I now want to consider some of the southern genera thought by some botanists to be related to and probably derived from north temperate genera. I propose an alternative hypothesis — that they may have evolved independently in the southern hemisphere.

With regard to the derivation of herbaceous genera in general it is worthwhile to consider what the vegetation pattern of the world may have been before this growth form suited to cooler, seasonal climates became widespread. The fossil evidence suggests that at these earlier times rain forest extended from the equator half way to the poles or even further and we can speculate that within such forests there would have been the woody ancestors of the herbaceous genera. Most of the latter would have originated in the northern hemisphere because of the greater area of land there and the more seasonal climates, but there seems no reason why other herbaceous genera could not have been derived in the southern hemisphere also from woody stock at the fringes of the great rain forest belt.

New Zealand Umbelliferae

In an earlier review¹⁵⁰ I proposed two evolutionary lines within the more notable New Zealand genera of the main subfamily of the Umbelliferae (carrot family). One included Aciphylla (spaniards) and Anisotome and is distinguished by the inflorescences arising at the centre of leaf rosettes and terminating their growth and by their dioecism (separate male and female plants). The other group, comprising Scandia, page 200Gingidia and Lignocarpa, has mostly lateral inflorescences and is mostly gynodioecious (separate female and hermaphrodite plants). In the latter group I considered Scandia to be the most primitive genus as it is semi-woody with extended stems. Scandia rosaefolia is the more shrubby of the two species of the genus and it is restricted to the northern half of the North Island and there often to mild coastal habitats. It is not difficult to imagine this species, and perhaps even woodier, now extinct, relatives, existing in New Zealand in warm pre-Ice Age times. From Scandia the completely herbaceous montane to alpine Gingidia could have evolved and also the specialised Lignocarpa of shingle slips. Since this idea was proposed a probable hybrid between Aciphylla squarrosa and Gingidia montana has been discovered in Marlborough¹⁵¹ which suggests that the two lines are more closely related than I thought and that all the genera concerned may have been derived from woody ancestors with a long history in New Zealand. In eastern Australia there are two species assignable to Aciphylla, one to Anisotome and three to Gingidia, including a localised occurrence of G. montana. In view of the foregoing discussion it seems more likely to me that the Australian species have a New Zealand ancestry than vice versa.

In a recent review Webb¹⁵² gives a largely opposite interpretation. He considers the few Australian species of these genera to be primitive, while New Zealand has a range from primitive to specialised forms. In contrast to my view, he regards the woodiness of Scandia as a recent specialisation. With one exception, he suggests that primitive species of these genera migrated from Australia to New Zealand with consequent diversification and specialisation here. The exception is Gingidia montana, where the localised occurrence in northern New South Wales he considers to be a case of recent migration from New Zealand to Australia.

Of course, at an earlier time before high mountains existed in New Zealand, there could have been only primitive species of these genera in both Australia and New Zealand and migration could have been in either direction. With the formation of the high mountains in New Zealand more specialised forms would then have evolved here.

According to yet another view some would argue that long distance migrations have not been involved at all. The occurrence of the same or related alpine plants in both Australia and New Zealand would derive from the time when they shared the same region before it was sundered page 201by the drifting away of the New Zealand crustal complex from Australia.

With little direct evidence from the past we can only speculate.

Celmisia

As well as the 60 or so species of Celmisia in New Zealand there are two related genera in the subantarctic islands: the semi-woody Damnamenia and the large, tufted, herbaceous Pleurophyllum.¹⁵³ Pleurophyllum is considered to be intermediate in some respects between Celmisia and one group of the woody genus Olearia, also strongly represented in New Zealand. This fact, plus possible wild hybrids between Olearia and Celmisia, suggests a derivation, perhaps within New Zealand, of Celmisia from Olearia. Some of the New Zealand celmisias are described as subshrubs, but the five species in east Australia are herbaceous and in Wardle's words 'seem to represent end points of evolutionary pathways stemming from New Zealand'.

Hebe and Related Genera

New Zealand has perhaps 100 species of Hebe and there are only three species elsewhere. Two of these in southern South America are shared with New Zealand (H. elliptica, H. salicifolia) and the third is Hebe rapensis on Rapa Island in French Polynesia. The related but generally less woody Parahebe is also represented in eastern Australia and New Guinea and Chionohebe, probably derived from Parahebe, is a genus of small, high alpine cushion plants of which two of the New Zealand species extend to south-eastern Australia. Veronica, from which Raven¹⁴⁸ would consider the Hebe alliance to have been derived, is a mostly north temperate, herbaceous to slightly woody genus. It seems reduced and specialised and so is unlikely to have been ancestral to the woody hebes, some of which may become small trees with trunks sometimes 30 cm in diameter. There are differences too between the basic chromosome numbers of Veronica and the Hebe alliance which do not suggest a close relationship. There are indeed a few true veronicas in eastern Australia but, as Wardle¹⁴⁹ comments, 'it seems … possible that the handful of Australian true veronicas derive from recent northern immigrants that are not directly related to the Hebe alliance'.

Presumably Veronica was originally derived from woody, northern hemisphere ancestors, which may have been related to and perhaps also ancestral to the Hebe alliance.

page 202

Once again it is clear that much more research is required into a number of genera before we can attain any degree of certainty about the relative proportions of New Zealand alpines that have been derived from:

(a)	geologically recent immigrants from the northern hemisphere;
(b)	more ancient immigrants from the northern hemisphere; or
(c)	genera of southern hemisphere origins.