Forest Vines to Snow Tussocks: The Story of New Zealand Plants
We come now to the 'moa113 theory' elaborated by Greenwood and Atkinson,62 which has found favour with some but disbelief from others. Some of the negative reactions, I think, stem from the fact that the moas are extinct. The 'moa theory' has a science fiction ring to it, just as would a 'dodo theory' or, to extend into myth, a 'dragon theory'. At our present state of knowledge, however, the moa theory merits serious consideration along with the climatic theories.
Essentially, Greenwood and Atkinson propose that the divaricating shrub habit is probably unique to New Zealand and results from an environmental factor that was also unique to New Zealand, namely the moas. Species of similar appearance to divaricating shrubs elsewhere in the world either have relatively large leaves and/or are spiny. Spininess of shrubs is considered to be a defence against the browsing of softnosed mammals, but it would be ineffective against large, browsing birds with their hard beaks. With the exception of matagouri (Discaria toumatou) New Zealand divaricates are not spiny, although there is a tendency that way with some reduced shrubs in exposed habitats. Ma-page 145tagouri has spiny relatives in Australia and South America so probably evolved its spininess outside New Zealand.
Greenwood and Atkinson considered the essential features of divaricating shrubs to be: interlaced branching; small leaf size; stem toughness; and leaves which are smaller on outer than on inner branches. In their original publication they assumed that moas could not bite with a cutting action, but clamped, pulled and broke off portions of foliage This would have made it difficult for them to browse on divaricates as the twigs are wiry and difficult to pull off and, when broken, difficult to disentangle. It was further suggested that where the outermost twigs had very reduced leaves and a dead appearance they would be rejected as unpalatable. With regard to trees with divaricating juveniles, they suggest that the change to adult branching and foliage took place above the reach of moas.
Later work showed that the assumption about moa eating habits was incorrect. Burrows114 examined the plant material contained in fossilised moa gizzards and found that it mostly 'consisted of twigs, commonly 1.5–6.0 mm thick and 10–30 mm long. Some of these twigs are of Olearia virgata and Plagianthus regius (both divaricating species) and much material appears to have come from Coprosma, a genus which contains a large number of divaricating species. Most of the twigs were cleanly cut, probably by the shearing action of the moa beak, rather than broken'.
Greenwood and Atkinson115 and Lowry116 suggested that although the divaricate habit may not have prevented moa browsing, the many growing points separated by the wide branching angles would have enabled plants of this form to survive and quickly recover from browsing.
Greenwood and Atkinson note that divaricates are most abundant on fertile soils such as those of river flats and suggest that this is because plants on such sites would be more nutritious and therefore favoured by moas. On the other hand they point out that there are few divaricates on outlying islands not reached by moas and none on cliffs or as epiphytes on branches: 'moas would have experienced difficulty in browsing either on cliffs or up trees!'
In opposition to this view, McGlone and Webb110 suggest that the prevalence of divaricates on river flats and similar places derives from the fact that such sites are particularly frosty. They also ask why divar-page 146icates, if they were maintained on river flats by browsing, have not been replaced in the centuries since moa extinction by faster growing, larger-leaved shrubs and trees.
To throw yet another hat into the ring, river flats are also subjected to floods from time to time, so perhaps the divaricate habit is flood-resistant and in trees with divaricate juveniles the change to the adult state takes place above flood level? It is perhaps relevant here that McGlone and Webb believe that river beds, dunes and wind-induced scrub were probably the only places in the lowlands during the warmest phases of the interglacials, which could serve as refugia for divaricates. The 'flood' hypothesis undoubtedly raises a number of questions, such as why there are no divaricates within the flood zone of river cliffs, but it is not alone in this respect. Perhaps divaricate plants are tolerant of a number of environmental factors rather than just one.
The one thing that is clear in all of this is that the divaricating shrub story has not yet been fully told and that we need to know much more about the morphological, physiological and ecological details of the plants concerned.
Finally, not all small leaved shrubs in New Zealand are of divaricate form. The large genus Hebe has many scale-leaved 'whipcord' species, mostly found at higher altitudes, which are probably derived forms, as are the many needle-leaved species of Dracophyllum and some small-leaved but not divaricating species of Olearia, Coprosma and other genera. We need to keep these in mind when theorising about shrubs which are divaricating as well as small-leaved.