
ANDREA LUCIA TEIXEIRA DE SOULA,1, 2 GERALDO WILSON FERNANDES,2
JOSE EUGENIO CORTES FIGUEIRA,2 AND MARCEL OKAMOTO TANAKA'
Ann. Entomol. Soc. Am . 91(4): 404 -409 (1998)
ABSTRACT The phenology , general characteristics , and mortality factors acting upon the weevil
Collabismus clitellae Boheman in a population of Solanum lycocarpum St. Hil. (Solanaceae) were
investigated in southeastern Brazil , as well as its distribution and impact on the host plants. Mating
and oviposition of the weevils were observed in the beginning of the summer , with larval development until the autumn and emergence in spring . Most galls were found toward the base on plants
0.9-1.2 m high, a distribution perhaps the result of physiological differences between plants of
different height . Natural enemies also may influence this pattern ; logistic regressions showed that
greater gall size and increased height above the ground increased mortality caused by the woodpecker Colaptes campestris ( Vieillot ). Larger galls also were more frequently attacked by the fungus
Penicillium sp. Gall attack rates were correlated with S. lycocarpum stem mortality ( 43.4% of plants
analyzed), because galls could act as nutrient sinks or favor the breaking of stems. The possible effect
of C. clitellae on the population dynamics of S. lycocarpum is discussed.
KEY WORDS
tality
Collabismus clitellae, Coleoptera, Solanaceae, plant-insect interactions , gall mor-
GALL-INDUCING INSECTS are extremely diverse and have
interesting and complex life histories. Most of their life
is spent embedded within tissues of their host plants,
with which they have perhaps the most intimate relationship of all organisms ( Mani 1964 , Mattson et al.
1988, Fernandes 1990). The understanding of the natural history of Neotropical galling insects is tremendously poor, most studies having centered on the ecological aspects of gall distribution ( Fernandes et al.
1988; Fernandes and Price 1988, 1991). In addition,
there are few studies on the frequency and kind of
impacts caused by galling insects on their host plants,
especially for shoot galls (Sacchi et al. 1988).
The weevil Collabismus clitellae Boheman forms
galls on shoots of the lobeira tree, Solanum lycocarpum
St. Hil. (a synonym of S. grandiflorum RandP. variety
1Carvalho 19851) (Solanaceae), a woody shrub that
grows to a height of 3 m. This plant is a common weed
in central and southeastern Brazil , occurring along highways, pasture, grasslands , and disturbed areas (Carvalho
1985 ). This association also was recorded for S . grandiflorum in Minas Gerais State ( Bondar 1923, Lima 1956).
Here we describe the phenology, general characteristics , and mortality factors acting upon C. clitellae
in a population of S. lycocarpum in southeastern Brazil.
We also provide data on the distribution of galls within
I Departamento de Zoologia, IB, Universidade Estadual de Campinas, C. P. 6109, CEP 13083 - 970, Campinas, SP, Brasil.
2 Departamento de Biologia Gera] , ICB, Universidade Federal de
Minas Gerais , C.P. 486, CEP 30161-970, Belo Horizonte , MG, Brasil,
and between plants and discuss some aspects of the
impact of these galls on the host plant.
Materials and Methods
The study was conducted between May 1991 and
July 1992 in a population of S. lycocarpum located
along highway MG-010 (km 89), near Serra do Cip6,
Minas Gerais, 840 m above sea level. The dominant
vegetation is cerrado, which is regularly cut as a management practice for cow grazing. Plants of S. lycocarpum were cut at 5 cm above the ground every other
year, and were last pruned in January 1990. The population studied consisted of clones with several ramets
patchily distributed. Ramets were 16-18 mo old and
varied from single small shoots (<10 cm from the
ground) to 1.6 m high. Ramets up to 3 m high also were
found and examined in the study area.
We randomly collected 385 galls from 97 individuals
of S. lycocarpum. To identify possible patterns of gall
distribution within and between plants, we measured
plant height and gall height on the plant relative to the
ground. Because plant height was variable, the distribution of galls on different plant parts was analyzed
using the gall height/plant height ratio. The distribution on the plants was then measured by dividing gall
locations into basal (0-0.33), median (0.33-0.66), and
apical (0.66-1.00) parts of the plant. We also recorded
the diameter of the galled stems and whether they
were wilting or dead, to evaluate the impact of galls on
stem survival, using the gall attack rates ( see below).
0013-8746/98/0404-0409$02. 00/0 © 1998 Entomological Society of America
July 1998
DF. SouzA Er AL.: WEEVIL CALLS ON S. lycocarpum
In the laboratory, we measured the length and diameter of galls and estimated their volume with the
formula volume - length X (diameter/2)2. To test
whether this measure provided a statistically reliable
estimate of gall size, we randomly subsampled 67 galls
whose volume we measured by immersing each gall in
a graduated tube filled with water and recording the
volume of water displaced. As the relationship between both measurements was linear (y = 1.88 +
8.01x, r2 = 0.83, P < 0.001), the former method was
used to estimate gall size throughout this study.
The attack rate per stem was evaluated as the sum
of chambers in all galls found per stem; later, all galls
were dissected and the chambers were counted and
opened. This approach was needed because the number of chambers per gall and number of galls per stem
was highly variable. Furthermore, there is only 1 beetle larva per chamber, which would be a better estimate of the attack rates.
During oviposition by females of C. clitellae, we
randomly marked 175 egg batches on 48 plants. Of
these, we measured the stem diameter, plant height,
and the number of scars left by the females on the
epidermis of the stem oviposited upon. We also recorded the number of stems that were broken exactly
where the eggs were deposited.
The mortality factors acting upon C. clitellae in
relation to plant height, gall height in relation to
ground, and gall volume were analyzed with logistic
regressions (Hosmer and Lemeshow 1989, Steinberg
and Colla 1991). The odds ratio ('1') was calculated
according to an increase of c units of the independent
variable analyzed, with the following equation:
4,=ep'`
and 95% confidence limits set by the relation
CI = e(s*c' 1.96"EPs•c)
where (3 is the parameter estimate and EPp is the standard error of the estimate (Hosmer and Lemeshow
1989). A similar model was used to analyze stem survival
in relation to beetle attack rates and stem diameter. The
relationship between these factors and stem diameter
categories on the breaking off of stems was analyzed
using the G test (Sokal and Rohlf 1995). Voucher specimens of C. clitellae were deposited in the Museu de
Hist6ria Natural, Universidade Estadual de Campinas.
Results
Biology of C. clitellae. Mating and oviposition of C.
clitellae occur in the beginning of summer (NovemberDecember). Ovipositing females use their mouthparts to
perforate the epidermis of nonwoody stems (shoots) of
S. lycocarpum to lay their eggs. Each hole is quickly sealed
off after egg deposition, leaving a conspicuous sear on the
stem. We collected several newly oviposited stems for
dissection in the laboratory, and only 1 egg was deposited
in each hole. Therefore, the number of eggs laid corresponded to the number of scars left on the surface of the
stem. We observed 3-181 (29.3 ± 21.46 ( mean ± SD ] )
eggs per oviposition. Eggs were deposited around the
entire circumference of a stem, but in a spiral formation. Thus, the length of the oviposition bout varied
405
both with the number of eggs deposited by a female
and stem diameter.
Each larva of C. clitellae (Fig. 1A) was concealed
within a single chamber (Fig. 1E), developing between December and May. Galls continued to grow as
larvae fed on the galled tissue until larvae terminated
feeding and pupated, which occurred in May. Pupae
(Fig. 1B) were found from May to July, and adults
were found inside galls between July and September.
Emergence of adults from their galls (Fig. IC) took
place from late September to October.
The adults of C. clitellae (Fig. 1D) are pale yellow
with a dark spot on the elytra. This color matches the
pale yellow of the stem of their host plant. Adults are
very abundant in the field from October to December,
where they are found commonly feeding on newly
growing shoots, leaves, and flowers of S. lycocarpum,
as well as ovipositing on new shoots.
Newly formed galls are green, juicy, and soft (Fig.
1F). Their internal tissues have many fibers and a large
amount of water. They are glabrous, globular, or ellipsoid and may occur singly or in coalescence. As galls
age, their tissue becomes drier, woody, and brownish.
Galls were very variable in size, ranging from 1.26 to
16.85 cm (3.55 ± 1.96, n = 385) long and from 0.59 to
4.65 cm (2.41 ± 0.81, n = 385) wide. Their volume
varied from 0.34 to 139.84 cm3 (21.74 ± 24.59, n = 385).
The number of chambers also was highly variable,
ranging from 1 to 70 chambers per gall.
Gall Distribution Between and Within Plants. Galls
were distributed on plants that ranged between 40 and
200 cm high, only 3% being recorded on plants higher
than 160 cm and the greatest proportion (42%) on
plants 80-120 cm tall (Fig. 2a). Within the host plants,
they occurred mostly on the median and basal parts,
greater numbers in the latter region (G = 9.7, df = 3,
P = 0.02). Because only 15 galls were found on apical
stems, they were excluded from the analysis.
Of all galls collected, 249 (65%) were smaller than
20 cm3, containing from 1 to 8 chambers. Large galls
(>40 cm3) might have up to 68 chambers, more of
them occurring on the basal stems of the plants than
on the median region (G = 15.6, df = 3, P = 0.001, Fig.
2b). All galls located on the apical region were <40 cm3.
Mortality Factors. An unidentified species of the
fungus Penicillium was the major mortality factor acting upon C. clitellae during our study (35.9% of all dead
galls). The external morphology of the gall was not
altered by the presence of the pathogen, whereas its
growth inside the gall damaged larval chambers and
killed the galling larva. On 38.2% of the galls attacked
by Penicillium sp., we observed 1-3 larvae of an unidentified species of Cerambycidae (Coleoptera).
These larvae formed tunnels inside the pathogendamaged galls. The occurrence of the fungus was most
frequent in the largest galls, but was independent of plant
and gall height (Table 1). According to the analysis, an
increase in 10 cm3 in gall size increased the chance of a
gall being attacked by a factor of 1.22. For example, if the
chance of a gall being attacked in the field is 10%, an
increase of 10 cm3 will raise this chance to 12.2%.
406 ANNALS OF THE ENTOMOLOGICAL SOCIETY OF AMERICA Vol. 91, no. 4
L•
5mm
D
Fig. 1 . Several phases of the galls caused by C. clitellae on S. lycocarpum. ( A) Larva. ( B) Pupa. ( C) Adult excavating the
gall walls to emerge . ( D) Adult . ( E) View of a gall chamberlocated close to the gall external wall. (F) Gallon the basal portion;
holes were drilled by adults that left the gall successfully.
Galls also were attacked by the woodpecker Colaptes
campestris (Vicillot ) ( Piciformes : Picidae ). These birds
perforated the gall walls to reach the juicy larvae inside
the chambers, partially or even entirely destroying galls.
C. campestris was frequently observed on branches of the
host plant, and attacks were commonly recorded in the
field. Birds most frequently attacked the largest galls or
those located on the superior portion of the host plant
(Table 1). Analysis indicated that an increase of 10 cm in
gall height increased the chance of being attacked by C.
campestris by a factor of 1.36, and a 10 cm3 increase in gall
volume led to a factor of 1.17. Despite the fact that plant
height had an important effect on the decision of the
woodpecker to attack a given gall , plant height alone
was not a risk factor because the lower confidence
limit of the odds ratio was <1.0 (e.g., Hosmer and
Lemeshow 1989 ). Therefore, the most important factors influencing the pattern of attack by C. campestris
were gall size and gall height from the ground.
Galls also were attacked by 2 species of parasitic
hymenopteran ; both leave their coccons inside the
gall after emergence . This study did not focus on the
relationship between the parasitoids and gall distribution patterns within plants because parasitism rates
were much lower than other mortality factors (14.4%).
A large percentage of galls was killed by undetermined
factors, possibly including nematodes , ants, and plant
resistance (30.7% of galls attacked).
DE SoIJZA ET AL.: WEEVIL GALLS ON S. lycocarpum 407
July 1998
larvae per stem was an important factor influencing stem
survival; an increase in 10 chambers per stem (10 individuals) increased 2.4 times the probability of stem
death, whereas an increase of 1 cm in stem diameter
decreased the probability of death by 0.042 times.
Thus, thinner stems without galls are more likely to
survive, as would larger stems, even if galled.
Drilling by females during oviposition frequently resulted in stem breakage by mechanical factors such as
wind, rain, and by passing animals. Among the stems
analyzed, 24.4% were broken exactly at the site of oviposition. Breaking was frequently observed in the thinnest stem classes (G = 11.63, df = 4, P = 0.02, Fig. 3).
(a)
80-120 120-160 >160
Plant Height (cm)
Discussion
(b)
The mortality of C. clitellae from external agents is
independent of plant height. Thus, the greater occurrence of galls on smaller plants may be caused by
inherent factors of the interaction between the galling
insect and S. lycocarpum. First, plants or parts of the
plant may be in different phenological or developmental stages, which could influence the survival and
performance of C. clitellae larvae and consequently
the attack rates observed (e.g., Kearsley and Whitham
1989). The aging of host plants can lead to differences
in the nutritional quality of young and mature tissues
because of distinct chemical or mechanical defense
levels, which influences the distribution of herbivorous insects (Frankie and Morgan 1984, Whitham et al.
1984, Craig et al. 1986, Akimoto and Yamaguchi 1994).
Second, the pattern observed could result from female
behavior when choosing oviposition sites as attempts
to maximize the survival and performance of her offspring (Thompson 1988, Craig et al. 1989, Price 1991)
(but see Burstein and Wool 1993). A 3rd hypothesis is
that this distribution may result from the impact of galls
on plants, reducing their growth (Fernandes 1986,1987).
Whether these hypotheses explain the pattern observed will be clarified only with further experiments.
In addition to the limitation on the distribution
among plants of different heights, galls of C. clitellae
occur more frequently on the basal portions of the
attacked plants, where galls attain the largest sizes. An
0-20 20 - 40 40-60 >
Gall size (cm)
Fig.2. Number of galls in (a) plant height classes and (b)
gall size classes in different parts of plant. Dark bars represent
galls located in apical portion, hatched bars in median portion, gray bars in basal portion.
Impact on I-lost Plant. The number and size of galls
varied enormously among stems. Many galled stems
died, probably as a consequence of gall induction; death
began with the wilting and drying of the apical region. Of
76 randomly collected stems on 76 plants, 33 (43.4%)
were dead. The probability of stem death was positively
associated with attack rates and negatively associated
with stem diameter (Table 2). Therefore, the number of
Table 1. Logistic regression results for the explanatory variables of C. clitellae mortality
95% Cl
Parameter
Estimate
SE
Constant
-1.916
0.565
Plant height
Gall height
Gall size
- 0.003
0.004
0 . 020
0 . 006
0 . 008
0 . 005
P
Penicillium sp.°
0.001
0 .658
0 .602
<0 . 001
Odds ratio
Upper
Lower
1.09
1.22
0.86
0.89
1.35
1.11
0.90
1.76
1.37
0.61
1.06
1.003
0.97
1.04
1.22
Colaptes campestrist'
Constant
Plant height
Gall height
Gall size
-0.877
- 0.030
0.031
0.016
C = 10 for all variables analyzed.
°G= 13.04, df = 3, P = 0.005.
6G= 14.18, df = 3, P = 0.003.
0.703
0 , 010
0.013
0.008
0.212
0 .002
0.013
0 .039
0.74
1.36
1.17
Vol. 91, no. 4
ANNALS OF THE ENTOMOLOGICAL SOCIETY OF AMERICA
408
Table 2. Logistic regression results for the explanatory variables of S. lycocarpum stem mortality
95% Cl
Parameter Estimate
Constant
Diameter
No. chambers
2.477
-3.168
0.087
SE
P Odds ratio
Upper
1.059
1.070
0.0.35
Lower
0.019
-
-
-
0.003
0.014
0.042
2.387
0 .342
4.740
0.005
1.202
G = 14.21, df = 3, P = 0.001. C = 1 for the stem diameter; C = 10 for the number of chambers.
important factor determining these distribution patterns was the attack of the woodpecker C. campestris,
because the chances of being attacked increase with
gall size and height on the plant. Birds of the family
Picidae search for insect larvae, mainly beetles in
wood, which constitute the preferred food item of
many Brazilian species (Sick 1984). Generally, insectivorous birds hunt by visual search, and individuals of
this species should detect the galls most easily visualized at greater distances. Attack of insect galls by the
woodpecker Picoides pubescens on the Canadian goldenrod, Solidago canadertsis, also were intensive on
large galls on tall stems (Confer and Paicos 1985). The
attack rates recorded for C. campestris (19% of all
attacked galls) seems to be relatively high and may
assume a larger importance for those galls located on
the higher portions of the plants, where gall occurrence is lower than on the basal region.
Mortality caused by fungus also was higher on the
largest galls. In 38.2% of the attacked galls, we observed larvae of a cerambycid that probably fed on the
mycelia, and adults may be acting as a dispersal agent
of this pathogen. Fungal transport by several coleopteran families has been recorded. Vectors of plant
diseases have been recorded in Scolytidae (Levieux et
al. 1991), Curculionidae (Nevill and Alexander 1992),
Nitidulidae (Currie 1995), and Anobiidae (Hoover et
al. 1995). Thus, the behavior of this cerambycid could
influence the distribution of Penicillium sp. and ultimately gall survivorship.
Solanum lycocarpum stems died according to a balance between gall infestation rate and stem diameter.
The stem dries progressively from the apex, modifying
10080-
60-
or interrupting the development of galls located at
greater heights . This suggests the possibility of competition among C. clitellae galls . Those in the basal
parts of the plants may be draining resources otherwise available for galls in the apical parts (e.g., galls on
different leaf parts [ Whitham 1980]). These galls not
only contributed to stem breaking, they also reduced
plant fitness ; stem mortality caused by intense gall
attack was 43.4%. The direct effect of herbivores on
plants can be very strong, influencing the net production of inflorescences and seeds, modifying plant architecture or sex expression ( or both ), or inducing
shoot mortality ( Craig et al. 1986, Sacchi et al. 1988,
Whitham et al. 1991 ). The presence of galls can influence the physiology of the host plants , acting as
metabolic sinks for energy and mineral nutrients and
modifying tissues that otherwise could serve for
growth and reproduction (Abrahamson and Weis
1987, Price and Louw 1996 ). Indirect effects include
the susceptibility to other herbivore species because
of the modification of plant tissue and/or plant resistance caused by a specific herbivore (Whitham et al.
1991 , Price and Louw 1996 , McGeoch and Chown 1997).
The damage of herbivores to the reproductive characteristics of plants may depend on the phenology of
the plant ( Harper 1977). Because gall frequency was
higher on smaller plants , young dead plants were observed several times. In larger plants, the infestation
rates were smaller , and there was less damage to the
plants. Sacchi et al. (1988 ) observed that the sawfly
Euura lasiolepis influenced the production of inflorescences by Salix lasiolepis only during the year of
infestation , rather than in subsequent years. The artificial cut of S . lycocarpum keeps a high resource
availability to C. clitellae, maintaining young branches
highly susceptible to attack by the herbivore . Thus, the
pruning season of those agricultural and pasture fields
should influence the effect of C. clitellae on S. lycocarpum when adjusted to the life cycle of this insect.
40Acknowledgments
20 ^
We are most grateful for critical discussions of this work
0-
0.4-0.6 0 . 6-0.8 0 .8-1.0 1.0-1.2 1.2-1.4
Shoot Diameter (cm)
Fig. 3. Number of eggs deposited on stems in different
shoot diameter classes. 0, Broken stems; •, unbroken stems.
with R. P. Martins and H. R. Pimenta. We thank H. Burke for
the identification of the galling weevil, and the Parque Nacional da Serra do Cip6-IBAMA for logistical support. We
also thank C. Schaefer and an anonymous referee, for critically reviewing the manuscript.This study was funded by
Conselho Nacional de Pcsquisa.
July 1998
DE Souza Er AL.: WEEVIL CALLS ON S. lycocarpum
References Cited
Abrahamson, W. G., and A. E. Weis. 1987. Nutritional ecology of arthropod gall makers, pp. 235-258. In F. Slanskey,
Jr. and J. G. Rodriguez [ eds. 1, Nutritional ecology of
insects , mites, spiders and related invertebrates. Wiley,
New York.
Akimoto , S., and Y. Yamaguchi . 1994 . Phenotypic selection
on the process of gall formation of a Tetraneura aphid
(Pemphigidae). J. Anim. Ecol. 63: 727-738.
Bondar, G. 1923. Biologia do genero Collabismus. Archivos
da Escola Superior de Agricultura e Medicina Veterinaria
7: 23-25.
Burstein , M., and D. Wool. 1993. Gall aphids do not select
optimal galling sites (Smynthurodes betae; Pemphigidae).
Ecol. Entomol. 18: 155-164.
Carvalho, L.A.F. 1985. Flora fanerogamica da Reserva do
Parque Estadual das Fontes do Ipiranga (Sao Paulo, Brasil). Hoehnea 12: 67-85.
Confer, J. L., and P. Paicos. 1985. Downy woodpecker predation at goldenrod galls. J. Field Ornithol. 56: 56-64.
Craig, T. P., P. W. Price, and J . K Itami. 1986. Resource
regulation by a stem-galling sawfly on the arroyo willow.
Ecology 67: 419-425.
Craig, T. P., J. K Itamy, and P. W. Price. 1989. A strong relationship between oviposition preference and larval performance in a shoot-galling sawfly. Ecology 70: 1691-1699.
Currie, C. R. 1995. Dissemination of the mycoparasite Scyladium uredinicola by Epuraea obliquus (Coleoptera:
Nitidulidae). Can. J. Bot. 73: 1338-1344.
Fernandes, G. W. 1986. Ecologia evolutiva de galhas de insetos: teoria e pratica. Brazilian Embassy, Sector of Science and Technology, Washington, DC.
1987. Gall-forming insects: their economic importance and
control. Rev. Bras. Entomol. 31: 579-595.
1990. Hypersensitivity: a neglected plant resistance mechanism against insect herbivores. Environ. Entomol. 19:
1173-1182.
Fernandes, G. W., and P. W. Price. 1988. Biogeographical
gradients in galling species richness: tests of hypotheses.
Oecologia (Berl.) 76: 161-167.
1991 . Comparisons of tropical and temperate galling species richness: the roles of environmental harshness and
plant nutrient status, pp. 91-115. In P. W. Price, T. M.
Lewinsohn, G. W. Fernandes, and W. W. Benson [eds.],
Plant-animal interactions: evolutionary ecology in tropical and temperate regions . Wiley, New York.
Fernandes, G. W., E. Tameirao- Neto, and R . P. Martins. 1988.
Ocorrencia a caracterizapao de galhas entomSgenas na
vegetacao do Campus-Pampulha, UFMG, Belo Horizonte-MG. Rev. Bras. Zoo]. 5: 11-29.
Frankie, G. I., and D. L. Morgan. 1984. Role of host plant
and parasites in regulating insect herbivore abundance,
with an emphasis on gall -inducing insects , pp. 101-140. In
P. W. Price, C. N. Slobodchikoff, and W. S. Gaud [ eds. ],
A new ecology: novel approaches to interactive systems.
Wiley, New York.
Harper, J. L. 1977. Population biology of plants. Academic,
New York.
Hoover, K, D. L. Wood, J. W. Fox , and W. E. Bros. 1995.
Quantitative and seasonal association of the pitch canker
fungus, Fusarium subglutinans f. sp. pini with Conophthorus radiatae (Coleoptera: Scolytidae) and Ernobius
punctulatus (Coleoptera: Anobiidae) which infest Pinus
radiate. Can. Entomol. 127: 79 -91.
409
Hosmer, D. W., Jr., and S . Lemeshow. 1989 . Applied logistic
regression. Wiley, New York.
Kearsley, MJ.C., and T. G. Whitham . 1989. Developmental
changes and resistance to herbivory: implications for individuals and populations. Ecology 70: 422-434.
Levieux, J., P. Cassier , D. Guillaumin , and A . Roques. 1991.
Structures implicated in the transportation of pathogenic
fungi by the European bark beetle, Ips sexdentatus
Boerner: ultrastructure of a mycangium. Can. Entomol.
123: 245-254.
Lima, A.M.M.C. 1956. Insetos do Brasil 10° Tome: Coleopteros. Serie didatica n° 12. Escola Nacional de Agronomia. Rio de Janeiro, RJ.
Mani , M. S. 1964. Ecology of plant galls. Junk, The Hague.
Mattson , W. J., 1. Levieux, and C. Bernard- Dagan.
1988. Mechanisms of woody plant defenses against insects -search for pattern. Springer, New York.
McGeoch, M. A., and S. L. Chown. 1997. Evidence of competition in a herbivorous, gall-inhabiting moth (Lepidoptera) community. Oikos 78: 107-115.
Nevill, R. J., and S . A. Alexander. 1992 . Pathogenicity of
three fungal associates of Hylobius pales and Pissodes
nemorensis (Coleoptera: Curculionidae) to eastern white
pine. Can. J. For. Res. 22: 1438-1440.
Price, P. W. 1991 . Patterns in communities along latitudinal
gradients, pp. 51-69. In P. W. Price, T. M. Lewinsohn,
G. W. Fernandes, and W. W. Benson [ eds. ], Plant-animal
interactions: evolutionary ecology in tropical and temperate regions. Wiley, New York.
Price, P. W., and S . Louw. 1996 . Resource manipulation
through architectural modification of the host plant by a
gall-forming weevil Urodontus scholtzi Louw (Coleoptera: Anthribidae). Afr. Entomol. 4: 103-110.
Sacchi, C. F., P. W. Price, T. P. Craig, and J. K Itami. 1988.
Impact of shoot galley attack on sexual reproduction in
the arroyo willow. Ecology 69: 2021-2030.
Sick, H. 1984. Ornitologia Brasileira, vol. 1. Editora Universidade de Brasilia, Brasilia.
Sokal, R. R., and F. J. Rohlf. 1995 . Biometry. Freeman, New
York.
Steinberg, D., and P. Colla. 1991 . LOGIT: a supplementary
module for SYSTAT. SYSTAT, Evanston, IL.
Thompson, J. N. 1988. Evolutionary ecology of the relationship between oviposition, preference and performance of
offspring in phytophagous insects . Entomol. Exp. Appl.
47: 3-14.
Whitham, T. G. 1980. The theory of habitat selection: examined and extended using Pemphigus aphids. Am. Nat.
115: 449-465.
Whitham, T. G., A. G. Williams, and A. M. Robinson. 1984.
The variation principle: individual plants as temporal and
spatial mosaics of resistance to rapidly evolving pests, pp.
15-51. In P. W. Price, C. N. Slobodehikoff, and W. S. Gaud
[ ads.], A new ecology: novel approaches to interative
systems. Wiley, New York.
Whitham, T. G., J. Maschinski, K C. Larson, and K N. Paige.
1991. Plant responses to herbivory: the continuum from
negative to positive and underlying physiological mechanisms , pp. 227-256. In P. W. Price, T. M. Lewinsohn,
G. W. Fernandes, and W. W. Benson ] eds. J, Plant-animal
interactions : evolutionary ecology in tropical and temperate regions. Wiley, New York.
Received for publication 5 August 1997; accepted 12 December 1997.