The Impact of the Weed Chrysanthemoides monilifera ssp. rotundata on Coastal Leaf Litter Invertebrates

In coastal areas of Australia, there are extensive infestations of the environmental weed Chrysanthemoides monilifera ssp. rotundata (bitou bush). This study looked at the impact of long-term infestations on the abundance and assemblage composition of leaf litter invertebrates. Assemblages were compared in weed infested and native shrublands along the New South Wales coastline over 12 months. The total abundance was not significantly reduced in the weedy habitat but the abundance of mites, thrips, spiders, ants, and centipedes was reduced at many sites. The invertebrate assemblages also differed between habitats, with the C. monilifera supporting a lower diversity of beetles. However, the millipedes, amphipods, earthworms, pseudoscorpions and isopods appeared to respond positively to the invasion, occurring in higher abundance and detected more frequently in the weedy areas. This has been partially attributed to a change in microclimate within the C. monilifera infestations. It is generally moister and darker, which these invertebrates tend to prefer. Secondly, C. monilifera produces less leaf litter of higher quality, and possibly higher palatability than the native sclerophyllous vegetation, which may encourage species that consume litter.     

The use of ants and other soil and litter arthropods as bio-indicators of the impacts of rainforest clearing and subsequent land use

The present study investigated the impacts of rainforest clearance, and associated subsequent land␣use for pasture, on assemblages of soil and litter arthropods in eastern subtropical Australia. We assessed the utility of soil and litter arthropods as potential bio-indicators of cleared and forested habitats. Arthropods were sampled from 24 sites (12 sites each in rainforest and pasture) using two methods (extraction from litter, pitfall traps). Responses of taxa were analysed at various levels of taxonomic resolution, including ‘coarse’ arthropods (all arthropods sorted to Order/Class), ant genera and ant species. Multivariate analyses of arthropod composition indicated that an increase in the level of taxonomic resolution did not provide a commensurate increase in the sensitivity of assemblage response. Indicator values (IndVals), computed for each taxon, showed that a number of arthropod taxa may have potential as bio-indicators of habitat change. However the use of many of these, especially many ant species found in our study, may be unreliable because even after extensive numbers of sites were sampled, most species showed patchy distributions. To overcome this problem, we generated ‘composite indices’, by combining information from sets of indicator taxa. The utility of these composite indices is discussed.     

Rainforest litter invertebrates decimated by high severity burns during Australia's gigafires

Climate change is increasing the frequency of extreme fires. In 2019–2020, extreme fires burned 97 000 km² of native vegetation in south-eastern Australia, affecting many areas of rainforest, which has historically burned less frequently. One year post-fires, we surveyed litter macroinvertebrates in 52 temperate rainforest sites. Sites had experienced increasing levels of fire severity (unburnt, medium severity and high severity). We asked how fire severity affected: (1) litter macroinvertebrate habitats; (2) the abundance of litter macroinvertebrate taxa per unit area; and (3) abundance relative to litter habitat (volumetric density). We also estimated the loss of litter macroinvertebrates across rainforests in the study region. High severity burns supported only a fifth of the litter volume and canopy cover as unburnt sites, lower soil moisture and higher herb cover. Medium burns were intermediate. Macroinvertebrate abundance declined with burn severity: high severity burns supported only 26% of the abundance in unburnt sites; medium severity burns supported 80% of that in unburnt sites. Patterns were similar for all taxa, with millipedes declining most. High severity fires resulted in up to 1.90 million fewer macroinvertebrates per hectare; 0.53 million fewer per hectare of medium burn rainforest. Across the study region, we estimate that 60 billion fewer litter macroinvertebrates persisted in temperate rainforests alone. Volumetric densities of many litter macroinvertebrate taxa in high severity burns were marginally higher than in unburnt sites, suggesting nutrients may be more available post-fire, or that persisting individuals become concentrated in the leaf litter. For less desiccation-tolerant groups (e.g., amphipods), density declines with increasing severity may reflect the combined impact of low soil moisture and reduced litter cover. Many taxa persisted following high severity fires, but declines were substantial, and taxa differed in their vulnerability. Longer-term monitoring is required to understand the recovery trajectory and impacts on ecological function.     

不同演替阶段热带森林地表凋落物和土壤节肢动物群落特征

为了解不同演替阶段热带森林土壤节肢动物群落结构特征及其与地表凋落物的关系, 2001年9月采用样线调查法对西双版纳23年次生林、35年次生林、季节雨林地表凋落物及其中的土壤节肢动物进行了调查。所获数据表明, 地表凋落物数量(现存量干重)和质量(N和C/N)总体上表现为35年次生林最好, 23年次生林次之; 蜱螨目和弹尾目为3林地地表凋落物土壤节肢动物群落优势类群, 膜翅目蚂蚁、马陆目、鞘翅目、双翅目和半翅目为常见类群。土壤节肢动物个体密度和个体相对密度均表现为35年次生林>季节雨林>23年次生林。群落的丰富度指数以季节雨林最高, 多样性和均匀度指数显示为23年次生林最高, 35年次生林的优势度指数最高, 3林地土壤节肢动物群落类群组成相似性达到较好水平。相关分析表明, 3种不同演替阶段热带森林土壤节肢动物个体密度与林地地表凋落物现存量呈正相关, 而现存凋落物N元素储量与土壤节肢动物的相关性仅表现在23年次生林和季节雨林。研究认为, 热带森林土壤节肢动物群落的发展与森林植被演替密切相关, 其群落个体数量和多样性受森林地表凋落物数量、质量的调控, 但其他环境因素如捕食效应、人为干扰等影响亦不可忽视。     

Litter and soil arthropod colonization in reforested semi‐deciduous seasonal Atlantic forests

The recovery of soil ecological processes during the restoration of tropical forests is greatly influenced by arthropods that live in the litter and soil. However, these communities present complex dynamics, and their colonization patterns are not well understood. In this study, we examined the response patterns of litter and soil arthropods to the ecosystem regeneration process by assessing reforestation sites from two regions of São Paulo State, Brazil, and we compared the data obtained from these sites with data from mature forests. We assessed the arthropod communities using similarity indices and high‐level taxa abundance, with the level of forest succession and the locations of the restoration areas as factors. Forest succession correlated with the species composition as communities from the reforestation sites gradually became more similar to communities from the mature forests, while their quantitative patterns were minimally related. Forest maturation positively affected the richness of the litter community and the abundance of some minor groups, such as Protura, Diplura, and Symphyla. The region influenced the species composition but did not influence the manner in which the communities changed during the maturation process. We also found a convergent soil colonization pattern as arthropod communities from different sites became more similar during forest succession. This finding is consistent with both empirical data and theoretical predictions from the specialized literature, although the subject has been poorly explored until now. We conclude that reforestation allows the colonization of soil and litter fauna in a biased manner.     

Biodiversity recovery during rainforest reforestation as indicated by rapid assessment of epigaeic ants in tropical and subtropical Australia

There is growing interest in the potential for reforestation to assist the recovery of rainforest biodiversity. There is also a need to identify taxonomically tractable groups for use as cost‐effective indicators when monitoring the status of biodiversity within reforested sites. Insects are an important component of terrestrial biodiversity but often require considerable resources to sample at species level. Ant genera and generic‐based functional groups have been suggested as possible indicators of environmental disturbance. Here we ask to what extent the development of biodiversity is indicated by epigaeic ant genera and functional groups, across different types of reforestation in tropical and subtropical Australia. In each region, we used pitfall traps to sample the ants in replicate sites of: unmanaged regrowth, monoculture and mixed species plantations and ‘ecological restoration’ plantings, together with reference sites in pasture and rainforest. We recorded 35 epigaeic ant genera (and 4623 individuals) from 50 tropical sites, and 39 genera (and 9904 individuals) from 54 subtropical sites, with 47 genera overall. Community composition of both genera and functional groups differed between pasture and rainforest, although many genera were widespread in both. Reforested sites were intermediate between pasture and rainforest in both regions, and showed a gradient associated with decreasing grass and increasing tree and litter cover. Older monoculture plantations and ecological restoration plantings had the most rainforest‐like ant assemblages, and mixed‐species cabinet timber plots the least, of the reforested sites. We conclude that ground‐active ant genera and functional groups sampled in rapid surveys by pitfall‐trapping showed only a modest ability to discriminate among different types of reforestation. Species‐level identification, perhaps together with expanded sampling effort, could be more informative, but would require resourcing beyond the scope of rapid assessments.     

Cross-taxon congruence of α and β diversity among five leaf litter arthropod groups in Colombia

In this study α and β diversity patterns of five leaf litter arthropod groups (ants, predatory ants, oribatid mites, spiders and other arachnids) were described and compared in 39 sampling patches of a transformed landscape in southwestern Colombia, that represented five vegetation types: secondary forest, riparian forest, giant bamboo forest, pasture and sugarcane crop. It was also assessed whether some taxa could be used as diversity surrogates. A total of 6,765 individuals grouped in 290 morphospecies were collected. Species richness in all groups was lower in highly transformed vegetation types (pasture, sugarcane crop) than in native ones (forests). In contrast, there were no clear tendencies of β diversity among vegetation types. Considering sampling patches, 0.1–42% of the variation in α diversity of one taxonomic group could be explained from the α diversity of another, and 0.2–33% of the variation of β diversity of a given taxon was explained by that in other groups. Contrary to recent findings, we concluded that patterns of α diversity are more congruent than patterns of β diversity. This fact could be attributed to a sampling effect that promotes congruence in α diversity and to a lack of a clear regional ecological gradient that could promote congruent patterns of β diversity. We did not find evidence for an ideal diversity surrogate although diversity patterns of predatory ants had the greatest congruencies. These results support earlier multi-taxon evaluations in that conservation planning should not be based on only one leaf litter arthropod group.     

Linking ecological function to species composition in ecological restoration: Seed removal by ants in recreated woodland

Evaluations of ecological restoration typically focus on associating measures of structural properties of ecosystems (e.g. species diversity) with time since restoration efforts commenced. Such studies often conclude a failure to achieve restoration goals without examining functional performance of the organism assemblages in question. We compared diversity and composition of ant assemblages and the rates of seed removal by ants in pastures, 4‐ to 10‐year old revegetated areas and remnants of Cumberland Plain Woodland, and an endangered ecological community in Sydney, Australia. Ant assemblages of forest remnant sites had significantly higher species richness, significantly different species composition and a more complex functional group structure in comparison with ant assemblages of pasture and revegetated sites, which did not differ significantly. However, the rates of seed removal by ants in revegetated sites were similar to those in forest remnants, with the rates in pasture sites being significantly lower. Approximately, one‐third of all ant species were observed to remove seeds. Forest remnant sites had significantly different assemblages of seed removing ant species from those in pasture and revegetated sites. These results demonstrate that similar ant assemblages of unrestored and restored areas can function differently, depending on habitat context. Evaluation of restoration success by quantifying ecosystem structure and function offers more insights into ecosystem recovery than reliance on structural data alone.     

Responses of ground-active beetle assemblages to different styles of reforestation on cleared rainforest land

Despite increasing efforts to re-establish forest cover in landscapes that have been previously cleared, the relative ability of different styles of reforestation to contribute to conservation and support forest biota is poorly known, particularly for invertebrates. We investigated the use of different types of reforested habitat by ground-active rainforest beetle assemblages on land, which had been previously cleared of rainforest, in the tropics and subtropics of eastern Australia. Between five and ten replicate sites within each of five reforestation styles were selected in each region: un-managed regrowth, young mono-species timber plantations, young mixed-species timber plantations, ecological restoration plantings, and old mono-species timber plantations, together with reference sites in pasture and in intact rainforest. Ground-active beetles were sampled using pitfall traps, and assemblages were compared among site-types. In both regions, beetle assemblages in all styles of reforestation were intermediate in species composition between pasture and rainforest. The similarity of beetle assemblages to intact rainforest increased with the age and structural complexity of reforested sites. The most rainforest-like beetle assemblages were from older reforestation sites (38–70 year plantations in tropics, and 30–40 year regrowth in subtropics) and in younger (6–22 years) but floristically and structurally diverse ecological restoration plantings in both regions. Assemblages in younger (5–20 year) sites of regrowth, mono-species timber plantations, and mixed-species timber plantations were more similar to pasture than rainforest. We conclude that achieving high canopy cover and sufficient structural complexity are important factors associated with the restoration of rainforest-like beetle assemblages to reforested sites.     

DNA metabarcoding uncovers dispersal-constrained arthropods in a highly fragmented restoration setting

Degraded areas are often restored through active revegetation; however, recolonization by animals is rarely engineered. Recolonization may be rapid for species with strong dispersal abilities. However, poor dispersers, such as many flightless arthropods, may struggle to recolonize newly restored sites. Actively reintroducing or “rewilding” arthropods may therefore be necessary to facilitate recolonization and restoration of arthropod communities and the ecological functions they perform. However, active interventions are rare. The purpose of this study was twofold. First, we asked whether potential source remnant arthropod communities were dispersal-constrained and struggling to recolonize restoration sites. Second, we tested whether reintroducing entire arthropod communities from remnant populations would help dispersal-constrained species establish during farmland ecological restoration in southern Australia. Rewilding was conducted in summer 2018 by transplanting leaf litter, soil, and entire communities contained within it from remnant source populations into geographically isolated restoration sites, which were paired with untreated controls (n = 6 remnant, rewilding transplant, and control sites). We collected leaf litter and extracted arthropod communities 19 months after the initial rewilding event, then sequenced mite, springtail, and insect communities using a metabarcoding approach. Within all groups, community similarity decreased with spatial distance between sites, suggesting significant dispersal barriers. However, only mite communities showed a strong response to rewilding, which was expressed as increased compositional similarity toward remnant sites and greater species richness relative to controls. Our results demonstrate that many arthropod species may struggle to recolonize geographically isolated restoration sites and that full community restoration requires active interventions via rewilding.     

Secondary succession, soil formation and development of a diverse community of oribatids and saprophagous soil macro-invertebrates

Investigations into different stages of secondary succession (from a wheat field to a beechwood on Threstone; Northern Germany) demonstrated the formation of a carbon rich top soil in later successional stages. Parallel to changes in plant species and soil formation, there were also changes in species composition and diversity of saprophagous macro-invertebrates (Lumbricidae, Diplopoda, Isopoda) and oribatid mites (Acari: Oribatida). Diversity of diplopod and isopod species increased after cessation of cultivation, but in a late successional stage (ca 50 y-old fallow, ash-dominated wood) species number of diplopods and isopods declined strongly. In comparison with the other soil invertebrate groups, species composition of earthworms among the sites was more similar. Accumulation of soil C was assumed to be related to wood formation and occurrence of woody debris and recalcitrant leaf litter of beech trees. Incorporation of recalcitrant litter materials by earthworm species living in the upper mineral soil presumably contributed significantly to accumulation of soil C. Accumulation of soil C was accompanied by the development of an oribatid mite community rich in species. In early successional stages oribatids predominantly colonized the litter layer, while most oribatid mites of the beechwood inhabited the upper mineral soil. Maximum diversity of oribatid mites in the beechwood is assumed to be related to instability of the mineral soil caused by earthworm activity. Changes in species composition and diversity are discussed considering succession theory. Even soil invertebrates of similar trophic groups appear to respond very differently to successional changes. It is concluded that conservation strategies to maintain high diversity of soil invertebrates are most likely to be successful if a wide range of habitats of different successional stages is included.     

Trees as templates for tropical litter arthropod diversity

Increased tree species diversity in the tropics is associated with even greater herbivore diversity, but few tests of tree effects on litter arthropod diversity exist. We studied whether tree species influence patchiness in diversity and abundance of three common soil arthropod taxa (ants, gamasid mites, and oribatid mites) in a Panama forest. The tree specialization hypothesis proposes that tree-driven habitat heterogeneity maintains litter arthropod diversity. We tested whether tree species differed in resource quality and quantity of their leaf litter and whether more heterogeneous litter supports more arthropod species. Alternatively, the abundance–extinction hypothesis states that arthropod diversity increases with arthropod abundance, which in turn tracks resource quantity (e.g., litter depth). We found little support for the hypothesis that tropical trees are templates for litter arthropod diversity. Ten tree species differed in litter depth, chemistry, and structural variability. However, the extent of specialization of invertebrates on particular tree taxa was low and the more heterogeneous litter between trees failed to support higher arthropod diversity. Furthermore, arthropod diversity did not track abundance or litter depth. The lack of association between tree species and litter arthropods suggests that factors other than tree species diversity may better explain the high arthropod diversity in tropical forests.     

陆地生态系统中马陆的生态功能

马陆是陆地生态系统中物种多样性极高的大型土壤无脊椎动物类群。作为营腐生动物, 马陆在陆地生态系统中具有不可替代的重要功能。通过大量取食及随后的肠道过程, 马陆在很大程度上决定着陆地生态系统凋落物的破碎、转化和分解过程, 从而驱动碳和关键养分元素的循环周转。然而, 目前对马陆生态功能的研究还非常有限, 远远落后于其他土壤动物类群(如蚯蚓等)。本文初步总结了马陆的生态功能: (1)通过破碎、取食凋落物来加速凋落物的分解。马陆偏好取食半分解的凋落物, 其同化效率受到凋落物来源、温度和凋落物中微生物含量的影响。(2)主要通过取食和排泄等活动影响养分循环。但对于马陆如何影响土壤碳循环, 存在两种不同的观点: 一是马陆粪球的分解速率比凋落物更快, 加速了碳的循环; 二是马陆粪球更难分解, 有助于碳的固存和稳定。马陆破碎凋落物后, 凋落物释放氮素进入土壤。此外, 马陆的活动也影响土壤磷的循环, 提高土壤中有效磷的含量。(3)调控微生物特性, 与蚯蚓也有互作关系。通过以上三个方面的总结, 展望了未来马陆的主要研究方向, 以期引起更多思考和研究。     

Community structures of Mesostigmata, Prostigmata and Oribatida in broad-leaved regeneration forests and conifer plantations of various ages

The community structures of Mesostigmata, Prostigmata, and Oribatida in the soil of broad-leaved regeneration forests and conifer plantations of various ages were assessed alongside soil and plant environmental variables using three response metrics (density, species richness, and species–abundance distribution). The density and species richness of mites recovered swiftly after clear-cutting or replanting. Oribatid mites dominated the soil mite communities in terms of densities and species richness for both forest types. Soil mite communities in broad-leaved forests was related to forest age, the crown tree communities index, and forest-floor litter weight. In contrast, soil mite communities in the conifer plantation sites were related to various indices of understory plants. The development of the understory plants was synchronized with the silvicultural schedules, including a closed canopy and thinning. Such a conifer plantation management may affect indirectly the community of mites.     

Spore dispersal of a resupinate ectomycorrhizal fungus, Tomentella sublilacina, via soil food webs

Patterns of fungal spore dispersal affect gene flow, population structure and fungal community structure. Many Basidiomycota produce resupinate (crust-like) basidiocarps buried in the soil. Although spores are actively discharged, they often do not appear to be well positioned for aerial dispersal. We investigated the potential spore dispersal mechanisms of one exemplar of this growth form, Tomentella sublilacina. It is a widespread ectomycorrhizal fungus that sporulates in the soil organic horizon, can establish from the spore bank shortly after disturbance, but also can be a dominant species in mature forest stands. We investigated whether its spores could be dispersed via spore-based food webs. We examined external surfaces, gut contents and feces from arthropod fungivores (mites, springtails, millipedes, beetles, fly larvae) and arthropod and vertebrate predators (centipedes, salamanders) from on and around T. sublilacina sporocarps. Spore densities were high in the guts of many individuals from all fungivore groups. Centipede gut contents, centipede feces and salamander feces contained undigested invertebrate exoskeletons and many apparently intact spores. DAPI staining of spores from feces of fungivores indicated that 7-73% of spores contained intact nuclei, whereas spores from predators had lower percentages of intact nuclei. The spiny spores often were lodged on invertebrate exoskeletons. To test the viability of spores that had passed through invertebrate guts we used fecal droppings of the millipede Harpaphe haydeniana to successfully inoculate seedlings of Pinus muricata (Bishop pine). These results indicate the potential for T. sublilacina spore dispersal via invertebrates and their predators in soil food webs and might help to explain the widespread distribution of this species. It is likely that this is a general mechanism of dispersal for fungi producing resupinate sporocarps, indicating a need to develop a fuller understanding of the linkages of soil food webs and spore dispersal.     

Hay-bait traps are a useful tool for sampling of soil dwelling millipedes and centipedes

Some species of centipedes and millipedes inhabit upper soil layers exclusively and are not recorded by pitfall trapping. Because of their sensitivity to soil conditions, they can be sampled quantitatively for evaluation of soil conditions. Soil samples are heavy to transport and their processing is time consuming, and such sampling leads to disturbance of the soil surface which land-owners do not like. We evaluated the use of hay-bait traps to sample soil dwelling millipedes and centipedes. The effectiveness of this method was found to be similar to the effectiveness of soil sampling. Hay-bait traps installed for 8–10 weeks can substitute for direct soil sampling in ecological and inventory studies.     

Mites in the mist: How unique is a rainforest canopy-knockdown fauna?

Abstract We intensively sampled the parasitiform mite faunas of three subtropical rainforest canopy habitats (leaves, bark, hanging humus) and three forest floor habitats (leaf litter, fungal sporocarps, arthropod associates) in the Green Mountains section of Lamington National Park, Queensland, and compared them to 423 specimens collected from the canopy by pyrethrin knockdown (PKD). In total, 165 species (80% new to science) were identified, including 58 from PKD. Few species occurred in more than one habitat, and complementarity averaged 96 ± 1%. About half of the species from PKD were found in canopy habitats, but less than 10% occurred in forest floor habitats. Thus, the canopy fauna is composed primarily of canopy specialists, not of forest floor mites, and habitat specificity is a major component of acarine diversity. An Incidence-based Coverage estimator (generated by the EstimateS program) proved useful in predicting asymptotes for collector's curves. Conservative extrapolations from this study suggest that more than 2000 species of mites live in subtropical rainforest in the Green Mountains.     

Current and Potential Roles of Soil Macroinvertebrates (Earthworms,Millipedes, and Isopods) in Ecological Restoration

Soil macroinvertebrates have a considerable impact on soil functions important to the restoration process, such as decomposition; yet, soil organisms have received relatively little attention in restoration ecology in terms of their applicability as agents of restoration. Here, we review how large obligate detritivores (earthworms, millipedes, and isopods) have been used to accomplish restoration goals, assess restoration progress, and function as bioindicators. Patterns of detritivore community succession and how these communities are themselves restored during restoration of perturbed ecosystems are also discussed. We conclude with a discussion of increased utilization of these taxa in future and ongoing restoration projects as well as the outreach activities that should be associated with use of these organisms.     

Essential and alternative prey in a ponerine ant: variations according to the colony life cycle

We studied the prey specialization of Plectroctena minor, a ponerine ant known to capture mostly millipedes. We compared the prey spectrum of the hunting workers from large colonies with that of the founding queens. The hunting workers captured all kinds of tested prey, but hunted mostly millipedes. Founding queens, which avoided relatively large prey, including the millipedes tested, captured mostly isopods under experimental conditions. We also verified that the presence of millipedes in the diet of the larvae of large colonies was necessary for the production of winged females and strongly enhanced the production of workers, permitting us to assert that P. minor is a predatory species specialized in the capture of millipedes. In contrast, the presence of millipedes had no impact on the production of males. We thus assert that millipedes constitute the 'essential prey' of P. minor, while other arthropod taxa are therefore 'alternative prey'.     

Potential of ants and beetles as indicators of rainforest restoration: characterising pasture and rainforest remnants as reference habitats

Assessment of habitat restoration requires baseline information on the communities present in both converted and intact forms of the focal ecosystem to enable comparisons with restored sites. Ants and beetles are commonly used in ecological monitoring programmes, as they display assemblage‐level responses to habitat change and can be a more direct measure of the recommencement of some ecosystem functions than the presence of more obvious biota such as plants. However, as these taxa differ substantially in ecological traits, their response patterns and utility as potential bioindicators may vary. Using pitfall traps, we compared assemblages of ant and beetle species between two reference habitats, pasture and remnant rainforest in subtropical eastern Australia. The assemblage composition of both groups differed significantly between rainforest and pasture but only beetles showed accompanying differences in species richness and abundance, which were both significantly lower in pasture. We identified ant and beetle species characteristic of either pasture or rainforest remnants, which may be used as bioindicators in future monitoring programmes. These species, however, displayed patchy distributions, suggesting that the use of individual species as bioindicators is likely to be unreliable. These findings support the use of ‘composite habitat indices’, which combine information from sets of indicator species. Given that patterns of change in species composition were similar between ants and beetles, either is an appropriate focal taxon for future monitoring programmes. Beetles, however, displayed some limitations as no species were indicative of the disturbed pasture habitat. Ants and beetles are likely to respond in different ways to different aspects of habitat change; thus, using both together could strengthen assessments of rainforest degradation or recovery.