ArticlePDF Available

Abstract and Figures

For tens of millions of years the ratite moa (Aves: Dinornithiformes) were the largest herbivores in New Zealand’s terrestrial ecosystems. In occupying this ecological niche for such a long time, moa undoubtedly had a strong influence on the evolution of New Zealand’s flora and played important functional roles within ecosystems. The extinction of moa in the 15th century ce therefore marked a significant event in New Zealand’s biological history, not only in terms of biodiversity loss, but in the loss of an evolutionarily and ecologically distinct order of birds. Understanding the full extent and magnitude of this loss, and its implications for New Zealand ecosystems, depends upon a detailed knowledge of moa diets. Over the past 100 years, periodic discoveries of preserved moa gizzard content and coprolites (ancient preserved dung) have gradually begun to shed light on the diets of moa and their roles within New Zealand ecosystems. Here, we review how the study of such samples has shaped our understanding of moa diets through time. We then provide a synthesis of current knowledge about moa diets, including summarising 2755 records of plant remains from 23 moa gizzard contents and 158 moa coprolites. A clear picture is now emerging of distinct differences between the feeding ecologies of moa species, which together with differences in habitat preferences facilitated niche partitioning. Such insights provide empirical data to inform the debate surrounding the role of moa herbivory in the evolution of distinctive plant traits within the New Zealand flora. These data also help identify specific ecological functions and roles that have been lost due to the extinction of moa, and resolve to what extent these could be replaced via surrogate taxa.
Content may be subject to copyright.
1Wood et al.: The diets of moaNew Zealand Journal of Ecology (2020) 44(1): 3397 © 2020 New Zealand Ecological Society.
DOI: https://dx.doi.org/10.20417/nzjecol.44.3
REVIEW
The diets of moa (Aves: Dinornithiformes)
Jamie R Wood1* , Sarah J Richardson1 , Matt S. McGlone1 and Janet M Wilmshurst1,2
1Manaaki Whenua – Landcare Research, PO Box 69040, Lincoln 7640, New Zealand
2School of Environment, The University of Auckland, Private Bag 92019, Auckland 1142, New Zealand
*Author for correspondence (Email: woodj@landcareresearch.co.nz)
Published online: 7 February 2020
Abstract: For tens of millions of years the ratite moa (Aves: Dinornithiformes) were the largest herbivores in
New Zealand’s terrestrial ecosystems. In occupying this ecological niche for such a long time, moa undoubtedly

ecosystems. The extinction of moa in the 15th century 
biological history, not only in terms of biodiversity loss, but in the loss of an evolutionarily and ecologically
distinct order of birds. Understanding the full extent and magnitude of this loss, and its implications for
New Zealand ecosystems, depends upon a detailed knowledge of moa diets. Over the past 100 years, periodic
discoveries of preserved moa gizzard content and coprolites (ancient preserved dung) have gradually begun
to shed light on the diets of moa and their roles within New Zealand ecosystems. Here, we review how the
study of such samples has shaped our understanding of moa diets through time. We then provide a synthesis
of current knowledge about moa diets, including summarising 2755 records of plant remains from 23 moa

             
partitioning. Such insights provide empirical data to inform the debate surrounding the role of moa herbivory

ecological functions and roles that have been lost due to the extinction of moa, and resolve to what extent these
could be replaced via surrogate taxa.
Keywords: birds, extinct species, herbivory, New Zealand, palaeoecology, plant communities, seed dispersal
Introduction
New Zealand’s native fauna has a complex evolutionary history.
Isolated for tens of millions of years, the ancient vicariant fauna
of New Zealand, an emergent part of the continent Zealandia
(Mortimer et al. 2017), was supplemented by the dispersal of
new species from nearby landmasses (such as Australia and
New Caledonia) throughout the Tertiary Period (Tennyson
2010; Gibbs 2016). Adaptation, speciation and extinction of
fauna inhabiting Zealandia’s dynamic landscapes was driven
by natural processes such as volcanism, tectonic uplift and
climate change (Fleming 1979; Tennyson 2010; Gibbs 2016).
In the near-absence of terrestrial mammals, except for several
        
& Holdaway 2002; Worthy et al. 2006; Hand et al. 2013;
2015; 2018), groups such as birds, reptiles, amphibians, and
invertebrates have dominated New Zealand’s terrestrial faunal
communities for much of its history (Lee et al. 2009; Jones
et al. 2009; Gibbs 2010; Tennyson 2010; Worthy et al. 2011;

the 13th century  (Wilmshurst et al. 2008), initiated a period
of unprecedented and sustained ecological transformation in
New Zealand. Hunting (Anderson 1989a; Perry et al. 2014),
the removal of forests by burning (McGlone 1983; McWethy

(Rattus exulans
contractions, population declines and extinctions within
New Zealand’s native fauna (Worthy & Holdaway 2002;
Tennyson & Martinson 2006; Wood 2013).
Of all New Zealand’s recently extinct species perhaps
none have evoked as much interest and attention as the moa

had a long evolutionary history in New Zealand, and molecular
dating suggests that moa were present on New Zealand at
the time of (or shortly after) its separation from Gondwana
(Mitchell et al. 2014). Currently, the oldest fossils of moa
are from Early Miocene lacustrine deposits in Central Otago
(Tennyson et al. 2010), and demonstrate that by c. 20 million

birds. However, the radiation of moa into the nine species
          
   
ago (Bunce et al. 2009). This was driven initially by the uplift
of the Southern Alps creating a diversity of habitats and new
2 New Zealand Journal of Ecology, Vol. 44, No. 1, 2020
ecological niches, with subsequent speciation resulting from the
separation of the North and South Islands (Bunce et al. 2009).
Moa were the largest herbivores in New Zealand’s terrestrial
ecosystems at the time of human settlement, with adult body
masses ranging from c. 17 to c. 242 kg; the next largest being
the extinct South Island goose (Cnemiornis calcitrans) at c.
18 kg (Tennyson & Martinson 2006). On insular landmasses
around the world large herbivores played important roles within
prehistoric ecosystems (Hansen & Galetti 2009), and this was
almost certainly the case in New Zealand. For example, based
on the relationship between herbivore mass and dry matter
intake (Nagy 2005; Müller et al. 2013), the largest moa would
have consumed several kilograms of plant matter each day,

cycling (Tanentzap et al. 2013). Consumption of fruit, seeds
and spore-bearing tissues meant that moa also dispersed a range
of New Zealand plants (Clout & Hay 1989; Lee et al. 2010;
Wood et al. 2012a), and fungi (Boast et al. 2018).
In addition to their ecological roles, moa may also have left

Over the years a variety of New Zealand plant traits have been
attributed to co-evolution with moa, as they are suggestive of
reducing the impact of moa browsing. These include interlacing

Atkinson 1977; Bond et al. 2004), heteroblasty (Greenwood
& Atkinson 1977; Mitchell 1980), deciduousness (Batcheler
1989), toxins (Greenwood & Atkinson 1977; Batcheler 1989),
spines or spine-like structures (e.g. enlarged stinging hairs),
leaf loss and photosynthetic stems, mimicry and reduced
visual apparency (Burns 2010; Fadzly & Burns 2010), tough
  
status (Atkinson & Greenwood 1989). Alternative climatic or
abiotic causes have also been suggested in most cases (e.g.
Wardle 1963; McGlone & Webb 1981; McGlone & Clarkson
1993; Day 1998; Howell et al. 2002; McGlone et al. 2004).
However, testing which selection pressures gave rise to these

and in many cases it could have been a combination of factors
including moa browsing, insect herbivory and abiotic drivers.
Although moa have now been extinct for > 500 years
(Perry et al. 2014) their legacy continues to have a strong
        
still stand in forests today, and the entire range of ecological
consequences associated with moa extinction may not yet be
entirely realised (Wood & Wilmshurst 2016). Understanding
the contribution of moa to the co-evolution of New Zealand
plant traits, and the niches of moa in New Zealand’s terrestrial
ecosystems, relies upon a good understanding of what moa ate.
Moreover, with the introduction of a suite of large mammalian
herbivores to New Zealand over the past 150 years (King
2005), information on moa diets can help inform debates such
as ecological replacement (Wood & Wilmshurst 2019), and
rewilding (Wood et al. 2017a). Fortunately, relative to other
recently extinct large herbivores around the world, there is
good evidence for the diets of moa in the form of analysed
preserved gizzard content and coprolites (ancient preserved
dung; Fig. 1). However, reports of these analyses are widely

types of analyses performed (e.g. pollen, macrofossil, DNA)

Here, we attempt to resolve these issues and bring together all
the existing direct evidence for moa diet, thereby providing
    
species. We also compare these diets to examine how niche
       
re-examine the roles of moa in New Zealand’s pre-human
ecosystems based on available diet evidence.
In search of moa diets
Early evidence and speculation about the foods of moa
For several decades after Owen (1840) concluded that
New Zealand had once been home to an ostrich-like bird,

and description of moa remains. The ecology and behaviour
of moa were rarely considered during the late 19th century
and mentions of moa diets in publications from this period are
relatively uncommon (Worthy 1990). In perhaps the earliest
published thoughts on the topic, Owen (1844) singled out
several morphological features of moa (e.g. robust cervical
vertebrae and strong neck muscles) and suggested that these
were adaptations for “dislodging the farinaceous roots of the
ferns that grow in characteristic abundance over the soil of
New Zealand”. Soon after this F. G. Moore, perhaps swayed
by Owen’s opinion, claimed in correspondence with Gideon
Mantell (11 August 1849) that the diets of moa had consisted
of ferns and small lizards (Anderson 1989b). The importance
of ferns in sustaining moa was also mentioned by Hochstetter
(1867), and again by Owen (1883), who wrote that moa
ate “the peculiarly nutritious roots of the common ferns of
New Zealand… with buds, foliage, or other parts of trees…”
         
        
cave near Mt Nicholas (Lake Wakatipu). White reported
that they “consisted of undigested fragments of what looked
like the stalk of the fern” (White 1875). In another cave near
Queenstown, White (1875) found more coprolites, and these
too he claimed “contained undigested vegetable fragments,
some of which seemed to be branches and stalks of fern broken
into short pieces of three-quarters of an inch in length”. It
seems likely that White’s description of the content of these
specimens could have been biased by the earlier mentions
of ferns having been an important part of moa diets. Recent
examination by JRW of the specimens collected by White
(now held in the collections of Te Papa Tongarewa) revealed
that the “undigested fragments of what looked like the stalk
of the fern” were simply the twigs of woody dicotyledonous
species adhering to the surface of the droppings.
Ferns featured prominently in early suppositions of
moa diets, although not everyone supported this view. In
1890 Vincent Pyke told the Otago Philosophical Institute
“Hochstetter and others allege, without authority, that the
moa lived on fern roots, which seems to me absurd when the
natural capacity of the bird is considered” (Otago Witness
  
He believed that moa were birds of non-forested habitats, and
postulated that “seeds of the Phormium tenaxCordyline
Australis [sic]… large species of AciphyllaCoprosma, and
many other plants, had been at one time the favourite food of
the Dinornis, whilst the roots of the Aciphylla, of the edible
fern (Pteris esculenta) [bracken], and several other plants,
might have provided an additional supply of food when the
seeds of the former were exhausted. Moreover, I have no doubt
Dinornis, like those of the Apteryx,
were omnivorous, so that they did not despise animal food,
and thus lizards, grasshoppers, and other insects might also
3Wood et al.: The diets of moa
have constituted part of their diet.” Evidence contradicting the
fern-diet hypothesis was soon to come. In 1891, the Otago
Witness newspaper reported the discovery of moa bones in a

also contained well-preserved gizzard content. The article
noted that “The chewed grass from a well-preserved stomach
is particularly interesting”. That same year an accumulation
of moa gizzard stones was found eroding out of peat deposits
on Swampy Summit near Dunedin. Hamilton (1891) reported
that amongst the stones was some light-coloured vegetable
material that appeared to be the remains of plants consumed
by the moa. This material included “vast numbers of seeds
of Leucopogon and Coprosmas (?) [sic], and short twigs and

Figure 1. Sources of direct evidence for moa diets. A and B, In situ gizzard content found adjacent to the skeleton of an immature female
South Island giant moa (Dinornis robustus), Pyramid Valley swamp, North Canterbury (images from Burrows et al. 1981); C, moa gizzard
content from Scaife’s Lagoon, west Otago (Otago Museum Av3647); D, Moa coprolites; 1. Heavy-footed moa (Pachyornis elephantopus)
from Dart River Valley, West Otago; 2. Little bush moa (Anomalopteryx didiformis) from Mt Nicholas Station, West Otago (image from
Wood et al. 2012b); 3. South Island giant moa from Dart River Valley, West Otago; 4. Upland moa (Megalapteryx didinus) from Dart
River Valley, West Otago; 5. Upland moa from Euphrates Cave, Northwest Nelson (image from Wood et al. 2012a).
Despite Owen’s clear support for the idea that moa were
forest herbivores adapted for consuming ferns, subsequent
interpretation of morphological and dietary evidence had


by the early 20th century the predominant view was that moa
had been grazers of non-forested habitats. In demonstration of
this, Buick (1931) wrote that moa “browsed upon the hillside
grasses as the cattle do, and devoured leaves and berries when
it could reach them”. Subsequent discoveries of moa gizzard
content appeared to support the idea of a grass-dominated diet.
For example, moa gizzard contents discovered in a swamp
near Whanganui in 1936 were reported to contain “triturated
grass and occasional seeds of supplejack” (Northern Advocate
4 New Zealand Journal of Ecology, Vol. 44, No. 1, 2020
1936). Moa gizzard contents found at Pyramid Valley swamp
in North Canterbury were described as “like a large plum-
pudding in the yellow peat…often found close at hand, with
seeds, grass and other vegetable matter amongst the gizzard

material in the gizzards was grass, not, however in a condition

   
in moa gizzard contents (Otago Witness 1891; Northern

already been mentioned in the literature that moa were birds
of non-forest habitats (Haast 1871). It is possible, therefore,
       
digested plant material could well have been unconsciously
swayed by the prevailing views of the day, in the same way
          
from moa coprolites during the late 19th century. Subsequent,
more detailed analyses would reveal a distinct lack of grasses
within the moa gizzard contents and create a new paradigm
of moa ecology.
Detailed analysis of gizzard content establishes new
paradigm for moa diet
Since 1891 several specimens of preserved moa gizzard content
had been found, yet these had not been examined in any great
detail. Except for the specimen from Swampy Summit near
Dunedin (Hamilton 1891), all gizzard content samples had been
found in association with skeletons of moa that had been mired
in anoxic non-acidic bogs. In such situations the preservation
conditions are ideal both for moa bones and the plant matter
within their gizzards, which are often found together in situ.

not have simply grazed herbaceous vegetation came with Ruth
Mason’s assessment of gizzard content samples from Pyramid
Valley (Falla 1941). Within South Island giant moa (Dinornis
robustus  
c. 200 Coprosma rhamnoides seeds, and in an eastern moa
(Emeus crassus     
Prumnopitys taxifolia, Myoporum laetum, and Nertera (Falla
1941). Mason also reported (in Gregg 1972) that another
E. crassus gizzard content sample, as well as a sample from a
coastal moa (Eurypateryx curtus) gizzard, contained mainly
         
browsing rather than grazing feeding strategy. Supporting
evidence that moa were principally browsers came with the
detailed gizzard content analyses performed by Burrows
(1980a; 1980b; 1989) and Burrows et al. (1981). These gizzard
content samples came both from Pyramid Valley in North
Canterbury and Scaife’s Lagoon in West Otago and contained
broadly similar assemblages of plant remains, dominated by
twigs, leaves and seeds of a diverse range of tree and shrub
taxa. Importantly, Burrows et al. (1981) reported that the
twigs found in moa gizzards had sheared ends, indicating
that the birds had cut them with their beaks rather than having

a new paradigm of moa having been browsers of trees and
shrubs, most gizzard content samples assessed up until this
time (14 out of 17) had been attributed to just one genus;
Dinornis, which we now know included just one species in
the South Island (Dinornis robustus; Bunce et al. 2003). The
comparative rarity of gizzard content samples from Emeus
and Euryapteryx, and lack of any from the other three moa
genera (Anomalopteryx, Megalapteryx and Pachyornis), were
recognised by Burrows et al. (1981) as major limitations to
understanding the full dietary breadth of moa and gaining

New methodological approaches unlock the potential of
coprolites
During the 20th century, analysis of gizzard content samples
provided the greatest insights into moa diets. Consisting of
relatively undigested plant material, these samples typically
included numerous seeds, leaves and stem fragments that
      
material. Wood (2007a) expanded upon the existing body

    
samples from heavy-footed moa (Pachyornis elephantopus)
and two further samples of E. crassus, for which just two
gizzard specimens from Pyramid Valley had previously been
studied. A total of 23 moa gizzard content samples which can
be directly associated to individual moa skeletons have now
been examined, yet several additional samples in museum
collections still await analysis.
By the 1980s more moa coprolites had been discovered
than moa gizzard content samples, yet this alternative source
of moa diet data had largely been unutilised. Being more
digested than gizzard content, the content of coprolites was
     
been found in situ with moa skeletons, allowing them to
be associated with a moa species (Falla 1941; Gregg 1972;
Burrows et al. 1981). Coprolites, on the other hand, were never
found in association with skeletons and so it was impossible to

which moa species might have deposited them. Neville Moar

of the content of a putative moa coprolite when he undertook
pollen analysis on a specimen collected from a rock overhang
at Shepherd’s Creek in the Waitaki Valley. Although the results
were entirely consistent with what is now known about the
diets of moa from dryland localities (Wood and Wilmshurst
2013), Moar was aware of the potential for the coprolite to
         
overhang, and of the multiple pathways through which the
pollen may have been incidentally ingested. Accordingly,
he concluded in a letter to Trotter that “It is not possible to

unpubl. correspondence).
       
analysis of putative moa coprolites, assessing the macrofossil
(seed, leaf, twig) and microfossil (pollen, phytoliths, starch)

Fiordland. The results supported the idea that moa browsed
trees and shrubs, but also provided evidence for grazing of
grasses and herbs around the margin of a nearby lake. With at
least three moa species having formerly inhabited the area –
D. robustus, little bush moa (Anomalopteryx didiformis), and
upland moa (Megalapteryx didinus) – it was not possible to

2004).
A breakthrough for coprolite analyses came with the
development of techniques for extracting, amplifying and
sequencing DNA molecules from ancient specimens in the
       
(of the extinct ground sloth) was performed in the late 1990s
(Poinar et al. 1998), and less than a decade later the potential

was demonstrated (Wood 2007b; Wood et al. 2008). Within
5Wood et al.: The diets of moa
the following decade, aDNA was used to identify c. 100 moa
coprolites to species, and in combination with conventional
diet proxies (e.g. pollen, macrofossils) provided further insights
into moa diets, niche partitioning and even parasite faunas
(Wood and Wilmshurst 2013; Wood et al. 2012a; 2012b;
2013a; 2013b; Boast et al. 2018). Evidence from coprolite
analyses supported the idea that some moa species browsed
trees and shrubs, as had previously been demonstrated by
gizzard content analyses (e.g. Burrows et al. 1981). However,
the coprolite analyses also revealed that a broader range of
feeding ecologies existed between (and even within some)
moa species, which included the grazing of small herbs in
non-forest communities. Moreover, coprolite analyses provided
answers to the age-old question of whether introduced ungulates
were performing the same ecological roles as moa had within
New Zealand ecosystems (Caughley 1983; 1988); in short,
they weren’t (Wood et al. 2008; 2013b; Wood & Wilmshurst
2019). Over 2000 moa coprolites are now known from more

Wilmshurst 2014).
Inferring diet from indirect evidence
In addition to the direct evidence of consumed plant species and
plant tissue types provided by coprolites and gizzard content,
other sources of information can also provide complementary
       
isotopes. The stable isotope ratios of animal body tissues,

been widely used (particularly carbon and nitrogen isotopes)
in palaeodietary analyses of extinct species (e.g. Hildebrand

1315N) can provide general
insights into the trophic level of an extinct species and details
about habitats and sources of food (e.g. marine v. terrestrial).
Although stable isotopes cannot provide dietary insights at the
resolution of individual prey species, in situations where the
range of potential prey is constrained, mixed-models can be

that contributed the diet (e.g. Bocherens et al. 2005).
Bone isotopes have been used in a relatively limited
capacity to study moa diets. While useful for discriminating
    1315   
13C and
15
factors, such as shading (e.g. forest vs non-forest conditions)

of diet if they are not controlled for by using samples with
restricted geographic and temporal ranges (e.g. Wood et al.
13C values of moa bones,
which became more depleted through time over the past 40000

        15N values
of P. elephantopus     
between eastern and western South Island sites which, rather
      
higher aridity in the east (Rawlence et al. 2012). Huynen et al.

(e.g. caves, dunes) than between moa species, though these
      
regions that the bones were from, rather than taphonomy.
Even within discrete environmental or climatic regions there
can be discrepancy between the diet interpretation based on
isotopes and the evidence provided by coprolites or gizzards,
revealing a complexity in interpreting isotope signals that
remains to be resolved (Rawlence et al. 2016). An interesting
application of bone isotopes to understanding moa diet was
the assessment of chick bones by Huynen et al. (2014), who
found isotopic values consistent with feeding on insects, as
is known for chicks of other ratites (e.g. Milton et al. 1993).
The second source of indirect evidence comes from
anatomy. Richard Owen’s descriptions of moa as specialist
fern consumers (Owen 1844) provides an early example of
how anatomical, and in particular osteological, features have
been used to infer aspects of moa feeding ecology. While
anatomical features cannot provide direct evidence of what
moa ate, they can provide insights into adaptations related to
feeding that can then be used to make inferences about diets.
Worthy & Holdaway (2002) provided an assessment and
comparison of the structure of moa skulls and beaks, revealing
       
feeding strategies. For example, the distinct mandibular groove
in Dinornis spp. appears to have been an adaptation for gripping
food (Worthy & Holdaway 2002), whereas the robust bill
and large temporal fossa of Anomalopteryx, combined with a
sharp edge on the mandible at the base of the gape suggest the
species was capable of a powerful secateur-like bite (Worthy
& Holdaway 2002).
Techniques for making such inferences have advanced
greatly in recent years (Wood & De Pietri 2015). Attard et al.
(2016) used three-dimensional FEA, a technique for modelling
and visualising stresses and strains within complex three-
dimensional structures (Young et al. 2012), to examine the
       
Their analyses supported niche partitioning between moa
genera, revealing better structural performance by skulls
    
indicated that A. didiformis was adapted for tugging (pulling
backwards) and support the contention that it performed
unilateral clipping (clipping twigs with one side of beak in a
secateur-like fashion), coastal moa (Euryapteryx curtus) had
a relatively weak skull adapted for plucking soft leaves and
fruits, crested moa (Pachyornis australis) was adapted for
both lateral shaking and pulling downwards (in a dorsoventral
direction), M. didinus for tugging and D. robustus for lateral

were noted, the authors did not discuss the types of plants

A synthesis of direct evidence for moa diets:
Methods
Assembly of moa herbivory database
We assembled data on moa herbivory within a spreadsheet (see
Supplementary Materials Appendix S1). Each row contained
a single observation, with columns for: moa species, unique
      
gizzard), locality, proxy (e.g. pollen, seed, leaf, DNA), high
taxonomic level identity (where it was not possible to resolve
the identity to family, genus or species, e.g. monocotyledon),

       

Plants database (Allan Herbarium 2000).
The completed database (Appendix S1) includes 2755
records from 23 gizzard content samples and 158 coprolites
(Figs. 2, 3). Data included are as follows:
6 New Zealand Journal of Ecology, Vol. 44, No. 1, 2020
Figure 2. Locations and numbers of gizzard content samples (white circles) and coprolites (black circles) for each moa species that

Figure 3. Cumulative number
of plant genera recorded from
moa gizzard content and
coprolite samples since 1941,
in total, and by each proxy
type (macros includes seeds,
leaves, stems and bark).
7Wood et al.: The diets of moa
Gizzards. Pyramid Valley, North Canterbury (macrofossils from
Falla 1941; Gregg 1972 and Burrows et al. 1981); Scaife’s
Lagoon, West Otago (macrofossils from Burrows et al. 1981;
Wood 2007a); Styx mire, East Otago (macrofossils from Wood
2007a); Treasure Downs, North Canterbury (macrofossils
from Wood 2007a).
Coprolites. Dart River Valley, West Otago (macrofossils from
Wood et al. 2008; pollen, macrofossils and rbcl DNA sequence
data from Wood et al. 2013b); Euphrates Cave, Northwest
Nelson (pollen, macrofossils and rbcl DNA sequence data
from Wood et al. 2012a); Takahe Valley, Fiordland (pollen
and macrofossils from Horrocks et al. 2004); Shepherd’s
Creek, Waitaki (pollen from Trotter 1970); Central Otago
rock overhangs (macrofossils from Wood et al. 2008; pollen
from Wood & Wilmshurst 2013); Mount Nicholas, West Otago
(pollen from Wood et al. 2012b; JRW unpubl. pollen data);
Old Man Range, Central Otago (JRW unpubl. pollen data);
Borland Burn, Western Southland (JRW unpubl. pollen data).
Data excluded from the database, and reasoning for these
exclusions, are as follows: phytoliths from the Takahe Valley
coprolites (Kondo et al. 1994; Horrocks et al. 2004), due to
their relatively poor taxonomic resolution and discrimination
of plant taxa; pre-1920s accounts of gizzard or coprolite
content, as the observers were unlikely to have had access to
comprehensive comparative material for identifying the plant

(Rawlence et al. 2011) due to the indistinct boundary between
the putative gizzard content and surrounding peat matrix; 18S
rDNA sequence data from Dart River and Euphrates Cave
coprolites (Boast et al. 2018), as the universal primers used
provided relatively poor discrimination of plant taxa and, as

held inherent biases in the taxa they detected. The database
was solely for herbivory on plants. Therefore, the limited
information on fungi consumption by moa (Boast et al. 2018)
was also excluded from the database but is discussed in later
sections of this paper.
Data in the database are raw counts unless otherwise stated
(in the notes column). A count value of 0.01 denotes that an
item was noted as being present in a sample, but its abundance

was provided (e.g. values of > 100 or > 600 in Burrows et al.
1981), the minimum value (i.e. 100 or 600) was used.
We took a conservative approach with attributing gizzard

“?Dinornis” or “probably Dinornis” (Burrows 1980b) were


Records, Pyramid Valley’, held by the Canterbury Museum)
to re-assess the identity of moa gizzard content samples from
Pyramid Valley, and to determine the sex of the D. robustus
individuals that had gizzard content associated with them.
 
Dinornis by Burrows et al. (1981) had an unclear association
     
positioned halfway between two adjacent skeletons of
D. robustus and E. curtus. A sketch plan of the excavation
       
assigned to F”, F being the E. curtus skeleton. Another note on
the same specimen read “Large gizzard, many stones, green,
?No. 121D, depth 53”, D being the D. robustus skeleton. Due
to this uncertain association, we have listed this specimen as

Apart from these instances, all other Pyramid Valley
Dinornis gizzard content samples examined by Burrows
D. robustus. The
following are details of their associated skeletons (excavation

 
DNA analysis of Allentoft et al. 2010) of individuals: 122B
(Av28434) immature female*; 76D (XXIID, Av8468) adult
female*; 76K (XXIIK, Av8471) adult female*; 76M (XXIIM,
Av8473) adult female; 108D (Av15025) adult female*; 108E
(Av15024) immature; 89B (XVIIB, Av8462) adult female
(skeleton on display in Canterbury Museum); XA (Av13899)
immature female*.

notes that appear to relate to samples that have not previously
    
notebook of the 1949 excavation of Pyramid Valley held by
Canterbury Museum, the D. robustus skeleton excavated from
square 70G is noted to have had an associated sample of “Crop
content very well preserved and no sign of semi digestion…
Twigs, seeds & leaves of a small shrub”. However, due to a
lack of detailed assessment of the content, these samples are
not included in the database.
Statistical analyses
The database was restructured into a dataframe with one row for
each specimen and columns for: (1) moa species, (2) sample id,
(3) sample type, (4) locality; followed by one column for each
unique combination of plant taxonomic hierarchy that enabled
us to capture the data regardless of the taxonomic resolution
 

family when that was the only taxonomic data available for the
row) and proxy type (e.g. NA-Araliaceae-Pseudopanax-ferox-
seed, NA-Araliaceae-Pseudopanax-ferox-pollen, Monolete
fern-NA-NA-NA-spore etc.) using the spread function of the
tidyr package (Wickham & Henry 2018) in R v.3.3.2 (R Core
Development Team 2017). This dataframe was manipulated
depending on the analysis being performed (e.g. subset for
particular proxies or localities, or quantitative data converted
 
(MDS) analyses were performed using the vegan package
(Okansen et al. 2010) in R using default settings.
Plant avoidance assessment


Garibaldi Plateau, (3) north Canterbury limestone forests (a
compilation of Cheviot and Pyramid Valley). For each site, we

We then compiled a list of native plant species present at
 
as comprehensive as possible to account for species turnover
between the time that the moa coprolites were formed and the
present day. For the Dart Valley we pooled Mark (1977) and
unpublished species lists downloaded from the New Zealand
Plant Conservation Network (NZPCN) website (NZPCN
2019; Dart Valley track by BD Rance; Dart-Rees Track by
BD Rance; Mt Aspiring National Park by JW Barkla and M
      
species lists (Druce lists 189, 198, 306) accompanying Druce
et al. (1987) downloaded from NZPCN (2019). For north
Canterbury limestone forests, we merged species lists from
ecologically comparable sites to our two gizzard content
8 New Zealand Journal of Ecology, Vol. 44, No. 1, 2020
localities: (1) the Tiromoana Scenic Reserve at Mt Cass (citizen

the Waipara Gorge Scenic Reserve (citizen science records

near Motunau (Molloy 1981, 1983, 1986). These three sites
include both limestone outcrop vegetation and alluvial valley
bottoms where tall forest occurs, as these habitats would both
have been available to moa at Cheviot and Pyramid Valley.
For each site, we removed all plant species from the
compiled lists if that species, or a higher taxon containing
that species, was recorded in the moa diets from that site.
    
moa diets in the Dart Valley, we removed all members of the
Apiaceae from the list of native species from the Dart Valley.
The resulting lists for each site are an estimate of the taxa
not eaten by moa – or not detected in the diets of moa. To
minimise the chance that non-detection of a plant taxon in a

rather than avoidance we generated a list of plant species that
would have been locally available but not present in moa diets

known to have been eaten by moa, based on gizzard content

remained (See Supplementary Materials Appendix S2).
Results and discussion
Moa diets
Evidence from gizzard content samples and coprolites
demonstrate that adult moa were strictly herbivorous. No
animal remains have been found in any sample, except for rare
invertebrate fragments in coprolites that were likely co-ingested
with plant material (Wood et al. 2008). It has previously been
suggested that moa chicks could have fed on insects, and this
has been supported by limited bone isotope data for moa chicks
(Huynen et al. 2014). Moreover, several small coprolites exist
from a former moa nesting site at Sawers’ rock shelter, some of
which do include visible insect remains (Wood & Wilmshurst
2014). Unfortunately, poor DNA preservation at this site has
       
(Boast 2016), and coprolite morphotypes attributed to other
bird species (e.g. laughing owl, Ninox albifacies) and reptiles
are also present in this site (Wood & Wilmshurst 2014). Below
we provide a species-by-species summary of moa diets.
Little bush moa (Anomalopteryx didiformis)
Gizzards: (0) None.
Coprolites: (5) Dart River Valley, Otago (3) (Wood et al. 2013b);
Mt Nicholas Station, Otago (2) (Wood et al. 2012b) (Fig. 2).
Plant taxa occurring in more than one sample: Ground ferns
(monolete spore types) (5), Fuscospora (5), Dacrydium
cupressinum (5), Hymenophyllum (4), Lophozonia menziesii
(4), Coprosma (4), Asteraceae (3), Myrsine (3), Nothofagaceae
(3), Ophioglossum (3), Poaceae (3), Phyllocladus alpinus (3),
Podocarpus (3), Prumnopitys taxifolia (3), Muehlenbeckia
(3), Cyperaceae (2), Prumnopitys ferruginea (2), Acaena (2),
Rubiaceae (2).
Most frequently recorded items excluding pollen/spores:
Nothofagaceae leaves (3), Rubiaceae DNA (3).
Diet and habitat notes: Anomalopteryx didiformis had powerful
jaw muscles and a robust beak that appears to have been
adapted for cutting twigs (Attard et al. 2016). Their sharp-
edged mandible that overlapped the premaxilla at the base
of the gape would have had a secateur-like action (Worthy &
Holdaway 2002). Although limited coprolite samples exist
for this species, the dominance of leaf cuticle and DNA of
trees and shrubs in these samples and the distribution of this
moa species being restricted to forest indicates that it was
most likely a browser of plants within the forest understorey.
Based on the relatively enriched 13C isotope values for A.
didiformis bones (Rawlence et al. 2016) it has been suggested
that this species may have fed around forest margins (Worthy
         
depleted 13C isotope values within forest understoreys (van
der Merwe & Medina 1991). However, this is at odds with
the lack of non-forest plant species in coprolites of this moa
species. An alternative explanation is that fallen fruit and leaves
from the canopy, where 13C isotope values are enriched, may
have contributed a large component of the diet of A. didiformis
(Rawlence et al. 2016).
South Island giant moa (Dinornis robustus)
Gizzards: (9) Pyramid Valley, North Canterbury (8) (Burrows
et al. 1981); Scaife’s Lagoon, Otago (1) (Burrows et al. 1981)
(Fig. 2).
Coprolites: (20) Dart River Valley, Otago (19) (Wood et al.
2008; 2013b); Mt Nicholas Station, Otago (1) (Wood et al.
2012b) (Fig. 2).
Plant taxa occurring in more than one sample: Coprosma
(29), Fuscospora (20), ground ferns (monolete spore types)
(18), Asteraceae (15), Cyperaceae (15), Poaceae (14),
Muehlenbeckia (14), Lophozonia menziesii (13), Ophioglossum
(11), Acaena (11), Gonocarpus (10), Dacrydium cupressinum
(10), Prumnopitys ferruginea (10), Rubus (9), Olearia virgata
(9), Prumnopitys taxifolia (8), Coprosma rotundifolia (7),
Carex secta (7), Leucopogon fraseri (7), Nothofagaceae (7),
Epilobium (7), Pimelea (7), Plagianthus betulinus (6), Myrsine
divaricata (6), Aristotelia fruticosa (6), Geranium (6), Gunnera
(6), Mentha (6), Myrsine (6), Ranunculus (6), Rubiacae (6),
Melicope simplex (6), Hydrocotyle (5), Brassicaceae (5),
Phyllocladus alpinus (5), Caryophyllaceae (5), Coriaria (5),
indeterminate monocotyledon (4), Apiaceae (4), Phormium
tenax (4), Corokia cotoneaster (4), GaultheriaPernettya
(4), Podocarpus (4), Polygonaceae (4), Cordyline australis
(3), Einadia cf. allanii (3), Muehlenbeckia australis (3), M.
axillaris (3), M. complexa (3), Rubus cf. squarrosus (3), Urtica
(3), Pseudopanax cf. ferox (2), P. ferox (2), Danthonioideae
(2), Hymenophyllum (2), Teucridium parviflorum (2),
Leptospermum scoparium (2), Lophomyrtus obcordata (2),
Halocarpus (2), Coprosma cf. rhamnoides (2), Stackhousia
minima (2), Violaceae (2)
Most frequently recorded items excluding pollen/spores:
Coprosma seeds (13), Olearia virgata stems (9), Coprosma
rotundifolia seeds (8), Rubus stems (8), Carex secta seeds
(7), Cyperaceae seeds (7), Myrsine divaricata leaves (6),
Plagianthus betulinus stems (6), Rubiaceae DNA (6), Melicope
simplex seeds (6), Prumnopitys taxifolia leaves (6), Rubus
seeds (5), Prumnopitys taxifolia seeds (5), Coriaria seeds
(5), Nothofagaceae DNA (5), Phormium tenax seeds (4),
monocotyledon leaves (4), Corokia cotoneaster seeds (4),
GaultheriaPernettya seeds (4), Leucopogon fraseri leaves
(4), Myrsine divaricata seeds (4), Nothofagaceae leaves (4),
Polygonaceae DNA (4), Ranunculus cf. gracilipes seeds
(4), Rubus leaves (3), Cordyline australis seeds (3), Einadia
9Wood et al.: The diets of moa
cf. allanii seeds (3), Gonocarpus seeds (3), Muehlenbeckia
australis seeds (3), M. axillaris seeds (3), M. complexa seeds
(3), Ranunculus seeds (3), Rubus cf. squarrosus leaves (3),
Urtica seeds (3), Pseudopanax ferox seeds (3), Danthonioideae
DNA (2), Leucopogon fraseri seeds (2), Teucridium parviorum
seeds (2), Plagianthus betulinus bark (2), Plagianthus betulinus
seeds (2), Leptospermum scoparium capsules (2), Lophomyrtus
obcordata seeds (2), Coprosma cf. rhamnoides seeds (2),
Violaceae DNA (2)
Diet and habitat notes: Dinornis robustus had powerful jaw
muscles and a robust beak (Attard et al. 2016), which may
have supported a sharp cutting edge (Atkinson & Greenwood
1989). Gizzard content samples from D. robustus contain
twigs with sheared ends, demonstrating that the beak could act
D.
robustus coprolites from the Daley’s Flat: (1) a diet consisting
mainly of browsed forest trees and shrubs, dominated by
beech (Nothofagaceae) and Coprosma, (2) a diet dominated
by grazed herbs in non-forested habitats. Wood et al. (2013b)


than males) having fed on lower nutrient status trees and
shrubs while the males extended into non-forest habitats. This
interpretation is supported by the D. robustus gizzard content
samples from Pyramid Valley, which are exclusively from
    
the detection of sex-linked genetic markers (e.g. Bunce et al.
2003) from D. robustus coprolites may help test this hypothesis.
Eastern moa (Emeus crassus)
Gizzards: (4) Treasure Downs, North Canterbury (2) (Wood
2007a); Pyramid Valley, North Canterbury (2) (Falla 1941;
Gregg 1972) (Fig. 2).
Coprolites: (0) None.
Plant taxa occurring in more than one sample: Prumnopitys
taxifolia (4), Rubus (3), Bryophyte (2), Corokia cotoneaster (2),
Olearia (2), Carex (2), Eleocharis cf. acuta (2), Elaeocarpus
hookerianus (2), Leptospermum scoparium (2), Dacrycarpus
dacrydioides (2), Ranunculus gracilipes (2), Coprosma (2),
Veronica cf. pimeleoides (2)
Most frequently recorded items excluding pollen/spores:
Prumnopitys taxifolia seeds (3), Rubus seeds (3), Bryophyte
leaves (2), Corokia cotoneaster seeds (2), Olearia stems (2),
Carex seeds (2), Eleocharis cf. acuta seeds (2), Elaeocarpus
hookerianus seeds (2), Leptospermum scoparium capsules (2),
Dacrycarpus dacrydioides leaves (2), Prumnopitys taxifolia
leaves (2), Ranunculus gracilipes seeds (2), Coprosma seeds
(2), Veronica cf. pimeleoides leaves (2)
Diet and habitat notes: With the limited number of samples
known from E. crassus
this moa species appears to have been adapted for a diet of
soft plant tissues, mainly leaves and fruit of trees and shrubs.
The diet appears to have been broadly like that of E. curtus,
     
preferences and distributions (Worthy & Holdaway 2002).
Although the two species are found together at many sites, E.
crassus appears to have preferred coastal lowlands while E.
curtus is more dominant at inland sites in the eastern South
Island (Worthy & Holdaway 2002).
Coastal moa (Euryapteryx curtus)
Gizzards: (1) Pyramid Valley, North Canterbury (1) (Gregg
1972) (Fig. 2).
Coprolites: (1). Earnscleugh Cave, Otago (1) (Wood et al.
2008; Wood & Wilmshurst 2013) (Fig. 2).
Plant taxa occurring in more than one sample: Prumnopitys
taxifolia (2)
Diet and habitat notes: As with Emeus crassus, the limited
number of coprolite and gizzard content samples from
Euryapteryx curtus
the diet of this species. As with E. crassus there appears
to have been a bias towards leaves and fruits of trees and
shrubs. This is supported by the high stresses experienced by
E. curtus 
feeding strategies (Attard et al. 2016). With weakly constructed
mandibles and a relatively blunt bill tip (Worthy & Holdaway
2002), E. curtus appears to have been better adapted for
plucking fruit and leaves (Attard et al. 2016) rather than
cutting like secateurs.
         
content samples from Pyramid Valley (121D) may also be from
E. curtus. Positioned midway between sterna of E. curtus and
D. robustus, this sample was originally attributed to E. curtus
(excavation notes, Canterbury Museum) but later this changed
to D. robustus (excavation notes, Canterbury Museum and
Burrows et al. 1981). However, Burrows et al. (1981) noted
Dinornis samples so far
examined by having a relatively large number of Podocarpus
spicatus [Prumnopitys taxifolia] seed and a very large number
of leaves of the same species. There is also a relatively large
amount of Rubus sp. leaves and petioles”. This corroborates
what is known from other E. curtus gizzard content in a bias
towards softer plant material such as leaves and fruits.
Upland moa (Megalapteryx didinus)
Gizzards: (0) None.
Coprolites: (55) Euphrates Cave, Northwest Nelson (33)
(Wood et al. 2012a); Dart River Valley, Otago (21) (Wood et al.
2013b); Old Man Range, Otago (1) (Wood et al. 2008) (Fig. 2).
Plant taxa occurring in more than one sample.: Fuscospora
(54), ground ferns (monolete spore types) (50), Poaceae (49),
Asteraceae (45), Lophozonia menziesii (43), Ranunculus (41),
Dacrydium cupressinum (37), Cyperaceae (36), Coprosma
(35), Myosotis (34), Brassicaceae (30), Acaena (29),
Ophioglossum (27), Prumnopitys taxifolia (27), Apiaceae (25),
Cyathea colensoi (25), Gentiana (25), Caryophyllaceae (23),
Hymenophyllum (21), Astelia (19), Myrsine (19), Plantago (19),
Phyllocladus alpinus (19), Muehlenbeckia (19), Epilobium
(18), Podocarpus (18), Lactuaceae (17), Bryophytes (16),
Prumnopitys ferruginea (16), Phormium (12), Cyperaceae
cf. Scirpus (12), Ericaceae (12), Rubiaceae (12), Gonocarpus
(11), Fuchsia excorticata (11), Leucopogon fraseri (10),
Mentha (10), Bulbinella (9), Gaultheria (9), Nothofagaceae
(9), Pseudopanax colensoi (8), Polygonaceae (8), Ranunculus
cf. gracilipes (8), Urtica (8), Hydrocotyle (7), Campanulaceae
cf. Pratia (7), Coriaria (7), Elaeocarpaceae (7), Geranium (7),
Pseudopanax (6), Neomyrtus pedunculatus (6), Oxalidales (6),
Griselinia (5), Veronica (5), Urticaceae cf. Urtica (5), Drosera
(4), Aristotelia fruticosa (4), Metrosideros (4), Anisotome (3),
Oreomyrrhus (3), Cyperaceae cf. Carex (3), Haloragaceae
(30), Loranthaceae (3), Halocarpus (3), Scrophulariaceae (3),
Pimelea (3), Pteridiophyta (3), Anthoceros (2), Colobanthus
(2), Ascarina lucida (2), GaultheriaPernettya (2), Peraxilla
(2), Lycopodium australinum (2), L. scariosum (2), Euphrasia
(2), Oxalidacae (2), Phymatasorus (2), Rubus (2), Acaena (2),
Donatia novae-zelandiae (2)
10 New Zealand Journal of Ecology, Vol. 44, No. 1, 2020
Most frequently recorded items excluding pollen/spores:.
Ranunculus seeds (17), Bryophyte leaves (16), Cyperaceae
cf. Scirpus seeds (12), Rubiaceae DNA (12), Gaultheria seeds
(9), Polygonaceae DNA (8), Ranunculus cf. gracilipes seeds
(8), Campanulaceae cf. Pratia seeds (7), Cyperaceae seeds (7),
Elaeocarpaceae DNA (7), Lophozonia menziesii leaves (6),
Nothofagaceae DNA (6), Oxalidales DNA (6), Leucopogon
fraseri seeds (5), Myrsine DNA (5), Urticaceae cf. Urtica seeds
(5), Asteraceae seeds (4), Griselinia DNA (4), Coprosma seeds
(4), Asteraceae DNA (3), Coriaria seeds (3), Cyperaceae cf.
Carex seeds (3), Leucopogon fraseri leaves (3), Haloragaceae
DNA (3), Loranthaceae DNA (3), Nothofagaceae leaves (3),

leaves (3), Myosotis DNA (2), Colobanthus seeds (2), Coriaria
DNA (2), Cyperaceae DNA (2), GaultheriaPernettya seeds
(2), Fuchsia excorticata seeds (2), Euphrasia DNA (2),
Oxalidaceae DNA (2), Poaceae seeds (2), Acaena DNA (2),
Veronica DNA (2), Urtica seeds (2)
Diet and habitat notes: Megalapteryx didinus had a widely-
varied diet, which included browsing of trees, shrubs and
herbs. Megalapteryx didinus is the only moa species for which
a preserved ramphotheca (the sheath over the beak) is known
(Worthy & Holdaway 2002), but this has not yet been described
in detail. Finite element analyses suggest that M. didinus was
better adapted for more precise twisting than other moa species
and shared broad functional similarities with emu (Dromaius
novaehollandiae) skulls (Attard et al. 2016).
Megalapteryx didinus may have exhibited some seasonal
diet variation, feeding in non-forested habitats most of year
but within forest during winter. Bones have been found in
mountainous regions that would have been snow-covered
during winter, suggesting the species may have had seasonal
      

Porphyrio hochstetteri) does. Megalapteryx didinus
was not restricted to high-altitude sites as its common name
(the upland moa) suggests, but instead was a specialist of steep
and rocky habitats. With sharp claws and slender legs this
agile moa may have specialised in feeding on plants growing
in sites that were inaccessible to other moa species.
Heavy-footed moa (Pachyornis elephantopus)
Gizzards: (2) Treasure Downs, North Canterbury (2) (Wood
2007a) (Fig. 2).
Coprolites: (10) Dart River Valley, Otago (8) (Wood et al.
2013b); Kawarau Gorge, Otago (1) (Wood et al. 2008; Wood
& Wilmshurst 2013); Roxburgh Gorge, Otago (1) (Wood et al.
2008; Wood & Wilmshurst 2013) (Fig. 2).
Plant taxa occurring in more than one sample: Coprosma
(12), Asteraceae (10), Cyperaceae (10), Poaceae (10),
Muehlenbeckia (10), Fuscospora (9), ground ferns (monolete
spore types) (8), Ophioglossum (7), Gonocarpus (6),
Lophozonia menziesii (6), Ranunculus (6), Hydrocotyle (5),
Brassicaceae (5), Leucopogon fraseri (5), Polygonaceae (5),
Apiaceae (4), Caryophyllaceae (4), Mentha (4), Myrsine
(4), Dacrydium cupressinum (4), Prumnopitys ferruginea
(4), P. taxifolia (4), Acaena (4), Myosotis (3), Gentiana (3),
Geranium (3), Epilobium (3), Plantago (3), Phyllocladus
alpinus (3), Podocarpus (3), Ranunculus cf. gracilipes (3),
Rubus (3), Urtica (3), Apiales (2), Bryophyte (2), Einadia
triandra (2), Chenopodiaceae (2), Eleocharis cf. acuta (2),
Polystichum vestitum (2), Ericaceae cf. Cyathodes empetrifolia
(2), GaultheriaPernettya (2), Ericaceae (2), Rubiaceae (2),
Veronica cf. pimelioides (2), Pimelea (2)
Most frequently recorded items excluding pollen/spores:
Coprosma seeds (6), Polygonaceae DNA (5), Leucopogon
fraseri seeds (4), Leucopogon fraseri leaves (4), Ranunculus
seeds (4), Ranunculus cf. gracilipes seeds (3), Urtica seeds
(3), Apiales DNA (2), Bryophyte leaves (2), Asteraceae seeds
(2), Einadia triandra seeds (2), Carex seeds (2), Eleocharis cf.
acuta seeds (2), Cyperaceae seeds (2), Polystichum vestitum
leaves (2), GaultheriaPernettya seeds (2), Gonocarpus seeds
(2), Rubiaceae DNA (2), Veronica cf. pimeleoides leaves (2)
Diet and habitat notes: Pachyornis elephantopus appears
to have been mainly a grazer. Coprolites of this species are
dominated by short-statured plants of non-forested habitats
and lianes. Matted grass leaves cf. Poaceae found between
a P. elephantopus    
swamp (Rawlence et al. 2016) could also represent grazing,

eaten by the moa. Large twigs were found in a gizzard content
sample from Styx swamp attributed to P. elephantopus on the
basis that this moa species was the most abundant at the site
(Wood 2007a). However, the gizzard content was actually not
associated with a particular skeleton and so this association
is tentative. Finite element analysis of the congeneric crested
moa (P. australis) revealed adaptation of skulls in this genus
for pulling down in a dorsoventral direction (Attard et al.
2016). Such a motion is often exhibited by birds such as geese
while grazing small herbs, with a small downwards tug being
used to break the plant before consuming. Such a motion also


lateral shaking. The P. elephantopus diet, being dominated by
non-forest plant taxa, would have been particularly well-suited
to glacial periods, when forest extent became restricted and
non-forest habitats expanded (Wood et al. 2016). In support
of this, genetic evidence indicates that the population size of
P. elephantopus may have increased around the height of the
last ice age (Rawlence et al. 2012), and this species tends to
dominate glacial loess bone deposits from the eastern South
Island (Worthy 1993).
Moa spp. (unidentied specimens)


Gizzards: (7) Pyramid Valley, North Canterbury (3) (Burrows
D.
robustus or E. curtus); Scaife’s Lagoon, Otago (3) (Burrows
et al. 1981; Wood 2007); Styx, Otago (1) (Wood 2007).
Coprolites: (67) Borland Burn, Southland (2) (Wood
unpublished); Dart River Valley, Otago (39) (Wood et al. 2008);
Euphrates Cave, Northwest Nelson (2) (Wood et al. 2012b);
Kawarau Gorge, Otago (2) (Wood et al. 2008); Old Man Range
(1) (Wood unpublished); Roxburgh Gorge, Otago (10) (Wood
et al. 2008); Sawers’ rock shelter, Otago (5) (Wood et al. 2008);
Shepherd’s Creek, Otago (1) (Trotter 1970); Takahe Valley,
Fiordland (5) (Horrocks et al. 2004).
Caveats with interpreting moa diet from gizzard and
coprolite analyses
When interpreting moa diets from gizzard and coprolite content,
it is prudent to consider the inherent biases of these two sample
types. Gizzard content samples are less digested than coprolites,

    
11Wood et al.: The diets of moa

seeds, which are preserved in gizzards, were destroyed by moa
digestive tracts and hence not preserved in moa coprolites
(Carpenter et al. 2018a). Having undergone more digestion,

those in gizzards and may be biased towards resistant objects
(such as small hard seeds). However, the content of gizzards
could be biased towards the plants that were growing around
the margins of the mire, and may therefore provide a biased
representation of the diets of free-ranging birds. Moreover,
gizzard samples are biased towards those moa species of
regions where miring sites are more frequent (lowlands east
      
for the contamination of gizzard content samples with plant
remains from the surrounding sediments (Wood 2007a). In
contrast, coprolites are discrete samples that can easily be
separated from the surrounding sediment (Wood & Wilmshurst
2016). However, coprolites are not found in association with
moa skeletons, as gizzards commonly are, and so attributing
them to a moa species requires aDNA analysis (Wood 2007a).
Moreover, many coprolite deposits may represent former moa
nesting sites (Wood 2008), and this could introduce seasonal
biases into dietary interpretations.
        
macrofossils, such as twigs, leaf fragments and seeds, provide
relatively reliable information on plant species consumed by
moa. However, while abundant items – such as the hundreds
of Coprosma seeds present in some gizzards (Burrows et al.
1981) – allow robust inference about diet, rare macrofossils
could represent accidental ingestion or sample contamination.
DNA metabarcoding of coprolites has allowed further

come from the potential for identifying consumed plant taxa

be particularly important where moa consumed easily-digested

plant DNA in moa coprolites is usually unclear, but based on
relative proportions of material comprising the coprolites is
likely to be dominated by consumed plant tissues rather than

throughput sequencing there is the potential for some of the
latter to be detected.
Pollen and spores in coprolites can represent either the

– i.e. wind-dispersed pollen and spores that had settled on
the plants consumed by moa or the water they drank. Trees
that produce large amounts of wind-dispersed pollen (e.g.
Nothofagaceae, Podocarpaceae) are frequently present in
   
      
In contrast, pollen from plant taxa that have limited pollen

are more likely to represent plants that were eaten by moa. So,
while pollen can provide important insights into moa diet (e.g.
consumption of Phormium, Astelia, and Fuchsia; Wood et al.
2012b) a certain degree of judgement is required in assessing

by moa. Wood et al. (2012b) proposed a method for separating

in moa coprolites using a combination of rank abundance
and trait scores. The method assumed that if all pollen-types
were environmental by-catch then their relative abundance

et al. 2011). Any pollen types that were more abundant than

intentionally consumed by moa.
  
some taxonomic bias. For example, plant macrofossils can often
be resolved to species, while pollen is typically resolved to
genus or family. The amount of taxonomic resolution provided
by DNA sequences depends partly upon the length of the
sequence, which in ancient specimens such as coprolites is
usually restricted to < 200 bp. Most plant taxa resolved from
moa coprolites have been to family or higher taxonomic ranks.
Niche partitioning
In addition to looking at the diets of moa species individually,

to examine niche partitioning. However, when interpreting the


(for example, there are no pollen data for gizzards, and not
all coprolites have DNA or pollen data), and that vegetation
communities vary between localities. Therefore, it is necessary
to subset the entire dataset into specimens with comparable
data from distinct ecological regions before analysis. With

examining niche partitioning in moa: gizzard content samples
from North Canterbury and coprolites from the Wakatipu
Basin in western Otago.
North Canterbury
Plant macrofossils in gizzard content samples from the two
swamp sites in North Canterbury (Pyramid Valley and Treasure
Downs) show relatively good separation between moa species
(Fig. 4). Gizzard 121D from Pyramid Valley, which was
excavated halfway between the skeletons of D. robustus and
E. curtus, clusters with D. robustus gizzards, as do the other
         
Burrows et al. (1981) noting that the content of gizzard 121D
Dinornis samples”, and reinforces the
uncertainty about the identity of this sample. Gizzard content
samples from Treasure Downs formed a cluster distinct
from those of Pyramid Valley, despite the two sites being in
relatively close geographic proximity. Gizzard content
samples from E. crassus, which was the only species to occur
at both localities, were also quite distinct between localities
(Fig. 4). This highlights the fact that variability between

 
between moa species when pooling samples from many

Wakatipu Basin
Two localities in the Wakatipu Basin have yielded moa
coprolites from multiple species: Daley’s Flat in the Dart River
Valley and Mt Nicholas Station. Pollen and plant macrofossils
       
but plant DNA has only been sequenced for some. With the
addition of the Mt Nicholas coprolites the patterns of moa
species separation evident in the Dart River coprolites (as
shown by Wood et al. 2013b) remain unchanged. The ordihulls
(representing breadth of diet composition) of A. didiformis
and P. elephantopus do not overlap, while D. robustus and M.
didinus have broad niche ranges (Fig. 5). Based on the plant
taxa present in the coprolites, A. didiformis is interpreted as
having fed within the forest, P. elephantopus in non-forest
12 New Zealand Journal of Ecology, Vol. 44, No. 1, 2020
Figure 4. MDS plot of moa gizzard content samples
from North Canterbury sites: diamonds, Pyramid
Valley Swamp; triangles, Treasure Downs Swamp,
Cheviot. Colours denote moa species: red, Dinornis
robustus; blue, Pachyornis elephantopus; dark green,
Euryapteryx curtus; orange, Emeus crassus; light
blue, unattributed to an individual moa skeleton
  Dinornis robustus gizzard
content samples represent those from immature birds.
Figure 5.
triangles, Dart River Valley. Colours denote moa species: red, Dinornis robustus; green, Pachyornis elephantopus; purple, Megalapteryx
didinus; black, Anomalopteryx didiformis.
13Wood et al.: The diets of moa
D. robustus and M. didinus across both these
locally available habitat types. Within the forest, D. robustus
appears to have favoured beech and Coprosma, while the
diets of M. didinus and A. didiformis included a more diverse
range of plants.
General patterns of moa diets

and food requirements, historically moa have been treated as

of moa vs other New Zealand avian herbivores (e.g. Clout &
Hay 1989; Lee et al. 2010; Forsyth et al. 2010) or in terms of


Batcheler 1989). Clearly there are aspects of their biology that
set moa apart from New Zealand’s other native herbivores,
and in some circumstances treating moa as a single entity is
      
and coprolites (i.e. those that have not been attributed to a

diet’. In the database assembled for this paper, 7 gizzards and

can still contribute important insights into the overall diets
of moa. With all available moa diet data now assembled, we
discuss some key observations about the general diets of moa.
Preferred plants
A selection of plant taxa occurred frequently in moa diets,
irrespective of moa species (Fig. 6). Those plants that were
favoured by moa are more likely to have experienced selection
pressure promoting the evolution of anti-browse defence
structures. A detailed review of the potential role of moa in

is overdue but out of scope for this paper. However, some of


The most frequently occurring plant genus, both in terms
of number of specimens (122, or 67.4%) and localities (12, or
75%) was Coprosma. Coprosma was also detected in specimens
from all moa species (Fig. 6), and across a range of proxies
(pollen, seeds, leaves). Rubiaceae DNA sequences also likely
represented Coprosma
Coprosma being a favoured food genus for moa, Wood &
Wilmshurst (2017) observed increases in the abundance of
Figure 6. Summary of the occurrence of plant genera (and the grass family Poaceae) in moa gizzard content and coprolites based on

moa species.
14 New Zealand Journal of Ecology, Vol. 44, No. 1, 2020
Figure 7. Frequency of occurrence of (a) plant families and (b) genera in moa gizzard and coprolite specimens. Taxon names are shown

Coprosma
the period after moa extinction. Together with Myrsine, which
was one of the most frequently occurring plant genera in moa
diet data (Fig. 7), these plant taxa contain high percentages
of divaricating forms, suggesting that perhaps divarication
did not reduce moa browse as has previously been suggested
(e.g. Bond et al. 2004).
A range of other frequently occurring plant taxa were
mainly found as pollen, and as taxa that produce large amounts
       
importance in terms of moa diet. Such taxa include Fuscospora,
Dacrydium, and Podocarpus (Fig. 7).
Surprisingly, the 19th century presumptions that moa were
fern specialists are supported to some extent here. Ferns do
appear frequently in the moa diet data: monolete spores (from
ground ferns) occurred in 91 specimens from 7 localities,
trilete spores 5 specimens from 2 localities. Hypolepis in two
specimens from one locality, Hymenophyllum in 31 specimens
Ophioglossum in 49 specimens from three
localities and Cyathea in 33 specimens from six localities.
Moreover, Polystichum leaves were observed in moa gizzard
content from Treasure Downs (Wood 2007a), and Wood &
Wilmshurst (2017) observed increases in the abundance of

period after moa extinction. Only one genus of moa, Emeus,
lacks any evidence for fern consumption (Fig. 6).
The high occurrence frequency of Muehlenbeckia
(Polygonaceae) in moa coprolites and gizzard content,
represented by seeds, pollen and DNA, is also an interesting
observation. This genus of lianoid plants appears to have been a
favoured food of moa. Another genus of native lianoids, Rubus
(bush lawyers), has been detected in the diet of every moa
species examined (Fig. 6). The relatively frequent occurrence
of mistletoe (Loranthaceae) pollen – resolved to Peraxilla by
Wood et al. (2013b) – and DNA in moa coprolites from sites
within beech forest – four from Takahe Valley (Horrocks et al.
2004), and six from Daley’s Flat (Wood et al. 2013b) – is also
worthy of note. At Daley’s Flat coprolites containing mistletoe
belonged to Megalapteryx and Anomalopteryx, and the Takahe
Valley coprolites are likely from the former species; a coprolite

being from M. didinus (see supporting information of Huynen
et al. 2010). These two species were browsers of short trees
and shrubs in the forest understorey, suggesting that mistletoes
may have formerly been more common at lower height tiers
in the past and have now been largely removed at these
levels by browsing ungulates. An alternative explanation is
   

year. However, pollen was not abundant enough in any of the

masse. Rather, consumption of nutrient-rich foliage that had
fallen to the ground, and leaves growing on plants within the
reach of the birds seem the most likely scenario.
Wood et al. (2017b) examined the degree of overlap
between the diets of D. robustus and M. didinus with extant
avian herbivore species currently in New Zealand. They found

been recorded in the diets of other birds). These were mainly
small shrubs, herbs and ferns (Wood et al. 2017a), including
the herbs Einadia, Myosotis, Ceratocephala and Myosurus.
The latter two of these are spring-annual herbs, which are
discussed in the section below.
What plants did moa avoid eating?
Moa had diverse diets, consuming a wide range of plant taxa.
However, by focussing only on the plants that moa ate we
may be missing a vital part of the story, that is, which plants
were avoided by moa? Understanding avoidance is a key
piece of evidence for the co-evolution of defence traits against

avoidance in extinct species, such as moa, is not a simple
matter. It requires not only a comprehensive assessment of
moa diet at a particular locality but also the plants that were
available for those moa to consume (i.e. the contemporaneous
composition of the local vegetation community). Our
assessment focussed on three localities where these two types
     
Garibaldi Plateau and north Canterbury limestone areas. At
these three localities, we resolved 25 plant species which we
strongly suspect were growing in the presence of moa, but
for which there was no evidence that moa ate them in any of
the coprolite or gizzard content samples from throughout the
South Island (Appendix S2).
15Wood et al.: The diets of moa
Several orchids were present but appear to have not
been eaten by moa (e.g. Corybas trilobus, Microtis unifolia,
Pterostylis montana and Thelymitra longifolia). The near
absence of orchids in moa diets (recorded only by 2 rbcL
clones from one coprolite at Euphrates Cave) seems odd,
given their ubiquity and apparently soft, edible foliage. Their
rarity in moa diets may be due to their growth habit of many
orchid species, having thin vertical stalks close to the ground,
yet a number of plants with similar growth habits are found
in moa diets (e.g. Ophioglossum, Ceratocephala pungens).
However, reduced visual apparency may explain the lack of
another species in moa diets, Parsonsia heterophylla, which
can be locally common but whose thin branches and leaves

cluster of plant taxa that were absent from moa diets include
small, long-leaved monocots, such as Luzula banksiana, L.
picta (Juncaceae), Arthropodium candidum (Asparagaceae)
and Libertia ixioides (Iridaceae). The tough leaves of these
plants may have reduced the likelihood that moa would have

with perceived defences against moa, including Discaria
toumatou (spines), Pseudowintera colorata (mottled leaves and
distasteful chemical compounds), Sophora (toxins) and Kunzea
(Leptospermum is also very rare in moa diets; small, tough
leaves, leaf volatiles and low-nutrient status). The absence of
Scheera digitata
plant is highly palatable and a preferred browse plant by deer
(Forsyth et al. 2002), and another araliad genus (Pseudopanax)
appears frequently in studies of moa diets.
Ecosystem functions provided by moa
Moa were the largest herbivores in New Zealand’s terrestrial
ecosystems at the time humans settled New Zealand, c. 750
years ago (Wilmshurst et al. 2008). Fossil evidence from the
early Miocene Manuherikia Group lacustrine sediments in
Central Otago suggest that their role as largest herbivores had
been established by at least 20 million years ago (Tennyson
et al. 2010), and it has been hypothesised that this niche
may have been occupied by moa ancestors not long after the
extinction of herbivorous dinosaurs (Mitchell et al. 2014).
Having been the largest herbivores for such a long-time it might
be expected that there has been some degree of co-evolution
between plants and moa, and it has been suggested that a

to moa browsing (Greenwood & Atkinson 1977; Atkinson &
Greenwood 1989; Batcheler 1989). Coprolites and gizzard
content may provide some insights into whether plants with
such traits were browsed by moa, but cannot provide answers
to these long-term evolutionary questions. However, coprolites
and gizzard content can provide insights into the ecosystem
processes that moa may have facilitated.
Pollination
There is no evidence to suggest that moa acted as pollinators.

which contain relatively high-levels of pollen from bird- or
    Phormium,
Fuchsia, and Astelia (Wood et al. 2012a). Even with wind-


Coprosma (Wood et al. 2013b). Birds covered in pollen as a


transfered pollen (Wood et al. 2012c). However, feeding on
    

pollinators.
Propagule dispersal
Moa are likely to have played a major role in seed dispersal.
Seeds are common in moa coprolites, and are often found
intact (Wood et al. 2008). Carpenter et al. (2018a) compared
the seeds present in moa gizzards and coprolites to show that
large seeds (> 3.3 mm) appear to have been destroyed during
passage through the digestive tract. This likely explains why
       
seen on other landmasses with extinct herbivores (Barlow
2000), and in extant associations such as with the southern
cassowary (Casuarius casuarius) in the rainforests of north
Queensland, Australia (Stocker & Irvine 1983). The largest

 Hemiphaga novaeseelandiae; Clout & Hay 1989)
and probably other birds such as weka (Gallirallus australis;

the identity of intact seeds from moa coprolites and those
recorded in droppings of other native birds such as kea (Nestor
notabilisStrigops habroptilus;
Butler 2006) and forest passerines (Williams & Karl 1996),
      
Therefore, while some plant species, especially herbs with small
indehiscent fruits (Wood et al. 2013b), may have experienced
reduced dispersal following moa extinction, this was likely
supplemented to some extent by other bird species, and later
by introduced birds and mammals (O’Donnell & Dilks 1994;
Williams & Karl 1996; Lee et al. 2010). A group of plants whose
seeds appear relatively frequently in coprolites from dryland
Central Otago and which are now quite rare are the spring
annual herbs including Ceratocephala pungens and Myosurus
minimus (Wood et al. 2008). These plants may be declining
as a result of reduced dispersal and habitat maintenance (soil

of moa (Rogers & Overton 2007; Lee et al. 2010) but more
work is required to demonstrate this conclusively.
Although we have dealt so far only with plants in the
diet of moa, it has been demonstrated through DNA analyses

fungi, including representatives of Cortinarius, Inocybe,
Tomentella, Lepiota, Geastrum and Lycoperdaceae (Boast
et al. 2018). These fungal taxa appear relatively commonly in
moa coprolites, occurring in 12 of the 19 (63.2%) specimens
examined by Boast et al. (2018). Cortinarius and Inocybe are
ectomycorrhizal species, and fruiting bodies of several species
within Cortinarius are sequestrate and colourful; traits that have
been suggested could be adaptations to enhance consumption
and spore dispersal by birds or reptiles (Beever 1999; Beever
& Lebel 2014). Although the results of Boast et al. (2018)
provide an indication that moa could have been consumers and
dispersers of key fungi such as ectomycorrhiza in New Zealand


is because the fungal DNA could originate from ingestion of
soil or plant matter containing hyphae of the fungi.
Nutrient cycling
Although moa did not change the total nutrient supply in
terrestrial ecosystems (as seabirds do, for example), herbivory
16 New Zealand Journal of Ecology, Vol. 44, No. 1, 2020
and defecation by moa would have accelerated nutrient cycling
rates, moved nutrients across the landscape, and concentrated
nutrients where they defecated. However, demonstrating that
    
communities is a challenging prospect. Tanentzap et al. (2013)

moa-derived nutrients on plants, applying hen (Gallus gallus
domesticus) manure to experimental plots as an analogue for
moa dung. In terms of nitrogen, the application rates used
(100 kg N ha) equated to 0.526–1.124 kg dry weight of moa
dung m (calculated using the %N values for moa coprolites
provided by Tanentzap et al. 2013), cf. accumulation rates of <
0.00001 kg dry weight of cattle dung m day for rangelands
(Tate et al. 2003). If moa dung was comparable to sheep dung
and c. 60% water by weight (Araújo 2010), then this would
translate to 1.3–2.8 kg m. While such rates may have been
achievable in discrete sites with a high frequency of moa
occupation (e.g. rock shelters), they are exceedingly high
compared to most natural situations. Moreover, in standardising

levels in the hen manure compared with moa coprolites meant
that experimental P application rates were equivalent to 7–32
kg of moa dung m. Despite these unrealistically high rates
of application, which were acknowledged by the authors,
native plant communities showed little response. Only fast-
growing lianes (e.g. Muehlenbeckia) responded positively to
the excessive nutrient application, along with non-native plant
species that would not have co-occurred with moa.
Other roles of moa dung in the environment
Although the contribution of moa dung to nutrient cycling
is unclear, it may have had other roles as food or habitat
      
ecosystems. For example, coprophilous fungi such as
Sporormiella are known to have grown on moa dung and
used the incidental ingestion of their spores by moa as part
of their life-cycle (Wood et al. 2011). New Zealand has 15
endemic species of dung beetle (Stavert et al. 2014a), and
experimental work has shown that they tend to be generalist

both carrion and dung (Stavert et al. 2014a; 2014b). With an

and introduced mammals (Stavert et al. 2014b), the relatively
large dung of moa would likely have once contributed to
the diet of New Zealand’s dung beetle fauna (Stavert et al.

(Splachnaceae) within the genus Tayloria. Dung mosses grow
on animal dung and carcasses, and their spores are dispersed by

that Tayloria would have grown on moa dung, evidence for

for moss growth are unlikely to favour the preservation of
coprolites. Finally, moa dung may have acted as a reservoir of
parasites within the environment (Wood et al. 2013a), which

bird species through the ingestion of infected faecal matter
(e.g. Dolnik et al. 2010).
Ecological replacement
Can the ecological roles left vacant following the extinction
       
gizzard content and coprolites, the two main roles of moa
within ecosystems appear to have been herbivory and seed
dispersal. We can explore each of these separately.
It is clear that moa dispersed seeds of a wide range of
plants. Although dispersal distances were unlikely to have
been as far as for volant birds, moa may still have ranged over
several kilometres. However, although the extinction of moa
may have led to reduced seed dispersal for some plants, there
appear to be no instances of plants that relied on moa solely
for their dispersal. Seed dispersal is commonly facilitated by
New Zealand fauna, and the seed dispersal roles once provided
by moa continue to be performed by extant native bird species,
     Prosthemadera novaeseelandiae),
silvereye (Zosterops lateralis) weka, waterfowl, reptiles
(Kelly et al. 2010; Wotton et al. 2016), invertebrates (Duthie
et al. 2006), and also some introduced species, e.g. blackbird
(Turdus merula; Kelly et al. 2010).
Herbivory was another ecological process performed by
moa. In non-forest habitats extant grazing birds may provide
some degree of surrogacy for moa, and in fact may have
been the main herbivores in such plant communities even
in the presence of moa. For example, waterfowl alone can

native wetland turf communities (Lee et al. 2010; Korsten
et al. 2013). Perhaps the most important herbivore niche left
vacant following the extinction of moa was that of a browser
of forest understorey trees and shrubs. While native fauna
         
to some extent, they do not browse twigs and leaves in the
understorey in the same way that some moa species once
did. As the largest herbivores in New Zealand forests during



open-understorey forests created and maintained by deer may
have appeared similar to those in which moa browsed hundreds
of years ago. At a broad scale this seems like a sensible idea,
          
browser. Yet the comparison does not hold up under closer


and hooves to their higher population densities and fecundity.
Accordingly, their ecological impacts appear to far exceed those

on native ecosystems has been comprehensively dismissed
based on multiple lines of evidence, including an assessment

content of moa coprolites and deer faeces, and pollen analysis
of forest soil cores to examine vegetation responses to moa
extinction (Atkinson & Greenwood 1989; Caughley 1989;
Duncan & Holdaway 1989; Forsyth et al. 2010; Tanentzap
et al. 2013; Wood & Wilmshurst 2017; 2019; Wood et al. 2008;
2012a; 2013b). Visions of extant large ratite species such as,
emu (Dromaius novaehollandiae) or ostrich (Struthio camelus),

of moa (Nicholls 2006; Bond et al. 2004) are also misguided.
Again, while these species may perform some of the seed

a large avian tree and shrub browser that is now missing from

means that extant large ratites could not consume enough
nutrition to survive if feeding on many of the divaricating
understorey shrubs once consumed by moa (Bond et al. 2004).
In addition to seed dispersal and herbivory, moa likely also
played a relatively minor role in a number of other ecosystem
processes, and some of these may also be being replaced by
17Wood et al.: The diets of moa
extant species. For example, it is possible that the dung of
ungulate herbivores may be replacing some of the roles of moa
dung. This has been demonstrated for fungal habitat (Wood
et al. 2011), but more research is required on aspects such as
their contribution to invertebrate foodwebs and nutrient cycling.
However, through an increased understanding of moa diversity,
diets and ecology, it is now clear that no single extant species

2013b). This is perhaps not surprising, given that moa were
c. 60 million years evolutionarily distinct from their nearest
living relatives; the comparatively small and volant tinamou
(Aves: Tinamidae) of South America (Mitchell et al. 2014).
Future directions
That coprolites and gizzard content samples have now been
found for all six genera of moa is quite remarkable. The ecology
of fauna that became extinct following initial human settlement
is perhaps better resolved in New Zealand than on any other
landmass globally. The relatively recent human settlement, a
temperate climate and geologically dynamic landscape with
many deposits suitable for preserving palaeofaunal remains
(e.g. caves, spring-bogs) and a long history of palaeofaunal
studies have contributed towards this understanding. However,
before we can achieve a more complete understanding of the
diets and ecology of our largest herbivores, there are still
some major gaps.

from 6 of the 9 moa species. There is still no direct evidence
of diet for P. australis, Mantell’s moa (Pachyornis geranoides)
and North Island giant moa (Dinornis novaezealandiae). Based
on similarity in habitat preferences and morphology it could
be assumed that D. novaezealandiae might have had a broadly
similar diet to D. robustus. However, P. australis was unique
in that it had a highly-restricted distribution and appears to
have been a specialist of subalpine habitat. Moreover, samples
from P. geranoides, the moa species with the smallest body
size, mean of 27 kg (Tennyson & Martinson 2006), would
undoubtedly provide interesting insights into the breadth of
moa diets. The lack of samples from D. novaezealandiae and
P. geranoides
content samples have yet been found in the North Island. Given
the high diversity of plants restricted to the North Island, the
discovery of such samples has the potential to greatly expand
the number of plant taxa known to have been eaten by moa,
which despite the wealth of data summarized in this paper has
not yet reached an asymptote (Fig. 3).
Second, there is a bias in the number of samples from those

been found. Dinornis robustus and M. didinus numerically
dominate studied samples, while just a handful of samples
from E. crassus and E. curtus means that the full dietary
breadth of these species is obscure. Additional samples from
these species, and additional samples from all species from
more localities, will help resolve the degree of diet overlap
between moa species and lead to a better understanding of
niche partitioning.
Finally, although moa played an important ecological
role as New Zealand’s largest terrestrial herbivores, they
were still just one part of a diverse avifauna. Coprolites of
a variety of shapes and sizes, likely representing a range
       
moa coprolites (Wood & Wilmshurst 2014). Applying the
techniques learnt through the study of moa diets to coprolites

et al. 2008; Wood et al. 2012c; Boast et al. 2018), will help
to broaden our understanding of New Zealand’s pre-human
ecosystems, former animal-plant interactions, and the legacy
of avian extinctions (Lee et al. 2010).
Acknowledgements

sketches from Pyramid Valley excavations and information
about the associated moa skeletons. This research was funded
by Strategic Science Investment Funding to the Crown
Research Institute from the Ministry of Business, Innovation
and Employment’s Science and Innovation Group
References

Plant Names Database. Manaaki Whenua Landcare



RN 2010. Highly skewed sex ratios and biased fossil
deposition of moa: ancient DNA provides new insight on
New Zealand's extinct megafauna. Quaternary Science
Reviews 29: 753–762.
Anderson AJ 1989a. Mechanics of overkill in the extinction
of New Zealand moas. Journal of Archaeological Science
16: 137–151.
Anderson AJ 1989b. Prodigious birds: Moas and moa-hunting
in New Zealand. Cambridge, Cambridge University
Press. 260 p.
Araújo GGL, Voltolini TV, Chizzotti ML, Turco SHN, de
Carvalho FFR 2010. Water and small ruminant production.
Revista Brasileira de Zootecnia 39 (supplement): 326–336.
Archey G 1941. The moa: a study of the Dinornithiformes.
Bulletin of the Auckland Institute and Museum 1. 135p.
Atkinson IAE, Greenwood RM 1989. Relationships between
moas and plants. New Zealand Journal of Ecology 12
(supplement): 67–95.

          

and prediction of biomechanical performance in avian
giants. Proceedings of the Royal Society of London B
283: 20152043.
Barlow CC 2000. The ghosts of evolution: Nonsensical fruit,
missing partners, and other ecological anachronisms. New
York, Basic Books. 260 p.
Batcheler CL 1989. Moa browsing and vegetation formations,
with particular reference to deciduous and poisonous
plants. New Zealand Journal of Ecology 12 (supplement):
57–66.
Beever RE 1999. Dispersal of New Zealand sequestrate
fungi. International Union of Microbiological Societies.
9th International Congress of Mycology, Sydney, 16–20
August, 1999. Abstracts p. 190.

of bird dispersal characteristics of fruit bodies. Auckland
Botanical Society Journal 69: 170–178.
Boast AP 2016. Ancient DNA of New Zealand’s extinct
fauna. Unpublished MPhil thesis, University of Adelaide,
Adelaide, Australia.
18 New Zealand Journal of Ecology, Vol. 44, No. 1, 2020
Boast AP, Weyrich LS, Wood JR, Metcalf JL, Knight R, Cooper
A 2018. Coprolites reveal ecological interactions lost with
the extinction of New Zealand birds. Proceedings of the
National Academy of Sciences USA 115: 15461551.
Bocherens H, Drucker DG, Billiou D, Patou-Mathis M,
Vandermeersch B 2005. Isotopic evidence for diet and
subsistence pattern of the Saint-Césaire I Neanderthal:
review and use of a multi-source mixing model. Journal
of Human Evolution 49: 7187.
Bocherens H, Drucker DG, Bonjean D, Bridault A, Conrad
NJ, Cupillard C, et al. 2011. Isotopic evidence for dietary
ecology of cave lion (Panthera spelaea) in North-Western
Europe: Prey choice, competition and implications for
extinction. Quaternary International 245: 249261.
Bond WJ, Lee WG, Craine JM 2004. Plant structural defences
against browsing birds: a legacy of New Zealand’s extinct
moas. Oikos 104: 500508.
Buick TL 1931. The mystery of the Moa. New Plymouth,
Thomas Avery & Sons. 357 p.
Bunce M, Worthy TH, Ford T, Hoppitt W, Willerslev E,
Drummond A, Cooper A 2003. Extreme reversed sexual
size dimorphism in the extinct New Zealand moa Dinornis.
Nature 425, 172–175.
Bunce M, Worthy TH, Phillips MJ, Holdaway RN, Willerslev

PJJ, Cooper A 2009. The evolutionary history of the extinct
ratite moa and New Zealand Neogene paleogeography.
Proceedings of the National Academy of Sciences USA
106: 20646–20651.
Burns KC 2010. Is crypsis a common defensive strategy in
plants? Speculation on signal deception in the New Zealand

Burrows CJ 1980a. Diet of New Zealand Dinornithiformes.
Naturwissenschaften 67: 151–153.
Burrows CJ 1980b. Some empirical information concerning the
diet of moas. New Zealand Journal of Ecology 3: 125–130.
Burrows CJ 1989. Moa browsing: evidence from the Pyramid
Valley mire. New Zealand Journal of Ecology 12
(supplement): 51–56.
Burrows CJ, McCulloch BJ, Trotter MM 1981. The diet of moas
based on gizzard contents samples from Pyramid Valley,
North Canterbury and Scaifes Lagoon, Lake Wanaka,
Otago. Records of the Canterbury Museum 9:309–336.
Butler DJ 2006. The habitat, food and feeding ecology of
       
MSc thesis of Richard Gray. Notornis 53: 55-79.
Carpenter J, Wood JR, Wilmshurst JM, Kelly D 2018a. An
avian seed dispersal paradox: New Zealand’s extinct
megafaunal birds did not disperse large seeds. Proceedings
of the Royal Society of London B 285: 20180352.
Carpenter JK, Kelly D, Moltchanova E, O'Donnell CF 2018b.
Introduction of mammalian seed predators and the loss

New Zealand tree Elaeocarpus dentatus. Ecology and
Evolution 8: 5992–6004.
Caughley G 1983. The deer wars. Auckland, Heinemann. 187 p.
Caughley G 1988. Plant–herbivore interactions. In: Seminar
2000: The future of New Zealand’s wild animals?
Proceedings of a national seminar organised by the
New Zealand Deerstalkers Association. Pp. 51–52.
Caughley G 1989. New Zealand plant-herbivore ecosystems
past and present. New Zealand Journal of Ecology 12
(supplement): 3–10.
Clout MN, Hay JR 1989. The importance of birds as browsers,
pollinators and seed dispersers in New Zealand forests.
New Zealand Journal of Ecology 12 (supplement): 27–34.
Day JS 1998. Light conditions and the evolution of heteroblasty
(and the divaricate form) in New Zealand. New Zealand
Journal of Ecology 22: 43–54.
  
foraging ecology on the prevalence and intensity of
coccidian infection in wild passerine birds. Ardea 98:
97–103.

of Tertiary calcareous rocks in the mountains of western
Nelson, New Zealand. New Zealand Journal of Botany
25: 41–78.
 
Waikari. Records of the Canterbury Museum 4: 330–8.
Duthie C, Gibbs G, Burns KC 2006. Seed dispersal by weta.
Science 311: 1575.
Fadzly N, Burns KC 2006. Hiding from the ghost of herbivory
past: evidence for crypsis in an insular tree species.
International Journal of Plant Sciences 171: 828–833.
Falla RA 1941. The avian remains. Preliminary report on
excavations at Pyramid Valley Swamp, Waikari, North
Canterbury. Records of the Canterbury Museum 4:
325–353.
Fleming CA 1979. The geological history of New Zealand
and its life. Auckland, Auckland University Press. 141 p.
Forsyth DM, Coomes DA, Nugent G, Hall GMJ 2002. Diet
and diet preferences of introduced ungulates (Order:
Artiodactyla) in New Zealand. New Zealand Journal of
Zoology 29: 323–343.
Forsyth DM, Wilmshurst JM, Allen RB, Coomes DA 2010.
Impacts of introduced deer and extinct moa on New Zealand
ecosystems. New Zealand Journal of Ecology 34: 48–65.
       
the dominance of birds in their evolutionary history?
New Zealand Journal of Ecology 34: 152–157.
Gibbs GW 2016. Ghosts of Gondwana. Nelson, Potton &
Burton. 367 p.
Greenwood RM, Atkinson IAE 1977. Evolution of divaricating
plants in New Zealand in relation to moa browsing.
Proceedings of the New Zealand Ecological Society 24:
21–33.
Gregg DR 1972. Holocene stratigraphy and moas at Pyramid
Valley, North Canterbury, New Zealand. Records of the
Canterbury Museum 9: 151–158.
Haast J 1871. Moas and moa hunters. Address to the
Philosophical Institute of Canterbury. Transactions and
Proceedings of the New Zealand Institute 4: 66–107.
Hamilton A 1891. Notes on moa gizzard-stones. Transactions
and Proceedings of the New Zealand Institute 24: 172–175.
Hand SJ, Worthy TH, Archer M, Worthy JP, Tennyson AJD,
     
Noctilionoidea) indicate a long history for endemic bats
in New Zealand. Journal of Vertebrate Paleontology 33:
1442–1448.
Hand SJ, Lee DE, Worthy TH, Archer M, Worthy JP, Tennyson

EM, Lindqvist JK 2015. Miocene fossils reveal ancient
roots for New Zealand’s endemic Mystacina (Chiroptera)
and its rainforest habitat. PloS ONE 10: e0128871.
Hand SJ, Beck RMD, Archer M, Simmons NB, Gunnell GF,

Worthy TH 2018. A new, large-bodied omnivorous bat
(Noctilionoidea: Mystacinidae) reveals lost morphological
19Wood et al.: The diets of moa
and ecological diversity since the Miocene in New Zealand.

Hansen DM, Galetti M 2009. The forgotten megafauna.
Science 324: 42–43.
Hildebrand GV, Farley SD, Robbins CT, Hanley TA, Titus K,
Servheen C 1996. Use of stable isotopes to determine diets
of living and extinct bears. Canadian Journal of Zoology
74: 2080–2088.
Hochstetter F 1867. New Zealand: its physical geography,
geology and natural history with special reference to the
results of Government expeditions in the provinces of
Auckland and Nelson. Stuttgart, J.G. Gotta. 515 p.
Horrocks M, D’Costa D, Wallace R, Gardner R, Kondo R
2004. Plant remains in coprolites: diet of a subalpine
moa (Dinornithiformes) from southern New Zealand
Emu 104: 149–156.
Horrocks M, Salter J, Braggins J, Nichol S, Moorhouse R,
Elliott G 2008. Plant microfossil analysis of coprolites of
the critically endangered kakapo (Strigops habroptilus)
parrot from New Zealand. Review of Palaeobotany and
Palynology 149: 229–245.
Howell CJ, Kelly D, Turnbull MH 2002. Moa ghosts exorcised?
New Zealand’s divaricate shrubs avoid photoinhibition.
Functional Ecology 16: 232–240.
Huynen L, Gill BJ, Millar CD, Lambert DM 2010. Ancient DNA
reveals extreme egg morphology and nesting behaviour in
New Zealand’s extinct moa. Proceedings of the National
Academy of Sciences USA 107: 16201–16206.
Huynen L, Gill BJ, Doyle A, Millar CD, Lambert DM 2014.
     
(Aves: Dinornithiformes) from the genus Euryapteryx.
PLoS ONE 9: e108995.
Jones MEH, Tennyson AJD, Worthy JP, Evans SE, Worthy
TH 2009. A sphenodontine (Rhynchocephalia) from the
Miocene of New Zealand and palaeobiogeography of the
tuatara (Sphenodon). Proceedings of the Royal Society of
London B 276: 1385–1390.
Kelly D, Ladley JJ, Robertson AW, Anderson SH, Wotton
DM, Wiser SK 2010. Mutualisms with the wreckage of an
avifauna: the status of bird pollination and fruit-dispersal in
New Zealand. New Zealand Journal of Ecology 34: 66–85.
King CM ed. 2005. The handbook of New Zealand mammals.
Melbourne, Oxford University Press. 630 p.
Kondo R, Childs C, Atkinson I 1994. Opal phytoliths of New
Zealand. Lincoln, Manaaki Whenua Press. 80 p.
Korsten AC, Lee WG, Monks A, Wilson JB 2013. Understanding
the role of birds in sustaining indigenous turn communities
in a lacustrine wetland in New Zealand. New Zealand
Journal of Ecology 37: 206–213.
Lee MSY, Hutchinson MN, Worthy TH, Archer M, Tennyson

geckos reveal long-term conservatism of New Zealand’s
lizard fauna. Biology Letters 5: 833–837.
Lee WG, Wood JR, Rogers GM 2010. Legacy of avian-
dominated plant-herbivore systems in New Zealand.
New Zealand Journal of Ecology 34: 28-47.
McGlone MS 1983. Polynesian deforestation of New Zealand:
a preliminary synthesis. Archaeology in Oceania 18: 11–25.
McGlone MS, Clarkson BD 1993. Ghost stories: moa, plant
defences and evolution in New Zealand. Tuatara 32: 1–21.

evolution of divaricating plants. New Zealand Journal of
Ecology 4: 20–28.
McGlone MS, Dungan RJ, Hall GMJ, Allen RB 2004. Winter

Journal of Botany 42: 1–19.
McWethy DB, Whitlock C, Wilmshurst JM, McGlone MS,

SC, Cook ER 2010. Rapid landscape transformation in
South Island, New Zealand, following initial Polynesian
settlement. Proceedings of the National Academy of
Sciences USA 107: 21343–21348.
Mark AF 1977. Vegetation of Mount Aspiring National Park
New Zealand. Wellington,National Parks Authority. 79 p.
Milton SJ, Dean WRJ, Linton A 1993. Consumption of termites
by captive ostrich chick. South African Journal of Wildlife
Research 23: 58–60.
Mitchell ND 1980. A study of the nutritive value of juvenile and
adult leaves of Pseudopanax crassifolius. New Zealand
Journal of Ecology 3: 159.
Mitchell KJ, Llamas B, Soubrier J, Rawlence NJ, Worthy TH,
Wood JR, Lee MSY, Cooper A 2014. Ancient DNA reveals

bird evolution. Science 344: 898–900.
Moar NT, Wilmshurst JM, McGlone MS 2011. Standardizing
names applied to pollen and spores in New Zealand
Quaternary palynology. New Zealand Journal of Botany

Molloy BPJ 1981. Proposed reserves: Coringa Farm Settlement.
Botanical Report. Botany Division Report no. 350,


limestone outcrop. Botanical Report. Botany Division
 
Research.
Molloy BPJ 1986. Cranky Tom Stream and Boundary Creek:
Botanical Report. Botany Division Report no. 584,

Mortimer N, Campbell HJ, Tulloch AJ, King PR, Stagpoole
VM, Wood RA, Rattenbury MS, Sutherland R, Adams
CJ, Collot J, Seton M 2017. Zealandia: Earth's hidden
continent. GSA Today 27: 27–35.
Müller DWH, Codron D, Meloro C, Munn A, Schwarm A,
Hummel J, Clauss M 2013. Assessing the Jarman–Bell
Principle: Scaling of intake, digestibility, retention time
     
Comparative Biochemistry and Physiology Part A:
Molecular & Integrative Physiology 164:129–140
Nagy KA 2005. Field metabolic rate and body size. The Journal
of Experimental Biology 208: 1621–1625.
Nicholls H 2006. Restoring nature’s backbone. PLoS Biology
4: e202.
       
Whangarei, 19 May 1936; pg. 3, col. 2.
NZPCN 2019. New Zealand Plant Conservation Network
  
observation_site_search.aspx (Accessed 11 October 2019).
O’Donnell CFJ, Dilks PJ 1994. Foods and foraging of
forest birds in temperate rainforest, South Westland,
New Zealand. New Zealand Journal of Ecology 18:
87–107.
Oksanen J, Blanchet FG, Kindt R, Legendre P, O’Hara RB,
Simpson GL, Solymos P, Stevens MHH, Wagner H 2010.


Otago Witness 1890. The moa. Dunedin, 18 September 1890;
pg. 31, col. 2.
Otago Witness 1891. Local & General. X, Dunedin, 17
September 1891; pg. 11, col. 1.
20 New Zealand Journal of Ecology, Vol. 44, No. 1, 2020
Owen R 1840. On the bone of an unknown struthious bird
from New Zealand, meeting of November 12, 1839.
Proceedings of the Zoological Society London for 1839,
Part VII, no. lxxxiii: 169–171.
Owen R 1844. Memoir on the genus Dinornis. Transactions
of the Zoological Society of London 3: 235–275.
Owen R 1883. On Dinornis (Part XXIII): containing a
description of the skeleton of Dinornis parvus, Owen.
Transactions of the Zoological Society of London 11:
233–256.
Perry GLW, Wheeler AB, Wood JR, Wilmshurst JM 2014.
A high-precision chronology for the rapid extinction of
New Zealand moa (Aves, Dinornithiformes). Quaternary
Science Reviews 105: 126–135.
Poinar HN, Hofreiter M, Spaulding WG, Martin PS,
Stankiewicz BA, Bland H, Evershed RP, Possnert G,
Paabo S 1998. Molecular coproscopy: Dung and diet
of the extinct ground sloth Nothrotheriops shastensis.
Science 218: 402–406.
R Core Development Team 2017. R: a language and
environment for statistical computing. Vienna, Austria,
R Foundation for Statistical Computing.

NT, Worthy TH 2011. New palaeontological data from

deposit, North Canterbury, South Island, New Zealand.
Journal of the Royal Society of New Zealand 41: 217–236.
Rawlence NJ, Metcalf JL, Wood JR, Worthy TH, Austin JJ,

change on the megafaunal moa of New Zealand in the
absence of humans. Quaternary Science Reviews 50:
141–153.
Rawlence NJ, Wood JR, Bocherens H, Rogers K 2016.
Dietary interpretations for the extinct moa (Aves:
Dinornithiformes) using coprolites and stable isotopes:
complimentary or contradictory? Quaternary Science
Reviews 142: 173–178.

herbs in rare non-forest ecosystems of New Zealand.
New Zealand Journal of Botany 45: 317–327.

for the analysis of patterns of extinction in the New Zealand
avifauna before European contact. Evolutionary Ecology
Research 5: 759–779.
Stavert JR, Gaskett AC, Scott DJ, Beggs JR 2014a. Dung
beetles in an avian-dominated island ecosystem: feeding
and trophic ecology. Oecologia 176: 259–271.
Stavert JR, Drayton BA, Beggs JR, Gaskett AC 2014b. The
volatile organic compounds of introduced and native
dung and carrion and their role in dung beetle foraging
behaviour. Ecological Entomology 39: 556–565.
Stocker GC, Irvine AK 1983. Seed dispersal by cassowaries
(Casuarius casuarius) in north Queensland’s rainforests.
Biotropica 15: 170–176.
Tanentzap AJ, Lee WG, Monks A 2013. Increased nitrogen
cycling facilitates native forest regeneration: potential
for restoring extinct ecological processes? Ecological
Applications 23: 36-45.
Tate KW, Atwill ER, McDougald NK, George MR 2003.
Spatial and tempral patterns of cattle feces deposition on
rangeland. Journal of Range Management 56: 432–438.
Tennyson AJD 2010. The origin and history of New Zealand’s
terrestrial vertebrates. New Zealand Journal of Ecology
34: 6–27.
Tennyson AJD, Martinson P 2006. Extinct birds of
New Zealand. Wellington, Te Papa Press. 160 p.
   
SJ 2010. Moa’s ark: Miocene fossils reveal the greay
antiquity of Moa (Aves: Dinornithiformes) in Zealandia.
In: Boles WE, Worthy TH eds. Proceedings of the VII
International Meeting of the Society of Avian Paleontology
and Evolution. Records of the Australian Museum 62:
105–114.
Trotter MM 1970. Archaeological investigations in the
Aviemore area, South Island. Records of the Canterbury
Museum 8: 439–453.

isotope ratios and foodwebs in Amazonia. Journal of
Archaeological Science 18: 249–259.
Wardle P 1963. Evolution and distribution of the New Zealand
 
Journal of Botany 1: 3–17.
White T 1875. Notes on moa caves, etc., in the Wakatipu
District. Transactions and Proceedings of the New Zealand
Institute 8: 97–102.
Wickham H, Henry L 2018. tidyr: Easily tidy data with


Williams PA, Karl BJ 1996. Fleshy fruits of indigenous and
adventive plants in the diet of birds in forest remnants,
Nelson, New Zealand. New Zealand Journal of Ecology
20: 127–145.
Wilmshurst JM, Anderson AJ, Higham TFG, Worthy TH
2008. Dating the late prehistoric dispersal of Polynesians
to New Zealand using the commensal Pacific rat.
Proceedings of the National Academy of Sciences USA
105: 7676–7680.
Wood JR 2007a. Moa gizzard content analyses: further
information on the diet of Dinornis robustus and Emeus
crassusPachyornis
elephantopus (Aves: Dinornithiformes). Records of the
Canterbury Museum 21: 27–39.
Wood JR 2007b. Pre-settlement paleoecology of Central
Otago’s semi-arid lowlands, with emphasis on the pre-
settlement role of avian herbivory in South Island dryland
ecosystems, New Zealand. Unpublished PhD thesis,
University of Otago, Dunedin, New Zealand.
Wood JR 2008. Moa (Aves: Dinornithiformes) nesting material
from the semi-arid interior of South Island, New Zealand.
Journal of the Royal Society of New Zealand 38: 115–129.
Wood JR 2013. New Zealand, 500 years ago. In: MacLeod
N ed. Grzimek’s animal life encyclopedia: Extinction.
Farmington Hills, Gale. Pp. 595–604.
Wood JR, De Pietri VL 2015. Next-generation paleornithology:
technological and methodological advances allow new
insights into the evolutionary and ecological histories of
living birds. The Auk 132: 486–506.
Wood JR, Wilmshurst JM 2013. Pollen analysis of coprolites
reveals dietary details of heavy-footed moa (Pachyornis
elephantopus) and coastal moa (Euryapteryx curtus)
from Central Otago. New Zealand Journal of Ecology
37: 151–155.
Wood JR, Wilmshurst JM 2014. Late Quaternary terrestrial
vertebrate coprolites from New Zealand. Quaternary
Science Reviews 98: 33–44.
Wood JR, Wilmshurst JM 2016. A protocol for subsampling
Late Quaternary coprolites for multi-proxy analysis.
Quaternary Science Reviews 138: 1–5.
21Wood et al.: The diets of moa
Wood JR, Wilmshurst JM 2017. Changes in New Zealand forest
plant communities following the prehistoric extinction
of avian megaherbivores. Journal of Vegetation Science
28: 160–171.
        
asynchronous herbivores on New Zealand montane
vegetation communities. PLoS ONE 14: e0214959.
Wood JR, Rawlence NJ, Rogers GM, Worthy TH, Austin JJ,
Cooper A 2008. Coprolite deposits reveal the diet and
ecology of New Zealand's extinct megaherbivore moas
(Aves: Dinornithiformes). Quaternary Science Reviews
27: 2593–2602.
Wood JR, Wilmshurst JM, Worthy TH, Cooper A 2011.
Sporormiella as a proxy for non-mammalian herbivores
in island ecosystems. Quaternary Science Reviews 30:
915–920.

NJ, Cooper A 2012a. High-resolution coproecology:
Using coprolites to reconstruct the habits and habitats of
New Zealand's extinct upland moa (Megalapteryx didinus).
PLoS ONE 7: e40025.
Wood JR, Wilmshurst JM, Worthy TH, Cooper A 2012b.
First coprolite evidence for the diet of little bush moa
(Anomalopteryx didiformis); an extinct forest ratite from
New Zealand. New Zealand Journal of Ecology 36:
164–170.
Wood JR, Wilmshurst JM, Holzapfel AS, Worthy TH, Cooper A

plant and the potential role of coprolites in conservation
paleobiology. Conservation Biology 26: 1091–1099.
Wood JR, Wilmshurst JM, Rawlence NJ, Bonner KI, Worthy
TH, Kinsella JM, Cooper A 2013a. A megafauna’s
microfauna: gastrointestinal parasites of New Zealand’s
extinct moa (Aves: Dinornithiformes). PLoS ONE 8:
e57315.
Wood JR, Wilmshurst JM, Richardson SJ, Rawlence NJ,
      
lost herbivore community structure using coprolites of
four sympatric moa species (Aves: Dinornithiformes).
Proceedings of the National Academy of Sciences USA
110: 16910–16915.
Wood JR, Wilmshurst JM, Newnham R, McGlone MS 2016.
Evolution and ecological change during the New Zealand
Quaternary. In: Shulmeister J ed. Landscape and
Quaternary environmental change in New Zealand. Paris,
Atlantis Press. Pp. 235–291.
Wood JR, Perry GLW, Wilmshurst JM 2017a. Using
palaeoecology to determine baseline ecological
requirements and interaction networks for de-extinction
candidate species. Functional Ecology 31: 1012–1020.
   
isotopes indicate a high trophic position for New Zealand’s
extinct South Island adzebill (Aptornis defossor)
(Gruiformes: Aptornithidae). New Zealand Journal of
Ecology 41: 240–244.
Worthy TH 1990. An analysis of the distribution and relative
abundance of moa species (Aves: Dinornithiformes).
New Zealand Journal of Zoology 17: 213–241.
Worthy TH 1993. A review of fossil bird bones from the loess
deposits in eastern South Island, New Zealand. Records
of the Canterbury Museum 10: 95–106.
Worthy TH, Holdaway RN 2002. The lost world of the moa.
Christchurch, Canterbury University Press. 716 p.

knowledge of the biology of moa (Aves: Dinornithiformes):
a new morphological analysis and moa diagnoses revised.
New Zealand Journal of Zoology 39: 87–153.
Worthy TH, Tennyson AJD, Archer M, Musser AM, Hand SJ,
Jones C, Douglas BJ, McNamara JA, Beck RMD 2006.
Miocene mammal reveals a Mesozoic ghost lineage on

National Academy of Sciences USA 103: 19419–19423.


Moa’s Ark. Copeia 2011: 72–76.
        
Early Miocene frogs (Anura: Leiopelmatidae) from
New Zealand. Journal of the Royal Society of New Zealand
43: 211–230.
Wotton DM, Drake DR, Powlesland RG, Ladley JJ 2016. The
role of lizards as seed dispersers in New Zealand. Journal
of the Royal Society of New Zealand 46: 40–65.
         
depend on declining frugivorous parrot for seed dispersal.
Biological Conservation 147: 133–142.

Upchurch P, Barrett PM 2012. Cranial biomechanics of
Diplodocus (Dinosauria, Sauropoda): Testing hypotheses
of feeding behaviour in an extinct megaherbivore.
Naturwissenschaften 99: 637–643.
Received 12 August 2019; accepted 30 August 2019
Editorial board member: Hannah Buckley
Supplementary material
Additional supporting information may be found in the

Appendix S1. Moa herbivory database.
Appendix S2.
localities (Dart River, Garibaldi Plateau and North Canterbury
limestone areas) but which were absent from moa gizzard
        
which their genera have not been recorded in moa diets from
elsewhere.
The New Zealand Journal of Ecology provides supporting
information supplied by the authors where this may assist
readers. Such materials are peer-reviewed and copy-edited
but any issues relating to this information (other than missing

... For example, if sporocarps were consumed prior to sporulation, spores would either be absent or potentially too undeveloped to survive digestion, preservation or laboratory procedures [100]. All moa species with gizzard or coprolite records consumed brightly coloured fleshy fruit, suggesting that they were visual foragers and would have been attracted to similarly coloured fungi [101]. Other ratites, such as emu (Dromaius novaehollandiae) [102] and southern cassowary (Casuarius casuarius) [103] consume agaricoid fungi, and upland moa, giant moa (Dinornis spp.) and little bush moa (Anomalopteryx didiformis) foraged on forest floors [86,88,90,101,104]. ...
... All moa species with gizzard or coprolite records consumed brightly coloured fleshy fruit, suggesting that they were visual foragers and would have been attracted to similarly coloured fungi [101]. Other ratites, such as emu (Dromaius novaehollandiae) [102] and southern cassowary (Casuarius casuarius) [103] consume agaricoid fungi, and upland moa, giant moa (Dinornis spp.) and little bush moa (Anomalopteryx didiformis) foraged on forest floors [86,88,90,101,104]. While our data demonstrate that at least one moa species consumed sequestrate fungi, it is likely that most other moa species did too, given their generalist diets and high degree of dietary overlap between species [86,88,101]. ...
... Other ratites, such as emu (Dromaius novaehollandiae) [102] and southern cassowary (Casuarius casuarius) [103] consume agaricoid fungi, and upland moa, giant moa (Dinornis spp.) and little bush moa (Anomalopteryx didiformis) foraged on forest floors [86,88,90,101,104]. While our data demonstrate that at least one moa species consumed sequestrate fungi, it is likely that most other moa species did too, given their generalist diets and high degree of dietary overlap between species [86,88,101]. Further analyses of moa coprolites will be required to reveal whether sequestrate fungi consumption was a common and widespread activity among moa. ...
Article
Full-text available
Mycovores (animals that consume fungi) are important for fungal spore dispersal, including ectomycorrhizal (ECM) fungi symbiotic with forest-forming trees. As such, fungi and their symbionts may be impacted by mycovore extinction. New Zealand (NZ) has a diversity of unusual, colourful, endemic sequestrate (truffle-like) fungi, most of which are ECM. As NZ lacks native land mammals (except bats), and sequestrate fungi are typically drab and mammal-dispersed, NZ’s sequestrate fungi are hypothesized to be adapted for bird dispersal. However, there is little direct evidence for this hypothesis, as 41% of NZ’s native land bird species became extinct since initial human settlement in the thirteenth century. Here, we report ancient DNA and spores from the inside of two coprolites of NZ’s extinct, endemic upland moa (Megalapteryx didinus) that reveal consumption and likely dispersal of ECM fungi, including at least one colourful sequestrate species. Contemporary data from NZ show that birds rarely consume fungi and that the introduced mammals preferentially consume exotic fungi. NZ’s endemic sequestrate fungi could therefore be dispersal limited compared with fungi that co-evolved with mammalian dispersers. NZ’s fungal communities may thus be undergoing a gradual species turnover following avian mycovore extinction and the establishment of mammalian mycovores, potentially affecting forest resilience and facilitating invasion by exotic tree taxa.
... In the past 15 years, Jamie Wood has inspired and led a renaissance of the study of moa coprolites with a focus on paleoecology rather than ichnotaxonomy (Wood et al., 2020 and references therein). There has been more limited study of other moa trace fossils, including bromalites and gastroliths (e.g., Trigg, 2001;Wood, 2007Wood, , 2008Carpenter et al., 2018;Carpenter, 2019;Wood et al., 2020). Overall, the study of moa trace fossils, with the notable exception of coprolites, has been proportionally low relative to the number of trace fossils available for study. ...
... They have an extensive Late Pleistocene and Holocene fossil/subfossil record in both the North and South Islands of New Zealand. Diverse taxonomic and DNA studies in the last 20 years have resulted in a consensus taxonomy of nine moa species in six genera within three families (e.g., Worthy, 1990;Huynen et al., 2003;Baker et al., 2005;Bunce et al. 2009;Worthy and Scofield, 2012;Wood et al., 2020: Edwards et al., 2024 Table 1). Anthropogenic pressures, including hunting, habitat destruction and introduction of new species, drove the moa to extinction by the first half of the 15 th Century (Holdaway and Jacomb, 2000;Allentoft et al., 2014;Holdaway et al., 2014;Perry et al., 2014). ...
... Geography, Taphonomy and Age Vertebrate trace fossils of moas are non-randomly distributed between the North and South Islands of New Zealand (Table 2). Notably, tracks mainly occur on the North Island, and coprolites and consumulites are restricted to the South Island (Wood et al., 2020). These distribution patterns reflect the geological and environmental histories of the two islands. ...
Article
Full-text available
Moa tracks were first discovered in New Zealand in 1866, but until the 1990s they received only intermittent study. Now 13 moa track localities are known on the North Island; five were found in the last 20 years. Moa tracks were first found on the South Island in 2019, and two sites are now known. Moa trace fossils are non-randomly distributed between the North and South Islands of New Zealand. Tracks mainly occur in the North Island, coprolites and consumulites are restricted to the South Island, and gastroliths, ara-moa (pathways) and nests are present on both islands. All moa tracks pertain to the ichnofamily Moapidae nov. We recognize four ichnonotaxa from the Late Pleistocene to Holocene of the North Island (Turanganuipus worthyi igen. et isp. nov., Moapus tennysoni igen. et isp. nov., Dinornipus oweni igen. et isp. nov., Gisbornepus angustus igen. et isp. nov.), one ichnotaxon from both the North and South Islands (Tutaenuipus woodi igen. et isp. nov.), and one ichnotaxon from Late Pliocene of the South Island (Aotearoapus lockleyorum igen. et isp. nov.) We also recognize two other morphotypes from the North Island, one of which is a probably a new ichnotaxon. There needs to be more ichnotaxonomic study of moa tracks and a focus on finding and collecting trackways with multiple footprints. All moa consumulites are gizzardalites (new term), and the majority are associated with gastroliths. Twenty-three gizzardalites can be directly associated with individual moa skeletons. Future study may discriminate other types of moa consumulites. Moa coprolites were first noted in 1875, and there has been a renaissance of study by Jaimie Wood and colleagues since 2007. Over 2000 Holocene moa coprolites are now known from more than 30 different localities across the South Island, and an ichnotaxonomy of moa coprolites is needed. Moa gastroliths occur on both the North and South Islands but also on D'Urville Island, Stewart Island, and Great Barrier Island; they are principally semi-round, often white quartz pebbles. There are at least 12 occurrences of moa nests, which may have been the size of emu nests. Nesting took place in late Spring-early Summer.
... Further, in NZ, mistletoes are known to be highly palatable to introduced brushtail possums (Trichosurus vulpecula) and red deer (Cervus elaphus) compared with other plants such as beech (Sweetapple, 2008;Crouchley et al., 2011). Further, palaeoecological data has shown that moa also regularly fed on mistletoe tissues (Wood et al., 2020(Wood et al., , 2021. It is therefore possible that kākāpō and other NZ species preferentially forage on Santalales taxa due to their nutritional content. ...
... This inference is concerning, as out of NZ' 12 endemic Santalales taxa, one is already extinct, and eight are classified as declining or threatened; an apparent result of prior forest clearance, browsing by invasive mammals, and a paucity of native dispersers or pollinators (Ecroyd, 1996;Ladley and Kelly, 1996;de Lange et al., 2018). For example, NZ's beech mistletoes (Peraxilla tetrapetala, P. colensoi or Alepis flavida) are supported by historic and palaeoecological evidence to have been considerably more abundant, and to have occurred closer to the ground, in the past (Wood et al., 2020). As kākāpō are thought to have been abundant and widespread in NZ's presettlement forests (Boast, 2021), our data suggests that they may have been a key pollinator or dispersers of these Santalales taxa prior to their near extinction. ...
Article
Full-text available
Threatened animal taxa are often absent from most of their original habitats, meaning their ecological niche cannot be fully captured by contemporary data alone. Although DNA metabarcoding of scats and coprolites (palaeofaeces) can identify the past and present species interactions of their depositors, the usefulness of coprolites in conservation biology is untested as few endangered taxa have known coprolite records. Here, we perform multilocus metabarcoding sequencing and palynological analysis of dietary plants of >100 coprolites (estimated to date from c. 400–1900 A.D.) and > 100 frozen scats (dating c. 1950 A.D. to present) of the critically endangered, flightless, herbivorous kākāpō (Strigops habroptilus), a species that disappeared from its natural range in Aotearoa-New Zealand (NZ) after the 13th C. A.D. We identify 24 orders, 56 families and 67 native plant genera unrecorded in modern kākāpō diets (increases of 69, 108 and 75% respectively). We found that southern beeches (Nothofagaceae), which are important canopy-forming trees and not an important kākāpō food today, dominated kākāpō diets in upland (c. >900 m elevation) habitats. We also found that kākāpō frequently consumed hemiparasitic mistletoes (Loranthaceae) and the holoparasitic wood rose (Dactylanthus taylorii), taxa which are nutrient rich, and now threatened by mammalian herbivory and a paucity of dispersers and pollinators. No single dataset or gene identified all taxa in our dataset, demonstrating the value of multiproxy or multigene datasets in studies of animal diets. Our results highlight how contemporary data may considerably underestimate the full dietary breadth of threatened species and demonstrate the potential value of coprolite analysis in conservation biology.
... D. novaezealandiae, fossils have also been found in subalpine herb fields and grasslands, in addition to forest and coastal environments 41,44 . Coprolites and gizzard content of E. curtus gravis (and D. robustus on the South Island) indicate a broad diet that included fruit, twigs and leaves from small herbs and trees 42 . Our modelling indicates that the range collapse and extinction of all species of moa resulted from low but sustained harvest rates, which nevertheless exceeded the limited reproductive capacities of these species. ...
Article
Full-text available
Human settlement of islands across the Pacific Ocean was followed by waves of faunal extinctions that occurred so rapidly that their dynamics are difficult to reconstruct in space and time. These extinctions included large, wingless birds called moa that were endemic to New Zealand. Here we reconstructed the range and extinction dynamics of six genetically distinct species of moa across New Zealand at a fine spatiotemporal resolution, using hundreds of thousands of process-explicit simulations of climate–human–moa interactions, which were validated against inferences of occurrence and range contraction from an extensive fossil record. These process-based simulations revealed important interspecific differences in the ecological and demographic attributes of moa and established how these differences influenced likely trajectories of geographic and demographic declines of moa following Polynesian colonization of New Zealand. We show that despite these interspecific differences in extinction dynamics, the spatial patterns of geographic range collapse of moa species were probably similar. It is most likely that the final populations of all moa species persisted in suboptimal habitats in cold, mountainous areas that were generally last and least impacted by people. We find that these refugia for the last populations of moa continue to serve as isolated sanctuaries for New Zealand’s remaining flightless birds, providing fresh insights for conserving endemic species in the face of current and future threats.
... Roughly 2000 kilometers east of the Thylacine's native Tasmania lies New Zealand, a land dominated by avifauna, the greatest among which were the Dinornithiformes or Moa: large flightless birds of the ratite clade (Worthy & holdaWay 2002). Moa lived in relative peace on New Zealand as megaherbivorous browsers (Wood et al. 2020), and were preyed upon only by Haast's eagle, Harpagornis moorei (BrathWaite 1992;Worthy & holdaWay 2002), until the arrival of the Austronesian peoples circa 1250-1300 CE (irWin & Walrond 2016). ...
Article
Full-text available
The Moa (Aves: Dinornithiformes) are an extinct group of the ratite clade from New Zealand. The overkill hypothesis asserts that the first New Zealand settlers hunted the Moa to extinction by 1450 CE, whereas the staggered survival hypothesis allows for Moa survival until after Europeans began to arrive on New Zealand. Alleged Moa sightings post-1450 CE may shed light on these competing hypotheses. A dataset of 97 alleged Moa sightings from circa 1675 CE to 1993 CE was constructed, with sightings given subjective quality ratings corresponding to various statistical probabilities. Cumulative probabilities of Moa persistence were calculated with a conservative survival model using these probabilistic sighting-records; a method recently applied to sightings of the Thylacine. Cumulative persistence probability fell sharply after 1408 CE, and across pessimistic and optimistic variations of the model, it was more likely than not that the Moa were extinct by 1770 CE. Probabilistic sighting-record models favour the overkill hypothesis, and give very low probabilities of Moa persistence around the time of European arrival. Eyewitness data on Moa sightings are amenable to scientific study, and these methods may be applied to similar animals.
... modern refers to species that have gone extinct since 1500 AD. Data for this map come from multiple sources(Rhodin et al. 2015;Wood 2020;Worthy et al. 2016;Hansford and Turvey 2018;Hume and Walters 2012) 2010). Stable isotope studies have confirmed that these turtles had a terrestrial, herbivorous diet (White et al. 2010), but their role in shaping the floras of these islands is almost entirely unconsidered. ...
Chapter
Island plants are predicted to have weak or absent defenses as part of the island plant syndrome. Evidence supporting the weak island defense prediction stems largely from observations of intense damage from invasive mammalian herbivores on islands. However, this evidence is misleading because most oceanic island plants have not evolved with native mammalian herbivores, and so should not have evolved defenses against them. In contrast, many islands have been home to other native vertebrate megafaunal herbivores, including flightless birds, tortoises, and turtles, many of which are now extinct or rare and therefore easy to overlook as agents of selection for island plant defenses. We review the evidence that island megaherbivores have selected for spinescence in island plants, supplementing published data with new estimates of spinescence for island floras varying in historical legacies of megafaunal herbivores. While the proportions of spinescent species are generally low, there are many spinescent island plants, likely functioning in defense against extant herbivores or persisting as defense anachronisms, no longer functioning due to the losses of native island megaherbivores. Future research exploring the evolvability of spinescence, including rates of losses or gains as herbivory selection pressure shifts, will be particularly enlightening for assessing island plant defenses in response to complex and variable historical legacies of megafaunal herbivory.
... The birds studied, several moa (Dinornithiformes, now extinct) species and kākāpō, are herbivores. Most are diet studies, examining the plant remains in coprolites (Kondo et al. 1994;Horrocks et al. 2004;Wood et al. 2020), but ancient DNA techniques have allowed detection of a number of parasites in moa coprolites (Wood et al. 2013;Boast et al. 2018Boast et al. , 2023. ...
Article
New Zealand’s kākāpō parrot, once widespread, is now critically endangered due to habitat loss and introduced mammalian predators. Prior to major population decline, a unique kākāpō cestode, Stringopotaenia psittacea, was found in the 1880s and first described in 1904. Here we report the discovery of eggs of this cestode in kākāpō coprolites of pre-human settlement age from the Honeycomb Hill cave system, north-west Nelson. Analysis of 52 samples, including coprolites of post-human settlement age, from nine sites within six South Island locations across a wide geographic range, yielded only eight infected samples in this single cave system. Results suggest that prior to human settlement, S. psittacea was not widespread within and between kākāpō populations, in marked contrast to other parasite types of the extinct moa spp. Intense management of the last remaining kākāpō has endangered or possibly caused the extinction of this cestode. This is the first confirmed record of S. psittacea since its discovery in 1884.
... It is possible that strong selection towards mimicry of the highly distasteful Pseudowintera colorata, which is commonly known as peppertree for its strong peppery flavour and which was only rarely consumed by moa (Wood et al. 2020), has resulted in convergence to a single morphology in Alseuosmia (A. pusilla) where the two are sympatric and where P. colorata is common. Further north, weaker selection towards models less well-defended than P. colorata may have resulted in mimetic polymorphism. ...
Article
Full-text available
Alseuosmia (Alseuosmiaceae) is an endemic New Zealand genus of small trees and shrubs, which is unusual in that some taxa appear to morphologically mimic unrelated species. The taxonomy of the group has long been debated with the extreme morphological diversity in A. banksii causing much of the confusion. Here we use ddRADseq to examine the genetic relationships between the species, with a particular focus on the morphological forms of A. banksii. Our analyses revealed that for species in the northern part of the distribution, genetic relationships largely matched geography rather than species’ boundaries based on morphology, and that hybridisation between morphs appears to be common. A diversity of morphologies is present within these northern Alseuosmia, including multiple forms that appear to mimic unrelated genera, and these may constitute a single gene pool. Further south, two species (A. turneri and A. pusilla) were genetically distinct in sympatry. We suggest maintaining the current taxonomy until further research can be undertaken.
... Coprolites, or ancient feces, are increasingly under investigation by researchers interested in records of past economy, environment, and evolution (Hunt et al. 2012;Qvarnström et al. 2016;Shillito et al. 2020). A variety of techniques are employed in coprolite analysis (e.g., Miller 1984;Poinar et al. 1998;Kühn et al. 2013;Linseele 2013;Camacho et al. 2018;Égüez and Makarewicz 2018;Sistiaga et al. 2014;Perrotti and van Asperen 2019;Zhang et al. 2019;Wood et al. 2020), and many studies apply multiple techniques to different coprolites in an assemblage (Reinhard and Bryant 1992;di Lernia 2001;Delhon et al. 2008;Shahack-Gross 2011;Marinova et al. 2013;Pineda et al. 2017;Baeten et al. 2018;Landau et al. 2020). Yet the full benefits of the multi-proxy approach will be realized when different complementary analyses are applied to each individual coprolite investigated, making the most of this finite archaeological resource (Fuks and Dunseth 2021). ...
Article
Full-text available
Archaeological dung pellets are time capsules of ancient herbivore diets and gut flora, informing on past agropastoral activity, ecology, and animal health. Improving multi-proxy approaches is key to maximizing this finite archaeological resource. Through experiments with standard pretreatments used in radiocarbon ( ¹⁴ C) dating, we address a fundamental problem in maximal multi-proxy analysis: How to chronometrically date individual caprine pellets while conserving as much as possible for additional analyses? We applied acid-alkali-acid (AAA) or acid-only pretreatments to 37 samples of ancient and recent sheep/goat dung pellets from sites in the Negev desert, Israel, measuring weight-loss due to pretreatment. Shavings of outer surfaces and remaining inner pellets of four pairs were dated and compared. We found that (i) sample-specific factors affect pretreatment survivability, including preservation quality and initial sample size; (ii) given sufficient start weight, AAA can be used to pretreat sheep/goat coprolites; (iii) 100 mg appeared a desirable minimum sample weight before pretreatment; and (iv) shavings of coprolites’ outer surface produced ¹⁴ C dates equivalent to dates obtained from inner coprolites. Whereas standard coprolite analysis protocols discard shavings removed from outer surfaces to avoid contamination, our findings indicate their efficacy for ¹⁴ C dating. This offers an important addition to workflows for multi-proxy coprolite analysis.
Preprint
Full-text available
Human settlement of islands across the Pacific Ocean was followed by waves of faunal extinctions that occurred so rapidly that their dynamics are difficult to reconstruct in space and time. These extinctions included large, wingless birds endemic to New Zealand called moa. We reconstructed the range and extinction dynamics of six genetically distinct species of moa across New Zealand at a fine spatiotemporal resolution, using hundreds of thousands of process-explicit simulations of climate-human-moa interactions, which were validated against inferences of occurrence and demographic change from an extensive fossil record. This statistical-simulation analysis revealed important interspecific differences in the ecological and demographic attributes of moa that influenced the timing and pace of their geographic and demographic declines following colonization of New Zealand by Polynesians. Despite these interspecific differences in extinction dynamics, the spatial patterns of geographic range collapse of moa species were similar. The final populations of all moa species persisted in suboptimal habitats in cold, mountainous areas that were generally last and least impacted by people. These isolated refugia for the last populations of moa continue to serve as sanctuaries for New Zealand’s remaining flightless birds, providing novel insights for conserving endemic species in the face of current and future threats.
Article
Full-text available
Large herbivores facilitate a range of important ecological processes yet globally have experienced high rates of decline and extinction over the past 50,000 years. To some extent this lost function may be replaced through the introduction of ecological surrogate taxa, either by active management or via historic introductions. However, comparing the ecological effects of herbivores that existed in the same location, but at different times, can be a challenging proposition. Here we provide an example from New Zealand that demonstrates an approach for making such comparisons. In New Zealand it has been suggested that post-19th Century mammal introductions (e.g. deer and hare) may have filled ecological niches left vacant after the 15th Century AD extinction of large avian herbivores (moa). We quantified pollen assemblages from fecal samples deposited by these two asynchronous herbivore communities to see whether they were comparable. The fecal samples were collected at the same location, and in a native-dominated vegetation community that has experience little anthropogenic disturbance and their contents reflect both the local habitat and diet preferences of the depositing herbivore. The results reveal that the current forest understory is relatively sparse and species depauperate compared to the prehistoric state, indicating that deer and moa had quite different impacts on the local vegetation community. The study provides an example of how combining coprolite and fecal analyses of prehistoric and modern herbivores may clarify the degree of ecological overlap between asynchronous herbivore communities and provide insights into the extent of ecological surrogacy provided by introduced taxa.
Article
Full-text available
Understanding the mutualistic services provided by species is critical when considering both the consequences of their loss or the benefits of their reintroduction. Like many other Pacific islands, New Zealand seed dispersal networks have been changed by both significant losses of large frugivorous birds and the introduction of invasive mammals. These changes are particularly concerning when important dispersers remain unidentified. We tested the impact of frugivore declines and invasive seed predators on seed dispersal for an endemic tree, hinau Elaeocarpus dentatus, by comparing seed dispersal and predation rates on the mainland of New Zealand with offshore sanctuary islands with higher bird and lower mammal numbers. We used cameras and seed traps to measure predation and dispersal from the ground and canopy, respectively. We found that canopy fruit handling rates (an index of dispersal quantity) were poor even on island sanctuaries (only 14% of seeds captured below parent trees on islands had passed through a bird), which suggests that hinau may be adapted for ground‐based dispersal by flightless birds. Ground‐based dispersal of hinau was low on the New Zealand mainland compared to sanctuary islands (4% of seeds dispersed on the mainland vs. 76% dispersed on islands), due to low frugivore numbers. A flightless endemic rail (Gallirallus australis) conducted the majority of ground‐based fruit removal on islands. Despite being threatened, this rail is controversial in restoration projects because of its predatory impacts on native fauna. Our study demonstrates the importance of testing which species perform important mutualistic services, rather than simply relying on logical assumptions.
Article
Full-text available
Often the mutualistic roles of extinct species are inferred based on plausible assumptions, but sometimes palaeoecological evidence can overturn such inferences. We present an example from New Zealand, where it has been widely assumed that some of the largest-seeded plants were dispersed by the giant extinct herbivorous moa (Dinornithiformes). The presence of large seeds in preserved moa gizzard contents supported this hypothesis, and five slow-germinating plant species (Elaeocarpus dentatus, E. hookerianus, Prumnopitys ferruginea, P. taxifolia, Vitex lucens) with thick seedcoats prompted speculation about whether these plants were adapted for moa dispersal. However, we demonstrate that all these assumptions are incorrect. While large seeds were present in 48% of moa gizzards analysed, analysis of 152 moa coprolites (subfossil faeces) revealed a very fine-grained consistency unparalleled in extant herbivores, with no intact seeds larger than 3.3 mm diameter. Secondly, prolonged experimental mechanical scarification of E. dentatus and P. ferruginea seeds did not reduce time to germination, providing no experimental support for the hypothesis that present-day slow germination results from the loss of scarification in moa guts. Paradoxically, although moa were New Zealand's largest native herbivores, the only seeds to survive moa gut passage intact were those of small-seeded herbs and shrubs.
Article
Full-text available
Over the past 50,000 y, biotic extinctions and declines have left a legacy of vacant niches and broken ecological interactions across global terrestrial ecosystems. Reconstructing the natural, unmodified ecosystems that preceded these events relies on high-resolution analyses of paleoecological deposits. Coprolites are a source of uniquely detailed information about trophic interactions and the behaviors, gut parasite communities, and microbiotas of prehistoric animal species. Such insights are critical for understanding the legacy effects of extinctions on ecosystems, and can help guide contemporary conservation and ecosystem restoration efforts. Here we use high-throughput sequencing (HTS) of ancient eukaryotic DNA from coprolites to reconstruct aspects of the biology and ecology of four species of extinct moa and the critically endangered kakapo parrot from New Zealand (NZ). Importantly, we provide evidence that moa and prehistoric kakapo consumed ectomycorrhizal fungi, suggesting these birds played a role in dispersing fungi that are key to NZ’s natural forest ecosystems. We also provide the first DNA-based evidence that moa frequently supplemented their broad diets with ferns and mosses. Finally, we also find parasite taxa that provide insight into moa behavior, and present data supporting the hypothesis of coextinction between moa and several parasite species. Our study demonstrates that HTS sequencing of coprolites provides a powerful tool for resolving key aspects of ancient ecosystems and may rapidly provide information not obtainable by conventional paleoecological techniques, such as fossil analyses.
Article
Full-text available
A new genus and species of fossil bat is described from New Zealand's only pre-Pleistocene Cenozoic terrestrial fauna, the early Miocene St Bathans Fauna of Central Otago, South Island. Bayesian total evidence phylogenetic analysis places this new Southern Hemisphere taxon among the burrowing bats (mystacinids) of New Zealand and Australia, although its lower dentition also resembles Africa's endemic sucker-footed bats (myzopodids). As the first new bat genus to be added to New Zealand's fauna in more than 150 years, it provides new insight into the original diversity of chiropterans in Australasia. It also underscores the significant decline in morphological diversity that has taken place in the highly distinctive, semi-terrestrial bat family Mystacinidae since the Miocene. This bat was relatively large, with an estimated body mass of ~40 g, and its dentition suggests it had an omnivorous diet. Its striking dental autapomorphies, including development of a large hypocone, signal a shift of diet compared with other mystacinids, and may provide evidence of an adaptive radiation in feeding strategy in this group of noctilionoid bats.
Article
Full-text available
The adzebills (Aptornithidae) were an ancient endemic lineage of large flightless Gruiformes that became extinct shortly after Polynesian settlement of New Zealand. The diet and ecology of these enigmatic birds has long been a matter for conjecture, but recent stable isotope analyses of bones of the North Island adzebill (Aptornis otidiformis) have indicated that adzebills may have been predatory. Here, we add to our understanding of adzebill ecology by providing the first stable isotope analyses of South Island adzebill (A. defossor) bones from two Holocene deposits. We interpret the results within frameworks of stable isotope measurements on bones of faunal species with known diets and from the same deposits (thereby mitigating regional effects on isotope values). Our results strongly support the hypothesis that adzebills had a high trophic position. Considered alongside the unique skeletal adaptations of adzebills and the distribution of the species during the Holocene, we suggest that adzebills were most likely terrestrial predators restricted to dry podocarp forests and may have specialised in dismantling rotting logs to obtain invertebrates and/or excavating burrowing animals such as tuatara or the chicks of burrow-nesting birds.
Article
Jaws and dentition closely resembling those of the extant tuatara (Sphenodon) are described from the Manuherikia Group (Early Miocene; 19-16 million years ago, Mya) of Central Otago, New Zealand. This material is significant in bridging a gap of nearly 70 million years in the rhynchocephalian fossil record between the Late Pleistocene of New Zealand and the Late Cretaceous of Argentina. It provides the first pre-Pleistocene record of Rhynchocephalia in New Zealand, a finding consistent with the view that the ancestors of Sphenodon have been on the landmass since it separated from the rest of Gondwana 82-60 Mya. However, if New Zealand was completely submerged near the Oligo-Miocene boundary (25-22 Mya), as recently suggested, an ancestral sphenodontine would need to have colonized the re-emergent landmass via ocean rafting from a currently unrecorded and now extinct Miocene population. Although an Early Miocene record does not preclude that possibility, it substantially reduces the temporal window of opportunity. Irrespective of pre-Miocene biogeographic history, this material also provides the first direct evidence that the ancestors of the tuatara, an animal often perceived as unsophisticated, survived in New Zealand despite substantial local climatic and environmental changes.
Article
A 4.9 Mkm2 region of the southwest Pacific Ocean is made up of continental crust. The region has elevated bathymetry relative to surrounding oceanic crust, diverse and silica-rich rocks, and relatively thick and low-velocity crustal structure. Its isolation from Australia and large area support its definition as a continent- Zealandia. Zealandia was formerly part of Gondwana. Today it is 94% submerged, mainly as a result of widespread Late Cretaceous crustal thinning preceding supercontinent breakup and consequent isostatic balance. The identification of Zealandia as a geological continent, rather than a collection of continental islands, fragments, and slices, more correctly represents the geology of this part of Earth. Zealandia provides a fresh context in which to investigate processes of continental rifting, thinning, and breakup.
Article
A key rationale for pursuing de‐extinction is the potential to restore lost processes and function to modern ecosystems. However, understanding and providing for the ecological requirements of the candidate species will also play a key role in determining the ultimate success of a de‐extinction. This assessment is challenging for prehistoric extinct species, where empirical studies or observations on their ecology are not available or are incomplete. A healthy, stable and self‐sustaining population of a resurrected species needs to be embedded in an ecological interaction network that consists not only of interactions between a resurrected species and its external environment (e.g. habitat, diet), but also those where the resurrected species is the environment (e.g. microbiota and parasites). Palaeoecology can provide information on all of these interactions for extinct species, and this information can help guide and inform the selection of suitable de‐extinction candidates or traits for resurrection. Ecological interaction network analyses offer a complementary tool to existing frameworks for determining the suitability of de‐extinction candidates, allowing palaeoecological information to be used to identify and quantify the potential implications of removing or adding resurrected species/functions to an ecosystem. Although palaeoecological data and understanding clearly have an important role in informing and modelling the potential ecological functions provided by extinct species, they can only ever provide an incomplete picture and therefore would only complement, rather than replace, observational or experimental data on the resurrected organisms themselves. A lay summary is available for this article.