The annual range of temperatures in the southern Andes is similar to that i
n New Zealand, although they are cooler at comparable latitudes. South of 4
5 degreesS in Chile the precipitation regime is also similar to New Zealand
, but northwards and also east of the Andes there is an increasing period o
f summer dryness. There are corresponding similarities in vegetation, espec
ially in evergreen Nothofagus forests, tussock grassland in eastern rain sh
adows, and cushion bogs. Deciduous Nothofagus species are absent from New Z
ealand, but dominate widely in drier and colder environments of the souther
n Andes.
In the southern Andes deciduous Nothofagus forms the subalpine forest and k
rummholz although, in the northern part of the region, there is also Arauca
ria araucana growing as erect, emergent trees. Volcanic eruptions may lower
the tree limit so that it intersects with evergreen Nothofagus forest.
In New Zealand, evergreen Nothofagus species. which are less cold-resistant
than the deciduous Andean species, form most tree limits, and a belt domin
ated by tall Chionochloa tussocks intervenes between forest and low-growing
alpine vegetation. The absence of very hardy tree species in the New Zeala
nd mountains may be related to isolation and late uplift.
Extrapolation to tree limit indicates southern Andean warmest-month mean te
mperatures from around 7 degreesC in northern localities to only 5.7 degree
sC at 55 degreesS, compared with 10 degreesC which prevails in New Zealand.
Since tree limit altitudes in the Southern Andes exceed those of New Zeala
nd by at least 250 m, this indicates that in temperature terms New Zealand
tree limits are 550 m lower than those of the southern Andes, corresponding
with summer isotherms 3.3 degreesC warmer. Vascular species were listed fr
om equivalent vegetation types in the southern Andes and New Zealand, inclu
ding forest, rain-shadow grassland, alpine, coastal, swamp, lake-edge, and
bog vegetation, and also from communities near the upper tree limit without
close equivalents, notably deciduous forest in the Andes and Chionochloa g
rassland in New Zealand. Phytogeographic categories were assigned to all sp
ecies, genera, and families. These were grouped as shared-austral, shared w
ith wider distributions, "realm-endemic" confined to either South America o
r Australasia, and non-shared with wider distributions.
About 90% of the 465 species listed from the southern Andes and 522 listed
from New Zealand are realm-endemics. Forty species or closely related pairs
of species are shared, nearly half of these being coastal. Among genera, r
ealm-endemics are the largest element overall and exceed 30% in forests of
both regions and in New Zealand subalpine, grassland, alpine, successional,
and bog vegetation. Shared genera that extend to north temperate regions f
orm the next most important, reaching over 40% in southern Andean swamp and
New Zealand coastal samples. Shared genera with austral distributions reac
h 21-32% in all New Zealand vegetation classes and southern Andean bog, com
pared with 5-20% in other southern Andean vegetation classes. They contribu
te disproportionately to the physiognomic similarities between the two regi
ons, notably in Nothofagus forests and bogs. Non-shared genera with austral
or subtropical distributions are most numerous in forest samples from each
region.
Non-shared genera with north temperate distributions constitute only 2% in
New Zealand, contrasting with the southern Andes where they constitute 19%
overall and 20-29% in subalpine, grassland, alpine, successional, and coast
al vegetation. Distribution of families among phytogeographic categories is
similar to that of genera, except that more are shared-widespread and fewe
r are shared-austral and realm-endemic.
Sequencing of the gene rbcL for pairs of taxa in shared-austral genera and
for populations of Hebe salicifolia yields divergence estimates ranging fro
m 130 +/- 75 to 7 +/- 4 million years. Divergences among tree species in sh
ared genera may predate the separation of Australasia from Antarctica and S
outh America, but not the separation of New Zealand from Australia. Taxa of
open habitats show the most recent divergences, and their presence in both
regions demands some mode of transoceanic dispersal. Possibly as late as t
he Pliocene, this may have involved a partly vegetated Antarctica.