The ancient DNA revolution: the latest era in unearthing New Zealand’s faunal history

ABSTRACT

In the 25 years since the first DNA sequences were obtained from the extinct moa, ancient DNA analyses have significantly advanced our understanding of New Zealand's unique fauna. Here, we review how DNA extracted from ancient faunal remains has provided new insights into the evolutionary histories and phylogenetic relationships of New Zealand animals, and the impacts of human activities upon their populations. Moreover, we review how ancient DNA has played a key role in improving our ability to taxonomically identify fragmentary animal remains, determine biological function within extinct species, reconstruct past faunas and communities based on DNA preserved in sediments, resolve aspects of the ecology of extinct animals and characterising prehistoric parasite faunas. As ancient DNA analyses continue to become increasingly applied, and sequencing technologies continue to improve, the next 25 years promises to provide many more exciting new insights and discoveries about New Zealand's unique fauna.

KEYWORDS:

• Archaeology
• biogeography
• birds
• evolutionary history
• palaeontology
• phylogenetics
• Quaternary

Introduction

The unique evolutionary histories of New Zealand's vertebrate animals and their variety of unusual ecological adaptations have long been the subject of scientific intrigue and examination (Fleming 1979; Diamond 1990). Recently, molecular evidence has played an increasingly important role in helping reconstruct the history of New Zealand's fauna (Goldberg et al. 2008; Tennyson 2010; Gibbs 2016), and has allowed the identification of vicariant taxa with great antiquity, potentially dating to the separation of Zealandia and Gondwana (e.g. Allwood et al. 2010), together with lineages that dispersed to New Zealand more recently (e.g. Trewick 2000). Molecular studies have also helped resolve the timing of radiations within different faunal groups (e.g. Waters & Wallis 2001; Chinn & Gemmell 2004; Chapple et al. 2009), providing clues about the drivers of speciation for these taxa.

New Zealand is unusual among larger landmasses in that a relatively large proportion of its recent fauna has become extinct in just the last few hundred years, specifically since initial human settlement c.750 years ago (Worthy & Holdaway 2002; Wood 2013; Perry et al. 2014). In most cases these extinct species leave behind only fragmentary skeletal remains with which to piece together and understand their biology and evolutionary histories. Molecular analysis provides an important tool for studying these extinct organisms. With a temperate climate ideal for preserving DNA, numerous late Quaternary fossil deposits, and relatively recent human settlement, extinct animals from New Zealand have played an important role in the development of ancient DNA research globally. For example, moa (Dinornithiformes) remains yielded the first DNA sequences (Cooper et al. 1992) and the first mitochondrial genomes (Cooper et al. 2001) of extinct birds, the first sex-linked nuclear DNA sequences from an extinct species (Bunce et al. 2003; Huynen et al. 2003) and the first microsatellites from an extinct bird (Allentoft et al. 2009).

It is now 25 years since the first DNA sequences from extinct New Zealand species were published, shedding light on the evolutionary relationships between moa and kiwis (Cooper et al. 1992). Since then, the study of DNA from ancient specimens has opened a new window into New Zealand's past. In this article we review studies that have used ancient DNA to better understand New Zealand's faunal history. The results of these studies have overturned long-held views about the evolutionary histories of some New Zealand taxa, resolved the antiquity and uniqueness of various endemic animals, and have even surprisingly revealed the existence of previously unknown prehistoric species.

A primer on ancient DNA

Several detailed reviews have been written on the topic of ancient DNA research (e.g. Hofreiter et al. 2001; Pääbo et al. 2004; Willerslev & Cooper 2005; Rizzi et al. 2012; Shapiro & Hofreiter 2014; Leonardi et al. 2017; Grealy et al. 2017), and so we restrict ourselves here to providing a brief summary of the key points. Ancient DNA is a term used to describe genetic material preserved in degraded post-mortem biological remains, including those of extinct species. Without the enzymatic repair mechanisms present in living cells, DNA begins to degrade after death (Lindahl 1993); nucleases cleave the DNA backbone into fragments (Briggs et al. 2007), microorganisms digest DNA during the decomposition process (Pääbo 1989; Lindahl 1993), hydrolysis of amino groups and oxidative reactions result in depurination causing further strand cleavage (Lindahl & Nyberg 1972; Briggs et al. 2007) and Maillard reactions create crosslinks that inhibit DNA amplification (Willerslev & Cooper 2005). DNA extracted from ancient specimens is therefore typically characterised by low endogenous concentration (i.e. high proportion of contaminant DNA), short fragment lengths, presence of miscoding lesions, and polymerase chain reaction (PCR) inhibition.

The first study to report an ancient DNA sequence involved the New Zealand born researcher Allan Wilson, and used dried tissues from a museum specimen of the extinct quagga (Equus quagga quagga) (Higuchi et al. 1984). With the subsequent development of the PCR technique (Mullis & Faloona 1987) researchers were provided with a powerful new tool for studying DNA from intractable specimens, including ancient remains (e.g. Pääbo 1989; Pääbo et al. 1989). The limits of ancient DNA retrieval began to be tested, and studies soon reported successful sequencing of DNA from geologically ancient (e.g. Miocene and Cretaceous) specimens (e.g. Golenberg et al. 1990; DeSalle et al. 1992; Woodward et al. 1994). Although these so-called ‘antediluvian’ sequences are now known to have been the results of contamination (Lindahl 1993; Hofreiter et al. 2001), they raised awareness about the need for authenticity criteria within the rapidly growing field of ancient DNA research. Such criteria include performing DNA extractions and pre-amplification steps in an ultra-clean environment that is physically isolated from any PCR work, independent replication of results, and checking for damage patterns characteristic of ancient DNA (Cooper & Poinar 2000; Gilbert et al. 2005; Skoglund et al. 2014).

In the mid-2000s, the field of ancient DNA research was boosted by the development of next-generation sequencing technologies. Prior to this time most workflows had relied upon PCR amplification using locus-specific primers, followed by Sanger sequencing, which despite having a high accuracy is a relatively restrictive approach because of the typically short DNA fragment lengths (i.e. often < 100 bp) in ancient specimens. The 2005 release of the Roche 454 Next-Generation Sequencing (NGS) platform (Margulies et al. 2005) was a major advance for the field, and facilitated the sequencing of 13 million base pairs of mammoth DNA (Poinar et al. 2006). High-throughput DNA sequencing, such as is now available via a suite of different platforms (e.g. Illumina, PacBio, Ion Torrent), allows sequencing of DNA libraries for a range of applications relating to ancient fauna (Figure 1), including palaeogenomics (Heintzman et al. 2015), metagenomics (Slon et al. 2017) and metabarcoding (Grealy et al. 2015). Methods for efficiently retrieving very short DNA fragments (e.g. Dabney et al. 2013), and techniques such as hybridisation capture, which can enrich libraries in DNA from the species of interest (e.g. Mitchell, Llamas, et al. 2014), are also particularly well suited to maximising the potential of ancient DNA.

Figure 1. Examples of substrates from which ancient DNA sequences of New Zealand fauna have been recovered. A, Naturally desiccated tissues (e.g. moa; Cooper et al. 1992); B, bones (e.g. yellow-eyed penguin; Boessenkool et al. 2009); C, museum skins (e.g. laughing owl; Wood et al. 2017); D, sediments (e.g. cave sediment; Haile et al. 2007); E, eggshell (e.g. moa; Oskam et al. 2011); F, feathers (e.g. moa; Rawlence et al. 2009); G, coprolites (e.g. moa; Wood et al. 2008).

DNA from ancient fauna has now been recovered from a wide range of remains and substrates, including bone (Dabney et al. 2013), hair (Rasmussen et al. 2010), feathers (Rawlence et al. 2009), muscle (Cooper et al. 1992), insect exoskeletons (King et al. 2009), parasite eggs (Loreille et al. 2001), coprolites (Wood et al. 2008), sediment (Willerslev et al. 2003), ice cores (Willerslev et al. 2007) and avian eggshell (Oskam et al. 2010). Coupled with different analytical approaches, ancient DNA provides a powerful tool in molecular biology, and has been used for a wide range of applications such as modelling DNA kinetics (Allentoft et al. 2012), exploring the relationships between environmental change and biodiversity (Willerslev et al. 2007, 2014), detailing the diet of extinct animals (Miller et al. 2005; Wood et al. 2008; Willerslev et al. 2014), examining past population structuring and barriers to gene flow (e.g. Bunce et al. 2009; Rawlence et al. 2012), characterising sexual dimorphism and sex biases (Bunce et al. 2003; Huynen et al. 2003; Allentoft et al. 2009), inferring archaic and modern hominid dispersal patterns (Green et al. 2010; Thomson et al. 2014), clarifying phylogenetic relationships (Austin et al. 2013; Orlando et al. 2013; Mitchell, Llamas, et al. 2014), assessing population genetics (Brace et al. 2012) and elucidating evolutionary developmental biology (Huynen et al. 2014).

Evolutionary histories and phylogenetic relationships

Ancient DNA from many extinct New Zealand avian taxa has contributed to our understanding of their evolutionary histories and deep phylogenetic relationships, and very often resulted in taxonomic changes. Of these taxa, the moa have been the most intensively studied, and in-depth reviews of how ancient DNA has contributed to our understanding of moa are provided by Allentoft & Rawlence (2012) and Worthy & Scofield (2012). Here, we examine other New Zealand taxa for which ancient DNA has provided insights into their deep history.

Kiwi (Apterygidae)

Until recently, the flightless ratite birds have long been regarded as a key example of a faunal assemblage split by ancient plate tectonic processes, with divergent taxa on different Southern Hemisphere continents following the Cretaceous break up of Gondwana (Cracraft 1972). Under such a model, flightlessness was assumed to have evolved only once in ratites, and species inhabiting the same landmass (e.g. kiwi and moa) were thought to be most closely related to one another (Cracraft 1972). However, the first ancient DNA sequences from moa appeared to contradict this, showing instead that kiwi were more closely related to the Australian ratites (emu and cassowary) (Cooper et al. 1992). The idea that ratites evolved from a single flightless ancestor was ultimately shown to be incorrect by Phillips et al. (2010), who demonstrated that the volant tinamou were sister to moa and nested within ratites. The exact phylogenetic relationships of kiwi were finally resolved by Mitchell, Llamas, et al. (2014), who sequenced mitochondrial genomes of Madagascan elephant birds (Aepyornithidae) and showed that they were actually the closest relatives to the kiwi. Calibration of the molecular tree indicated that the kiwi and elephant bird lineages had diverged c. 50 million years ago (Mitchell, Llamas, et al. 2014), at least 5 million years after separation of New Zealand from Gondwana, and therefore the ancestral kiwi must have been volant when it arrived in New Zealand. Rather than evolving from a large flightless bird, as had previously been thought, this suggests that kiwi were always small and had never become large and herbivorous (like other flightless ratites) because that niche had already been occupied in New Zealand by moa (Mitchell, Llamas, et al. 2014). The hypothesis is supported by the presence of small fossil kiwi bones (Worthy et al. 2013) and contemporaneous large moa bones and eggshell (Tennyson et al. 2010) from the Early Miocene of New Zealand.

Chatham Island duck (Anas chathamica)

The phylogenetic relationships of the enigmatic extinct Chatham Island duck (formerly Pachyanas chathamica) were recently resolved through sequencing its mitochondrial genome (Mitchell, Wood, et al. 2014). The study revealed that the Chatham Island species was basal in the clade comprising the New Zealand (Anas chlorotis), Auckland Island (A. aucklandica) and Campbell Island (A. nesiotis) teals, and therefore was not a unique genus but part of the wider Anas radiation. Calibration of the molecular phylogeny indicated a lineage age of c. 0.69–1.80 million years, consistent with the age of other lineages endemic to the Chatham Islands and just post-dating the estimated time of emergence of the Chatham Islands (Mitchell, Wood, et al. 2014). This study also suggested that divergence between the New Zealand brown teal and the two sub-Antarctic teal species pre-dated the last few glacial cycles, raising interesting questions about how these ducks persisted at high latitudes during ice age events.

Imber’s petrel (Pterodroma imberi)

Tennyson et al. (2015) used ancient mitochondrial DNA in combination with morphology to assess bones from the Chatham Islands that had previously been suggested could represent an undescribed extinct species (Cooper & Tennyson 2008). The DNA results indicated that the Chatham Islands bones were from a distinct taxon, which the authors described as Imber's petrel (Pterodroma imberi), yet were not able to identify the nearest related species with strong support (Tennyson et al. 2015). Calibration of the phylogenetic tree for cytochrome b sequences indicated a divergence date for Imber's petrel of c. 1.37 million years, and for the other two Chatham Island endemic species in the tree from their sister taxa of c. 1.98 million years (Chatham petrel, P. axillaris) and 0.56 million years (magenta petrel, P. magentae). As with the Chatham Island duck, these divergence dates are in line with other endemic Chatham Island fauna.

Scarlett’s shearwater (Puffinus spelaeus)

Ancient DNA extracted from bones of the extinct Scarlett's shearwater (Puffinus spelaeus) confirmed that it was a distinct taxon, and was sister to the extant fluttering shearwater (P. gavia) (Tennyson & Shepherd 2017). These two species, together with Hutton's shearwater (P. huttoni), formed a distinct New Zealand endemic clade that had radiated within the past 2 million years (Tennyson & Shepherd 2017). The three species likely occupied different feeding niches and breeding colony sites (Tennyson & Shepherd 2017).

Haast’s eagle (Aquila moorei)

The world's largest eagle, Haast's eagle (Aquila moorei, formerly Harpagornis moorei) inhabited New Zealand at the time of human settlement, but had likely become extinct prior to European arrival. Bunce et al. (2005) used ancient DNA extracted from bones of Haast's eagle to demonstrate its surprising yet well supported position within a clade of small eagles (with little eagle [Hieraaetus morphnoides] and booted eagle [H. pennatus]). The relatively recent divergence date for the base of this clade (0.7–1.8 million years) suggested a rapid size change had occurred to bring about the order of magnitude difference in body mass between species. The causes of this size reduction may have been a release from competition for the ancestral eagle arriving in New Zealand and adaptation to predate large avian taxa such as moa (Bunce et al. 2005). It is also worth noting here that the taxonomy of the booted eagles (Aquilinae), including Haast's eagle, is still much debated. The checklist of New Zealand birds follows the grouping of booted eagles into a single genus (Aquila) (Worthy 2010), yet the most recent molecular phylogeny recognises nine genera and places Haast's eagle within the genus Hieraaetus (Lerner et al. 2017).

Adzebills (Aptornithidae)

Houde et al. (1997) sequenced a 673 bp fragment of the mitochondrial 12S rRNA gene from South Island adzebill (Aptornis defossor). The phylogenetic placement of adzebills within the Gruiformes had long been a matter of debate (Houde et al. 1997) yet this study found good support for the family being sister to the Rallidae (rails, gallinules and coots). However, more recent work has added further detail to the phylogenetic relationships of taxa within the Gruiformes (e.g. Fain et al. 2007; Garcia-R et al. 2014a, 2014b; Prum et al. 2015) and so the exact relationships of adzebills likely require re-examination.

Rails, gallinules and coots (Rallidae)

Trewick (1997) sequenced ancient DNA from extinct flightless rail species from New Zealand; the Chatham Island coot (Fulica chathamensis), New Zealand coot (F. prisca), North Island takahē (Porphyrio mantelli), Chatham Island rail (Cabalus modestus) and Dieffenbach's rail (Gallirallus dieffenbachii), and assessed these within a phylogenetic framework that included other rail taxa from New Zealand and around the world. The results indicated a close relationship between the New Zealand and Chatham Island coots and Eurasian coot (F. atra), suggesting that the former two species had recently colonised New Zealand and rapidly become large and flightless (Trewick 1997). The two flightless Chatham Island rails, C. modestus and G. dieffenbachii, were found to be sister taxa, and were suggested to represent different dispersal events of the same ancestral taxon to the Chatham Islands (Trewick 1997). However, with the inclusion of additional taxa to the phylogenetic analysis, the close sister relationship of these two rails was no longer supported (Garcia-R et al. 2014a; Trewick et al. in press).

Trewick (1997) found the North Island takahē to be genetically more similar to the pūkeko (P. porphyrio) than to the South Island takahē (P. hochstetteri), suggesting again that these species reflected two different colonisation events. However, analyses by Garcia-R & Trewick (2015) indicate that a sister relationship between the North Island and South Island takahē could not be ruled out.

Garcia-R et al. (2014a) used mitochondrial and nuclear DNA sequences to explore the phylogeny and evolutionary history of rails. Their dataset included the first ancient DNA sequence from the extinct Hawkin's rail (Diaphorapteryx hawkinsi), and provided sequences from two further specimens of Dieffenbach's and Chatham Island rails. The results demonstrated that Diaphorapteryx and Wallace's rail (Habroptila wallacii) formed a clade basal to Gallirallus (Garcia-R et al. 2014a). This clade, with a divergence of around 14 million years between the two remaining species (Garcia-R et al. 2014a), possibly reflects an old radiation with lots of subsequent extinction.

Chatham Island parrot (Nestor chathamensis)

A combination of osteology, morphometrics and ancient DNA was used by Wood et al. (2014) to describe Nestor chathamensis, a new species of extinct parrot from the Chatham Islands. A partial mitochondrial genome sequence from the taxon revealed that it was the sister taxon to the New Zealand kākā (N. meridionalis), and diverged approximately 0.53–3.29 (mean 1.74) million years ago; a timing comparable to that seen in several other endemic Chatham Island taxa.

Laughing owl (Ninox albifacies)

The endemic laughing owl, which became extinct in the early 20th century (Tennyson & Martinson 2006), was for more than a century regarded distinct enough from other owls as to warrant placement within its own genus (Sceloglaux). The phylogenetic relationships of the laughing owl were examined by Wood et al. (2017) through sequencing a mitochondrial genome from a museum skin specimen. The study found that the laughing owl was nested within the genus Ninox, sister to the clade that contained two Australian taxa and New Zealand's other native owl, the morepork (Ninox novaeseelandiae). This new understanding of the phylogenetic relationships of the laughing owl allowed Wood et al. (2017) to propose the hypothesis that it could represent the earliest of a double invasion of New Zealand by the same owl lineage from Australia.

New Zealand owlet-nightjar (Aegotheles novaezealandiae)

Mitochondrial DNA was extracted and sequenced from bone of the extinct New Zealand owlet-nightjar (Aegotheles novaezealandiae) by Dumbacher et al. (2003). The sequence was phylogenetically assessed against sequences from all extant owlet-nightjar species and the New Zealand species was found to be sister to the New Caledonian owlet-nightjar (A. savesi). Together they formed the basal clade within the genus (Dumbacher et al. 2003). The great antiquity of the lineage in New Zealand is supported by the presence of owlet-nightjar (Aegotheles sp.) bones in Early Miocene fossil deposits from Otago (Worthy et al. 2007).

New Zealand wrens (Acanthisittidae)

The New Zealand wrens are the sister taxon to all other perching birds (Passeriformes), and represent the most diverse endemic passerine family in the recent New Zealand avifauna. However, five of the seven species present at the time of human settlement are now extinct (Tennyson & Martinson 2006), meaning that ancient DNA is important for fully assessing the evolutionary history of this group.

Cooper & Cooper (1995) sequenced a short section of mitochondrial DNA from moa, kiwi and wrens, including the extinct bush wren (Xenicus longipes), in order to explore whether a genetic bottleneck had occurred in New Zealand sometime in the past. Their results indicated a widespread loss of diversity in all groups during the mid-Cenozoic, a feature now known to be present in many New Zealand endemic taxa. This genetic bottleneck event, known as the Oligocene bottleneck, corresponds to a time when much of the current New Zealand landmass was under the sea (Cooper & Cooper 1995).

Mitchell et al. (2016) sequenced a mitochondrial genome from the bush wren, as well as from two additional extinct wren species (Lyall's wren [Traversia lyallii] and South Island stout-legged wren [Pachyplichas yaldwyni]), to examine the evolutionary history of the family. Two unexpected results were revealed by this study. First, the extant rock wren (Xenicus gilviventris) was found to be genetically closer to the stout-legged wren than to the congeneric bush wren, leading to the authors synonymising Pachyplichas with Xenicus. This scenario, of a stout and flightless representative of a genus having been formerly placed within its own genus, mirrors that of the takahē (formerly Notornis, but now with pūkeko in Porphyrio). Secondly, Lyall's wren was found to have a deep divergence to all other wren species. The timing of this divergence pre-dated the Oligocene drowning event (Mitchell et al. 2016), and thus at least two lineages (Dendroscansor was not assessed) of New Zealand wrens persisted through the marine highstand.

New Zealand wattlebirds (Callaeidae)

Shepherd & Lambert (2007) analysed mitochondrial and nuclear DNA from two historical huia (Heteralocha acutirostris) toe pads, as part of a study on the phylogenetic relationships of the endemic New Zealand wattlebirds. The study resolved that the family was a monophyletic clade within Corvida, and suggested a mid-Cenozoic (mean 33.7 million years) divergence from its sister taxon within the phylogeny, the satinbirds (Cnemophilidae). This result indicated a dispersal, rather than vicariant, origin for this group in New Zealand. Although the phylogenetic analysis could not resolve the relationships within Callaeidae, most of the analyses placed huia or kōkako (Callaeas) as the most basal lineage (Shepherd & Lambert 2007).

Subsequently, Anmarkrud & Lifjeld (2017) sequenced mitochondrial genomes of huia and the extinct South Island kōkako (Callaeas cinerea). Comparison of the South Island kōkako sequence with sequences available for the North Island kōkako (C. wilsoni) showed them to be highly similar, including an identical cytochrome b region. However, no phylogenetic analyses were undertaken using these new sequences.

Piopio (Oriolidae: Turnagrinae)

The phylogenetic relationships of the extinct New Zealand piopio (South Island piopio [Turnagra capensis] and North Island piopio [T. tanagra]) have been contested. Christidis et al. (1996) placed the piopio as the sister taxon to bowerbirds (Ptilonorhynchidae) based on analysis of cytochrome b sequences, but this relationship had less than 50% bootstrap support. Subsequent studies generating sequences from additional genetic markers (Johansson et al. 2011; Zuccon & Ericson 2012), including two mitochondrial genomes (Gibb et al. 2015; Anmarkrud & Lifjeld 2017) suggest a nested position within orioles (Oriolidae) (Zuccon & Ericson 2012), a taxon not included in the analysis of Christidis et al. (1996), but with a divergence date of c. 22 million years (Gibb et al. 2015). The piopio therefore provide a further example of volant dispersal of an avian lineage to Zealandia.

Crows and jays (Corvidae)

Scofield et al. (2017) sequenced mitochondrial genomes from bones of the extinct New Zealand (Corvus antipodum) and Chatham Island (C. moriorum) ravens and revealed them to be sister to the clade comprising the Australian raven (C. coronoides), little raven (C. mellori) and forest raven (C. tasmanicus). Divergence between these New Zealand and Australian clades occurred in the late Pliocene, coinciding with the proliferation of non-forested glacial habitats which the New Zealand crow likely inhabited (Scofield et al. 2017). The relatively shallow genetic divergence between the two New Zealand taxa (calibrated to the late Pleistocene) led the authors to suggest that differentiation was at subspecies level, rather than full species.

The human factor

The settlement of New Zealand by Polynesians some 750 years ago (Wilmshurst et al. 2008; Anderson 2017) initiated a period of rapid and dramatic ecological change. Hunting, burning of forest, and introduction of commensal species, resulted in declines and fragmentation of populations of many animal species, which in many cases ultimately led to extinction (Worthy & Holdaway 2002; Wood 2013). Ancient DNA has provided valuable new insights into the prehistoric settlement period in New Zealand, such as different aspects of the biology and ecology of commensal fauna. Moreover, ancient DNA studies discovered previously unknown cryptic extinctions and population declines in marine faunas that followed human settlement (e.g. for penguins and sea lions), and subsequent range expansions, providing a new perspective on lineages traditionally thought to be long-established endemics of some areas (e.g. yellow-eyed penguin on the South Island).

Faunal hunting practices

Although this review does not specifically deal with moa, it is worth noting two studies that have used ancient DNA to shed light on faunal-hunting practices of prehistoric New Zealanders. Avian eggshell is relatively common in New Zealand archaeological and palaeontological sites. It has been demonstrated that eggshell is a relatively good source of ancient DNA (Oskam et al. 2010), allowing identification of eggshell fragments to species and thus taxonomic assessment of the past use of eggs. Oskam et al. (2011) further explored this using moa eggs, which were an important item in the diets of New Zealand's early settlers. They extracted mitochondrial (control region) and nuclear (microsatellite) DNA from 56 moa eggshell fragments excavated from an oven feature at Wairau Bar. Their results revealed the fragments represented 31 different eggs, from three genera: Emeus, Euryapteryx and Dinornis. In a subsequent study, 50 different eggs were revealed in a single Wairau Bar site, of six genera: Anomalopteryx, Dinornis, Emeus, Euryapteryx, Megalapteryx and Pachyornis (Oskam et al. 2012).

Oskam et al. (2012) used genetic sex markers to reveal that male moas were more abundant than females in South Island archaeological deposits. This is in contrast to the female-skewed sex ratios of natural moa populations (Allentoft et al. 2010) and, together with the large number of eggs at the same sites, may reflect a preferential hunting method of raiding nests (in ratites males are the incubating sex).

Impact of humans on marine fauna

Yellow-eyed penguins (Megadyptes spp.). The presence of bones attributed to the yellow-eyed penguin (Megadyptes antipodes) in natural and archaeological deposits from New Zealand led to the belief that this species had been present on the mainland since before human settlement. This idea was overturned by Boessenkool et al. (2009) who examined the morphology and ancient DNA of prehistoric, historic and modern yellow-eyed penguin samples from New Zealand and the sub-Antarctic region. They demonstrated that a previously unknown species of Megadyptes, genetically distinct from and slightly smaller than the extant M. antipodes, had inhabited the mainland New Zealand coast prior to human arrival. Following human settlement this penguin (which the authors named the Waitaha penguin, M. waitaha) had been hunted to extinction, and was subsequently replaced by an expanded population of the sub-Antarctic M. antipodes. While ancient DNA analyses revealed that M. antipodes had occasionally visited the New Zealand mainland in low numbers, the extinction of the Waitaha penguin left a vacant ecological niche that apparently allowed the sub-Antarctic species to become resident on the mainland (Boessenkool et al. 2009). Radiocarbon dating subsequently revealed that extinction of M. waitaha and its subsequent replacement by M. antipodes occurred within a century (Rawlence, Perry, et al. 2015). The timing of the colonisation of South Island by M. antipodes coincides with re-colonisation events in other marine bird and mammal taxa in southern New Zealand, and it has been suggested that these may have been related to climatic and human demographic changes during the Little Ice Age (Waters et al. 2017).

Blue penguins (Eudyptula spp.). It was recently demonstrated using various lines of biological and behavioural evidence, including genetics, that two species of blue penguin breed around the New Zealand coast: the widespread New Zealand endemic Eudyptula minor and E. novaehollandiae, which is restricted to the Otago coast, but also occurs in Australia (Grosser et al. 2015, 2017). Using ancient DNA and radiocarbon dating of pre-human, archaeological and historical Eudyptula remains, Grosser et al. (2016) demonstrated that the Otago E. novaehollandiae population represents a cryptic colonisation event from Australia, which occurred between AD 1500–1900. This colonisation event was hypothesised to have been promoted by post-settlement hunting and predation pressure bringing about the extirpation of E. minor from the Otago region, and opening a niche into which dispersing individuals of E. novaehollandiae could occupy and establish (Grosser et al. 2016).

Leucocarbo shags. Fossil and archaeological remains suggest that southern Leucocarbo shags on the South Island (formerly Stewart Island shag, L. chalconotus) had a much wider distribution at the time of Polynesian arrival than it does at present. Rawlence, Kennedy, et al. (2015) used ancient mitochondrial DNA and radiocarbon dating to assess the impacts of human hunting on this seabird. The study found two distinct phylogeographical lineages, which were redesignated as distinct species: the Foveaux shag (L. stewarti), restricted to the south (Foveaux Strait area) and the Otago shag (L. chalconotus) restricted to the southeast (Otago coast) of New Zealand. Ancient DNA data revealed that the Otago shag had formerly been more widespread around the South Island's east coast, and revealed dramatic population extinctions, range retractions and lineage loss soon after Polynesian arrival (Rawlence, Kennedy, et al. 2015). In contrast, the Foveaux shag population was found to have been relatively stable since Polynesian arrival. Archaeological estimates suggest that the South Island supported several thousand people by the end of the prehistoric period (c. 1450 AD), with much of the population concentrated along the east coast (Rawlence, Kennedy, et al. 2015); perhaps with a corresponding greater hunting impact on the Otago than Foveaux shags (Rawlence, Kennedy, et al. 2015). Rawlence et al. (in press) used ancient mitochondrial DNA and morphology to show that Leucocabo king shags (L. carunculatus), formerly widespread around the southern coasts of the North Island and the northern parts of the South Island but now restricted to the northeastern South Island, experienced population lineage extinctions. The authors revealed that northern populations comprised a separate species; the Kohatu shag (L. septentrionalis), which is now extinct.

New Zealand sealion (Phocarctos hookeri). Collins, Rawlence, Worthy, et al. (2014) confirmed the presence of former breeding colonies of sea lion (Phocarctos hookeri) around the New Zealand coast through a combination of radiocarbon dating and ancient DNA analysis of bones from the northern South Island. In a further study, Collins, Rawlence, Prost, et al. (2014) used ancient DNA analysis of the prehistoric bones to reveal that they in fact represented a previously unknown extinct mainland endemic sea lion lineage (Phocarctos hookeri ‘NZ’), whose extinction in turn facilitated a subsequent northward range expansion of a lineage previously restricted to the sub-Antarctic region (P. hookeri ‘subantarctic’). The endemic mainland lineage was apparently eliminated soon after human arrival, then replaced by the genetically divergent sub-Antarctic lineage within just a few centuries (Collins, Rawlence, Prost, et al. 2014). The scenario parallels that of the Megadyptes penguins described above (Boessenkool et al. 2009), where P. hookeri bones were abundant in early middens yet absent from later archaeological sites; consistent with rapid extinctions due to over-hunting. Moreover, mitochondrial genomes sequenced from prehistoric (AD 1450–1650) bones of P. hookeri from the Chatham Islands revealed that there was another large population of a sea lion lineage unique to the Chatham Islands (P. hookeri ‘Chathams’) at the time of Polynesian arrival (Collins, Rawlence, Prost, et al. 2014; Rawlence et al. 2016). This population was driven to extinction within 200 years, again probably due to over-hunting (Collins, Rawlence, Prost, et al. 2014; Rawlence et al. 2016). Current surviving Phocarctos are thus remnants of a once more diverse and widespread sea lion clade.

Fur seals (Arctocephalus spp.). Salis et al. (2016) analysed DNA of modern and ancient fur seal (Arctocephalus spp.) remains from New Zealand and the sub-Antarctic, and found that both areas contained historical genetic diversity that seems to have suffered from a population bottleneck since human arrival.

Impact of humans on terrestrial avifauna

Kiwi (Apterygidae). Using DNA extracted from 46 ancient bone specimens Shepherd & Lambert (2008) explored the impact of post-settlement population declines on the genetic diversity of large kiwi species (including southern brown kiwi [Apteryx australis], North Island brown kiwi [A. mantelli], Ōkārito brown kiwi [A. rowi] and great-spotted kiwi [A. haastii]). The results indicated that the long known reduction in the geographic range of brown kiwi had resulted in significant loss of genetic variation within the South Island brown kiwi and rowi (Shepherd & Lambert 2008). The authors found no support for a putative extinct taxon, the eastern South Island brown kiwi, being a monophyletic clade; specimens attributed to this taxon were nested within the diversity of South Island brown kiwi (Shepherd & Lambert 2008). However, subsequent analyses have recognised at least two distinct extinct eastern South Island brown kiwi mitochondrial clades (‘Mt. Somers’ and ‘Mt. Cookson’) (Weir et al. 2016).

In contrast to the high levels of genetic diversity within past populations of these large kiwi species, Shepherd et al. (2012) found much lower diversity of mitochondrial DNA haplotypes in both recent and ancient specimens of little spotted kiwi (Apteryx owenii). Interestingly, the results also recognised a distinct (but now extinct) genetic lineage from the Waikato region, and demonstrated that little spotted kiwi mitochondrial lineages had persisted on the South Island until more recently than had been thought (Shepherd et al. 2012).

Swans (Cygnus spp.). Rawlence et al. (2017) sequenced a fragment of mitochondrial control region to assess relationships between a number of prehistoric and recent black swan specimens from New Zealand and the Chatham Islands. They identified a similar pattern to that seen in the marine taxa discussed above, where an endemic taxon (the poūwa, Cygnus sumnerensis) disappeared shortly following human settlement and was subsequently replaced by a more recent arrival, in this case the Australian black swan (C. atratus).

Kākāpō (Strigops habroptilus). Bergner et al. (2016) used microsatellites and mitochondrial DNA from both contemporary and museum specimens (dating back to 1884) to examine the demographic history of kākāpō. Their results indicated that kākāpō had experienced a population bottleneck associated with European settlement. The study also found no evidence for an earlier population decline associated with initial human settlement (Bergner et al. 2016). However, this is at odds with the wider distribution of fossil remains of kākāpō compared with the distribution of the species in 1840 (e.g. figure 1 in Bergner et al. 2016) and probably reflects inadequate sampling for observing such a decline (e.g. the study did not sample prehistoric specimens).

Kea (Nestor notabilis). Dussex et al. (2015) used kea (Nestor notabilis) specimens, spanning from the prehuman to present, to study how humans had impacted on the genetic diversity of this species. Surprisingly, although kea were persecuted in large numbers during the 19th and 20th centuries, the authors found no evidence for a loss of genetic diversity associated with this population decline. However, the study did find a loss of genetic diversity within kea prior to human arrival. This may have been due to a reduction in population size associated with a decreasing area of preferred habitat, driven by post-glacial reafforestation (Dussex et al. 2015).

Introduced prehistoric fauna

Ancient DNA analyses have provided insights into the origins and biology of the two species that were commensal with Māori in pre-European New Zealand: the kiore or Pacific rat (Rattus exulans) and the kurī or dog (Canis familiaris). Moreover, molecular analyses have also helped clarify the origins of two other possible pre-European commensal species in New Zealand, the chicken (Gallus gallus domesticus) and pig (Sus scrofa).

Kiore (Rattus exulans). Radiocarbon dates indicating an early (c. 2000 year old) arrival of the kiore, or Pacific rat, in New Zealand (Holdaway 1996) have since been refuted using multiple lines of evidence (e.g. Anderson 1996, 2000; Wilmshurst et al. 2008) and it is now widely accepted that the kiore was introduced to New Zealand by the first Polynesian settlers during the 13th century AD (Wilmshurst et al. 2008). Its close association with people has meant that genetic studies (including ancient DNA) of this species have provided proxy data for prehistoric human migration patterns across Polynesia (Matisoo-Smith 1994; Matisoo-Smith & Robins 2009). This idea was explored further within the New Zealand region by Matisoo-Smith et al. (1997), who modified existing ancient DNA extraction protocols to allow the recovery of ancient DNA from archaeological Pacific rat bones ranging from 400–2000 years old, including from the Chatham Islands. Work by Matisoo-Smith et al. (1999) applied the method to new samples from throughout the Pacific, including 15 archaeological samples from the Chatham Islands. The study compared a 200 bp fragment of mitochondrial control region, identifying the same unique SNP and revealing a lack of variation in Chatham Island rat populations consistent with an early limited introduction. The result also demonstrated a direct descendent relationship between kiore on the Chatham Islands today and those present there some 450 years ago. This pattern was in contrast to samples from the Kermadecs, where there were highly variable mitochondrial DNA lineages suggesting multiple introduction events (Matisoo-Smith et al. 1999).

Matisoo-Smith (2002) compared ancient DNA sequences from five archaeological kiore samples with 50 recent samples collected throughout the Pacific. Two major haplotypes were identified in prehistoric samples from New Zealand, and the South Island samples were more similar to extant populations. One sample yielded a sequence similar to those obtained from East Polynesia, suggesting a possible source area for settlement of New Zealand. Mitochondrial control region sequences obtained from 87 archaeological rat bones and 33 historic museum specimens collected from Island South-East Asia and the Pacific demonstrated the presence of at least two haplogroups in New Zealand, one widespread throughout East Polynesia and the other which was elsewhere restricted to Lapita island groups (Matisoo-Smith & Robins 2004).

Kurī (Canis familiaris). Dogs (kurī) served important functional (including for consumption, skins and cloaks) and cultural roles in the prehistoric Pacific region, including New Zealand. Purebred kurī became extinct shortly after European settlement by hybridisation with dogs of European origin, and so their study is now restricted to the archaeological and ancient DNA records.

Table 1. Taxonomic summary of ancient DNA recovered from New Zealand fauna and prehistorically introduced taxa. Moa are excluded from this table, but are reviewed by Allentoft & Rawlence (2012) and Worthy & Scofield (2012).

Family or higher taxonomic group	Genus	Species († denotes extinct)	Common name	References
NATIVE FISH
Eleotridae	Gobiomorphus	sp.	Bullies	Matisoo-Smith et al. (2008)
NATIVE BIRDS
Acanthisittidae	Xenicus (=Pachyplichas)	yaldwyni †	South Island stout-legged wren	Mitchell et al. (2016)
	Traversia	lyalli †	Lyall's wren	Mitchell et al. (2016)
	Xenicus	longipes †	Bush wren	Cooper & Cooper (1995); Mitchell et al. (2016)
Accipitridae	Aquila	moorei	Haast's eagle	Bunce et al. (2005)
		sp.		Grealy et al. (2015)
Aegothelidae	Aegotheles	novaezealandiae †	New Zealand owlet-nightjar	Dumbacher et al. (2003)
Anatidae	Cygnus	sumnerensis †	Poūwa	Rawlence et al. (2017)
	Pachyanas	chathamica †	Chatham duck	Mitchell, Wood, et al. (2014)
Anatidae		sp.		Oskam et al. (2010); Grealy et al. (2015)
Apterygidae	Apteryx	sp.	Kiwi	Grealy et al. (2015)
	Apteryx	australis	Southern brown kiwi	Shepherd & Lambert (2008); Shepherd et al. (2013)
	Apteryx	haastii	Great spotted kiwi	Shepherd & Lambert (2008)
	Apteryx	mantelli	North Island brown kiwi	Shepherd & Lambert (2008); Hartnup et al. (2009); Shepherd et al. (2013)
	Apteryx	owenii	Little spotted kiwi	Hartnup et al. (2009); Shepherd et al. (2013)
	Apteryx	rowi	Ōkārito brown kiwi	Shepherd & Lambert (2008)
Aptornithidae	Aptornis	defossor	South Island adzebill	Houde et al. (1997)
	Aptornis	spp. †		Grealy et al. (2015)
Callaeidae	Callaeas	cinerea †	South Island kōkako	Anmarkrud & Lifjeld (2017)
	Heteralocha	acutirostris †	Huia	Shepherd & Lambert (2007); Anmarkrud & Lifjeld (2017)
Corvidae	Corvus	antipodum †	New Zealand raven	Scofield et al. (2017)
	Corvus	moriorum †	Chatham Island raven	Scofield et al. (2017)
Nestoridae	Nestor	chathamensis †	Chatham parrot	Wood et al. (2014)
	Nestor	notabilis	Kea	Dussex et al. (2015)
Phalacrocoracidae	Leucocarbo	carunculatus	King shag	Rawlence et al. (In Press)
Phalacrocoracidae	Leucocarbo	chalconotus	Otago shag	Rawlence, Kennedy, et al. (2015)
	Leucocarbo	septentrionalis †	Kohatu shag	Rawlence et al. (In Press)
	Leucocarbo	stewarti	Foveaux shag	Rawlence, Kennedy, et al. (2015)
Procellariidae	Pterodroma	imberi †	Imber's petrel	Tennyson et al. (2015)
	Puffinus	spelaeus †	Scarlett's shearwater	Tennyson & Shepherd (2017)
Psittacidae	Cyanoramphus	novaezelandiae	Red-crowned parakeet	Willerslev et al. (2003)
	Cyanoramphus	sp.	Parakeet	Holdaway et al. (2010)
Rallidae	Cabalus	modestus †	Chatham Island rail	Trewick (1997); Garcia-R et al. (2014a)
	Diaphorapteryx	hawkinsi †	Hawkin's rail	Garcia-R et al. (2014a)
	Fulica	chathamensis †	Chatham Island coot	Trewick (1997)
	Fulica	prisca †	New Zealand coot	Trewick (1997)
	Gallirallus	dieffenbachii †	Dieffenbach's rail	Trewick (1997); Garcia-R et al. (2014a)
	Gallirallus	sp.		Grealy et al. (2015)
	Porphyrio	mantelli †	Moho (North Island takahē)	Trewick (1997); Garcia-R & Trewick (2015)
Spheniscidae	Eudyptula	minor	Little blue penguin	Grosser et al. (2016)
	Eudyptula	novaehollandiae	Fairy penguin	Grosser et al. (2016)
	Megadyptes	antipodes	Yellow-eyed penguin	Boessenkool et al. (2009); Rawlence, Perry, et al. (2015)
	Megadyptes	waitaha †	Waitaha penguin	Boessenkool et al. (2009); Rawlence, Perry, et al. (2015)
	Pygoscelis	adeliae	Adélie penguin	Lambert et al. (2002); Ritchie et al. (2004); Subramanian et al. (2009)
Strigidae	Ninox	albifacies †	Laughing owl	Wood et al. (2017)
Strigopidae	Strigops	habroptilus	Kākāpō	Wood et al. (2012c); Grealy et al. (2015); Bergner et al. (2016)
Turnagridae	Turnagra †		Piopio	Christidis et al. (1996); Johansson et al. (2011); Zuccon & Ericson (2012); Gibb et al. (2015); Anmarkrud & Lifjeld (2017)
INTRODUCED BIRDS
Phasianidae	Gallus	gallus domesticus	Domestic chicken	Wood, Herrera, et al. (2016)
NATIVE MAMMALS
Otariidae	Arctocephalus	forsteri	New Zealand fur seal	Salis et al. (2016)
	Arctocephalus	tropicalis	Sub-Antarctic fur seal	Salis et al. (2016)
	Phocarctos	hookeri	New Zealand sea lion	Collins, Rawlence, Worthy, et al. (2014); Collins, Rawlence, Prost, et al. (2014); Rawlence et al. (2016)
INTRODUCED MAMMALS
Canidae	Canis	lupus familiaris	Domestic dog (kurī)	Greig et al. (2015); Wood, Crown, et al. (2016)
Muridae	Rattus	exulans	Pacific rat (kiore)	Matisoo-Smith (1994); Matisoo-Smith et al. (1997); Matisoo-Smith et al. (1999); Matisoo-Smith (2002); Matisoo-Smith & Robins (2004)
Suidae	Sus	scrofa	Pig	Allen et al. (2001)

Greig et al. (2015) used hybridisation-capture and high-throughput sequencing to generate complete mitochondrial genomes for 14 early New Zealand kurī specimens from Wairau Bar. Five different closely-related haplotypes were identified within these, corresponding to sequenced specimens from China, South East Asia and the Pacific, but not Taiwan or the Philippines. This result supported a past movement of dogs into the Pacific (including New Zealand) by a southwest route via Indonesia (Greig et al. 2015).

The diets of New Zealand kurī had previously been inferred from microscopic analyses of coprolites, yet these methods did not provide a robust way to distinguish kurī coprolites from those of human origin (e.g. Horrocks et al. 2002). Wood, Crown, et al. (2016) extracted and sequenced ancient DNA from five coprolites from an archaeological site in Auckland, confirming their identity as being from kurī. Using combined microscopic and high-throughput ancient DNA metabarcoding approaches, they revealed that the diet of some kurī at the site had comprised dominantly of marine fish, shellfish and plant material (including commensal crops such as paper mulberry [Broussonetia papyrifera] and bottle gourd [Lagenaria siceraria]).

Chicken (Gallus gallus domesticus). Wood, Herrera, et al. (2016) radiocarbon dated chicken bones from archaeological layers in three South Island midden deposits that were stratigraphically contemporary with bones of extinct and/or extirpated bird species to test the possibility that chickens had been introduced to New Zealand prehistorically. However, the results indicated that the bones were not prehistoric, but dated to the early post-contact period (late 18th or early 19th centuries) and may have been descended from stock brought to New Zealand by James Cook's second expedition (Wood, Herrera, et al. 2016). Concordant with this interpretation, DNA sequences obtained from the bones showed that they belonged to the E-haplogroup, a now widespread group that only entered the Pacific region with European colonisation (Thomson et al. 2014).

Pig (Sus scrofa). As with chickens, there is currently no clear evidence that pigs were introduced to New Zealand during prehistory (Allen et al. 2001). While excavation of an archaeological site at Long Beach, near Dunedin, recovered the tooth of a pig from a pre-European layer with a calibrated date of around 1600 AD (Hamel & Leach 1979), this specimen was likely to have European origins. Specifically, DNA extracted from the specimen (Allen et al. 2001) suggested that the tooth was not of Polynesian origin, but rather that it ended up in the layer via site disturbance (Allen et al. 2001).

Identifying the unidentifiable

An important role of ancient DNA has been in providing the ability to identify faunal remains that could not be readily identified by other means. These have included remains with no diagnostic characters (e.g. bone fragments, coprolites), and remains for which no comparative material exists (e.g. feathers of long extinct species). Some examples relevant to New Zealand fauna are provided below.

Bones and bone fragments

Large numbers of small bone fragments are typically recovered from palaeontological and archaeological deposits. Many retain few or no diagnostic characters, and so their usefulness for traditional morphological-based bone identification, and assessments of faunal assemblages, has been limited. Recently, ancient DNA metabarcoding has been applied to the study of such bone fragments. In this technique, hundreds of fragmented or morphologically indistinguishable bones are ground together into a powder from which DNA is extracted (Murray et al. 2013). Selected barcoding regions are then amplified and sequenced, and through comparison with reference sequence databases it is possible to construct a faunal assemblage list. This ‘bulk-bone’ method has been used to supplement existing palaeofaunal records from fossil deposits, revealing rare, endangered and extinct species or haplotypes over different temporal and spatial scales (Murray et al. 2013; Grealy et al. 2015). In New Zealand the method has been applied to the late Holocene ‘Finsch's Folly’ tomo in South Canterbury (Grealy et al. 2015). Here, the analysis identified eight bird families, including two that had not previously been recorded from the deposit based on morphological identification of bones (Dinornithidae, although emeid moa bones were common, and Accipitridae).

In other cases, ancient DNA analyses have provided insights into the limits of morphological identification of complete bones. For example, Matisoo-Smith & Allen (2001) used ancient DNA sequences from archaeological rat bones, including New Zealand specimens, to demonstrate that current methods for identification of these bones were inadequate due to a significant size overlap between the bones of Pacific rat and other local rat species. There are many other cases where it can be difficult or impossible to resolve closely related taxa based on osteological (particularly post cranial) characters, and in such instances ancient DNA analyses may provide important insights into their past distributions or reveal cryptic extinct taxa. Some good examples include the New Zealand parakeets (Cyanoramphus spp.) (e.g. Holdaway et al. 2010) and lizards.

Feathers

Feathers require dry conditions for preservation, and have been found in a number of Holocene prehuman and archaeological rockshelter and cave sites in New Zealand; mainly in the South Island (e.g. Rawlence et al. 2013). Moa feathers are perhaps the most common of these. Rawlence et al. (2009) extracted the first ancient DNA from moa feathers, and by amplifying and sequencing a diagnostic fragment of the mitochondrial control region were able to identify feathers from four different moa species. Interestingly, their results showed that a dark, white-tipped, feather type that had previously been attributed to upland moa (Megalapteryx didinus) was also present in the plumage of heavy-footed moa (Pachyornis elephantopus), and would have given the bird a spotted appearance similar to that seen in spotted kiwi (Rawlence et al. 2009).

Ancient DNA was also used to resolve the identity of five small violet-tipped contour feathers that were among a collection of moa feathers donated to Otago Museum in 1943. The feathers did not match any species present in the extant, or known extinct, New Zealand avifauna. The results of the ancient DNA analysis showed that the feathers were from an undetermined kingfisher species, most closely allied with southeast Asian and American taxa (Rawlence et al. 2013). It was suggested that the feathers could have belonged to a Chinese goldminer who occupied the cave where the moa feathers were found (as kingfisher feathers were used in art and clothing), and a radiocarbon date from one of the feathers confirmed them as overlapping with the goldrush period (Rawlence et al. 2013).

A further application of ancient DNA to feather analysis is in the study of feather cloaks. Hartnup et al. (2009) used ancient DNA to identify putative moa feathers in a Māori cloak held by the Hawkes Bay Museum and Cultural Trust. Their results demonstrated that the feathers were in fact from emu (Dromaius novaehollandiae) (Hartnup et al. 2009). In another study, Hartnup et al. (2011) sampled DNA from 113 kiwi feather cloaks held by museums around the world, in order to determine the species and provenance of the birds used for making of the cloaks. Of 849 feather samples that DNA sequences were obtained for, 847 (99.8%) were from North Island brown kiwi (Apteryx mantelli) and two (0.2%) were from little spotted kiwi (A. owenii) (Hartnup et al. 2011). The feathers were skewed towards males (59.2%), and the authors suggested that the incubating males may have been more vulnerable to hunting. Moreover, Hartnup et al. (2011) used a reference dataset of modern haplotype distributions to show that 15% of the cloaks included feathers of birds that originated from different geographic areas to where the cloaks were made.

Museum specimens

Museums hold significant collections of prehistoric specimens, which, as discussed in the sections above, provide an invaluable resource for studying the past patterns of biodiversity as well as the evolutionary histories and phylogenetic relationships of extinct species. However, ancient DNA analyses can also provide details about specimens that add value to the museum collections themselves. For example, Boessenkool et al. (2010) analysed ancient DNA from a large number of 19th century museum skins of yellow-eyed penguin and identified eight mislabeled specimens, for which collection localities were at odds with their geographic origin based on DNA sequences. Shepherd et al. (2013) used ancient DNA to resolve the species identities of six unidentified kiwi skeletons in the collections of Museum of New Zealand Te Papa Tongarewa. Moreover, the results also provided information on the potential geographic provenance for one of the specimens (Shepherd et al. 2013). In another study, Rawlence, Kennedy, et al. (2014) used ancient DNA analysis to correct the identification of two 19th century shag specimens in Canterbury Museum.

Biological function

The analysis of ancient DNA provides the potential for studying biological functions and developmental biology of extinct species. For example, the sequencing of the DNA responsible for haemoglobin in the extinct woolly mammoth (Mammuthus primigenius) (Campbell et al. 2010) allowed the haemoglobin itself to be resurrected and studied; revealing that it had adaptations for increased efficiency in cold climates. In a local example, Huynen et al. (2014) explored the molecular basis of winglessness in moa by reconstructing tbx5, a highly conserved gene region related to hox genes, essential for the development of forelimbs in many animals. When placed into chicken embryos the moa tbx5 was shown to be fully functional, suggesting that other genes must also have played a role in the unique complete loss of forelimbs in moa.

Faunal reconstruction using sedimentary ancient DNA

Ancient DNA can be extracted from a variety of environmental substrates such as sediments, soils and ice (Rawlence, Lowe, et al. 2014). In New Zealand, faunal DNA has been recovered from ancient sediments, allowing past faunal communities to be reconstructed in the absence of physical remains such as bones. Willerslev et al. (2003) provided the first such results from New Zealand when they sequenced DNA of upland moa (Megalapteryx didinus), heavy-footed moa (Pachyornis elephantopus) and red-crowned parakeet (Cyanoramphus novaezelandiae) from prehuman sediment from a Central Otago rock shelter.

Vertical leaching of DNA through sediments has the potential to introduce younger DNA into older sediment layers, thereby limiting the usefulness of sedimentary DNA for reconstructing faunal changes through time. This issue was explored by Haile et al. (2007) using rock shelter sediments from the Hawkes Bay region. This study detected DNA of Mantell's moa (Pachyornis geranoides), little bush moa (Anomalopteryx didiformis) and North Island giant moa (Dinornis novaezealandiae) in the sediments, but also found DNA of post-European sheep (Ovis aries) within the same layers indicating vertical leaching, perhaps as a result of urine movement through the sediment. However, the authors concluded that the DNA record for animals that did not produce large amounts of urine (e.g. birds) was stratigraphically intact (Haile et al. 2007).

In addition to rock shelter sediments, ancient faunal DNA has also been recovered from New Zealand lake sediments. Matisoo-Smith et al. (2008) extracted DNA from a sediment core taken from the Hawkes Bay region and recovered 12S sequences belonging to native Gobiomorphus fishes throughout the core.

Palaeoecology

Within the past decade, coprolites (ancient dung) have offered unique insights into the diets, habitats, niches and ecological interactions of New Zealand's prehuman avifauna. Ancient DNA analyses have played a vital role in these studies, through allowing both the identification of the depositing species and consumed material. Most of the New Zealand studies have focused on moa (Wood et al. 2008; Wood, Wilmshurst, Worthy, et al. 2012; Wood, Wilmshurst, Wagstaff, et al. 2012; Wood, Wilmshurst, Richardson, et al. 2013), but kākāpō (Strigops habroptilus) coprolites have been identified using ancient DNA (Wood, Wilmshurst, Holzapfel, et al. 2012). The high morphological diversity of late Quaternary coprolites from New Zealand (Wood & Wilmshurst 2014) suggest that there is a wealth of dietary and ecological information about other prehistoric faunal species still to be uncovered.

Palaeoparasitology

DNA analyses have been used to identify the remains of ancient parasites from around the world, but almost exclusively of human parasites (e.g. Loreille et al. 2001; Oh et al. 2010; Myšková et al. 2014). Wood, Wilmshurst, Rawlence, et al. (2013) used remains from New Zealand to demonstrate that this approach could also be used for studying parasites of other fauna. The authors sequenced DNA of a diverse range of gastrointestinal parasite taxa from moa coprolites (including nematodes, trematodes and apicomplexans), highlighting the potential for ancient DNA palaeoparasitology to provide ‘real-time’ insights into the processes of co-evolution of host/parasite relationships and co-extinction (Wood, Wilmshurst, Rawlence, et al. 2013).

Longevity of DNA molecules

DNA from ancient New Zealand faunal remains has also provided insights into different aspects of DNA preservation and degradation. Allentoft et al. (2012) analysed the abundance of mitochondrial DNA fragments in samples of 158 radiocarbon dated moa bones, and used these data to determine the decay rate of DNA within the samples. The study estimated that a 242 bp fragment of DNA had a half-life of 521 years, which was considerably longer than previous estimates of DNA longevity at similar environmental conditions. Interestingly, the data also allowed an estimate of the upper limit of DNA retrieval at ideal conditions (–5 °C) of 6.8 million years (Allentoft et al. 2012), providing further evidence for dismissing claims of antediluvian DNA.

Evolutionary rates

Ancient DNA analysis allows DNA sequences from individuals at different time points to be compared, providing a real-time view of evolutionary processes. Ancient DNA preserved in Adélie penguin (Pygoscelis adeliae) remains from New Zealand's Ross Dependency have been used for this purpose, allowing evolutionary rates within this species to be quantified. Lambert et al. (2002) concluded that the actual observed rate within this penguin species was several times higher than estimates based on phylogenetic analyses, a result supported by subsequent analyses (Ritchie et al. 2004; Subramanian et al. 2009). However, evolutionary rate estimates based on time-structured DNA sequence data have since been the topic of considerable ongoing scientific debate (e.g. Emerson & Hickerson 2015; Ho et al. 2015).

The future of genetic research on ancient New Zealand fauna

In the 25 years since the first ancient DNA was sequenced from an extinct New Zealand bird species, genetic studies have provided important new insights into the ecology, evolutionary histories, phylogenetic relationships and biology of many taxa in our prehistoric fauna. Moreover, ancient DNA has resolved past population changes due to human impacts, many of which have clear implications for conservation (e.g. Rawlence, Kennedy, et al. 2015; Waters & Grosser 2016; Bergner et al. 2016). The incredible insights provided by ancient DNA analyses have resulted in increased application of such techniques over time. As recently as a decade ago ancient DNA research was the restricted realm of a few research groups internationally, but such work is fast becoming more commonplace. Moreover, costs associated with ancient DNA research (particularly high-throughput sequencing) continue to decrease, making the techniques more accessible to a wider range of researchers. Nevertheless, specialised facilities and rigorous approaches to minimising contamination will always be a prerequisite of ancient DNA research. A small suite of laboratories dedicated to ancient DNA research, and designed based on specific criteria (e.g. Knapp et al. 2012) and following strict protocols (e.g. Cooper & Poinar 2000), are now (2017) operating across New Zealand, with researchers (including postgraduate students) working on a variety of faunal-related projects.

Advancement of techniques and current limitations

From Sanger sequencing to high-throughput platforms, and single gene phylogenies to palaeogenomics, methods used in ancient DNA research have advanced dramatically; so much so that techniques now routinely used are almost unrecognisable compared to those that produced the first ancient DNA sequences in the 1980s. A broad range of DNA extraction protocols, library preparation methods and sequencing methods are now available, different combinations of which are ideally suited to specific sample types and research questions. Some of these applications (e.g. studying methylation patterns) have yet to be attempted on ancient New Zealand taxa, but promise to yield exciting new insights into our past. Moreover, as techniques continue to develop (McCormack et al. 2017), and costs reduce, ancient DNA research is likely to drastically improve our understanding of the history and ecology of a much wider range of taxa in New Zealand's fauna, such as invertebrates, reptiles and amphibians.

There are currently a few limitations in this area of research. Preservation of DNA is a key one, but the temperate climate, numerous caves and recent extinction events mean that New Zealand has a large number of well-preserved remains for DNA studies. Moreover, new and improved techniques continue to be developed for retrieving and enriching DNA from marginal specimens (e.g. Dabney et al. 2013; Gansauge & Meyer 2014; Gansauge et al. 2017). The comprehensiveness of reference sequence databases is another limiting factor, and can affect the ability to resolve taxa in studies that use DNA metabarcoding of sediments or bone fragments (e.g. Grealy et al. 2015). Comprehensive reference databases will be particularly important as past communities of new and diverse groups of organisms, such as invertebrates, begin to be explored.

New applications

As new analytical approaches become available, the types of information that can be gained through analysis of ancient DNA from ancient fauna are likely to increase, as are the number of potential applications of such work. De-extinction is one such example of a new research area that ancient DNA data can feed into. Although original ideas of resurrecting extinct species via cloning are now largely regarded as impossible (Shapiro 2017), current work is exploring the potential for resurrecting traits of extinct species in close relatives via gene editing technologies (Shapiro 2017). Such methods may also assist with reintroduction of genetic diversity to rare and threatened species (Johnson et al. 2016), and in future may play a role in ensuring the survival of some New Zealand endemic fauna. Ancient DNA research will play a key role in identifying important functional genes and characterising baseline genetic diversity to underpin these applications.

Minimising damage to specimens

As a final point in this review, and with a view to the future, it is worth considering that ancient remains are a finite resource, yet the extraction of ancient DNA from such specimens is an inherently destructive process. As the number of ancient DNA studies continues to increase, careful study design will become ever more important in helping to minimise non-essential sampling and ensure the protection of specimens for future generations. It is a promising trend that, over time, the amount of sample required for ancient DNA extraction has decreased markedly, from the original chunks of bone weighing several grams to current methods using powder from fine drill holes. The development and application of a ‘soaking out’ approach for DNA extraction using lysis buffers (e.g. Tennyson et al. 2015), which can leave no visible damage to specimens and thus retain morphological characters, also offers great potential for the future. However, several aspects of this method require further assessment before its widespread application can be advocated for, including the potential increased risk of contamination from the surfaces of specimens, the repeatability of this technique on the same specimen, and potential impacts on other analyses (e.g. radiocarbon dating, isotope analyses).

To some extent, re-sampling of specimens held in public collections could be greatly minimised through the immortalisation of ancient DNA extracts (e.g. Meyer & Kircher 2010; Gansauge & Meyer 2013) and depositing vouchers from these into a national archive. The immortalised libraries could then be, in theory, re-amplified indefinitely to provide a resource suitable for most (though not all) subsequent research applications.

Acknowledgements

We thank Jon Waters, Janet Wilmshurst, Alexander Boast, and two reviewers for helpful comments on this manuscript.

Disclosure statement

No potential conflict of interest was reported by the authors.

Additional information

Funding

T. L. Cole was supported by a University of Otago Postgraduate Scholarship. J. R. Wood was supported by Core Funding for Crown Research Institutes from the New Zealand Ministry of Business, Innovation and Employment’s Science and Innovation Group.