When I was looking for the original paper for my post on moa feather colour & reductionism, I found a whole lot of other equally interesting stuff. As one does. (It's just so easy to wander off down some interesting side path & get completely distracted from the original task...) One of those 'other' papers was on fossil moa poo - coprolites - & what they have to tell us about moa feeding ecology (Wood et al., 2008).
Moa have generally been regarded as browsers on trees & shrubs, occupying ecological niches that on other land masses were filled by mammals. These assumptions about diet have been based partly on the fact that moa were big, & partly on analysis of gizzard contents. They've led to various hypotheses about the selection pressure that moa would have placed on the plant species that they ate. These included the idea that the divaricating growth habit of many NZ plants is an adaptation which reduced exposure to being snacked on by these big birds. (Otago Museum has a wonderful display of moa skeletons; I spent a happy hour or so viewing them when I was down in Dunedin at Biolive.) But how can we tell what moa were actually eating? (For example, the gizzard contents come from birds that died & were preserved in swamps, so they're going to be biased towards lowland/wetland habitats.)
That's where the coprolites come in. It's not actually all that unusual to study fossil poo - such work has enhanced our understanding of what early H.sapiens was eating, for example. Wood & his colleagues analysed 116 moa coprolites to find out whether they were focusing on trees & shrubs, or whether their diet was rather wider than that. This wasn't just out of idle curiosity. We don't really know the full impact of moa herbivory on the evolution of plant growth forms & reproductive habits. Nor can we be sure of the ecological impact of their extinction. Attempting to answer both questions could give us useful information about our modern ecosystems & help in reconstructing their pre-human state.
First, find your coprolites. The team excavated a range of sites in Otago, covering two main ecological zones, & came up with nearly 2000 pieces of fossil poo. Then they used a combination of old & new techniques - examining plant macrofossils, and studying aDNA in the ancient faeces - to work out what the poo producers had been eating. The ancient DNA data let the researchers identify which species of moa had generated the coprolites, while seeds & leaf cuticles) in the faeces indicated dietary preferences. The seeds in particular allowed identification of plants to species level in many cases.
They were able to identify 24 coprolites as coming from 4 species of moa: the South Island giant moa Dinornis robustus, upland moa Megalapteryx didimus, heavy-footed moa Pachyornis elephantopus, and stout-legged moa Euryapteryx gravis, & bones of 3 of these species were found in the same study sites as the coprolites themselves. (92 other fossil faeces were attributed to moa becuase of their size & shape.)
It turns out that the big birds were eating a much wider range of plants than previously thought, including trees & tall shrubs, lianes such as Muehlenbeckia, subshrubs & dicot herbs, & monocot herbs, from around 30 different taxa. Many of these species were less than 1m high, which is somewhat unexpected given that many moa species where much taller (& hence the earlier assumptions about them feeding on trees & shrubs). Perhaps surprisingly, different moa species within the same habitat type appeared to eat a very similar range of plant species (Wood et al., 2008), nor was there much difference in the types of plants eaten by sympatric species. You could argue that seeds aren't really a good proxy for diet as the birds could well have been eating other parts of some plants, not as easily identified. But the team noted that the gizzard studies showed that the majority of plant species identified were represented by seeds. Seeds don't,however, indicate whether the different species were feeding preferentially on different parts of the plants.
What I found really interesting (apart from the fact that we now have the technical ability to tease all this information out of a pile of petrified poo!) was the differences in feeding ecology between the moa species the researchers identified, and modern mammalian browsers & grazers. While the moa were grazing & browsing, they were focusing on different species from those eaten by introduced mammals - eating a high proportion of plant taxa that are avoided by ungulates, including several that are toxic to mammals. This suggests that the ecological impact of moa & mammals would be quite different. On the other hand, in some more open habitats moa diets seem to have resembled those of living ratites (emu, ostrich, & rhea). On the surface this could imply that if we wanted to 'restore' ancient NZ ecosystems, we could use ostrich or emu as moa surrogates. Unfortunately, it seems that some woodland moa were much better than the living ratites at eating woodier plant parts, so perhaps the living ratites wouldn't do the trick in at leat some habitat types.
But barring the advent of some sort of Jurassic Park technology, maybe ostriches trampling through our forests is the closest we're likely to get to pre-human New Zealand :-)
J.R. Wood, N.J. Rawlence, G.M. Rogers, J.J. Austin, T.H. Worthy & A. Cooper (2008) Coprolite deposits reveal the diet and ecology of the extinct New Zealand megaherbivore moa (Aves, Dinornithiformes). Quaternary Science Reviews 27: 2593-2602