Rural R&D for Profit round 4 project (2019-2023)
Achievements of Rural R&D for Profit round 4 project: Underpinning agricultural productivity and biosecurity by weed biological control
Pre-release monitoring of vegetation at selected future release sites of biocontrol agent(s)
Optimising sampling methods for flaxleaf fleabane infestations
During March-April 2020, we developed a sampling method to allow rapid assessment of fleabane population conditions at the long-term monitoring sites. This involved harvesting 126 reproductively mature fleabane plants from a variety of microhabitats throughout the ACT, including roadsides, drainage ditches, native grassy woodlands, stream embankments, urban footpaths and other urban green spaces. Statistical analyses revealed that a combination of (i) plant canopy volume (calculated from height, canopy diameter and foliage cover), (ii) number of inflorescences, which can be rapidly and non-destructively measured in the field and are strong, robust and positive predictors of plant biomass (a measure of plant productivity) and (iii) number of capitulae (a measure of plant reproductive effort) was the most effective assessment method.

Schematic diagram showing (A) the typical habit of a flaxleaf fleabane (Conyza bonariensis) plant, with a shallow root system with a short, stocky tap root; and a short primary stem that branches into multiple lateral stems that bear one or more inflorescences (denoted by yellow crosses); small grey circles clustered on each inflorescence represent capitulae (i.e., flower heads bearing multiple florets). For each focal plant positioned at the centre of each 1 m2 plot (denoted by the yellow circle in image C), we measured (B) height (cm) from the soil surface to the tip of the tallest stem and (C) maximum canopy diameters (cm) in two directions (D1 and D2), parallel to the soil surface. Percentage planar foliage cover of the target plant (yellow dotted outline in image C) was visually estimated from the two-dimensional space bound by D1 and D2 (the blue square in image C).
Pre-release monitoring of flaxleaf fleabane populations
In November 2020, we identified six mixed-cropping sites suitable for the establishment of fleabane population monitoring and future releases of candidate biocontrol agent(s), located across the Riverina Irrigation Region of NSW. The aim of this field trip was to identify the different microhabitat characteristics of irrigated mixed cropping systems in which crop weeds, especially fleabane, can persist during fallow periods and pose significant challenges to effective weed control across land tenures. The identified microhabitats that will be targeted during future monitoring include roadsides, fence lines, field margins, draining ditches and irrigation embankments.
Flaxleaf fleabane population sampling
Between February and May 2021, we established 45 transects (20 m length) at 9 sites located throughout the Murrumbidgee Irrigation Area, between the townships of Griffith and Leeton, NSW. These transects were established in each of four different habitats comprising the highest densities of fleabane infestations: along roadsides, drainage ditches, field margins and irrigation embankments.
We randomly sampled 10 fleabane plants along each transect, thus resulting in data being collected for a total of 450 plants. For each plant we measured height, foliage cover and number of inflorescences, which were previously shown to be robust surrogates for productivity (i.e., biomass) and reproductive output. A 1 m × 1 m quadrat was also placed around each focal plant, in which we counted the number of fleabane plants (as a measure of population density) and visually estimated overall fleabane % foliage cover. We also estimated the % cover of several abiotic (e.g., bare soil cover) and biotic (e.g., herb and grass cover) variables.
We found that fleabane plants were significantly larger along drainage ditches and irrigation embankments compared with roadsides and field margins. Reproductive output per plant was more than two-times higher along drainage ditches and field margins than either roadsides or irrigation embankments, and declined with increasing cover of other plant species, especially grasses. Population density (i.e., number of individual plants per 1m2) tended to be higher along roadsides and irrigation embankments than either drainage ditches or field margins. The overall % cover of fleabane populations did not vary amongst the four habitat categories but was positively associated with bare soil cover and negatively associated with the cover of herbs and wireweed. This suggests that fleabane cover may be suppressed under high competition with other herbaceous species, or that fleabane could benefit from disturbances (e.g., roadside slashing or spraying) that create bare soil patches not suitable for the establishment of other herbaceous species.

Photographic examples of each of the four main habitats in which flaxleaf fleabane populations were sampled in 2021.
Approved release of the pathogen from biosecurity containment and establishment of a laboratory culture
Flaxleaf fleabane was endorsed as a candidate for biological control in Australia in November 2017 by the Invasive Plants and Animals Committee (now the Environment and Invasives Committee). Comprehensive host-specificity testing for the rust fungus, Puccinia cnici-oleracei (ex. Conyza) obtained from Colombia, South America, began in quarantine in February 2019, during the previous RRnDFP2 project, although testing remained incomplete and was finalised during this current RRnDFP4 project. The application to release the fleabane rust fungus was submitted to the federal regulators on 30th June 2020.
The Department of Agriculture, Water and the Environment (DAWE) approved the release of the rust fungus Puccinia cnici-oleracei (ex. Conyza) for the biological control of flaxleaf fleabane on 17 June 2021. Leaves infected with the fungus were removed from the quarantine facility on 21 June, in the presence of DAWE officers. Since then, a culture on live plants has been established in a non-quarantine controlled-environment room at the CSIRO Black Mountain site in Canberra.
A viable, healthy culture of the rust fungus P. cnici-oleracei (ex. Conyza) was maintained on Conyza bonariensis (flaxleaf fleabane) seedlings within controlled temperature (CT) rooms at CSIRO’s Black Mountain Laboratories in Canberra (culture maintenance ongoing). Infected leaves with mature pustules were then removed from the recipient plants, dried flat to preserve the spores, and stored in the dark within airtight plastic containers with silica desiccant at 4 ˚C. In this way, we accumulated a stockpile of over 1,000 severely infected fleabane leaves for provision to registered community participants in the pilot mass-release program which ran from September 2022 to March 2023.

Examples of dried flaxleaf fleabane leaves infected with the biocontrol agent.
Experimental optimisation of pathogen release methods in the field
Several experiments were undertaken with the aim of optimising release methods for the rust fungus in a field context and initiating infection by the fungus on naturalised flaxleaf fleabane populations under prevailing climate conditions in the Australian environment. It had been determined that the best way to inoculate plants was to attach a dried infected leaf to moistened paper towel, which was then suspended above the target plants and covered with a plastic bag.

Release of the rust fungus at the Ginninderra Lake site in the ACT.
In December 2021, the fungus was experimentally released on potted flaxleaf fleabane seedlings and maintained for several months under field conditions to monitor for patterns of disease incidence and severity, and effects of fungal infection on plant performance. After a month in January 2022, infection monitored across the treatments provided evidence that maintaining a still, humid microclimate at the time of inoculation is critical for successful infection by the fungus from the lab-reared inoculum to seedlings.
Multiple infection events occurred between January and May 2022, whereby the first set of lesions that developed in January produced spores that spread to nearby healthy leaves that subsequently became infected, and so on. Infection progressed over the entire body of each plant, with lesions eventually being detected on stems and flower heads. These observations confirmed that the fungus was able to readily infect fleabane plants growing outdoors over multiple months, under variable climate conditions.

Depiction of disease progression for the fungus on Conyza bonariensis
At the conclusion of the experiment in May 2022, seedlings were harvested, dried, and measured. The reproductive output (measured as the number of flower heads per inflorescence) declined significantly with increasing percentage of leaves on each inflorescence infected by the fungus. On average, severely infected inflorescences produced 50-60% fewer flower heads than non-infected inflorescences. These results indicate that, under high infection severity of developing inflorescences, the fungus can reduce the overall reproductive output of host flaxleaf fleabane plants.

Contrast of the inflorescences where (A) has severe infection by the fungus and stunted flower heads while (B) no infection by the fungus. The graph depicts the relationship between number of flower heads (capitulae) (y axis) and % leaves per inflorescence infected with the fungus (x axis).
In February 2022 and again in November 2022, a total of 24 release plots were set up in the Irrigation Research and Extension Committee (IREC) research station, near Whitton, to test release methods further across different habitat conditions. When returning to the sites to monitor the release plots, infection by the rust fungus was not detected on any focal plant within each of the release plots. Often by the time monitoring occurred, most plants had begun to senesce, with a high cover of dead leaves, making it difficult to determine if any prior infection had occurred between inoculation date and monitoring date.

Example of two transects in which the rust was released on flaxleaf fleabane plants in February 2022.
Further experimental releases of the fungus commenced in December 2022 on a naturalised flaxleaf fleabane population located at the CSIRO Black Mountain site in the ACT. Two release methods were trialled; for the first method, flaxleaf fleabane plants were inoculated by suspending an infected leaf above each of the target plants and covered with a large, clear plastic bag or opaque plastic box. For the second method, 12 severely infected flaxleaf fleabane plants growing in 15 cm pots were installed at 12 plots with clusters of field grown plants. Neither bags nor moisture were applied to these plants, as the aim was to allow gradual spore release over a longer period under field conditions.
At one month post inoculation, all plants were inspected for signs of fungal infection. For the first method, 100 % of plants (presence/absence) displayed at least some level of infection. The severity of infection was generally low, ranging from 1-10 % of leaves infected per plant. For the second method, ~58 % of plots showed low levels of infection on resident plants.
By March 2023, approximately 3 months post release, many of the infected leaves that were detected in January-February had died off. We did, however, detect infection on other parts of the plants – namely stem lesions – including of nearby plants, which indicated some spread of the fungus throughout the local population.
The results from these experiments have provided insights into optimal methods for releasing the rust fungus on flaxleaf fleabane:
- Infection by the rust fungus cannot readily occur without creating a humid microclimate around a recipient plant during sporulation, using a plastic bag or other enclosure for at least 12 hours.
- It is likely that younger seedlings are more susceptible to infection by the rust fungus than mature plants, which are likely to be relatively more resistant to infection.
- There is a need to monitor infection incidence and severity at high frequency over short time periods within the first three to four weeks following release of the fungus. Where infection does occur, telia may only be visible for a few days, after which they release spores and cause the host leaf to die off and shed from the infected plant.
Community engagement a in pilot mass-release program of the pathogen across Australia
From September 2022 until February 2023, CSIRO researchers launched a pilot biocontrol agent release program in partnership with landholders/growers and other weed management stakeholders from the grains sector, with the aim of trialling the optimised release protocols for the rust fungus in a variety of landscape and climate contexts. Interest from potential participants in the program was elicited through a joint GRDC-AgriFutures-CSIRO media campaign and face to face workshops.
- https://grdc.com.au/news-and-media/audio/podcast/flaxleaf-fleabane-rust-fungus
- https://groundcover.grdc.com.au/innovation/plant-breeding/new-fungus-to-help-landholders-fight-fast-spreading-weed

Ben Gooden running a face-to-face workshop on the flaxleaf fleabane biocontrol project for IREC, Murrumbidgee Irrigation and local growers in the Riverina, 10th November 2022.
Fifty-five stakeholders were selected for participation in the program, comprising of private landholders/growers, professional agronomists, research institutes, local councils, and Landcare networks. Participants were sent their own biocontrol agent release kits which included a dried flaxleaf fleabane leaf (or multiple leaves) that were infected with the rust fungus, other tools to prepare and apply the agent, and release and monitoring instructions. Altogether, 366 biocontrol agent release kits were distributed nationwide, focussing on the south-eastern parts of Australia where flaxleaf fleabane infestations cause the greatest impacts on crop yield. From participants who returned to their sites to monitor for the fungus, only ~10 % were able to observe infection.
Several challenges were identified with regards to eliciting verified infection presence/absence data from participating stakeholders. In at least 10 % of cases, inspection of photographs at the time of release indicated the participants had released the fungus on a different plant species, most often tall fleabane (Conyza sumatrensis), which is resistant to infection by the rust. In ~6 % of cases, the host flaxleaf fleabane plants had become senescent by the time the participants had returned to inspect the plants for signs of infection. During the CSIRO-run experiments, we observed that pustules can appear on the leaf surface after about 2-4 weeks, after which time the host leaves die off very quickly (within a couple of days) after the fungus has sporulated. As such, the incidence of infection was likely much higher at the community release sites, but the low frequency of return visits meant that many visible symptoms were likely missed by the participants.

Map of release locations of the flaxleaf fleabane rust fungus in partnership with community stakeholders.
Future research aspirations for biocontrol of flaxleaf fleabane with the rust fungus
Based on the results obtained so far, it is predicted that, even where the rust fungus can establish successfully in the field, infection is unlikely to result in significant reductions in the population size of flaxleaf fleabane for several years. As such, biocontrol represents a longer term and self-sustaining means of gradually reducing weed invasion pressure across productive landscapes. Furthermore, the combination of several biocontrol agents may enable more robust control of focal weeds. In this way, further research into the biocontrol of flaxleaf fleabane with insects may provide enhanced biocontrol outcomes.
Initiate native range research to identify additional candidate agents for Fleabane
Native range surveys were conducted in Brazil (Santa Catarina; Parana; Sergipe; Alagoas; Pernambuco states), Colombia (Antioquia department) and USA. In addition to the insect species reported and prioritized (e.g., Trupanea bonariensis, Astteromyia modesta and Lixus caudiger) as part of the RRnD4P Round 2 project (2016-2020), the following insect species were recorded feeding on Conyza spp:
- Brazil: a new unidentified weevil species, a leaf miner, a stem feeding thrips, Clinodiplosis sp, (Cecidomyiidae) and a micro lepidoptera (Torticidae)
- Colombia: Eutreta rhinophora (Tephritidae), Caloreas cydrota (Choreutidae) (reported as Argyrotaenia in previous report), Lioptilodes sp. (Pterophoridae), Puto barberi (Putoidae) and Proba vitiscuttis (Miridae)
- USA: Neolasioptera erigerontis. Colonies of most of these species were maintained in the native ranges for various biology and host specificity testing (detailed below). Based on the details in the existing literature, barberi and P. vitiscuttis recorded in Colombia were considered generalists and thus considered to have no potential as biocontrol agent.

(A) larva of the micro lepidoptera recorded in Brazil, (B) Damage caused by the larval feeding – floral bud with no seeds
Undertake prelimary host-specificy testing to priortise candidates for importation
Stem-boring fly Trupanea bonariensis on flaxleaf fleabane
Host-specificity testing were carried out using the stem-galling fly (T. bonariensis) in Brazil by exposing representative plant species in the genera Conyza, Eupatorium, Porophyllum, Emilia, Symphyotrichum, Chromolaena, Bidens and Baccharis to T. bonariensis.
Only species in the genus Conyza developed galls induced by T. bonariensis. The species Conyza glandulitecta developed the highest number of galls, followed by C. sumatrensis var. sumatrensis, and with only one gall on the species C. bonariensis.

(A) Entomological cages containing Asteraceae species for no-choice tests, (B) making measurements of galls formed in Conyza glandulitecta, (C & D) galls formed by Trupanea bonariensis on Conyza species
Micro Lepidoptera (Torticidae) on flaxleaf fleabane
Host-specificity testing of the micro lepidopteran species (Torticidae) was conducted in Brazil using 43 Asteraceae species. This species was recorded in all field expeditions causing defoliating damage to the five species of Conyza, with a greater extent in C. bonariensis and C. sumatrensis var. sumatrensis. A test was carried out in the field, selecting and measuring the morphological characteristics of two adult Asteraceae plants Bidens pilosa and Praxelis clematidea. In addition, specificity tests were carried out using cut leaves. For studies with cut leaves, larvae were introduced into each petri dich containing leaves of test plants. The experiment was carried out in 43 species of Asteraceae.

A) Defoliating damage caused by the Torticidae larvae on leaves of Praxelis clematidea, (B) cages setup with Tortricidae caterpillars, (C & D) Petri dish test with Tortricidae caterpillars on leaves of 43 species of Asteraceae
At the end of the trial, observations on the damage caused by the larvae and emergence of adults were made. Larvae were found to reach the adult stage from the following species: B. pilosa, E. fosbergii, Erechtites valerianifolius, Sonchus asper, Elephantopus mollis, Erechtites hieracifolius, C. bonariensis, C. canadensis, C. glandulitecta, C. sumatrensis var. sumatrensis, C. sumatrensis var. leiotheca and C. primulifolia. This candidate agent does not have a sufficiently narrow host range relative to other candidate agents (T. bonariensis, A. modesta and L. caudiger) thus it was not prioritized for importation into Australia.
Leaf-galling fly Asteromyia modesta on flaxleaf fleabane
No-choice specificity tests were carried out in Brazil. The specificity tests were conducted in 73 plant species from 66 genera spanning 20 families. Conyza sumatrensis var. sumatrensis was considered as a control in these trials because of lower number of galls induced by A. modesta on C. bonariensis. All experiments were carried out in “pop up” entomological cages, and at the end of the trial, observations on total number of leaves per plant, number of leaves with galls and total number of galls per plant were recorded.
None of the 73 plant species tested developed galls except C. sumatrensis var. sumatrensis.
In addition to no-choice tests, choice tests were performed to determine host preference of A. modesta for the species within the genus Conyza. After 20 days exposure to the insect, the total number of galls and number of leaves with galls were greater on Conyza sumatrensis var sumatrensis than on the other species (C. bonariensis, C. canadensis, C. glandulitecta, C. primulifolia, C. sumatrensis var. leiotheca).

Seedlings of Conyza spp. caged for choice tests, (B) observations made after 20 days of exposure to Asteromyia modesta and (C) dgalls formed by Asteromyia modesta on Conyza sumatrensis var. sumatrensis
Establish and maintain colonies of candidate agents in the native range colony
Colonies of five insect species, Lixus caudiger, an unidentified new weevil (Curculionidae), Asteromyia modesta, Clinodiplosis sp. and Trupanea bonariensis were maintained during the project period (2019-2023). Insects from these colonies were used for various host specificity and biology studies. These colonies were used to ship agents to France (L. caudiger and A. modesta) and Australia (T. bonariensis and A. modesta). Colonies of these five species are still being maintained to import to Australia for host testing should the resources become available for ongoing research on these prioritized insect agents.
Importation of new candidate agent(s) into quarantine to and commence testing
Trupanea bonariensis:
A total of three shipments were made during November 2019, November 2020 and January 2021 to establish a colony under Australian quarantine condition. Hand carrying of this agent was not possible due to COVID-19 impact on travel. Adults emerging from these shipments were introduced into cages containing 3 or 4 potted plants provided as either C. bonariensis only, C. sumatrensis only or a combination of both plant species for oviposition. Number of cages set up were 9, 7 and 13, respectively during the first, second and third shipment.
Plants were exposed to adult T. bonariensis and the adults were collected and reintroduced into another cage after the exposure period of 3 to 15 days. This process of reintroducing adults to fresh plants continued until all imported adults were dead. A total of nine and seven cages of plants exposed to T. borarienesis were maintained during the first and second shipment. These cages with exposed plants were maintained and observations on possible gall formation were made.
No galls were noticed on plants post-exposure and dissections of plants failed to reveal any evidence of oviposition or larval development. It was unclear why the adults did not oviposit on Conyza plants exposed to them. This could be because pupae perhaps have been damaged during the transit or bulk processing of the packages with X-ray screening might have induced sterility.
Trupanea bonarensis is a promising biocontrol agent for C. bonariensis because of its high specificity and the damage it causes to plants. The difficulty to establish a colony is because of the inability to hand carry insects to Australian quarantine and the project team was unable to travel to Brazil to learn key techniques and knowledge about the agent to rear them under quarantine conditions.
Asteromyia modesta:
Asteromyia modesta midge was sourced from a colony established at the FURB laboratories (Blumenau, Brazil) and shipped to Australia and France. These colonies were established using field collected insects from C. sumatrensis. Because hand carrying of agents was impossible because of COVID-19 restrictions on international larval, shipments through international courier were attempted. Larvae/pupae derived from the clean culture, as leaf galls on clean C. sumatrensis leaf material, were shipped to Australia during May and October 2021, and to France during August 2021, September 2021 and February 2022.
A colony of A. modesta could not be established either in Australia or France. The shipments were considerably delayed during transit and the insect did not survive on arrival in Australia during the first shipment. During the second shipment, a total of 160 adults emerged from the imported materials, and these adults were setup in 10 cages each with 12 to 20 adults. However, there was no evidence of oviposition or other development across all the cages and a colony could not be established. It was presumed that possible exposure of the package containing insects to X-ray screening during the transit could have induced sterility in emerged adults. The shipments sent to France were also delayed and all larvae were dead on arrival.
Lixus caudiger:
French import permits for the stem-boring weevil Lixus caudiger were obtained and the agent was sent via courier by collaborators in Brazil in mid-May 2021. A colony of the stem-boring weevil were initiated using both C. bonariensis and C. sumatrensis plants. The adults L. caudiger fed on the leaves of both Conyza species from the margins to the leaf axis, and oviposition marks were observed two weeks after the arrival of the weevils. The eggs are yellowish, oval, about 0.8 x 1.6 mm. Newly hatched larvae seemed to bore directly into the stem and feed on pith and cambium tissue. Using these weevils, a colony of L. caudiger was successfully established in the laboratory, and the colony produced more than 300 individuals and these adults were used in the host specificity testing.

Observations performed on Conyza bonariensis (A&B) Adult feeding damage, (C&D) oviposition marks in the stem, and (E) female ovipositing in a stem.

The stem-weevil, Lixus caudiger (A) Adult resting on a Conyza leaf; (B) Pairs in copulation, and (C) Egg laid by L. caudiger.
No-choice tests: 14 species in the subfamily Asteroideae belonging to four different tribes and ten subtribes were tested in adult no-choice trials. Pairs of sexually mature adults were placed on a plant of each species and placed in screened cages. Survival of adults, herbivory and the presence and number of oviposition marks were recorded.
Adult feeding was observed on several but not all tested species. The feeding intensity was far smaller compared to C. bonariensis and other Conyza species. For instance, our results on crops and native Australian species showed no feeding or less feeding relative to the focal weed. Likewise, mortality of adults on those species was also always higher than on Conyza species. While herbivory was quite high on Erigeron karvinskianus, a closely related species to Conyza that is used as ornamental, no oviposition mark was observed; this was confirmed by dissecting the stems. It is unlikely that this species supports the L. caudiger larval development.
Egg laying was observed only on Conyza species, with no differences between the three different Conyza species (C. bonariensis, C. sumatrensis and C. canadensis) in terms of L. caudiger mortality, number of oviposition marks, and herbivory. These species seem to be equivalent hosts for L. caudiger in a no-choice setting. The high mortality observed at the end of the test on most species except Conyza and Erigeron, combined with low herbivory suggest that non-target effects should be minimal on other tested species.

No-choice test set-up for Lixus caudiger. (A) Series of test performed in CT room; (B) Example of size of the plant used in host testing
Choice tests: Choice trails were conducted to investigate if L. caudiger has any preference for E. karvinskianus and C. bonariensis. Experimental setup used were similar to no-choice tests, except that both plant species were maintained in a same cage.
Adult feeding was observed on both plant species; however, the feeding intensity was far greater on C. bonariensis than on E. karvinskianus. Moreover, the different plant parameters measured at the end of the test showed that the feeding of the weevil did not impact E. karvinskianus growth, and there was no difference in growth of this species between the plants that were damaged by adult feeding and control plants not damaged by adult feeding. In contrast, adult feeding and/or larval development seemed to impact the development of C. bonariensis. The number of flower heads per plant was significantly lower (by ~30%) in plants damaged by L. caudiger compared to the control plant.
The results from the choice and no-choice host-specificity testing are very promising. The host range of L. caudiger appears to be restricted to Conyza. All Conyza species included in the tests are introduced weeds in Australia. Further host specificity testing under quarantine conditions in Australia is required to screen other non-target plants in the host test list.