Content uploaded by Costas Zachariades
Author content
All content in this area was uploaded by Costas Zachariades on Aug 27, 2024
Content may be subject to copyright.
Biological Control 197 (2024) 105591
Available online 3 August 2024
1049-9644/© 2024 The Author(s). Published by Elsevier Inc. This is an open access article under the CC BY-NC-ND license (http://creativecommons.org/licenses/by-
nc-nd/4.0/).
Risk assessment of the laboratory host range and a molecular
characterisation determining the eld host range of Lixus aemulus, for the
biological control of Chromolaena odorata in South Africa
Rosie Mangan
a
,
*
, Milly Gareeb
b
, Marcus Boeno
c
, Chirley Gonçalves da Silva
d
, Blair Cowie
e
,
Aristˆ
onio Magalh˜
aes Teles
f
, Marcos Silveira
d
, Costas Zachariades
b
,
g
a
Centre for Biological Control, Rhodes University, Makhanda, South Africa
b
Agricultural Research Council, Plant Health and Protection, PO Box 1055, 3245 Hilton, South Africa
c
Departamento de Engenharia Florestal da Universidade Regional de Blumenau, Blumenau, Santa Catarina, Brazil
d
Laborat´
orio de Botˆ
anica e Ecologia Vegetal, Centro de Ciˆ
encias Biol´
ogicas e da Natureza, Museu Universit´
ario, Universidade Federal do Acre, Rio Branco, Brazil
e
School of Animal, Plant and Environmental Sciences, University of the Witwatersrand, Johannesburg, South Africa
f
Departamento de Botˆ
anica, Instituto de Ciˆ
encias Biol´
ogicas, Universidade Federal de Goi´
as, Campus Samambaia, Av. Esperança, s/n, Vila Itatiaia, Goiˆ
ania, GO 74690-
900, Brazil
g
School of Life Sciences, University of KwaZulu-Natal, Private Bag X01, Scottsville 3201, South Africa
HIGHLIGHTS GRAPHICAL ABSTRACT
•Laboratory tests revealed that Lixus
aemulus is host-specic to Chromolaena
odorata.
•L. aemulus may utilise closely related
invasive plants but not indigenous,
ornamental or crop species in South
Africa.
•Native range exploration conrms
L. aemulus present on C. odorata.
•Genetic assessments show Lixus sp.(p.).
adults collected on various hosts,
limited to the Eupatorieae tribe, as
L. aemulus.
•Limited establishment of L. aemulus in
South Africa may be attributed to a cli-
matic mismatch with Brazilian collec-
tion sites.
ARTICLE INFO
Keywords:
Climatic mismatch
Curculionidae
Asteraceae
Genetic diversity analysis
Native range exploration
ABSTRACT
Chromolaena odorata (Asteraceae: Eupatorieae) is a sprawling shrub native to the Americas, and a destructive
invader of much of the humid tropics and subtropics of the Old World. Opportunistic native-range exploration in
1995 identied a stem-boring weevil, Lixus aemulus, as a promising biological control candidate agent. Host-
specicity testing was conducted on L. aemulus in South Africa using laboratory no-choice and paired-choice
tests. Three invasive alien plants closely related to C. odorata may be utilized by L. aemulus but no indige-
nous, ornamental or crop species in South Africa was or is expected to be attacked by the weevil. A native-range
eld survey was conducted in Brazil to determine the exact identity of the host plant L. aemulus had been
* Corresponding author at: School of Computer Science and Statistics, O’Reilly Institute, Trinity College Dublin, Dublin 2, Ireland.
E-mail address: manganro@tcd.ie (R. Mangan).
Contents lists available at ScienceDirect
Biological Control
journal homepage: www.elsevier.com/locate/ybcon
https://doi.org/10.1016/j.biocontrol.2024.105591
Received 30 January 2024; Received in revised form 20 July 2024; Accepted 30 July 2024
Biological Control 197 (2024) 105591
2
collected in 1995, and to identify additional host-plant species. Genetic assessments of the Lixus sp.(p.). adults
collected on the three host plants (C. odorata,Chromolaena laevigata and Heterocondylus vitalbae) reveal these
individuals are L. aemulus and the weevil can be classed as an oligophage in its native range. Over 5,500 adults
were released in South Africa, but overall establishment has been poor. The most likely explanation appears to be
a climate mismatch between the region of South Africa invaded by C. odorata and the collection locality in Rio
Branco, Acre state, Brazil. Additionally, because the full extent of the native range of L. aemulus is unknown, it is
uncertain whether individuals can be sourced from an area whose climate resembles that of South Africa.
Furthermore, despite being oligophagous, L. aemulus may perform sub-optimally on the southern African
C. odorata biotype.
1. Introduction
Chromolaena odorata (L.) R.M.King &H.Rob. is a perennial shrub
species belonging to the tribe Eupatorieae of the Asteraceae family (King
and Robinson, 1987). The shrub is widely distributed in tropical
America and has since been introduced to Asia and Africa (Cock and
Holloway, 1982; Muniappan et al., 2005).This species harms biodi-
versity, ecotourism, agriculture and forestry, which justies its consid-
eration as one the worst invasive plant species in the subtropical regions
of South Africa (Macdonald and Jarman, 1985; Zachariades et al.,
2013). Classical biological control (CBC) of C. odorata in South Africa
was initiated in 1988 and surveys have been undertaken in numerous
Central, South and North American countries (Strathie and Zachariades,
2002). In 1997, a strategic plan was drawn up, advocating the release of
a suite of damaging, host-specic insects attacking different parts of the
plant, namely the leaves, stems and roots (Kluge et al., 1997). A stem
borer is considered a highly desirable element in the suite for C. odorata
because the weed has photosynthetically active stems that can produce
new leaf material after being defoliated (Zachariades et al., 2002). A
small culture of the stem-boring weevil, Lixus aemulus Petri, 1928
(Coleoptera: Curculionidae) was collected opportunistically in 1995
from a pubescent ‘hairy chromolaena’in Rio Branco, Acre state, Brazil
by Dr S. Neser (Kluge and Zachariades, 2006). The weevil was imported
into quarantine in South Africa in late 1995 under the Department of
Agriculture: Directorate of Plant Health and Quality Permit 14/2/2/1
(9/94/124). Lixus aemulus bred easily on the southern African biotype of
the plant (Zachariades et al., 1998; Zachariades et al., 2011), which
originates from the northern Caribbean region and differs substantially
from C. odorata plants from South America (Paterson and Zachariades,
2013; Shao et al., 2018).
Identifying an agent’s potential during the early stages of screening is
paramount to the success of a biocontrol programme. Failure to do so
can be costly in terms of resources and time, and contributes to the
perception that biocontrol is a hit-or-miss strategy with risks of
ecological side effects (Hoelmer and Kirk, 2005; McClay and Balciunas,
2005). Curculionids maintain a good record as biocontrol agents on
weeds, with two other Lixus species having been successfully used
against invasive alien plants in Australia: L. cardui Olivier, 1807 on
Onopordum spp. (Asteraceae), and L. cribricollis Boheman, 1835 on Emex
australis Steinh. (Polygonaceae) (Julien and Grifths, 1999). Other Lixus
species have been investigated as biocontrol agents (Schmidl, 1981;
Sobhian et al., 2003). Under laboratory conditions, L. aemulus larvae
caused signicant damage to the stems of C. odorata, resulting in a 94 %
reduction in seed output (Kluge and Zachariades, 2006).
Host-specicity testing is another crucial step in the process of
introducing natural enemies for classical biological control efforts,
providing the basic information upon which the safety of a proposed
biocontrol agent can be assessed (Heard, 2002; McClay and Balciunas,
2005). Any risks L. aemulus may pose must be identied and fully
assessed prior to release. Characteristically, the host-range experiments
examine aspects of host selection and how the agent utilizes the target
plant and non-target plants, as indicated by oviposition, adult feeding,
larval feeding, larval development, adult longevity and fecundity (Heard
and Van Klinken, 1998). The range of non-target plants attacked is often
broader in laboratory tests than in the eld (Louda et al., 2003; Mayhew,
1998; van Klinken, 2000). Laboratory host-range trials presented in this
paper indicate that L. aemulus is quite specic but is not monophagous,
as is true of most Lixus species (Volovnik, 2024). Additionally, the exact
identity of the host plant from which L. aemulus had been collected was
uncertain (Kluge and Zachariades, 2006).
The identication of host-specic species with potential as weed
biological control agents relies on studying the host range of insect
herbivores from the native range of the target weed. This encompasses
the diversity of host plants these insects utilize for feeding and repro-
duction within their native geographical location (Goolsby et al., 2006;
McFadyen, 1998; Sheppard et al., 2003). Native range exploration is
considered a technically difcult and time-consuming element of bio-
logical control programmes but is the foundation upon which the
research ultimately depends (Goolsby et al., 2006). Given that the
identity of the original host plant remained uncertain and laboratory
trials may have overestimated the realized host range, eld host-range
surveys conducted in 2015 in the Rio Branco area were desirable.
These surveys aimed to accurately determine the host plant’s identity
and assess whether other species within the Asteraceae are utilized by
the herbivore, thus offering a more realistic estimate of the herbivore’s
host range under natural conditions (Sheppard et al., 2006).
This work examines the laboratory host specicity of L. aemulus with
no-choice and paired-choice tests, to determine if and which non-target
species could potentially sustain L. aemulus populations in the eld in
South Africa. We used a DNA barcoding approach to determine whether
other Asteraceae plant species share L. aemulus, focusing on 14 sites
sampled in and around Rio Branco, Acre state, Brazil. Mitochondrial
DNA (mtDNA) markers were used to establish if individuals collected on
the three host plants; C. odorata,Chromolaena laevigata (Lam.) R.M.King
&H.Rob. and Heterocondylus vitalbae (DC.) R.M.King &H.Rob., together
with representatives of the South African laboratory culture, are
conspecic. In addition, the notion of a climatic mismatch was explored
by performing a climatic comparison between the native collection site
of L. aemulus in Brazil and Sub-Saharan Africa, particularly the release
sites of the weevil in South Africa.
2. Materials &methods
2.1. Biology and taxonomy of Lixus aemulus Petri, 1928
Following its collection and importation into South Africa, the beetle
was identied by Dr C.W. O’Brien of Florida A&M University, Talla-
hassee in 1997, as “Lixus aemulus Petri (Coleoptera: Curculionidae) (or
near)”(Zachariades, 2004). Lixus aemulus was originally described from
a single female collected in an unspecied region of Brazil (Petri, 1928).
A similar species was collected on C. odorata in Venezuela in 1999
(Zachariades et al., 2011). It lacked the pink cuticular wax that is present
on the head of L. aemulus (Fig. 1). The weevil, collected at three locations
across a 180 km distance in north-western Venezuela, was identied by
C.W. O’Brien in 2020 as Lixus sp. (all belonging to a single species),
distinct from L. aemulus. This illustrates the potential diversity of Lixus
species on C. odorata across the South American continent.
Life-history metrics for L. aemulus were recorded in the quarantine
R. Mangan et al.
Biological Control 197 (2024) 105591
3
glasshouse and laboratory facility of the Agricultural Research Council,
Plant Health and Protection (ARC-PHP), Cedara, KwaZulu-Natal prov-
ince, South Africa, at 23–31 ◦C and relative humidity of 45–80 %. In the
laboratory, L. aemulus adults feed on young C. odorata leaves, causing
minor damage (Zachariades et al., 2002). After a preoviposition period
of about one month, the female inserts single eggs into green (non-
woody) C. odorata stems containing pith (Fig. 1). The egg hatches after
about one week and the larva feeds inside the green stem before pu-
pating inside a chamber within the stem. Adult progeny cut a circular or
oval hole through the stem wall before emerging. In the laboratory,
adults emerge en masse in the spring. Adults typically live for over six
months (n =21 for which this was recorded), and the period from
oviposition to adult emergence from the stem is 3–4 months, although
where oviposition occurs later in the summer, overwintering occurs
within the stem over winter. In the laboratory, therefore, there are 1–2
generations per year.
2.2. Selection of test plants
The test plant list for L. aemulus (Supplementary Table S1) was
compiled using (i) standard centrifugal testing principles (Wapshere,
1974), (ii) host records for other Lixus species (Zachariades et al., 1999)
(iii) stem morphology and (iv) taxa on which other species of Lixus have
been recorded as agricultural pests. A test list was initially published by
Zachariades et al. (2002). The asteraceous tribe Eupatorieae, and closely
related tribes such as Heliantheae, which form a composite termed “the
Heliantheae Alliance”(Baldwin, 2009), are poorly represented in the
paleotropics. Furthermore, there are no native South African genera in
the same subtribe (Praxelinae) as C. odorata. There are only three native
genera of Eupatorieae in South Africa, viz. Adenostemma (subtribe
Adenostemmatinae –two species), Mikania (Mikaniinae –two species)
and Stomatanthes (subtribe Eupatoriinae –one species) (Retief, 2002).
Only one species of Adenostemma was included in the host-range list
because A. viscosum J.R.Forst. &G.Forst. has stems that are morpho-
logically similar to those of A. caffrum DC. –both are succulent with no
pith and high water content, such that the L. aemulus larvae cannot
survive. Mikania natalensis DC. was not tested as it was not found in the
eld; it is very similar to M. capensis DC, and may not be a separate
species (W.C. Holmes, Baylor University, 2013, pers. comm. to C.
Zachariades). The herbaceous perennial Stomatanthes africanus (Oliv. &
Hiern) R.M.King &H.Rob., the only native species to have been a former
congener of C. odorata (both Eupatorium), grows in high-altitude grass-
lands, and has extremely thin stems. Several less closely related Aster-
aceae were excluded if their stem morphology was judged unsuitable for
larval survival of L. aemulus (e.g. no pith, high water content, thin and
herbaceous). Conversely, several non-asteraceous crop plants (sweet
potato, Ipomoea batatas L. (Convolvulaceae), tomato, Solanum lyco-
persicum L. (Solanaceae) and squash and pumpkin, Cucurbita pepo L.
(Cucurbitaceae)) were included because they were judged to have
physically suitable stems for L. aemulus development. Finally, Lixus
brevirostris Boheman, L. incanescens Boheman, L. juncii Boheman and
L. scabricollis Boheman are known pests of beet (Beta vulgaris L.
(Amaranthaceae)); Hypolixus haerans (Boheman) and H. truncatulus (F.),
both formerly in Lixus, of amaranth (Amaranthus spp. (Amaranthaceae));
and L. algirus L. of faba bean (Vicia faba L. (Leguminosae)). Therefore
these plants, or plants closely related to them, were used in host-range
trials.
Fig. 1. A: Adult L. aemulus; B: L. aemulus oviposition holes in the stem of C. odorata. The female drills an inspection hole with her rostrum and if the stem is suitable
inserts the egg into the stem and plugs the hole; C: Stem with L. aemulus larval damage and exit hole chewed by adult progeny; D: Adult L. aemulus feeding damage.
R. Mangan et al.
Biological Control 197 (2024) 105591
4
2.3. No-choice adult trials
All laboratory culturing and experimental studies were conducted in
the quarantine glasshouse and laboratory facility of the ARC-PHP,
Cedara, at 23–31 ◦C and relative humidity of 45–80 %. Host-range tri-
als for L. aemulus were conducted by comparing adult mortality, feeding
damage, oviposition, and larval development on C. odorata to that of 29
non-target plant species (Supplementary Table S1). Four mating adult
pairs between 1 and 2 months old (peak egg-laying age for females) were
placed into a cage (30 ×40 ×70 cm) containing a single plant. Males
were distinguished from females by their shorter, thicker rostrum.
Adults were provided with sprayed water once a day. Plants were
replaced every 10 days, and the trial was terminated after 30 days. Each
time plants were replaced, a record was made of the number of live
adults remaining, but dead and missing adults were not replaced. Once
plants had been removed from the cage, they were examined for adult
feeding and oviposition. Leaves on which adults had fed were removed
from the plant and categorised according to the percentage of leaf
feeding These feeding percentages were converted to a total leaf area
removed by rst obtaining an average area (n =50) for a leaf of each
plant species, and then by multiplying the total percentage leaf area
eaten on each plant by the average species leaf area. Oviposition holes in
the stem were marked with paper tags and judged to be “plugged”i.e.
containing an egg, or “unplugged”i.e. exploratory boring only. Plants on
which oviposition holes were present were dissected after a period
equivalent to the development time on C. odorata. The numbers of live
and dead larvae, pupae and adults, and the number of eggs and adult
emergence holes, were recorded from dissected plants. Over the esti-
mated larval development period, some experimental plants died. These
were dissected, and total larval counts were made. Trials were arranged
in a series of “runs”containing four to nine non-target species and one
C. odorata plant to control for temporal variation in L. aemulus behav-
iour. All C. odorata were inoculated with L. aemulus adults to provide a
positive control. Replicates of a given non-target species were always
assigned to different runs.
2.4. Adult paired-choice trials
Paired-choice trials were conducted for the non-target species that
supported adult development from progeny production in no-choice
adult trials. One C. odorata plant and one non-target plant were placed
in a cage (50 ×50 ×90 cm). Four pairs of L. aemulus adults were
introduced onto a platform between the plants. On days 1, 2, 3, 5 and 10
after introduction, the position of the adults (on control plants, test
plants or elsewhere) was recorded. The plants were again replaced every
10 days, and after 30 days the trial was terminated. The same procedure
was followed, and the same variables were recorded as for the no-choice
trials. In addition, each time plants were replaced, their position in the
cages was altered by 90◦to control for directional movement of adults,
adults were placed back onto the platform between the plants, and adult
positions were again recorded on the same days after introduction as
above.
2.5. Risk assessment
For those non-target plant species that demonstrated susceptibility to
L. aemulus, the potential risks (feeding and reproduction) to non-target
species quantied (Wan and Harris, 1997) by measuring adult feeding
and survival, together with the production of progeny, on each non-
target test plant species, as a proportion of that on C. odorata. The
following performance procedures were used to assess the risk:
1) Feeding risk percentage was calculated as a product of plant pref-
erence (relative feeding damage) of the adults during paired-choice
tests, multiplied by feeding damage by the adults (relative dam-
age) during no-choice tests (Mangan and Baars, 2023).
Feeding Risk =Plant preference ×Feeding damage
where:
•Plant preference is the attractiveness of test plant species, deter-
mined by adult preference (relative feeding) in paired-choice trials.
Relative feeding =Feeding on non −target plant
Feeding on C.odorata
•Feeding damage reects the relative damage determined using the
mean percentage of damage per test species in proportion to that on
C. odorata (relative feeding damage) in no-choice trials.
Relative feeding damage =Feeding damage on non −target plant
Feeding damage on C.odorata
2) Reproductive risk was calculated as survival of adults (relative sur-
vival) in no-choice tests, multiplied by the host suitability of test
plants, and determined by the production of F
1
progeny (relative
numbers) in no-choice tests (Mangan and Baars, 2023).
Reproductive Risk ¼Adult survival £Progeny production
where:
•Adult survival is the relative suitability determined using the mean
survival on the test plants in proportion to that on C. odorata as
observed in no-choice trials.
Relative survival =Survival rate on non −target plant
Survival rate on C.odorata
•Progeny Production reects the number of F
1
progeny produced on
each test plant relative to those on C. odorata, as measured in no-
choice trials.
Relative progeny =Progeny on non −target plant
Progeny on C.odorata
2.6. Host-specicity statistical analysis
All statistical analyses were conducted in the R environment version
3.5.2 (R_Core_Team, 2021). A binomial generalised linear mixed model
(GLMM) with a logit link function was employed to examine the effects
of different test plants and experimental runs on survival outcomes,
incorporating a random intercept for each experimental run (1 | Run).. A
Gaussian linear mixed model (LMM) with an identity link function was
employed to examine the effects of different test plants and experi-
mental runs on adult feeding, also including a random intercept for
experimental variability (1 | Run). A generalised linear mixed model
(GLMM) using a negative binomial distribution was employed, with
glmmTMB function from the glmmTMB package in R (Brooks et al.,
2017), to analyse the effects of different test plants and controlled for
variability across experimental runs on total and plugged oviposition
holes (1 | Run). A Poisson GLMM with a log link function was employed
to examine the effects of different test plants and incorporated random
effects to account for variability between different experimental runs on
progeny production, specically including a random intercept for each
run (1 | Run). To facilitate direct comparisons with C. odorata, C.
odorata was designated as the reference level for the Test plant variable.
Fixed effects parameter signicance was assessed using a Likelihood
Ratio (LR) test (p <0.05) using the ’car’R package (Fox and Weisberg,
2019). To facilitate the interpretation of the model coefcients, the raw
beta coefcients were transformed into odds ratios for the binomial
models and incidence rate ratios for the Poisson and Negative Binomial
models. This transformation was performed by exponentiating the raw
R. Mangan et al.
Biological Control 197 (2024) 105591
5
coefcients (exp(coef(model)) in R).
To determine statistical differences during adult paired-choice tests,
a Gaussian (GLM) with robust standard errors was employed to analyse
the continuous variables ’position’and ’feeding’. The robust standard
errors were calculated using the HC3 estimator to adjust for potential
heteroscedasticity and provide reliable inference under the presence of
unequal variances across groups. Count data for oviposition holes and
progeny were analysed using Poisson regression models to address the
distributional properties of count data. Initial assessments of over-
dispersion were conducted by comparing the Pearson chi-squared sta-
tistic to the degrees of freedom of the model residuals. If signicant
overdispersion was indicated (Chi-squared/df >1), a negative Binomial
model was used to better accommodate the extra-Poisson variation.
After tting the model, an Analysis of Variance (ANOVA) using Type II
sums of squares with a Likelihood Ratio (LR) test was conducted to
determine the signicance of differences among test plants using the
‘car’package (Fox and Weisberg, 2019).
2.7. Native-range eld plant inspection
Chromolaena odorata plants were inspected for the presence of adult
weevils identical in appearance to those of the L. aemulus culture in
South Africa. When the weevil or suspected adult weevil damage was
present, all Asteraceae in the vicinity were inspected for the presence of
adults of the same appearance as those on C. odorata. If possible, 10
plants of each species were examined per site. Plant height and the
number of shoot tips (on which Lixus adults may be present) were
estimated and grouped into one of six categories for each plant: 1 =1–5;
2=6–10; 3 =10–20; 4 =20–50; 5 =50–100; 6 >100.
2.8. Insect collection
Adult weevils, similar in appearance to L. aemulus, were sampled
from the Rio Branco area of Acre state, Brazil during eld surveys
(Supplementary Table S2). Adult weevils were collected at 3 sites where
leaf damage was evident. Host plants included the target plant C. odorata
as well as a species in the same genus, C. laevigata, and H. vitalbae on
which characteristic L. aemulus adult weevil feeding damage was also
evident.Additionally, individuals from the L. aemulus laboratory culture
(originating in Rio Branco, Acre State) released as a biological control
agent against C. odorata in South Africa in 2011, were collected. Addi-
tional sequences from GenBank were included in the analysis as out-
groups: Lixus liformis (Fabricius, 1781) (Coleoptera: Curculionidae)
(host plant: Picris sp. ((Asteraceae)) and Lixus angustatus (Fabricius)
(Coleoptera: Curculionidae) (host plant: Carduus acanthoides L.
(Asteraceae)).
2.9. mtDNA extraction
All specimens were stored individually in 92 % ethanol. DNA was
extracted from 14 adult Lixus (including ve conrmed L. aemulus) using
the Qiagen DNeasy Blood and Tissue kit (Qiagen GmbH, Hilden, Ger-
many) following the manufacturer’s protocol. DNA quality and con-
centration were measured using a NanoDrop®ND-1000
Spectrophotometer (Labtech Int., UK). These extractions were then
stored at −20 ◦C until required.
2.10. mtDNA sequencing
The cytochrome oxidase I (COI) region of the mitochondrial genome
(mtDNA) was amplied for all samples. PCR reactions were carried out
in 20
μ
l volumes under the following conditions: 4
μ
l of genomic DNA,
8.48
μ
l of ddH20, 1.8
μ
l of MgCl2 (25 mM), 0.6
μ
l of each primer (10
μ
M;
Mtd6-curculio (5
′
GGRGGWTTTGGAAAYTGAYTARTTCC 3
′
) and Mtd9-
curculio (5
′
CCNGGDARAATTAAAATRTMWACTTC 3
′
); (Simon et al.,
1994), 0.4
μ
l dNTPs (10 mM each), 4.0
μ
l 10X buffer (Promega) and
0.12
μ
l of Taq polymerase. Thermal proles started with an initial
denaturing @ 94 ◦C for 60 s, followed by 32 cycles of 94 ◦C for 60 s, 50
◦C for 45 s and 72 ◦C for 60 s. The cycle ended with one nal extension of
240 s at 72 ◦C. The PCR products were puried, and sequencing re-
actions were conducted at Stellenbosch University DNA Sequencing
Unit. Sequences were deposited in GenBank (Accessions PP054253 to
PP054264).
2.11. mtDNA sequence analyses
Chromatogram contigs were assembled using CodonCode™Aligner-
Software (CodonCode Corp., Dedham, MA) and sequence alignments
were proofread manually and aligned using Se-Al 2.0 (Rambaut, 2001).
Phylogenetic analyses were conducted using MEGA 11 including the
neighbour joining and maximum likelihood methods. The Tamura-Nei
model and uniform rates of evolution assumption were employed for
both methods (Tamura et al., 2011). Molecular Phylogenetic analysis by
Maximum Likelihood tree using the GTR+G model of substitution of 3
control region haplotypes (411 bp). The evolutionary history was
inferred by using the Maximum Likelihood method based on the
Tamura-Nei model (Tamura and Nei, 1993). The tree with the highest
log likelihood (−2015.78) is shown. The percentage of trees in which the
associated taxa clustered together is shown next to the branches. Initial
tree(s) for the heuristic search were obtained by applying the
Neighbour-Joining method to a matrix of pairwise distances estimated
using the Maximum Composite Likelihood (MCL) approach.
The analysis involved 14 nucleotide sequences. Codon positions
included were 1st +2nd +3rd +Noncoding. There were a total of 411
positions in the nal dataset. Evolutionary analyses were conducted in
MEGA7 (Kumar et al., 2016).
The haplotype diversity (h) mean pairwise differences (MPD) and
nucleotide diversity (
π
) for each population were estimated using
DnaSp. Ver. 5 (Librado and Rozas, 2009). Population structure was
analysed using the Analysis of Molecular Variance (AMOVA) (Excofer
et al., 1992) and by calculating the FST values (Hudson et al., 1992)
between populations, using the Kimura two-parameter distance method
(Kimura, 1980). The statistical signicance was determined by per-
forming 1000 permutations of the original data set using Arlequin 3.0
(Excofer et al., 2005).
2.12. Climatic suitability assessment (‘climate matching’)
The climate prediction-modelling program CLIMEX (Dymex Simu-
lator: version 4.0.2) was used to compare the climatic similarity of the
Brazilian collection sites of L. aemulus, namely Rio Branco, Acre State,
Brazil, to Sub-Saharan Africa, more specically the release sites of the
weevil within South Africa. CLIMEX comparison between the Brazilian
collection locality and South African release sites of L. aemulus was
undertaken using the ‘match climates’feature. The climatic parameters
incorporated included mean annual rainfall, rainfall seasonality, mini-
mum, maximum, and average temperature, relative humidity as well as
soil moisture, all of which were equally weighted (=1) during the
matching procedure. From these climatic matches, composite match
indices were generated between the localities compared, typically
collection vs release sites, with 0 % indicating no match and 100 %
indicating a perfect match (Kriticos et al., 2015). In the case of biological
control programmes, these climatic matches broadly offer insight to-
ward the likelihood of agent establishment within the introduced range;
with values of <50 % deemed as ‘unsuitable’, 50–59 % deemed as ‘low’,
60–69 % deemed as ‘moderate’, 70–79 % deemed as ‘high’and ≥80 %
typically being optimal (Cowie et al., 2023).
R. Mangan et al.
Biological Control 197 (2024) 105591
6
3. Results
3.1. No-choice adult survival, feeding, and oviposition
The analysis revealed signicant differences in survival outcomes
across the test plants (
χ
2
=565.24, df =31, p <0.0001) and between
experimental runs (
χ
2
=7.02, df =1, p =0.008). These results indicate
that the type of test plant signicantly inuenced survival, suggesting
specic interactions between L. aemulus and certain plant species within
the trial. The variation between runs also suggests possible effects of
experimental conditions or temporal factors on mortality rates. C.
odorata, Ageratina riparia (Regel) R.M. King &H. Rob. (Asteraceae) and
Bidens pilosa L. (Asteraceae) supported the highest survival percentages
among the test plants (Table 1). In the binomial logistic regression
analysis, C. odorata was used as the reference plant against which the
survival rates on other plants were compared. The results demonstrate a
signicantly higher probability of insect survival on C. odorata, with an
odds ratio of 116.437. Conversely, the odds of survival on the alien
ornamental species Argyranthemum frutescens (L.) Sch.Bip. and the
invasive alien Campuloclinium macrocephalum (Less.) DC. (both Aster-
aceae) are substantially lower, with odds ratios of 0.006 and 0.009,
respectively, both at a signicance level of p <0.001 (Supplementary
Table S3). This indicates that the likelihood of insect survival is drasti-
cally reduced on these plants, being about 166.67 times lower on
A. frutescens and about 111.11 times lower on C. macrocephalum
compared to C. odorata. Analysis revealed signicant differences in
adult feeding across the test plants (
χ
2
=300.256, df =31, p <
0.000001). Almost all test plants exhibit signicant differences in
feeding behaviour relative to C. odorata, as indicated by the p-values
mostly being less than 0.001. Adult weevils removed similar areas of leaf
material on C. odorata and the invasive alien plant Ageratina adenophora
(Spreng.) R.M.King &H.Rob. (Asteraceae). Argyranthemum frutescens
and A. hybridus show large negative coefcients of −2.033 and −2.103
respectively, suggesting substantially reduced feeding, with signicant
p-values (p <0.001) (Supplementary Table S4). Total oviposition holes
were highest in A. adenophora and Xanthium strumarium L. (Asteraceae)
(Table 1).The analysis revealed no signicant differences in the number
of oviposition holes across the test plants (
χ
2
=35.049, df =31, p =
0.2818). The vast majority of plant species show extremely negative
coefcients (e.g., A. frutescens, A. riparia), suggesting a signicantly
lower number of oviposition holes compared to the host plant. However,
the corresponding p-values are generally very high (p =0.999), indi-
cating that these differences are not statistically signicant (Supple-
mentary Table S5). The experimental run had no signicant effect on the
number of oviposition holes (
χ
2
=0.000, df =1, p =0.9830). Chro-
molaena odorata obtained twice as many plugged holes (i.e., containing
eggs) as the next two highest species, A. conyzoides and B. pilosa. About
70 % of holes bored by L. aemulus on C. odorata were plugged. In
contrast, the number of probes on weed species such as X. strumarium
was high, but the proportion of those chosen for egg-laying (i.e., plug-
ged) was low (9.03 %). There was a strong dependency of progeny
production on the type of test plant (
χ
2
=1617.1, df =31, p <0.0001).
Chromolaena odorata supported twice as much progeny production as
non-target species. The host plant showed a positive and signicant
impact on progeny production, with an estimate of 32.385 and a sig-
nicant p-value (<0.001) (Supplementary Table S6). Of the six non-
target test species in which progeny were found, adults were obtained
from all except weed A. frutescens.
3.2. Paired-choice adult position in cage, feeding, oviposition, and
progeny production
In paired-choice trials, L. aemulus adults consistently favoured
C. odorata over non-target species. During observations, L. aemulus
adults were recorded 65–88 % of the time on C. odorata plants, 6–25 %
on the ve non-target species and 6–11 % elsewhere in the cage
(Table 2). Adults spent the least amount of time on C. macrocephalum,
B. pilosa and Senecio madagascariensis Poir. (Asteraceae). Lixus aemulus
adults displayed a signicant avoidance of S. madagascariensis for both
position (
χ
2
=165.08, df =1, p <0.0001) and feeding behaviours (
χ
2
=
35.994, df =1, p <0.0001) (Supplementary Table S7). All other paired-
choice trials exhibited similar position and feeding behaviour trends,
with signicant differences in feeding amounts between the target and
non-target test plant species.
Females bored more exploratory holes in A. adenophora than in
C. odorata, but the number of these used for oviposition, i.e., plugged,
was lower in A. adenophora, although not signicantly so (
χ
2
=0.32335,
df =1, p =0.5696) (Table 3,Supplementary Table S7). For the four
other test species, there were lower numbers of both exploratory and
plugged holes than the corresponding control plants. Of the non-target
species, A. conyzoides had the highest number of plugged oviposition
holes, at 50 % of the number on C. odorata.Senecio madagascariensis and
B. pilosa both had less than 10 % of the number of plugged holes
recorded on C. odorata. Total and plugged oviposition activities did not
signicantly differ between the target and non-target plant species in
any of the paired choice trials (Supplementary Table S7). Ageratum
conyzoides produced 77 % of the progeny (larvae, pupae, adults) that the
C. odorata control produced. The likelihood ratio tests revealed progeny
production did not differ signicantly (
χ
2
=0.11917, df =1, p =
0.7299). The other four non-target species produced less than 40 % of
the number produced by their respective C. odorata controls. Senecio
madagascariensis and B. pilosa, the only two species outside the Eupa-
torieae, contained the lowest numbers of progeny, with
χ
2
=15.545 and
6.8231 (df =1), which were both statistically signicant (p <0.001).
3.3. Risk assessment
The risk of feeding by L. aemulus on non-target plants is highest on
A. adenophora and A. conyzoides, where the chance of sustaining damage
relative to that of C. odorata was 29.05 and 6.44 % respectively when
C. odorata is available in close proximity (Table 4). All other plants
presented a less than 2 % chance of sustaining damage if growing close
to C. odorata. Ageratum conyzoides, A. adenophora and B. pilosa present a
reproductive risk of 46.08, 36.63, and 34 % respectively. In the absence
of C. odorata, these plants are nutritionally suitable to produce progeny.
3.4. Native range eld plant inspection
Four sites were inspected (RIOB 03, 06, 07, 08), of which Lixus sp(p).
were present at three (not found at RIOB 07). Site size varied from 720-
3000 m
2
. All plants sampled were in the correct life stage and of an
adequate size to realistically be attacked/utilised by the weevils. Lixus
adults were found on three species of Asteraceae in Acre state, viz.
C. odorata,C. laevigata and H. vitalbae (Table 5).
3.5. Sequence data, haplotypes and genetic diversity
After excluding the ambiguously called base pairs at the beginning
and ends of each sequence, a 411-bp portion of the COI mtDNA se-
quences was obtained for 14 Lixus collected on C. odorata (n =5),
C. odorata: South African biotype (n =5), C. laevigata (n =3), and
H. vitalbae (n =1). A total of three different haplotypes were identied
(Supplementary Table S8). Lixus sp. collected on H. vitalbae was the only
population to contain a single haplotype. Haplotype diversity ranged
from 0.400 (CO and COSA) to 0.667 (CL), with an average of 0.385 for
all populations (Supplementary Table S9). Nucleotide diversity (
π
)
ranged from 0.00105 (CO and COSA) to 0.00175 (CL), with an average
of 0.00106.
3.6. Genetic relationships between populations
Three distinct clades were resolved by the mitochondrial CO1
R. Mangan et al.
Biological Control 197 (2024) 105591
7
Table 1
Adult no-choice feeding damage, survival, oviposition, and progeny production of the weevil Lixus aemulus exposed to plant species in no-choice conditions. See Table
S1 for taxonomic details of plant species.
Test plant species n Adult
feeding
a
Relative
damage
b
%
Survival
c
Relative
survival
d
Total
oviposition
holes
e
Plugged
oviposition
holes
e
%
plugged
Progeny Relative progeny
production
f
Chromolaena odorata
w
13 2.12 ±
0.77
1.0 93.38 ±
4.03
1.0 0.60 ±0.30 0.42 ±0.25 70.35 32.38 ±
19.24
1.0
Adenostemma caffrum
i
3 0.79 ±
0.40
0.37 82.22 ±
9.35
0.88 0.30 ±0.30 0.13 ±0.02 8.41 0 0
Ageratina adenophora
w
3 1.75 ±
0.55
0.83 92.68 ±
6.55
0.99 0.62 ±0.45 0.13 ±0.02 20.98 12.0 ±
16.52
0.37
Ageratina riparia
w
3 0.43 ±
0.21
0.20 100.0 ±
0.00
1.0 0.06 ±0.04 0.003 ±0.01 4.79 0 0
Ageratum conyzoides
w
3 0.98 ±
0.31
0.46 89.4 ±
9.36
0.96 0.52 ±0.09 0.23 ±0.07 43.40 15.67 ±
13.05
0.48
Campuloclinium
macrocephalum
w
3 0.57 ±
0.18
0.26 47.8 ±
41.40
0.51 0.10 ±0.09 0.04 ±0.04 41.32 3.67 ±
5.51
0.11
Mikania capensis
i
3 0.15 ±
0.90
0.07 38.5 ±
36.50
0.41 0.00 ±0.00 0.00 ±0.00 −0 0
Stomatanthes africanus
i
1 0.50 0.23 81.25 0.87 0.14 0.00 0 0 0
Helianthus annuus
c
3 0.15 ±
0.07
0.07 37.74 ±
12.89
0.40 0.01 ±0.01 0.00 ±0.00 0 0 0
Helianthus tuberosus
c
3 0.93 ±
0.71
0.44 48.90 ±
18.04
0.52 0.15 ±0.13 0.10 ±0.10 68.84 0 0
Xanthium strumarium
w
3 1.18 ±
0.38
0.56 69.14 ±
45.7
0.74 0.80 ±0.50 0.07 ±0.07 9.03 0 0
Bidens pilosa
w
3 0.80 ±
0.42
0.37 97.14 ±
4.95
1.0 0.29 ±0.25 0.22 ±0.20 73.35 11.33 ±
9.45
0.34
Bidens formosa
w, o
3 0.00 ±
0.00
0.00 15.03 ±
16.9
0.16 0.06 ±0.09 0.00 ±0.00 0 0 0
Dahlia rosea
o
3 0.02 ±
0.04
0.01 8.09 ±
7.01
0.09 0.00 ±0.00 0.00 ±0.00 −0 0
Chrysanthemum ×
morifolium
o
3 0.33 ±
0.32
0.16 52.34 ±
17.0
0.56 0.003 ±0.01 0.00 ±0.00 0 0 0
Argyranthemum
frutescens
o
3 0.09 ±
0.90
0.04 31.2 ±
27.1
0.33 0.024 ±0.03 ±0.01 18.09 1.0 ±
1.73
0.03
Symphyotrichum novi-
belgii
o
3 0.04 ±
0.07
0.02 0.00 ±
0.00
0.00 0.00 ±0.00 0.00 ±0.00 −0 0
Senecio
madagascariensis
i,w
3 0.48 ±
0.27
0.22 70.94 ±
25.6
0.75 0.124 ±0.15 0.09 ±0.12 74.28 5.0 ±
4.36
0.15
Senecio tamoides
i
3 1.26 ±
0.31
0.59 66.30 ±
14.3
0.71 0.11 ±0.07 0.014 ±0.01 12.94 0 0
Lactuca sativa
c
3 0.00 ±
0.00
0.00 0.00 ±
0.00
0.00 0.00 ±0.00 0.00 ±0.00 −0 0
Cichorium intybus
c
3 0.10 ±
0.17
0.05 3.03 ±
5.25
0.03 0.05 ±0.08 0.00 ±0.00 0 0 0
Distephanus
angulifolius
i
3 0.31 ±
0.12
0.14 42.5 ±
20.84
0.45 0.005 ±0.01 0.002 ±0.004 50 0 0
Gymnanthemum
crataegifolium
i
3 0.06 ±
0.05
0.03 11.07 ±
9.86
0.12 0.004 ±0.01 0.004 ±0.006 100 0 0
Cynara scolymus
c
3 0.08 ±
0.04
0.04 0.00 ±
0.00
0.00 0.00 ±0.00 0.00 ±0.00 −0 0
Ipomoea batatas
c
3 0.01 ±
0.01
0.01 0.00 ±
0.00
0.00 0.00 ±0.00 0.00 ±0.00 −0 0
Solanum lycopersicum
c
3 0.00 ±
0.00
0.00 0.00 ±
0.00
0.00 0.00 ±0.00 0.00 ±0.00 −0 0
Amaranthus hybridus
c
3 0.02 ±
0.03
0.01 0.00 ±
0.00
0.00 0.00 ±0.00 0.00 ±0.00 −0 0
Beta vulgaris
c
3 0.00 ±
0.00
0.00 0.00 ±
0.00
0.00 0.00 ±0.00 0.00 ±0.00 −0 0
Cucurbita pepo
c
3 0.00 ±
0.00
0.00 0.00 ±
0.00
0.00 0.00 ±0.00 0.00 ±0.00 −0 0
Phaseolus vulgaris
c
3 0.05 ±
0.08
0.02 11.00 ±
19.2
0.12 0.00 ±0.00 0.00 ±0.00 −0 0
n, Number of replicates.
a
Adult feeding is expressed as leaf area (cm
2
) removed per adult per day of the trial.
b
Relative damage determined using the mean percentage of damage per test species in proportion to that on C. odorata.
c
% of adults surviving over the 30 days.
d
Relative suitability determined using the mean survival on the test plants in proportion to that on C. odorata.
e
Oviposition holes are expressed as number per female per day of the trial.
f
Relative progeny production determined using the mean progeny production on the test plants in proportion to that on C. odorata.
R. Mangan et al.
Biological Control 197 (2024) 105591
8
sequence analysis for 18 specimens, representing three different species
on a variety of host plants. One clade represents specimens collected on
C. odorata (both during the current survey and from the South African
laboratory culture), C. laevigata, and H. vitalbae during the native-range
exploration. Another clade denotes the outgroup L. liformis and the
nal clade represents L. angustatus (Fig. 2). The K2P genetic distances
between haplotypes within the 6 populations ranged from 0.001
(L. aemulus on C. odorata (SA biotype), Lixus sp. on C. odorata,
C. laevigata, and H. vitalbae) to 0.353 (L. angustatus on Picris sp.)
(Table 6). The overall mean within population was 0.133. Mean dis-
tances between populations ranged from 0.001 between L. aemulus on
C. odorata (SA biotype) and Lixus sp. on C. laevigata to 0.353 between
Lixus sp. on C. laevigata and L. angustatus on Picris sp. An AMOVA was
performed on the 14 specimens of Lixus sp(p). collected in Brazil and
L. aemulus from C. odorata (South African biotype) demonstrated that 95
% (Φ
ST
=0.9518, P >0.01) of the variance occurred within populations
(Supplementary Table S10). AMOVA also established low variance be-
tween populations (Φ
ST
=0.0482, P <0.01) indicating a lack of genetic
structure.
Table 2
Percentage of time spent by individuals and feeding damage of the adult L. aemulus during paired-choice tests.
Test plant nPosition (% time) Feeding (cm
2
/adult/day) Relative feeding
C. odorata Non-target Elsewhere C. odorata Non-target
Ageratina adenophora 3 0.65 ±0.14a 0.25 ±0.15b 0.10 ±0.02 1.37 ±0.42a 0.48 ±0.23b 0.35
Ageratum conyzoides 3 0.69 ±0.07a 0.23 ±0.06b 0.08 ±0.02 1.70 ±0.40a 0.24 ±0.12b 0.14
Bidens pilosa 3 0.88 ±0.04a 0.09 ±0.05b 0.08 ±0.01 1.66 ±0.01a 0.07 ±0.03b 0.04
Campuloclinium macrocephalum 3 0.88 ±0.04a 0.06 ±0.06b 0.06 ±0.03 1.84 ±0.01a 0.01 ±0.02b 0.01
Senecio madagascariensis 3 0.83 ±0.10a 0.06 ±0.03b 0.11 ±0.09 1.83 ±0.01a 0.01 ±0.02b 0.01
Means (±SD) within columns followed by the same letter are not signicantly different; Position: P <0.05, Likelihood Ratio Test; Adult feeding: P <0.05, Likelihood
Ratio Test.
Test species are listed alphabetically.
Table 3
Oviposition and progeny production of the adult L. aemulus during paired-choice tests.
Test plant nNo. oviposition holes/female/day Progeny
Total Plugged C. odorata Non-target
C. odorata Non-target C. odorata Non-target
Ageratina adenophora 3 0.5 ±0.1a 0.6 ±0.4a 0.4 ±0.1a 0.1 ±0.1a 30.7 ±14.6a 8.7 ±4.3b
Ageratum conyzoides 3 0.6 ±0.5a 0.4 ±0.2a 0.5 ±0.4a 0.2 ±0.1a 14.3 ±9.3a 11.0 ±8.5a
Bidens pilosa 3 0.7 ±0.3a 0.1 ±0.1a 0.4 ±0.1a 0.0 ±0.0a 34.7 ±26.2a 4.7 ±4.5b
Campuloclinium macrocephalum 3 0.6 ±0.5a 0.1 ±0.2a 0.4 ±0.4a 0.1 ±0.1a 22.0 ±5.6a 8.3 ±12.7a
Senecio madagascariensis 3 0.3 ±0.2a 0.1 ±0.1a 0.3 ±0.1a 0.02 ±0.03a 13.0 ±12.3a 0.3 ±0.6b
Means (±SD) within columns followed by the same letter are not signicantly different; Total and plugged oviposition: P <0.05, Likelihood Ratio Test; Total number of
progeny: P <0.05, Likelihood Ratio Test.
Test species are listed alphabetically.
Table 4
Risk assessment of non-target attack by L. aemulus, using its preference for and performance on test plants in host-specicity trials relative to that on C. odorata.
Test plant species Plant preference
a
Feeding damage
b
Feeding risk (%)
c
Adult survival (%)
d
Progeny production
e
Reproductive risk (%)
f
Chromolaena odorata
w
1.0 1.0 100 1.0 1.0 100
Ageratina adenophora
w
0.35 0.83 29.05 0.99 0.37 36.63
Ageratum conyzoides
w
0.14 0.46 6.44 0.96 0.48 46.08
Campuloclinium macrocephalum
w
0.01 0.26 0.26 0.51 0.11 5.61
Bidens pilosa
w
0.04 0.37 1.48 1.0 0.34 34.00
Senecio madagascariensis
i,w
0.01 0.22 0.22 0.75 0.15 11.25
a
Attractiveness of test plant species, determined by adult preference (relative feeding) in paired-choice trials (Table 2).
b
Host suitability of test plant species, determined by feeding damage (relative damage) in no-choice trials (Table 1).
c
Product of suitability indices for preference
a
and performance
b
.
d
Host suitability of test plant species, determined by adult survival in no-choice trials (Table 1).
e
Host suitability of test plant species, determined by the production of progeny (relative progeny production) in no-choice trials (Table 1).
f
Product of suitability indices for adult survival
d
and production of progeny
e
.
Table 5
Plants examined in Acre state, Brazil, for the presence of Lixus sp(p). Adults.
Means ±SD are given. See Table S1 for taxonomic details of plant species.
Plant species n
1
No of Lixus sp(p).
per plant
Total
Lixus
No. of plants with
Lixus (%)
Chromolaena
laevigata
30
(3)
0.27 ±0.69 8 5 (16.7)
Chromolaena
odorata
29
(4)
0.21 ±0.49 6 5 (17.2)
Clibadium
surinamense
1(1) 0 0 0
Heterocondylus
vitalbae
9(2) 0.33 ±0.71 3 2 (22.2)
Praxelis clematidea 10
(1)
0 0 0
Tilesia baccata 11
(2)
0 0 0
Tithonia diversifolia 1(1) 0 0 0
Vernonanthura
patens
28
(4)
0 0 0
Test species are listed alphabetically.
1
No. individual plants (no. sites).
R. Mangan et al.
Biological Control 197 (2024) 105591
9
3.7. Climatic suitability assessment (‘climate matching’)
Overall, much of Sub-Saharan Africa remains poor to moderately
matched climatically to the Brazilian collection sites of L. aemulus, with
only the western and central tropical regions presenting high (>70 %) to
optimal (>80 %) climatic matches (Fig. 3A). As for South Africa, the
climatic parameters for all the release sites of L. aemulus showed a mean
match of 58 ±2 % to the native Brazilian collection sites (Supplemen-
tary Table S11). Climatic matching indicated that rainfall and temper-
ature showed the greatest discrepancies amongst the climatic variables
between the collection and release sites, particularly for most inland
regions of South Africa. Release sites of L. aemulus located along the
KwaZulu-Natal coastline maintained the highest composite match
indices (climate matches) to Rio Branco, Brazil, ranging between low to
moderate, with matches of 59–66 % (Fig. 3B). The remaining release
sites in the Mpumalanga and Limpopo provinces were poorly matched to
the Brazilian collection sites, averaging low matches of 52 % and 50 %
respectively (Fig. 3B).
4. Discussion
A total of 29 test plant species, consisting mainly of Asteraceae but
including a few non-asteraceous crop species, was used for testing the
laboratory host specicity of L. aemulus. The preferred host of L. aemulus
is C. odorata, under both no-choice and paired-choice conditions in cages
in the laboratory. The ovipositing female is the host-selecting life stage,
which is typical of insect species whose larvae are endophagous, and
have ‘limited mobility’(Prager et al., 2014). The female bores an
exploratory oviposition hole, and only lays her egg if she nds the host
suitable. Once the egg has been laid it usually hatches (87.4 % (n =103)
for which this was recorded) and the progeny develops through to
adulthood. The plant species must be physically suitable before ovipo-
sition occurs. Stems must have a pithy centre for the larva to survive and
they must be of sufcient diameter. Although the adults were able to
feed and survive on a range of Asteraceae under no-choice conditions,
they caused negligible damage. Adult females were selective of the
species in which they laid eggs. Larval progeny were obtained in six of
the twenty-nine non-target species exposed to L. aemulus (A. frutescens,
A. adenophora, A. conyzoides, B. pilosa, C. macrocephalum, S. mada-
gascariensis) and adult progeny in ve (all but A. frutescens). Increased
selectiveness was evident under paired-choice conditions. No crop spe-
cies from other families on which other species of Lixus have been
recorded as pests were suitable for the survival of L. aemulus.
Native-range surveys conrmed that the weevil was present on
C. odorata. An equivalent number of Lixus adults were found on C. lae-
vigata and H. vitalbae. Several other species of Asteraceae surveyed had
Fig. 2. Maximum Likelihood tree using the GTR+G model of substitution of 3 control region haplotypes (411 bp) resolving three major lineages; Lixus aemulus
samples are indicated in blue, Lixus angustatus are indicated in yellow, and Lixus liformis are indicated in green. (For interpretation of the references to colour in this
gure legend, the reader is referred to the web version of this article.)
Table 6
Pairwise F
ST
values between (below diagonal) and within (diagonal) Lixus sp. populations based on the Kimura two-parameter distance between mtDNA haplotypes.
Species (host plant) L. aemulus (C.
odorata*)
Lixus sp. (C.
odorata)
Lixus sp. (C.
laevigata)
Lixus sp.
(H. vitalbae)
L. liformis
(C. acanthoides)
L. angustatus (Picris
sp.)
L. aemulus (C. odorata*)0.001
Lixus sp. (C. laevigata) 0.001 0.002
Lixus sp. (C. odorata) 0.001 0.001 0.001
Lixus sp. (H. vitalbae) 0.001 0.001 000.1 n/c
L. liformis
(C. acanthoides)
0.328 0.328 0.327 0.327 0.011
L. angustatus (Picris sp.) 0.353 0.353 0.353 0.353 0.233 0.261
FST values range from 0.0 (no differentiation) to 1.0 (complete differentiation).
n/c: not enough replicates to conduct within haplotype diversity.
*
Southern African biotype.
R. Mangan et al.
Biological Control 197 (2024) 105591
10
no adults present. Molecular analysis reveals low levels of genetic di-
versity and low levels of genetic differentiation across the individuals
collected on C. odorata (southern African biotype), C. odorata (in the
native range), C. laevigata and H. vitalbae for mitochondrial markers. The
low levels of genetic diversity observed appear to be correlated with low
Fst and Φ
ST,
suggesting low differentiation between populations of
L. aemulus and Lixus sp(p). This indicates that the weevils across all three
plant species collected in Brazil in 2015, as well as the SA laboratory
culture, collected opportunistically from a pubescent ‘hairy chromo-
laena’in 1995 (Kluge and Zachariades, 2006), belong to one inter-
breeding population and they are all one species, identied by C.
O’Brien as L. aemulus or near. This also conrms the results of laboratory
trials conducted in quarantine that indicate that L. aemulus is not
monophagous. The host range of L. aemulus is conned to certain plants
within the tribe Eupatorieae, and therefore it still has an acceptably
narrow host range for safe release in South Africa (Zachariades et al.,
2021). In Rio Branco, only adults and their non-target feeding damage
were found on C. laevigata and H. vitalbae. These plants may not have
been used as a larval host plant and laboratory host-range trials indi-
cated that adults feed on a broader range of species than the females
oviposit on. However, it seems unlikely that these two species were not
acting as larval host plants as adult L. aemulus are quite sedentary and
feeding and mating on both plant species, which were about 150 m from
the closest C. odorata plants.
Permission to release L. aemulus in South Africa was obtained in
2010, and releases of over 5,500 adults were made at 21 sites between
2011 and 2019. However, establishment of the weevil appears to have
been limited. In addition to concern surrounding the partial or complete
clearing of some ~20 % of L. aemulus release sites by 2016 (Zachariades
et al., 2021), much of South Africa’s climate was found to be poorly
matched to Rio Branco, Brazil. Considerable climatic mismatch, most
notably in rainfall and temperature, between South African regions
invaded by C. odorata and the collection sites of L. aemulus is likely to
have constrained the establishment and proliferation of the weevil
(Robertson et al., 2008). Hindrances to establishment, due to climatic
mismatches, have been a common occurrence in South African biocon-
trol programmes, particularly when agent collections have occurred in
high-rainfall, tropical regions of South America (Cowie et al., 2016).
Often substantial mismatches like this result in the persistence of agents
at low abundances, typically in localised to smaller areas, or more
severely a complete failure to establish (Cowie et al., 2016; Harms et al.,
2021). Broadly matching the South African regions invaded by
C. odorata to South America may offer useful insight for potential col-
lections in climatically better suited localities, with preliminary
matching suggesting that semi-arid regions, such as Remanso and
Caetit´
e, in the State of Bahia, appear as more suitable collection local-
ities. However limited information regarding the distribution of
L. aemulus in Brazil may pose difculties in assessing whether this agent
could in fact be collected from these regions (Zachariades et al., 2021).
Although climate remains one of the major constraints to biological
control programmes (Harms et al., 2021), with climatic mismatches
sought to be avoided, the biology of oligophagous species used for
biological control endeavours should also be taken into consideration.
Oligophagous species, such as L. aemulus, often display differential
feeding and performance amongst suitable host plants, as seen in this
study, which offers the possibility that the weevil may perform sub-
optimally on the southern African biotype of C. odorata.Chromolaena
odorata varies considerably in morphology (growth habit, leaf and stem
pilosity, inorescence structure and colour, etc.) as well as the odour of
its crushed leaves (indicating chemical differences) across its native
range, and the two major invasive biotypes differ from one another in
ecological characters such as susceptibility to burning and their
Fig. 3. Projected climatic match of Lixus aemulus collection sites in Rio Branco, Acre state, Brazil to (A) Sub-Saharan Africa, (B) South African release sites. Res-
olution of grid cells for South Africa is at a quarter degree square (QDS: ~25 km ×25 km) (data adapted from CLIMEX: Kriticos et al. (2015)).
R. Mangan et al.
Biological Control 197 (2024) 105591
11
performance at a given ambient temperature (Zachariades et al., 2009).
Lixus aemulus was collected opportunistically from a tropical region
with high rainfall at a time when the origin of the C. odorata population
invading southern Africa was unknown. Although the weevil was shown
to be adequately damaging and suitability host-specic for release
within South Africa, its poor establishment thus far justies concerns
regarding climatic and biotype mismatches. This situation advocates for
the prioritisation of biological control agents from regions of the native
range of C. odorata that are climatically more similar to the invaded
areas in South Africa.
5. Author statement
Host specicity: Laboratory: C.Z. conceived the experiments, C.Z.
and M.G. completed the laboratory work; eld: C.Z. and M.B. recorded
L. aemulus eld-range data, C.G.S., M.S. and A.M.T. assisted in the
identication of surveyed Asteraceae; R.M. and C.Z. conducted the
analysis and prepared the manuscript. All authors provided critical
feedback and helped shape the nal manuscript.
mtDNA: R.M. and C.Z. conceived the experiments, R.M. contributed
to sample preparation, conducted the analysis and prepared the manu-
script. All authors provided critical feedback and helped shape the nal
manuscript.
Climatic matching: B.C., R.M., and C.Z. conceived the experiments.
B.C. conducted the analysis. R.M., C.Z. and B.C prepared the manuscript.
All authors provided critical feedback and helped shape the nal
manuscript.
CRediT authorship contribution statement
Rosie Mangan: Writing –review &editing, Writing –original draft,
Visualization, Software, Project administration, Methodology, Investi-
gation, Formal analysis, Data curation, Conceptualization. Milly Gar-
eeb: Writing –review &editing, Data curation. Marcus Boeno: Writing
–review &editing, Data curation. Chirley Gonçalves da Silva: Writing
–review &editing, Data curation. Blair Cowie: Writing –review &
editing, Methodology, Formal analysis. Aristˆ
onio Magalh˜
aes Teles:
Writing –review &editing, Data curation. Marcos Silveira: Writing –
review &editing, Data curation. Costas Zachariades: Writing –review
&editing, Writing –original draft, Project administration, Methodology,
Investigation, Funding acquisition, Formal analysis, Data curation,
Conceptualization.
Declaration of competing interest
The authors declare that they have no known competing nancial
interests or personal relationships that could have appeared to inuence
the work reported in this paper.
Acknowledgements
Thanks to the Agricultural Research Council for permission to un-
dertake native range exploration and to the Department of Forestry,
Fisheries and the Environment: Natural Resource Management Pro-
grammes and the South African Research Chairs Initiative of the
Department of Science and Technology and the National Research
Foundation of South Africa for the provision of funding. Prof. Marcelo
Diniz Vitorino, Universidade Regional de Blumenau, is thanked for his
assistance in the logistics of organising the 2015 eldwork. Prof. Martin
Hill and Prof. Iain Paterson, Rhodes University, are thanked for funding
acquisition for the molecular work. We thank the anonymous reviewers
for helpful suggestions to improve this manuscript.
Appendix A. Supplementary material
Supplementary data to this article can be found online at https://doi.
org/10.1016/j.biocontrol.2024.105591.
References
Baldwin, B.G., 2009. Heliantheae alliance. In: Funk, V.A., Susanna, A., Stuessy, T.F.,
Bayer, R.J. (Eds.), Systematics, Evolution and Biogeography of Compositae.
International Association for Plant Taxonomy, Vienna, Austria, pp. 689–711.
Brooks, M.E., et al., 2017. glmmTMB balances speed and exibility among packages for
zero-inated generalized linear mixed modeling. R J. 9, 378–400.
Cock, M.J.W., Holloway, J.D., 1982. The history of, and prospects for, the biological
control of Chromolaena odorata (Compositae) by Pareuchaetes pseudoinsulata Rego
Barros and allies (Lepidoptera: Arctiidae). Bull. Entomol. Res. 72, 193–205.
Cowie, B.W., et al., 2016. Does climate constrain the spread of Anthonomus santacruzi, a
biological control agent of Solanum mauritianum, in South Africa? Biol. Control 101,
1–7.
Cowie, B.W., et al., 2023. Will climate affect the establishment and efcacy of Agnippe
sp. #1 (Lepidoptera: Gelechiidae), a promising biological control agent of Mesquite
in South Africa? BioControl 68, 681–695.
Excofer, L., et al., 1992. Analysis of molecular variance inferred from metric distances
among DNA haplotypes: application to human mitochondrial DNA restriction data.
Genetics 131, 479–491.
Excofer, L., et al., 2005. Arlequin (version 3.0): An integrated software package for
population genetics data analysis. Evol. Bioinforma. 1.
Fox, J., Weisberg, S., 2019. Nonlinear regression, nonlinear least squares, and nonlinear
mixed models in R. Population 150, 200.
Goolsby, J.A., et al., 2006. Maximising the contribution of native-range studies towards
the identication and prioritisation of weed biocontrol agents. Aust. J. Entomol. 45,
276–286.
Harms, N.E., et al., 2021. Climate mismatch between introduced biological control
agents and their invasive host plants: improving biological control of tropical weeds
in temperate regions. Insects 12, 549.
Heard, T.A., 2002. Host specicity testing of biocontrol agents of weeds. In: Denslow, J.
E., Hight, S.D., Smith, C.W. (Eds.), Proceedings, Hawaii Biological Control
Workshop. University of Hawaii, Honolulu, Pacic Cooperative Studies Unit,
pp. 21–28.
Heard, T.A., Van Klinken, R.D., 1998. An analysis of test designs for host range
determination of insects for biological control of weeds. In: M. Zalucki, White, G.,
(Ed.), Proceedings of the 6th Australasian Applied Entomological Research
Conference, University of Queensland, Brisbane, 29 September–2nd October 1998,
pp. 539–546.
Hoelmer, K.A., Kirk, A.A., 2005. Selecting arthropod biological control agents against
arthropod pests: can the science be improved to decrease the risk of releasing
ineffective agents? Biol. Control 34, 255–264.
Hudson, R.R., et al., 1992. Estimation of levels of gene ow from DNA sequence data.
Genetics 132, 583–589.
Julien, M.H., Grifths, M.W., 1999. Biological Control of Weeds-A World Catalogue of
Agents and their Target Weeds. CABI Publishing, Wallingford, UK.
Kimura, M., 1980. A simple method for estimating evolutionary rates of base
substitutions through comparative studies of nucleotide sequences. J. Mol. Evol. 16,
111–120.
King, R.M., Robinson, H., 1987. The Genera of the Eupatorieae (Asteraceae),
Monographys in Systematic Botany, The Missouri Botanical Garden. Allen Press, Inc.
Kluge, R.L. et al., 1997. Proposed three-year plan for the biological control of
chromolaena in South Africa. ARC-PPRI, Unpublished ARC-PPRI report, South
Africa.
Kluge, R.L., Zachariades, C., 2006. Assessing the damage potential of the stem-boring
weevil Lixus aemulus for the biological control of Chromolaena odorata. BioControl
51, 547–552.
Kriticos, D.J., et al., 2015. Exploring the effects of climate on plants, animals and
diseases. CLIMEX Version 4, 184.
Kumar, S., et al., 2016. MEGA7: molecular evolutionary genetics analysis version 7.0 for
bigger datasets. Mol. Biol. Evol. 33, 1870–1874.
Librado, P., Rozas, J., 2009. DnaSP v5: a software for comprehensive analysis of DNA
polymorphism data. Bioinformatics 25, 1451–1452.
Louda, S.M., et al., 2003. Invasiveness of some biological control insects and adequacy of
their ecological risk assessment and regulation. Conserv. Biol. 17, 73–82.
Macdonald, I.A.W., Jarman, M.L., 1985. Invasive alien plants in the terrestrial
ecosystems of Natal, South Africa. National Scientic Programmes Unit: CSIRO.
Mangan, R., Baars, J.-R., 2023. Risk assessment of the host range of Hydrellia lagarosiphon
for the biological control of Lagarosiphon major in Ireland. Biocontrol Sci. Technol.
33, 681–700.
Mayhew, P.J., 1998. Testing the preference–performance hypothesis in phytophagous
insects: lessons from chrysanthemum leafminer (Diptera: Agromyzidae). Environ.
Entomol. 27, 45–52.
McClay, A.S., Balciunas, J.K., 2005. The role of pre-release efcacy assessment in
selecting classical biological control agents for weeds—applying the Anna Karenina
principle. Biol. Control 35, 197–207.
McFadyen, R.C., 1998. Biological control of weeds. Annu. Rev. Entomol. 43, 369–393.
Muniappan, R. et al., 2005. Distribution and biological control of Chromolaena odorata.
pp. 223–233.
Paterson, I.D., Zachariades, C., 2013. ISSRs indicate that Chromolaena odorata invading
southern Africa originates in Jamaica or Cuba. Biol. Control 66, 132–139.
Petri, K., 1928. Bestimmungstabelle der mir bekannt gewordenen sudamerikanischen
Arten der Gattung Lixus Fabr. nebst Neubeschreibungen. Verhandlungen Und
R. Mangan et al.
Biological Control 197 (2024) 105591
12
Mitteilungen Des Siebenbburgischen Vereins Fur Naturwissenschaften Zu
Hermannstadt 78 (1), 63–132.
Prager, S.M., et al., 2014. Factors inuencing host plant choice and larval performance in
Bactericera cockerelli. PLoS One 9, e94047.
R_Core_Team, 2021. R: A Language and Environment for Statistical Computing. Austria,
Vienna.
Rambaut, A., 2001. Se-Al: Sequence Alignment Editor ver. 2.0. Oxford Dept. of Zoology,
University of Oxford.
Retief, E., 2002. The tribe Eupatorieae (Asteraceae) in southern Africa. In: Zachariades,
C., Muniappan, R., Strathie, L.W. (Eds.), Proceedings of the Fifth International
Workshop on Biological Control and Management of Chromolaena odorata. ARC-
PPRI, Pretoria, South Africa, pp. 81–89.
Robertson, M.P., et al., 2008. Climate matching techniques to narrow the search for
biological control agents. Biol. Control 46, 442–452.
Schmidl, L., 1981. Ragwort, Senecio jacobaea, in Victoria and renewed attempts to
establish the cinnabar moth, Tyria jacobaeae, for its control. 5th International
Symposium on Biological Control of Weeds, Brisbane (Australia), 22 Jul 1980.
Commonwealth Scientic and Industrial Research Organization.
Shao, X., et al., 2018. On the origin and genetic variability of the two invasive biotypes of
Chromolaena odorata. Biol. Invasions 20, 2033–2046.
Sheppard, A.W., et al., 2006. Top 20 environmental weeds for classical biological control
in Europe: a review of opportunities, regulations and other barriers to adoption.
Weed Res. 46, 93–117.
Sheppard, A.W. et al., 2003. What is needed to improve the selection, testing and
evaluation of weed biological control agents: workshop synthesis and
recommendations. In: Spafford, J.H., Briese, D.T. (Eds.), Improving the Selection,
Testing and Evaluation of Weed Biological Control Agents. Proceedings of the CRC
for Australian Weed Management Biological Control of Weeds Symposium and
Workshop, 13 September 2002, University of Western Australia, Perth, Australia.
Sobhian, R., et al., 2003. Observations on the host specicity and biology of Lixus salsolae
(Col., Curculionidae), a potential biological control agent of Russian thistle, Salsola
tragus (Chenopodiaceae) in North America. J. Appl. Entomol. 127, 322–324.
Strathie, L.W., Zachariades, C., 2002. Biological control of Chromolena odorata in South
Africa: developments in research and implementation. In: Zachariades, C.,
Muniappan, R., Strathie, L.W. (Eds.), Proceedings of the Fifth International
Workshop on Biological Control and Management of Chromolaena odorata, ARC-
PPRI, Pretoria, South Africa, 23–25 October 2000, pp. 74–79.
Tamura, K., et al., 2011. MEGA5: molecular evolutionary genetics analysis using
maximum likelihood, evolutionary distance, and maximum parsimony methods.
Mol. Biol. Evol. 28, 2731–2739.
Tamura, K., Nei, M., 1993. Estimation of the number of nucleotide substitutions in the
control region of mitochondrial DNA in humans and chimpanzees. Mol. Biol. Evol.
10, 512–526.
van Klinken, R.D., 2000. Host specifcity testing: why do we do it and how can we do it
better. In: Van Driesche, R.G., Heard, T.A., McClay, A., Reardon, R. (Eds.),
Proceedings of Session: Host-specicity Testing of Exotic Arthropod Biological
Control Agents—The Biological Basis for Improvement in Safety, USDA Forest
Service, Morgantown, West Virginia, USA, July 4–14, 1999, pp. 54–68.
Volovnik, S., 2024. Host plants of Palaearctic weevils of the Lixus Fabricius (Coleoptera:
Curculionidae). Arthropod Plant Interact. 18.
Wan, F.-H., Harris, P., 1997. Use of risk analysis for screening weed biocontrol agents:
Altica carduorum Guer. (Coleoptera: Chrysomelidae) from China as a biocontrol
agent of Cirsium arvense (L.) Scop. in North America. Biocontrol Sci. Technol. 7,
299–308.
Wapshere, A.J., 1974. A strategy for evaluating the safety of organisms for biological
weed control. Ann. Appl. Biol. 77, 201–211.
Zachariades, C., et al., 1999. The South African programme on the biological control of
Chromolaena odorata (L.) King &Robinson (Asteraceae) using insects. Afr. Entomol.
Memoir 1, 89–102.
Zachariades, C., 2004. Release of the stem-boring weevil, Lixus aemulus, from quarantine
in Cedara, for biological control of trifd weed, Chromolaena odorata, South Africa.
In: ARC-PPRI report, South Africa, pp. 1–25.
Zachariades, C., et al., 2011. Progress towards the biological control of Chromolaena
odorata (L.) R.M.King &H.Rob. (Asteraceae) in South Africa. Afr. Entomol. 19,
282–302.
Zachariades, C., et al., 2021. Biological Control of three Eupatorieae weeds in South
Africa: 2011–2020. Afr. Entomol. 29, 742–767.
Zachariades, C. et al., 1998. Promising new candidates for the biocontrol of Chromolaena
odorata. In: Ferrar, P., Muniappan, R., Jayanth, K.P. (Eds.), Proceedings of the Fourth
International Workshop on Biological Control and Management of Chromolaena
odorata. Bangalore, India, October 1996. pp. 79.
Zachariades, C. et al., 2002. Biology, host-specicity and effectiveness of insects for the
biocontrol of Chromolaena odorata in South Africa. In: Zachariades, C., Muniappan,
R., Strathie, L.W. (Eds.), Fifth International Workshop on Biological Control and
Management of Chromolaena odorata. Pretoria, South Africa, 23–25 October 2000,
ARC-PPRI. pp. 160–166.
Zachariades C. et al., 2009. Chromolaena odorata (L.) King and Robinson (Asteraceae).
Biological control of tropical weeds using arthropods.
Zachariades, C. et al., 2013. Recent spread and new records of Chromolaena odorata in
Africa. In: Zachariades, C., Strathie, L.W., Day, M.D., Muniappan, R. (Eds.),
Proceedings of the Eighth International Workshop on Biological Control and
Management of Chromolaena odorata and other Eupatoriea. Nairobi, Kenya, 1–2
November 2010. ARC-PPRI, Pretoria, pp. 20–27.
R. Mangan et al.