If the profusion of vines and epiphytes in the conifer broadleaf forests is the first thing that strikes the overseas botanist, the second is probably the prevalence, in various open habitats and some forest types, of densely twiggy shrubs with very small leaves and slender, wiry stems. Cockayne called them 'divaricating shrubs' because of the wide or divaricate branching angles (to 90° or more) of the twigs.¹⁰⁰ Another feature of many divaricates is the vigorous growth of lateral twigs from all, or almost all, of the leaf angle buds of each branch, even those nearest its tip. In 'normal' shrubs the growing tip of a branch grows strongly and forms a chemical inhibitor which prevents the lateral buds for some distance below from themselves growing into branches. With many divaricate shrubs, growth of a new branch, although vigorous at first, eventually becomes weak and as a result the no-longer-inhibited lateral buds grow out strongly until their activity too declines. It could almost be said that the branching of these shrubs is uncontrolled in comparison with more familiar growth patterns. The result of this distinctive branching behaviour is a densely interlaced, springy, more or less balllike shrub. Where these grow in windswept situations on hill tops or at the coast they form low hummocks which are so dense that a person can stand on them (Fig. 74), or even jump up and down on them, without breaking through the surface. At first acquaintance all divaricate shrubs tend to look the same, but, according to Greenwood and Atkinson,⁶² there are in New Zealand 54 species of this form (including the divaricating juveniles of some trees) belonging to 20 genera and 17 families (Figs. 75, 76, 77). A further 9 species are illustrated in Eagle's Trees and Shrubs of New Zealand, but have not yet been named formally. Nine species from an additional 6 genera (including Nothojagus, Neomyrtus, Lophomyrtus and Teucridium) and 4 families are termed semi-divaricating. Some of these have small leaves and slender twigs but their branching angles are relatively narrow.

page 134

Some divaricate shrubs are the persistent juvenile stages of forest trees, which undergo a dramatic change in branching pattern and an equally dramatic increase in leaf size when they change to the adult state (Fig. 78).

Two species have divaricate juveniles with leaves which are remarkably variable from long and narrow to short and round and from deeply lobed and dissected to entire — pokaka (Elaeocarpus hookerianus) (Fig. 78) and Pittosporum turneri. The non-divaricating vine Parsonsia heterophylla has similar juvenile leaves. Divaricating shrubs may themselves have juvenile forms. Melicope simplex and Pseudopanax anomalus have simple adult leaves but compound juvenile leaves. Several divaricate Pittosporums have strongly dissected or narrowly elongate juvenile leaves. Perhaps most remarkable of all, Aristotelia fruticosa has a juvenile which is not divaricate, but has a similar range of leaf forms to that of pokaka. Cockayne¹⁰¹ says: 'The various forms assumed by the (juvenile) leaves of A. fruticosa are almost beyond belief.'

There are differences of opinion about whether divaricating shrubs which are exactly equivalent to those of New Zealand occur elsewhere in the world,⁶² but it seems that only in New Zealand are they so prevalent in both the flora and the vegetation.¹⁰²

The majority of New Zealand divaricates belong to genera in which there are also local forest species of normal branching patterns, usually with much larger leaves. In the following list the numbers of divaricate and non-divaricate species are given for each genus.

page 136 page 137 page 138 page 139

Greenwood and Atkinson⁶² consider that 26, or nearly half, of the divaricate species they considered are largely restricted to one vegetation type: forest (10); forest margins (3); scrubland¹⁰⁴ (5); Shrubland¹⁰⁴ (6); grassland and other open vegetation (2). The other 28 species can be found in two or more vegetation types. Some of these are remarkable in their range. For example Coprosma rhamnoides at one extreme grows in shade and shelter on open forest floors and there forms attractive shrubs with leaves which tend to be arranged in horizontal layers. At the other extreme it may be found in the open on hill tops exposed to violent gales, where it forms or contributes to dense shrub hummocks shaped and channelled by the wind (Fig. 74). The fivefinger relative Pseudopanax anomalus is a similar case.

Within the genera concerned, the 'normal' species differ so markedly in appearance from the divaricates that it is difficult to believe that they are related. Apart from the differences in branching pattern the leaves of the normal species are often hundreds of times the area of those of the divaricates. Even more surprising is the fact that natural hybrids are not uncommon between divaricate and non-divaricate species of several genera — Melicytus, Aristotelia, Sophora, Melicope, Tseudopanax, Corokia, Pennantia (Fig. 79), Myrsine, Coprosma¹⁰⁵ (Fig. 80) (3 pairs of species) and Olearia. In some cases the first generation hybrids are partly fertile and give rise to bewildering arrays of second generation hybrids known as 'hybrid swarms'.

It is generally believed that the divaricates are derived and specialised forms and the fact that a number of them are able to hybridise with large-leaved relatives might indicate that this distinctive growth habit evolved in relatively recent geological times.

We now come to the problem of how and why divaricate plants became such a conspicuous feature of the New Zealand flora. If there were only one species of this form it might be dismissed as a chance aberration, but as there are many, and these are mostly unrelated to one another, some special environmental factor or combination of factors selecting for the habit is indicated. A contrary view is that of Went,¹⁰⁶ who suggests a virus-like transmission of chromosome segments, bearing divarication inducing factors, from species to species, genus to genus and family to family. Because of the difficulty of envisaging a mechanism for such transfers this idea has not received much support, although it does not seem so outlandish in these days of genetic engineering as it would have done some years ago.

page 140 page 141