234 ECOLOGICAL MANAGEMENT & RESTORATION VOL 7 NO 3 DECEMBER 2006 © 2006 Ecological Society of Australia NOTES & ABSTRACTS GENETIC ISSUES & SOLUTIONS 18.2 Molecular markers detect multiple origins of Agonis flexuosa (Myrtaceae) plants used in urban bushland restoration. Elizabeth A. Sinclair, 1 John D. Bussell, 2 Siegfried L. Krauss, 3 Richard Hobbs, 4 and Kingsley W. Dixon 3 ( 1 School of Environmental Science, Murdoch University, Murdoch, WA 6150, Australia, and Botanic Gardens and Parks Authority, Fraser Ave, West Perth, WA 6005 Australia; 2 ARC Centre of Excellence in Plant Energy Biology, The University of Western Australia, Crawley, Western Australia, 6009, Australia; 3 Botanic Gardens and Parks Authority, Fraser Ave, West Perth, Western Australia, 6005, Australia, and School of Plant Biology, The University of Western Australia, Crawley, Western Australia, 6009, Australia; 4 School of Environmental Science, Murdoch University, Murdoch, WA 6150, Australia. Email: esinclair@iinet.net.au). Key words: AFLP, Agonis flexuosa, genetic marker, plant source, revegetation. Introduction. Conservation issues faced by urban bushland managers include habitat loss and degradation associated with urban expansion, and substantial reductions to the range and abundance of many species. Habitat fragmentation commonly increases bushland susceptibility to deterioration in species diversity, composition and ecosystem processes, and provides increased opportunities for the invasion of weedy species. Consequently, urban bushland areas often require some form of intervention. In highly degraded sites where additional planting may be required, the selection of plant material becomes an issue, as poor choice for (introduced) source material may compromise the long- term genetic integrity of the local native populations. The principle of using local seed in native plant restoration / rehabilitation activities is widely regarded as best-practice by the restoration industry, and is currently specified by Natural Heritage Trust (NHT), regional natural resource management organizations, Greening Australia, local bush management groups, and all levels of government (Mortlock 2000; Sackville Hamilton 2001; Carr 2005). Poor sourcing of material may potentially result in failure of restoration efforts, loss of biodiversity, and/or poor integra- tion with the remaining native vegetation. Identification of appropriate seed-source populations is important, particu- larly in Western Australia where population genetic differ- entiation within species is often pronounced (Coates 2000). The greatest limitation to the application of the principle of using local seed is in the identification of the extent of what is local (Mortlock 2000), and this will vary on a species-by- species basis. In this context, molecular markers can deter- mine population genetic structure, and more specifically in the case described here, assist in the identification of non- local plants used in a previous restoration effort. Peppermint tree (Agonis flexuosa Willd.) is a common and widespread species, typically found in the coastal dunes and limestone areas of the south-west corner of Western Australia between Perth and Albany. It has a fairly continuous coastal distribution; however, native popula- tions are rare and highly fragmented in the Perth metro- politan area (northern most extent of its range). This species is also a favoured plant for urban gardens and street verges, making differentiation between native and non-native populations difficult at some locations. The Peppermint Tree is a relatively long-lived, perennial species that can grow as a tree or in mallee form. Here, we generate DNA fingerprints to estimate the genetic relationship between three natural occurrences of the species including one at Bold Park, a 437-ha native bushland remnant in the western suburbs of Perth, as well as an occurrence of specimens apparently planted in a restoration zone at Bold Park approximately 20 years ago. Methods. Leaf samples were sampled from a total of 38 plants from three native bushland sites: Dawesville (32′37′′S 115′38′′E; n = 12), Preston Beach Road (32′54′′S 115′42′′E; n = 11), and Bold Park (31′57′′S 115′45′′E; n = 15). The collection sites outside Bold Park were the nearest native populations seen as potential seed sources; locations where replanting had occurred were avoided. Within Bold Park, another 46 plants were sampled from a rehabilitated area north of the native plants. Local Bold Park plants were clearly identifiable as large, multistemmed (mallee-like) trees, whereas the apparently planted trees were young, small, and most were single stemmed. Collected material was kept fresh at 4°C until DNA extraction. DNA was extracted using a modified cethyltrimethyl ammonium bromide (CTAB) procedure (He et al. 2004). Amplified fragment length polymorphism (AFLP) DNA fingerprints (Vos et al. 1995) were generated for each sample using fluorescently labelled primers (primer combi- nations were M-CTA/E-ACT, M-CTA/E-AGG, M-CAG/E-ACC) and an ABI PRSIM 377 automated sequencer, as described previously (Krauss 1999), except that restriction digests and ligations were performed in half the stated volume. AFLP fragments between 60 and 500 bp were scored for presence (1) or absence (0) for each sample with the aid of ABI PRISM  software. A principle coordinate analysis (PCA) was performed in  (Peakall & Smouse 2001) to visually represent the relative degree of genetic similarity among individuals and the distinction, if any, among sampling locations. Pairwise Fisher’s exact tests (Raymond & Rousset 1995) were performed to statistically assess the significance of genetic differentiation between pairs of sampling locations using the program Tools For Population Genetic Analysis (, & Miller 1997);