Keywords: Invasive species Ecosystem function Insect pests Invasive plants Ecological restoration Biological control Natural ecosystems a b s t r a c tOf the 70 cases of classical biological control for the protection of nature found in our review, there were fewer projects against insect targets (21) than against invasive plants (49), in part, because many insect biological control projects were carried out against agricultural pests, while nearly all projects against plants targeted invasive plants in natural ecosystems. Of 21 insect projects, 81% (17) provided benefits to protection of biodiversity, while 48% (10) protected products harvested from natural systems, and 5% (1) preserved ecosystem services, with many projects contributing to more than one goal. In contrast, of the 49 projects against invasive plants, 98% (48) provided benefits to protection of biodiversity, while 47% (23) protected products, and 25% (12) preserved ecosystem services, again with many projects contributing to several goals. We classified projects into complete control (pest generally no longer important), partial control (control in some areas but not others), and ''in progress," for projects in development for which outcomes do not yet exist. For insects, of the 21 projects discussed, 62% (13) achieved complete control of the target pest, 19% (4) provided partial control, and 43% (9) are still in progress. By comparison, of the 49 invasive plant projects considered, 27% (13) achieved complete control, while 33% (16) provided partial control, and 49% (24) are still in progress. For both categories of pests, some projects' success ratings were scored twice when results varied by region. We found approximately twice as many projects directed against invasive plants than insects and that protection of biodiversity was the most frequent benefit of both insect and plant projects. Ecosystem service protection was provided in the fewest cases by either insect or plant biological control agents, but was more likely to be provided by projects directed against invasive plants, likely because of the strong effects plants exert on landscapes. Rates of complete success appeared to be higher for insect than plant targets (62% vs 27%), perhaps because most often herbivores gradually weaken, rather than outright kill, their hosts, which is not the case for natural enemies directed against pest insects. For both insect and plant biological control, nearly half of all projects reviewed were listed as currently in progress, suggesting that the use of biological control for the protection of wildlands is currently very active.

Summary Since the myrtle rust pathogen (Austropuccinia psidii) was first reported (as Puccinia psidii) in Brazil on guava (Psidium guajava) in 1884, it has been found infecting diverse myrtaceous species. Because A. psidii has recently spread rapidly worldwide with an extensive host range, genetic and genotypic diversities were evaluated within and among A. psidii populations in its putative native range and other areas of myrtle rust emergence in the Americas and Hawaii. Microsatellite markers revealed several unique multilocus genotypes (MLGs), which grouped isolates into nine distinct genetic clusters [C1–C9 comprising C1: from diverse hosts from Costa Rica, Jamaica, Mexico, Puerto Rico, and USA‐Hawaii, and USA‐California; C2: from eucalypts (Eucalyptus spp.) in Brazil/Uruguay and rose apple (Syzygium jambos) in Brazil; C3: from eucalypts in Brazil; C4: from diverse hosts in USA‐Florida; C5: from Java plum (Syzygium cumini) in Brazil; C6: from guava and Brazilian guava (Psidium guineense) in Brazil; C7: from pitanga (Eugenia uniflora) in Brazil; C8: from allspice (Pimenta dioica) in Jamaica and sweet flower (Myrrhinium atropurpureum) in Uruguay; C9: from jabuticaba (Myrciaria cauliflora) in Brazil]. The C1 cluster, which included a single MLG infecting diverse host in many geographic regions, and the closely related C4 cluster are considered as a “Pandemic biotype,” associated with myrtle rust emergence in Central America, the Caribbean, USA‐Florida, USA‐Hawaii, Australia, China‐Hainan, New Caledonia, Indonesia and Colombia. Based on 19 bioclimatic variables and documented occurrences of A. psidii contrasted with reduced sets of specific genetic clusters (subnetworks, considered as biotypes), maximum entropy bioclimatic modelling was used to predict geographic locations with suitable climate for A. psidii which are at risk from invasion. The genetic diversity of A. psidii throughout the Americas and Hawaii demonstrates the importance of recognizing biotypes when assessing the invasive threats posed by A. psidii around the globe.

Abstract. 1. Plants may compensate for the effects of herbivory, especially under favourable growing conditions, limited competition, and minimal topdown regulation. These conditions characterise many disturbed wetlands dominated by introduced plants, implying that exotic, invasive weeds in these systems should exhibit strong compensatory responses.2. The Australian native Melaleuca quinquenervia is highly invasive in the Florida Everglades, U.S.A., where it experiences limited competition or herbivory from native species, making it a likely candidate for compensation. The introduced biological control agent Oxyops vitiosa feeds exclusively on the seasonal flushes of developing foliage at branch apices, which represents &15% of the total foliar biomass.3. The hypothesis that M. quinquenervia compensates for folivory by O. vitiosa was tested in a series of field-based experiments. Trees experiencing folivory over four consecutive years maintained similar levels of foliar biomass after attack yet possessed twice the number of leaf-bearing terminal stems as undamaged trees. The biomass of these stems was similar among attacked and unattacked trees, indicating that herbivore-damaged trees produce greater quantities of smaller terminal branches. However, undamaged trees were 36 times more likely to reproduce than herbivore-damaged trees.4. In a separate herbivore exclusion study, a single bout of herbivory on previously undamaged M. quinquenervia trees caused an 80% reduction in reproductive structures the following season. Herbivore-damaged trees also possessed 54% fewer fruits than undamaged trees. An increase in the herbivory frequency (two bouts per year) or magnitude (100% simulated herbivory) did not result in a further reduction in fitness.5. It has been concluded that M. quinquenervia partially compensates for herbivory by producing new stems and replacing foliage, but this compensation results in a substantial reduction in reproduction.

No abstract