Contributed Paper Use of Pollen and Ancient DNA as Conservation Baselines for Offshore Islands in New Zealand JANET M. WILMSHURST, ∗ § NEVILLE T. MOAR, ∗ JAMIE R. WOOD, ∗ PETER J. BELLINGHAM, ∗ AMY M. FINDLATER,† JAMES J. ROBINSON,† AND CLIVE STONE‡ ∗ Landcare Research, P.O. Box 69040, Lincoln 7640, New Zealand †Department of Archaeology and Anthropology, University of Otago, P.O. Box 56, Dunedin 9054, New Zealand ‡Ng¯ atiwai Trust Board, P.O. Box 40 1332, Whangarei 0110, New Zealand Abstract: Islands play a key role globally in the conservation of endemic species. Many island reserves have been highly modified since human colonization, and their restoration and management usually occur without knowledge of their prehuman state. However, conservation paleoecology is increasingly being recognized as a tool that can help to inform both restoration and conservation of island reserves by providing prehuman vegetation baselines. Many of New Zealand’s mammal-free offshore islands are foci for biological diversity conservation and, like many islands in the Polynesian region, were deforested following initial human settlement. Therefore, their current restoration, replanting, and management are guided either by historic vegetation descriptions or the occurrence of species on forested islands. We analyzed pollen and ancient DNA in soil cores from an offshore island in northern New Zealand. The result was a 2000-year record of vegetation change that began >1200 years before human settlement and spanned 550 years of human occupation and 180 years of forest succession since human occupation ceased. Between prehuman and contemporary forests there was nearly a complete species turnover including the extirpation of a dominant conifer and a palm tree. The podocarp-dominated forests were replaced by a native but novel angiosperm-dominated forest. There is no modern analog of the prehuman forests on any northern New Zealand island, and those islands that are forested are dominated by angiosperms which are assumed to be climax forests. The pollen and DNA evidence for conifer- and palm-rich forests in the prehuman era challenge this climax forest assumption. Prehuman vegetation records can thus help to inform future restoration of degraded offshore islands by informing the likely rate and direction of successional change; helping to determine whether natural rates of succession are preferable to more costly replanting programs; and providing past species lists if restoration replanting is desired. Keywords: conservation paleoecology, Dacrydium, deforestation, extinctions, island restoration, pollen record, prehuman, Rhopalostylis Uso de Polen y ADN Antiguo como L´ ıneas de Base de Conservaci´ on en Islas Litorales en Nueva Zelanda Resumen: Las islas juegan un papel global clave en la conservaci´ on de especies end´ emicas. Muchas reservas insulares han sido modificadas significativamente desde la colonizaci´ on por humanos, y su restauraci´ on y manejo generalmente ocurren sin conocimiento de su condici´ on prehumana. Sin embargo, la paleoecolog´ ıa de conservaci´ on es reconocida cada vez mas como una herramienta que puede ayudar a informar la restauraci´ on y conservaci´ on de reservas insulares al proporcionar l´ ıneas de base de la vegetaci´ on prehumana. Muchas de las islas litorales de Nueva Zelanda libres de mam´ ıferos son polos para la conservaci´ on de la biodiversidad y, como muchas islas en la regi´ on Polinesia, fueron deforestadas despu´es de la colonizaci´ on por humanos. Por lo tanto, su restauraci´ on, reforestaci´ on y manejo son guiados ya sea por descripciones hist´ oricas de la vegetaci´ on o por la ocurrencia de especies en islas arboladas. Analizamos polen y ADN antiguo en muestras de suelo de una isla litoral en el norte de Nueva Zelanda. El resultado fue un registro de 2000 a˜ nos de cambios en la vegetaci´ on que comenzaron >1200 a˜ nos antes de la colonizaci´ on por humanos y abarcaron 550 a˜ nos de ocupaci´ on humana y 180 a˜ nos de sucesi´ on forestal desde que ces´ o la ocupaci´ on humana. §email wilmshurstj@landcareresearch.co.nz Paper submitted November 11, 2012; revised manuscript accepted May 6, 2013. 202 Conservation Biology, Volume 28, No. 1, 202–212 C  2013 Society for Conservation Biology DOI: 10.1111/cobi.12150