Hunter Valley Weeping Myall Woodland in the Sydney Basin Bioregion should remain listed as a Threatened Ecological Community until strong evidence emerges in support of delisting

Determinations for the listing of Ecological Communities made under the NSW Threatened Species Conservation (TSC) Act (1995) draw on the best information available at the time of listing. These data may either be qualitative or quantitative, and with or without analytical evaluation. Periodic reappraisal of the definition or status of a community may therefore be required as new data or analyses become available. In this respect, the contributions of Bell and Driscoll (2014, 2016) constitute a commendable commitment to the application of quantitative survey data to the question of how (or if) the Hunter Valley Weeping Myall Woodland should be defined. In our response to their proposition that Acacia pendula was introduced to the Hunter Valley following European settlement (Tozer and Chalmers 2015), we argued that: i) the definition of a community should not be based on the presence of an individual species; and ii) irrespective of the status of Acacia pendula, there is evidence to support the definition of Hunter Valley Weeping Myall Woodland as an assemblage characterised by semi-arid species occupying hot and dry parts of the Hunter Valley. We believe that there are several weaknesses in the counter-arguments presented by Bell and Driscoll (2016). We outline these below before briefly elaborating on how the ordination of floristic data may be used to progress toward a more refined understanding of the composition and distribution of this Critically Endangered Community. Bell and Driscoll (2016) contend that the assemblage of species listed in the Final Determination does not constitute a valid definition of Hunter Valley Weeping Myall Woodland because the species are widespread in a range of communities and thus show no fidelity to the community as listed. We concede that in the absence of a set of quantitative data sampled across the range and extent of the community, the nominated list of species is unlikely to be both optimal and comprehensive, and can be expected to evolve as suitable data become available. We point out, however, that the existence of such a sample is a prerequisite to the determination of fidelity, and thus the lack of fidelity alluded to by Bell and Driscoll (2016) is asserted rather than demonstrated. More importantly, their argument suggests a misunderstanding of the community concept as currently applied to vegetation classification. The contemporary application is informed by a continuum model, under which species are understood to vary in abundance along environmental gradients with some degree of independence (Begon et al. 2006). As a consequence of this: communities are expected to exhibit variability in the assemblage of species present in different locations ; boundaries between different communities are likely to be vague ; there is likely to be much overlap in the species membership between different communities, and ; communities can rarely be identified conclusively based on their dominant species alone.