Efﬁciency of biological control of Gonipterus platensis (Coleoptera: Curculionidae)
by Anaphes nitens (Hymenoptera: Mymaridae) in cold areas of the Iberian Peninsula:
Implications for defoliation and wood production in Eucalyptus globulus
Ana Raquel Reis
, Luis Ferreira
, Margarida Tomé
, Clara Araujo
, Manuela Branco
Centro de Estudos Florestais (CEF), Instituto Superior de Agronomia (ISA), Technical University of Lisbon (UTL), Tapada da Ajuda, 1349-017 Lisbon, Portugal
Altri Florestal, SA, Quinta do Furadouro, 2510-582 Olho Marinho, Portugal
Received 28 November 2011
Received in revised form 21 January 2012
Accepted 25 January 2012
Sustainable management of forest plantations and cost-effective control strategies depend on previous
estimations of the economic level of damage caused by the pests. The eucalyptus weevil, a key pest of
Eucalyptus plantations worldwide, is mainly controlled using classical biological control, using the mym-
arid egg-parasitoid Anaphes nitens (Girault). Nevertheless, in several temperate regions, the parasitoid
fails to reduce the weevil populations to economically sustainable levels. This study attempts to (i) relate
the efﬁciency of the parasitoid with climate variables, (ii) relate the level of damage caused by the weevil
with the rate of parasitism, (iii) estimate the implications of weevil damage on wood production. Weevil
density, damage caused by defoliation on the upper crown and parasitism rates were monitored in 2007,
in 34 Eucalyptus globulus stands. Elevation, temperature and precipitation were assessed by using the
Worldclim database. Using historic inventory data, wood production was projected to an age of 10 years,
prior to the arrival of the weevil, and compared with current data for the same stands.
Parasitism rate by A. nitens was a key element explaining weevil density and tree defoliation, r
and 0.41 (p< 0.001), respectively. Signiﬁcant relationships between parasitism rates and maximum tem-
perature of the winter months (MaxTw), r
= 0.55 and elevation r
= 0.59 (p< 0.001) were found. Other
climatic variables, such as temperatures of the warmest months and precipitation, were not signiﬁcantly
related to parasitism rates. An upper threshold limit for the efﬁciency of the parasitoid appears for
MaxTw of 10–11 °C. The mean percentage of parasitism was low 10.1% (±4.9) for MaxTw below 10 °C,
increasing to 70.9% (±3.8) above 11.5 °C. A reduction of the efﬁciency of A. nitens due to differences in
the climatic niches of both the host and the parasitoid is hypothesised. The lower temperature threshold
in particular, is of paramount importance for this host-parasitoid system. In consequence, in colder areas
MaxTw < 10 °C, a defoliation of 74.1% was attained. Wood volume (projected to the age of 10 years) was
estimated to decrease to 51% in the affected areas in 2004–2006, compared to the previous period of
1995–1998. Estimated loss in wood volume increased exponentially reaching 43% and 86%, for 75%
and 100% of tree defoliation, respectively. Therefore, considering the increase in the economic costs
calculated for these regions, due to the high defoliation caused by the weevil, research into alternative
control strategies is urgently needed.
Ó2012 Elsevier B.V. All rights reserved.
Pests and diseases seriously threaten the productivity and sus-
tainability of forest plantations. In particular, alien pests and
pathogens, either affecting native or exotic forest plantations, are
known to cause major economic losses in forestry worldwide
(Pimentel et al., 2002). Yet, information on the efﬁcacy of manage-
ment strategies as well as their economic impact is frequently
insufﬁcient (Kenis and Branco, 2010).
Exotic tree species are extensively used in forest plantations
worldwide, allowing high-yield wood production and monetary
revenue (Sedjo, 1999). In particular, Eucalyptus species are pre-
ferred for plantations in many regions for their fast growth, excel-
lent wood-ﬁbre characteristics and tolerance to a wide range of
environmental conditions (Campinhos, 1999; Bhattacharya et al.,
2003; Stape et al., 2004). In the Iberian Peninsula, Eucalyptus plan-
tations represent an important part of the agricultural landscape.
In Spain, 325000 and 175000 ha of Eucalyptus globulus Labill.
and Eucalyptus camaldulensis Dehnh., respectively, are planted
(Montoya, 1995), and in Portugal eucalyptus plantations represent
about 24% of forest plantations, occupying 740000 ha, mostly
0378-1127/$ - see front matter Ó2012 Elsevier B.V. All rights reserved.
Corresponding author. Tel.: +351 213653382; fax: +351 213653388.
E-mail address: email@example.com (M. Branco).
Forest Ecology and Management 270 (2012) 216–222
Contents lists available at SciVerse ScienceDirect
Forest Ecology and Management
journal homepage: www.elsevier.com/locate/foreco
E. globulus (AFN, 2010). Although exotic forest plantations initially
beneﬁt from the release of their co-evolved pests and pathogens, in
the longer term such freedom will cease, as eventually the natural
enemies will also reach the new regions of establishment of their
host plant and compromise plantation health and productivity.
Over the past century, several species of insects associated with
Eucalyptus trees have found their way all over the world, thus
becoming serious pests. For instance in Europe nine alien arthro-
pod pests are currently affecting Eucalyptus plantations (Kenis
and Branco, 2010).
The Eucalyptus weevil, referred as Gonipterus scutellatus Gyllen-
hal (Coleoptera: Curculionidae) in the literature, is a major pest of
Eucalyptus plantations worldwide, and both the larvae and the
adults feed on Eucalyptus leaves. However, recent taxonomic and
genetic studies revealed that this name refers to a complex of sev-
eral cryptic species, originating from different regions in Australia
(Newete et al., 2011; Mapondera et al., 2012). According to these
studies, the species found in the Iberian Peninsula is not G. scutell-
atus but G. platensis Marelli, which was described from Argentina
although being native from Tasmania. Further, the Gonipterus spe-
cies introduced in Italy in 1975 and in South-eastern France in
1977, is not G. platensis (Mapondera et al., 2012), thus indicating
the occurrence of multiple introductions in Europe of Gonipterus
In South Africa, Gonipterus spp. referred as G. scutellatus (Tooke,
1953), and now identiﬁed as a different, but still undescribed spe-
cies (Newete et al., 2011; Mapondera et al., 2012), was at the
beginning, considered a problem to eucalyptus forestry. Biological
control of the weevil started when an egg-parasitic wasp Anaphes
nitens Girault (Hymenoptera: Mymaridae) was ﬁrst introduced in
South Africa in 1926 (Tooke, 1953). From South Africa, after the ini-
tial success of the agent, the parasitoid was introduced into many
other regions over the world, usually with at least partial success
(e.g., Cordero Rivera et al., 1999; Hanks et al., 2000). Nevertheless,
in temperate regions of South Africa, Western Australia and the
Iberian Peninsula, the damage caused by the eucalyptus weevil re-
mains high, despite the presence of the parasitoid (Loch and Floyd,
2001; Valente et al., 2004; Tribe, 2005; Govender and Wingﬁeld,
2005; Echeverri-Molina and Santolamazza-Carbone, 2010). Both
in Portugal and in Galicia, Spain, G. platensis became a major
limiting factor of Eucalyptus wood production, prompting recent
research programs to cope with this insect pest through either
tree resistance (Basurco and Toval, 2004), chemical control
(Santolamazza-Carbone and Fernandez, 2004; Echeverri-Molina
and Santolamazza-Carbone, 2010), or seasonal augmentative
release of A. nitens, as conducted by Altri Florestal enterprise since
Despite the recognized loss of productivity caused by G. platen-
sis, little is known about the impact of defoliation on tree growth in
the Iberian Peninsula. Reliable estimates of tree growth reduction
due to a speciﬁc agent, face a major difﬁculty since they require
the control of other factors inﬂuencing tree susceptibility and tol-
erance, such as soil, climate as well as other biotic interactions.
Chemically treated control trees have been used by some research-
ers to estimate defoliation impact on tree growth (e.g., Salleo et al.,
2003; Loch and Matsuki, 2010). However, in Gonipterus a prefer-
ence of the adults to oviposit on leaves of chemically treated plants
in the weeks following treatment and the combined damage done
by both larvae and adults, mask the true defoliation impact when
compared with non-chemically treated plants (Palma and Valente,
2008). The factors responsible for the low efﬁciency of the parasit-
oid A. nitens in some regions are still poorly understood, although
some studies indicate differences in parasitism rate related to sea-
son, or host density (Cordero Rivera et al., 1999; Loch, 2008). In the
present study we assess the damage caused by G. platensis and its
relationship with the efﬁciency of the parasitoid, in relation to site
climate variables, in particular monthly temperatures and precipi-
tation. We further use recent historical data, collected previously
to the arrival of the weevil and current inventory data from the
same stands, to estimate the wood loss caused by G. platensis and
determine its possible economic impact.
2.1. Field surveys: site description
During 2007, 34 pure stands representative of the Eucalyptus
young plantations of the northern and central regions of Portugal,
were monitored. These plantations of E. globulus occupy 2300 ha
(Fig. 1) and present the highest levels of attack by G. platensis
observed in Portugal (Valente et al., 2004). Stands were pure
even-aged, varying between two and ﬁve years old. Tree mean
height varied between 2 and 5 m across stands. Further, each stand
was genetically highly uniform, since all plants originated from the
same family. Each of the stands selected was systematically sam-
pled based on a 500 1000 m grid (ca. 1 plot per each 50 ha),
therefore depending on total stand area, 1 to 6 plots resulted per
stand. According to this procedure, a total of 64 plots were sampled
across the 34 stands. The plots sampled were neither affected by
pests other than G. platensis, nor by pathogens, in a conspicuous
The elevation of the plots sampled ranged from 290 to 900 m.
Monthly mean, maximum and minimum temperature and precip-
itation were taken from the WORLDCLIM datasets (Hijmans et al.,
2005;http://www.worldclim.org). This data set was derived from
the average monthly temperatures and precipitation, measured
between 1950 and 2000, by a network of meteorological stations,
from which climate surfaces were estimated by very high resolu-
tion interpolation, at a resolution of about 1 km
. To interpolated
climate surfaces the model considers elevation as a variable
(Hijmans et al., 2005). This is most relevant for the present study,
given that the studied region is mountainous, where elevation has
a major inﬂuence on climate. The network of national meteorolog-
ical stations could not provide accurate information, since the
nearest meteorological stations were at dissimilar elevations and
at a mean distance from the sample points of about 50 km. No cli-
mate extrapolation model at national level was available. Based on
Worldclim estimates, minimum, mean and maximum tempera-
tures of the three warmest (MinTs, MeanTs, MaxTs) and of the
three coldest months (MinTw, MeanTw, MaxTw) were calculated.
2.2. Population density and defoliation level
The densities of adults, larvae and the number of egg masses of
G. platensis in each plot were sampled by choosing an initial point
at the centre of the plot, from which three trees, about 125 m apart,
were selected along a row, thus minimizing spatial autocorrelation.
The number of adult weevils per tree was estimated by bending
and shaking trees over a 2 m wide plastic sleeve, and collecting
fallen weevils. This was possible due to the relatively small tree
size and trunk ﬂexibility of the young trees. When a tree was
impossible to bend, a plastic sheet was placed around its base
and it was hit with a pole. Quick and strong strokes were found
to dislodge the weevils, preventing them from gripping thigh, or
ﬂying away. The relative density of the adults was scored according
to the following scale: (0) no weevils; (1) 1–20 weevils; (2) 21–30
weevils; (3) >30 weevils.
In each tree, the number of young branches with newly formed
leaves, preferred by the weevil for oviposition, was counted. On
two branches of the upper crown of the tree with newly formed
leaves, which were taken from opposite quadrants of the trees,
A.R. Reis et al. / Forest Ecology and Management 270 (2012) 216–222 217
the total number of larvae and egg capsules were counted by care-
fully inspecting all the leaves. The total number of larvae and egg
capsules per tree was then estimated by extrapolating of the total
number of branches with newly formed leaves. Population sam-
pling was made in February 2007, at the peak of the weevil egg lay-
ing activity, as observed during previous ﬁeld surveys.
In June, after the spring damage, defoliation levels were esti-
mated in the plots sampled. In each plot, 10 trees were selected,
the ﬁrst one located at the centre of the stand and the other nine
by following a line of trees, along which one tree was selected
approximately every 10 m. The percentage of defoliation was esti-
mated in the upper third of the canopy by visual observation and
comparison with calibrated photographs, considering four levels
of damage: (0) no defoliation; (1) 1–30% defoliation; (2) 30–75%
defoliation; (3) >75% defoliation. Only the upper third of the can-
opy, containing new ﬂushing foliage, was considered because it
represents the section of plant where severe defoliation has taken
place during the spring season.
2.3. Parasitism rates
In each plot, 30 egg capsules were collected from ﬁve trees,
located about 125 m apart and selected using the same procedure
as for the trees sampled for weevil density. Only fresh egg capsules,
recognised by their texture, taken from newly foliage, and with no
exit holes were collected. Egg capsules were taken to the labora-
tory to estimate parasitism levels, and kept at about 5 °C in a refrig-
erated chamber during transportation to avoid dehydration.
In the laboratory, each egg capsule was placed in a Petri
dish sealed with paraﬁlm, appropriately identiﬁed and stored in
a climate chamber (22 ± 0.5 °C 14hL: 20 ± 0.5 °C 10hD; RH: 55%).
The number of adults of A. nitens which emerged and the number
of weevil larvae which hatched were recorded for each egg capsule,
over 30 days. Egg capsules with no emergences, either larvae or
wasps, were excluded from the analysis. The rate of parasitism
was estimated as the percentage of egg capsules with A. nitens
emergences. The application of this variable is straightforward
and it is highly correlated with the mean number of wasps
emerged per egg mass (r
= 0.916, p< 0.001).
2.4. Wood production loss
Estimates of wood production loss were based on AltriFlorestal
enterprise forest inventory, using data from 31 pure, even-aged E.
globulus stands, distributed over the same region where the forest
plantation used in this study is located. Inventory data comprised
tree height and diameter at breast height (dbh) from all trees with-
in 900 m
plots. For each stand, inventory data referring to two
periods were considered: (1) 1995 to 1998, before the establish-
ment of G. platensis and (2) 2004 to 2006, after the establishment
of G. platensis. Stand age varied from 2 to 10 years.
Defoliation levels were assessed in 2007 by surveying the 31
inventory stands, using the same methods described in section
2.2. This expressed the intensity of damage at site level, as from
our previous ﬁeld surveys the populations of G. platensis and defo-
liation levels, were kept at similar levels within each site across
years. Of the 31 stands surveyed in 2007, 18 denominated U stands
(undamaged), were not visibly affected by the weevil having a zero
defoliation level, whereas 13 D stands (damaged) experienced
visible defoliation, ranging from medium to severe.
Fig. 1. Site location of 34 Eucalyptus globulus stands monitored in spring (2007) in northern and central Portugal.
218 A.R. Reis et al. / Forest Ecology and Management 270 (2012) 216–222
The inventory data from 1995–1998 and 2004–2006 were pro-
jected onto the rotation age (10 years) using an empirical growth
model, in order to estimate dominant height (mean tree height
of the largest 100 trees per ha) and merchantable volume at har-
vest. This model is operationally used by AltriFlorestal company
and was developed by M. Tomé (unpublished) (Table Appendix
1). Using the same inventory stands, along with the projections
of the estimated merchantable volume at 10 years in the two peri-
ods of time (1995–1998 and 2004–2006), we calculated the loss of
merchantable volume caused by the presence of G. platensis,asa
percentage. This loss of production was then related to the level
of damage observed in the same plots, in 2007 by curve ﬁtting.
2.5. Data analysis
The levels of density of the adult weevils, larvae and egg cap-
sules obtained, as explained in section 2.2., were extrapolated by
multiplying the mean number per branch by the number of
branches with newly foliage to obtain an estimation of density
per tree and then an average of density per tree, at plot level.
Average tree defoliation was expressed as percentage of defoli-
ation of the third upper crown. Parasitism rates were given by the
average percentage of parasitized egg capsules per plot sampled. A
Generalized Linear Model, type III mean squares, was used to ana-
lyze differences between stands, on the following predictor vari-
ables: adult density, larva density and egg capsule density. A
Poisson distribution with log link function was used and the results
were presented in Wald Chi-square (W), degrees of freedom (df)
and p-value. Pearson’s correlations were estimated to assess the
relationship between population density, parasitism rate and site
variables. Lack of ﬁt tests were performed for the linear regression
between these variables and defoliation. Non-parametric Spear-
man correlation (r
) was instead used, when lack of ﬁt of a linear
model was observed. Pairwise t-test were applied to compare
dominant height (dh) and merchantable volume projected to
10-years of age (Vm
), between the two periods: 1995–1998, be-
fore establishment of the weevil, and 2004–2006, after the estab-
lishment of the weevil; tests were applied independently on both
types of stands, currently defoliated (D) and undamaged (U). For
current defoliated stands, the loss of merchantable volume as
dependent variable in function of defoliation was ﬁtted through
curve estimation. Linear, power, logarithmic and exponential func-
tions were tested, the function providing the best ﬁt being selected.
All analyses were performed with SPSS 18.0 software for Windows
(SPSS Inc., Chicago, IL, USA). The results were considered statisti-
cally signiﬁcant at p< 0.05. Data are presented as means ± standard
3.1. Weevil density and defoliation damage
Signiﬁcant differences between stands were found for the mean
number of egg capsules (W= 100.57; df = 28; p< 0.001), weevil lar-
val density (W= 133.68; df = 33; p< 0.001) and weevil adult den-
sity (W= 184.60; df = 33; p< 0.001).
Average defoliation of the upper third of the canopy in the spring
of 2007 was 55.5% ± 3.3, n= 64. The intensity of damage measured
in June increased with the density of adults, larvae and egg capsules
estimated in February (Fig. 2). Additionally, adult density in Febru-
ary gave the best prediction of spring season defoliation levels
= 0.77, F
= 204.25, p< 0.001). The results showed no lack of
ﬁt for the linear regression model for this variable (F
p= 0.591). A lower correlation between larvae and late season defo-
liation was observed although it was still signiﬁcant (Fig. 2). A
signiﬁcant lack of ﬁt for the linear regression of spring defoliation
in function of both larvae (F
= 4.405, p= 0.004) and egg masses
= 6.109, p< 0.001) was observed. Yet, Spearman correlations
with defoliation were signiﬁcant, r
= 0.694; r
= 0.751 p< 0.001,
for larvae and egg masses, respectively.
Defoliation level increased with elevation (r
= 0.48, F
57.81, p< 0.001) and precipitation (r
= 0.21, F
= 16.49, p
< 0.001). On the other hand, defoliation was negatively correlated
with the average temperature of the three coldest months
= 33, F
= 30.16, p< 0.001, and of the three warmest
months (MaxTs) r
= 0.30, F
= 26.75, p< 0.001. The average defo-
liation was 74.1 ± 7.7% for sites exhibiting MaxTw below 10 °C,
whereas it decreased to 33.5 ± 4.8% for sites where MaxTw sur-
passed 11.5 °C.
3.2. Parasitism rates by nitens
The number of wasps emerging per egg capsule ranged from 0
to 14, with a mean of 2.22 ± 3.21. Parasitism rates varied from 0
to 100%, with a mean of 44.1 ± 3.9%. Signiﬁcant negative correla-
tions of 0.61, 0.42 and 0.64 (p< 0.001) were found between
the rates of parasitism and the density of adult weevils, egg cap-
sules and defoliation levels, respectively. Also the parasitism rate
correlated negatively with elevation (Table 1). Above 700 m eleva-
tion, the mean percentage of parasitism was 14.7 ± 3.9%, whereas
below 400 m it was 76.6 ± 10.2%.
The analysis of the climatic variables showed that the maxi-
mum temperature of the three coldest months was the variable
that could better explained the differences in parasitism rates
observed among plots, whereas correlations with temperatures of
the three warmest months and total precipitation, were not signif-
icant (Table 1). Maximum temperatures of the three coldest
months (MaxTw) between 10 and 11 °C evidenced a threshold lim-
it (Fig. 3). Below 10 °C of MaxTw, parasitism rates were extremely
low and independent of the temperature (F
= 0.221, p= 0.653),
whereas above this threshold parasitism rates increased signiﬁ-
cantly with temperature (F
= 39.257, p< 0.001). For sites with
Mean number of Gonipterus platensis
= 0.56, p <0.001
= 0.48, p <0.001
= 0.76, p <0.001
Fig. 2. Average defoliation (%) of the upper third of the tree canopy of Eucalyptus
globulus, observed in June 2007, in relation to the average number of egg masses,
larvae and adult of G. platensis estimated in February 2007, on 64 plots located in
northern and central Portugal.
Pearson correlation, r, between the parasitism rate and the climatic variables:
maximum and minimum temperatures of the three coldest (MaxTw, MinTw) and
warmest (MaxTs, MinTs) months, and total rainfall (Rain) and elevation.
MaxTw MinTw MaxTs MinTs Rain Elevation
r0.741 0.554 0.345 0.171 0.173 0.760
p-value <0.001 <0.001 0.052 0.177 0.172 <0.001
A.R. Reis et al. / Forest Ecology and Management 270 (2012) 216–222 219
MaxTw below 10 °C, the mean parasitism rate was 10.1 ± 4.9%,
ranging from 0 to 45%, whereas at MaxTw above 11.5 °C it was
70.9 ± 3.8%, ranging from 50 to 100% (Fig. 3).
3.3. Wood production loss
A signiﬁcant decrease in both dominant height (dh) and mer-
chantable volume, projected to 10 years of age, was observed in
the stands currently attacked by the weevil (D) in comparison
with the values obtained for the same stands before establishment
of the weevil (1995–1998). Losses in dh and merchantable volume
in currently attacked plantations were estimated to be on average,
about 29% and 51%, respectively (Table 2). In contrast, wood pro-
ductivity did not decrease in the stands currently not attacked by
the weevil (U); in fact a slight increase in merchantable volume
and dh was observed in recent stands in comparison with historical
ones, although the differences were not signiﬁcant (Table 2).
When the stands under current attack by the weevil (D) are
considered, an exponential increase of wood loss as a function of
the intensity of defoliation provide the best ﬁt (R
p< 0.001) (Fig. 4). For 50 and 75% of the defoliation of the upper
third canopy, the predicted loss in merchantable volume was 21
and 42%, respectively. This production loss increased substantially
to 86% when defoliation was 100%, as frequently observed in
stands located above 700 m of elevation in the study area.
Knowledge of the impacts of forest pests, as well as the efﬁcacy
of pest management strategies is essential for forest managers.
Classical biological control using co-evolved natural enemies is
one of the most promising management tools for the control of
exotic pests. Nevertheless, besides other inherent risks (van
Lenteren et al., 2003), the efﬁciency of the biological control organ-
ism is not always satisfactory. Over 80% of the organisms released
for the biological control of arthropods do not result effective (Hall
et al., 1980). Understanding the reasons for such failures will
contribute to improve the efﬁcacy of biological control programs.
The parasitoid A. nitens is the main agent used worldwide to con-
trol the eucalyptus weevil (Hanks et al., 2000; Paine et al., 2000;
Sanches, 2000). Nevertheless, in regions of Galicia (Spain) and of
northern and central Portugal, the damage caused by the weevil
is above the level of economic acceptance, despite successive
releases of the parasitoid by public and private forest enterprises
(Cordero Rivera et al., 1999; Valente et al., 2004). Inefﬁcient rates
of parasitism by A. nitens were also recorded in some areas of South
Africa (Tribe, 2005) and southwestern Australia (Loch, 2008). Pos-
sible reasons for the reduced efﬁcacy of the parasitoid have been
suggested, such as microclimatic conditions and ﬂuctuations in the
population dynamics of both the host and parasitoid (Cordero
Rivera et al., 1999; Santolamazza-Carbone and Fernandez, 2004;
Loch, 2008). Since A. nitens is highly host speciﬁc, it has been sug-
gested that high parasitism rates may lead to a local extinction of
the host and consequently to the collapse of the A. nitens popula-
tion, originating a release of the weevil from the parasitoid in the
next generation (Cordero Rivera et al., 1999). In the present study,
the mean parasitism rates varied between 0 and 100%. High para-
sitism rates were observed only at sites with low elevation, where
G. platensis was under control. In effect, mean parasitism reached
100% in two sites, located at 430 and 290 m elevation, in which
mean defoliation was low, 6.7 and 20%, respectively. Therefore,
the hypothesis of host-parasitoid population disruption by high
parasitism rates was not observed in this study.
In the present study, the rate of parasitism by A. nitens and con-
sequent efﬁcient control of G. platensis, were found to vary along an
altitudinal range, decreasing steeply with elevation, as previously
observed by Valente et al. (2004). Analysis of the climatic variables
showed that the parasitism rate was most strongly correlated with
temperature of the warmest hours of the day, during the three
coldest months (r= 0.74), whereas no signiﬁcant relationship with
temperatures during the summer months, or with precipitation
was found. Maximum temperatures of the three coldest months,
between 10 and 11 °C, were found to set the lower threshold for
the efﬁciency of the parasitoid. The fact that parasitism ranges
from 0 to 45% in regions with maximum winter temperatures be-
low 10 °C, but from 50 to 100% under temperatures equal, or above,
12 °C(Fig. 3) indicates that A. nitens is ineffective at low tempera-
tures, below 10 °C, such as those prevailing during the winter and
early spring months. Precipitation does not seem to be a relevant
variable since it did not correlate signiﬁcantly with the rate of
parasitism. Other regions where the parasitoid fails to efﬁciently
control the eucalyptus weevil also experience climates with cold-
temperate winters, conﬁrming that low temperatures might be a
limiting factor for this host-parasitoid system (Cordero Rivera
R2= 0.549, p<0.001
7 8 9 1011121314
Parasitism rate %
Maximum temperature of the three coldest months ºC
Fig. 3. Relationship between parasitism rate and maximum temperature observed
during the three coldest months of the year (December, January and February),
calculated from WorldClim monthly datasets, and from 64 plots located in northern
and central Portugal.
Dominant height (dh) and merchantable volume projected to 10-years of age (Vm),
±SE, in two types of stands, currently defoliated (D) and undamaged (U) by G.
platensis, and two periods: 1995–98 before establishment of the beetle; 2004–06,
after establishment, in the northern and central regions of Portugal. For the same
stand category, different letters indicate signiﬁcant differences, t-test (p< 0.05).
Defoliated stands - D Undamaged stands - U
% 1995–98 2004–06
± 0.5 15.4
± 1.4 28.6 19.9
± 0.5 21.4
± 0.6 +7.5
± 7.6 69.8
± 10.5 50.6 114. 8
± 9.0 115.3
± 6.8 +0.4
y = 5.428 e0.027x
R² = 0.834
30 40 50 60 70 80 90 100
Percentage decrease in projected
wood volume (%)
Fig. 4. Relationship between the level of damage by G. platensis and the decrease in
the projected merchantable volume (%) at age 10 years in Eucalyptus globulus stands
in northern and central regions of Portugal.
220 A.R. Reis et al. / Forest Ecology and Management 270 (2012) 216–222
et al., 1999; Tribe, 2005). However, this system is also considered
to be ‘‘unnatural’’, as A. nitens originates from South Australia
(Tooke, 1953), a region with a warm to mild climate, whereas G.
platensis is native from Tasmania (Mapondera et al., 2012), a region
with a colder climate. Adaptation of parasitoid and host to different
climatic niches, in particular regarding low-temperature thresh-
olds, may thus explain the failure of A. nitens to achieve high par-
asitism rates in colder areas, or times of the year. In agreement
with this hypothesis, G. platensis in our ﬁeld observations started
oviposition in mid-January and February, when average maximum
temperatures are below 10 °C, thus precluding A. nitens from
becoming active. The geographical mismatch between G. platensis
and A. nitens therefore seems to arise from different climatic
preferences and temperature thresholds. On the contrary, in other
regions such as South Africa, no geographical/ evolutionary mis-
match between A. nitens and its Gonipterus host exists, since both
are native to South Australia (Rolf Oberprieler, com. pers.).
Average crown defoliation decreased with higher winter tem-
peratures and increased with elevation, to an average of 75% at
elevations above 700 m. This result, as well as the negative rela-
tionship found between parasitism and weevil population density,
are most probably indirect effects of the low efﬁciency of the par-
asitoid, under colder conditions. In contrast, the positive correla-
tion detected between defoliation and precipitation may be due
to a higher tree primary productivity (higher new-ﬂush biomass
available for the weevil) in the summer.
Determination of the impact of high levels of defoliation on
wood production is crucial but difﬁcult and frequently con-
founded by several factors, depending on the particular insect-
plant system. In the present study we used an empirical growth
model applied to historical and current inventory data, before
and after the establishment of G. platensis in the study area, in
1999. Results predict a high impact of the defoliation caused by
the weevil on the productivity of E. globulus plantations, as pro-
jected to 10 years of age, leading to an average loss of 51% in mer-
chantable volume in the affected stands. Wood volume losses
increased exponentially with tree defoliation (R
= 0.84), 43%
and 86% losses being projected for 75% and 100% defoliation,
respectively. The high wood loss observed may be due, in part,
to the continuous feeding of the weevils during the spring and
autumn, as well as to the continuous growth of the trees, and it
further suggests that fast-growing trees may be at higher risk of
decreased wood production due to severe defoliation. Our results
apparently contradicts previous ﬁndings that defoliation by
Gonipterus weevils has a limited impact on growth of E. globulus
(Loch and Matsuki, 2010). In this study however, defoliation esti-
mates of 18–33% were calculated for the growing tip of the trees
affected, whereas mean defoliation of the control group remained
at 5–16%. Nevertheless, even for this moderate defoliation, Loch
and Matsuki (2010) estimated a signiﬁcant impact of defoliation
on tree growth over a 2.5-year period. Also, Pinkard et al.
(2006) found that as little as a 20% defoliation of three-year-old
E. globulus, resulted in signiﬁcant reductions of stem growth,
within one year only after defoliation. In our study, defoliation
was far more severe, ranging from 75 to 100% in the most seri-
ously affected areas. A value of 50% defoliation resulted in a
decrease of wood loss production of up to 21%.
Despite such high values, wood loss is probably underesti-
mated. The growth model projected wood volume to a tree age
of 10 years, considering the dominant height and stand basal area
measured at a given age, but the model disregards the effect of
defoliation on the future growth of the trees. However, consecutive
reduction of leaf surface, essential for tree growth, will also impact
negatively during several post-defoliation years (Kulman, 1971).
Secondly, successive defoliation events are likely to occur in the
years following the inventories, accounting for further reduction
Field monitoring, to anticipate future plantation damage may
thus be most helpful to decision makers. In the present study we
found that the population of adult weevils at the end of winter
was signiﬁcantly linearly related to the percentage of defoliation
at the end of spring. Therefore, ﬁeld monitoring of adult weevils
may constitute a planning tool of pest management decisions, for
this particular insect-plant system.
As observed in other regions, efﬁcient parasitism by A. nitens
results in low damage by the weevils at low elevations with war-
mer climates. However, in Portugal, about 1/3 of the eucalyptus
plantations are located in central and northern regions above
450 m elevation, in which the estimated wood loss is severe. In
light of the high economic cost uncovered by the present work,
we advocate research on control strategies alternative to classical
biological control with A. nitens, for these regions. Given the evi-
dent mismatch between the parasitoid and host, the logical strat-
egy will be to look for natural (‘‘co-evolved’’) parasitoid species,
native from the same geographical region.
We thank AltriFlorestal for supporting this study through a col-
laboration protocol with ISA-UTL and Dr. Rolf Oberprieler (CSIRO
Ecosystem Sciences, Canberra, Australia) for comments on an ear-
lier version of this article. Additional funding support was provided
by the Portuguese Science Foundation (FCT) through project PTDC/
Appendix A. Supplementary data
Supplementary data associated with this article can be found, in
the online version, at doi:10.1016/j.foreco.2012.01.038.
AFN (Autoridade Florestal Nacional) 2010. Inventário Florestal. Lisboa, Direcção
Nacional de Gestão Florestal, Inventário Florestal Nacional – Apresentação de
Basurco, F., Toval, G. 2004. Estimated genetic parameters and breeding values for
resistance against Gonipterus scutellatus in 13-year-old Eucalyptus globulus in
northwest Spain. In: Borralho et al. (2004). Eucalyptus in a Changing World.
Proc of IUFRO Conf. Aveiro 11–15 Outubro 2004.
Bhattacharya, S.C., Salam, P.A., Pham, H.L., Ravindranath, N.H., 2003. Sustainable
biomass production for energy in selected Asian countries. Biomass and
Bioenergy 25, 471–482.
Campinhos Jr, E., 1999. Sustainable plantations of high-yield shape Eucalyptus trees
for production of ﬁber: the Aracruz case. New Forests 17, 129–143.
Cordero Rivera, A., Santolamazza-Carbone, S., Andrés, J.A., 1999. Life cycle and
biological control of the Eucalyptus snout beetle (Coleoptera, Curculionidae) by
Anaphes nitens (Hymenoptera, Mymaridae) in north-west Spain. Agricultural
and Forest Entomology 1, 103–109.
Echeverri-Molina, D., Santolamazza-Carbone, S., 2010. Toxicity of synthetic and
biological insecticides against adults of the Eucalyptus snout-beetle Gonipterus
scutellatus Gyllenhal (Coleoptera: Curculionidae). Journal of Pest Science 83 (3),
Govender, P., Wingﬁeld, M.J. 2005. Overview on the entomological research in the
Mediterranean forest ecosystems of South Africa. In: Entomological Research in
Mediterranean Forest Ecosystems. Lieutier, F.; Ghaioule, D., (Eds.), pp: 43–52.
Hall, R.W., Ehler, L.E., Bisbabri-Ershadi, B., 1980. Rate of success in classical
biological control of arthropods. Journal of the Entomological Society of
America 26, 111–114.
Hanks, L.M., Millar, J.G., Paine, T.D., Campbell, C.D., 2000. Classical biological control
of the Australian beetle Gonipterus scutellatus (Coleoptera: Curculionidae) in
California. Biology Control 29, 269–375.
Hijmans, R.J., Cameron, S.E., Parra, J.L., Jones, P.G., Jarvis, A., 2005. Very high
resolution interpolated climate surfaces for global land areas. International
Journal of Climatology 25, 1965–1978.
Kenis, M., Branco, M., 2010. Chapter 5: Impact of alien terrestrial arthropods in
Europe. In: Roques, A., et al. (Eds.), Alien terrestrial arthropods of Europe.
BioRisk 4(1), 51–71.
A.R. Reis et al. / Forest Ecology and Management 270 (2012) 216–222 221
Kulman, H.M., 1971. Effects of insect defoliation on growth and mortality of trees.
Annual Review of Entomology 16, 289–324.
Loch, A.D., 2008. Parasitism of the Eucalyptus weevil, Gonipterus scutellatus
Gyllenhal, by the egg parasitoid Anaphes nitens Girault in Eucalyptus globulus
plantations in southwestern Australia. Biological Control 47, 1–7.
Loch, A.D., Matsuki, M., 2010. Effects of defoliation by Eucalyptus weevil, Gonipterus
scutellatus, and chrysomelid beetles on growth of Eucalyptus globulus in
southwestern Australia. Forest Ecology and Management 260, 1324–1332.
Loch, A.D., Floyd, R.B., 2001. Insect pests of Tasmanian blue gum, Eucalyptus globulus
globulus, in south-western Australia: history, current perspectives and future
prospects. Austral Ecology 26, 458–466.
Mapondera, T.S., Burgess, T., Matsuki, M., Oberprieler, R.G., 2012. Identiﬁcation and
molecular phylogenetics of the cryptic species of theGonipterus scutellatus
complex (Coleoptera: Curculionidae: Gonipterini). Australian Journal of
Montoya, J.M., 1995. El eucalipto. Mundi-Prensa, Madrid, pp. 125.
Newete, S.W., Oberprieler, R.G., Byrne, M.J., 2011. The host range of the Eucalyptus
Weevil, Gonipterus ‘‘scutellatus’’ Gyllenhal (Coleoptera: Curculionidae), in South
Africa. Annals of Forest Science 68, 1005–1013.
Paine, T.D., Dahlsten, D.L., Millar, J.G., Hoddle, M.S., Hanks, L.M., 2000. UC scientists
apply IPM techniques to new eucalyptus pests. California Agriculture 54, 8–13.
Palma, C., Valente, C., 2008. Efeito do gorgulho do eucalipto, Gonipterus scutellatus,
no crescimento das plantas. In: Branco, M., Valente, C., Paiva, M.R. (Eds.), Pragas
e doenças em Pinhal e Eucaliptal – Desaﬁos para a sua gestão integrada.
ISAPress, Lisboa, pp. 87–103.
Pimentel, D., McNair, S., Janecka, J., Wightman, J., Simmonds, C., O’Connell, C.,
Wong, E., Russel, L., Zern, J., Aquino, T., Tsomondo, T. 2002. Economic and
environmental threats of alien plant, animal and microbe invasions. pp. 307–
330 In: Pimentel, D. (Ed.), Biological Invasions. Economic and Environmental
costs of Alien Plants, Animal and Microbe Species. Boca Raton, USA, CRC
Pinkard, E.A., Baillie, C., Patel, V., Mohammed, C.L., 2006. Effects of fertilising with
nitrogen and phosphorus on growth and crown condition of Eucalyptus globulus
Labill. experiencing insect defoliation. Forest Ecology and Management 231,
Salleo, S., Nardini, A., Raimondo, F., Lo Gullo, M.A., Pace, F., Giacomich, P., 2003.
Effects of defoliation caused by the leaf miner Cameraria ohridella on wood
production and efﬁciency in Aesculus hippocastanum growing in northeastern
Italy. Trees: Structure and Function 17, 367–375.
Sanches, M.A., 2000. Parasitismo de ovos de Gonipterus scutellatus Gyllenhal, 1833 e
Gonipterus gibberus Boisduval, 1835 (Coleoptera, Curculionidae) por Anaphes
nitens Girault, 1928 (Hymenoptera, Mymaridae) em Colombo (Pará, Brasil).
Revista Arquivos do Instituto Biológico 67, 77–82.
Santolamazza-Carbone, S., Fernandez, A.M.F.J., 2004. Testing of selected insecticides
to assess the viability of the integrated pest management of the Eucalyptus
snout-beetle Gonipterus scutellatus in north-west Spain. Journal of Applied
Entomology 128, 620–627.
Sedjo, R.A., 1999. The potential of high-yield plantation forestry for meeting timber
needs. New Forests 17, 339–359.
Stape, J.L., Binkley, D., Ryan, M.G., 2004. Eucalyptus production and the supply, use
and efﬁciency of use of water, light and nitrogen across a geographic gradient in
Brazil. Forest Ecology and Management 193, 17–31.
Tooke, F.G.C., 1953. The Eucalyptus snout beetle Gonipterus scutellatus Gyll – a study
of its ecology and control by biological means. Entomology Memoirs,
Department of Agriculture, Union of South Africa 3, 1–282.
Tribe, G.D., 2005. The present status of Anaphes nitens (Hymenoptera: Mymaridae),
an egg parasitoid of the Eucalyptus snout beetle Gonipterus scutellatus, in the
Western Cape Province of South Africa. Southern African Forestry Journal 203,
Valente, C., Vaz, A., Pina, J., Manta, A., Sequeira, A., 2004. Control strategy against the
Eucalyptus snout beetle, Gonipterus scutellatus Gyllenhal (Coleoptera,
Curculionidae), by the Portuguese Cellulose Industry. In: Borralho, N.M.,
Pereira, J.S., Marques, C., Coutinho, J., Madeira, M., Tomé, M. (Eds.), Eucalyptus
in a Changing World. Procedings of IUFRO Conference, Aveiro, pp. 622–627.
van Lenteren, J., Babendreier, D., Bigler, F., Hokkanen, H.M.T., Kuske, S., Loomans,
A.J.M., Menzler-Hokkanen, I., van Rijn, P.C.J., Thomas, M.B., Tommasini, M.G.,
Zeng, Q.-Q., 2003. Environmental risk assessment of exotic natural enemies
used in inundative biological control. BioControl 48, 3–38.
222 A.R. Reis et al. / Forest Ecology and Management 270 (2012) 216–222