Soil-Vegetation Relationships in Kipukas — Mueller-Dombois and Lamoureux 
293 
However, such minimal area calculations are 
based on the more common species. From the 
records it is quite clear that the now extinct 
species were extremely rare. The smaller size of 
Kipuka Ki can be used, therefore, as one ex- 
planation for its smaller number of indigenous 
species. 
(b) greater age. Rock (1913) believed 
that Kipuka Ki was more recent in origin than 
Puaulu, because of the common assumption that 
an older area would have more species. His 
idea cannot be disproved from current evidence, 
but one observation points in the opposite direc- 
tion. The amount of organic carbon did not 
decrease in the lower profile of Kipuka Ki, 
whereas it did so in both soils of Kipuka Puaulu 
(Table 1). This may indicate vegetative activ- 
ity at an earlier date in Kipuka Ki as compared 
with Puaulu. 
(c) GREATER DIVERSIFICATION IN HABITATS. 
This factor in Kipuka Puaulu is not expected 
from observations made so far. Both kipukas 
have similar topographic variations and deep, 
rich soils. Also, the distribution of tree species 
is not as likely to be affected by small-scale 
environmental variations as is that of herba- 
ceous plants. 
(d) DIFFERENT HISTORY OF DISTURBANCE. 
Little definite information is available on 
differences in disturbance-history. We know 
only that three important disturbance factors 
operated in both kipukas: fire, cattle grazing, 
and pig damage. Current pig damage appears 
to be less in Kipuka Ki. Past cattle grazing also 
was probably less devastating here. It is pos- 
sible, however, that fire eliminated a few trees, 
either directly, or indirectly by competition of 
more aggressive plants that followed the fire in 
both kipukas. In this connection, the chance of 
the smaller, isolated kipuka to be restocked with 
rare species would be less than that of the 
larger one, which also may have provided a 
greater chance of survival of rare tree species 
simply because of its larger size. 
(e) DIFFERENCES IN RAINFALL AND PRO- 
DUCTIVITY. It was interesting to find that the 
current moisture distribution downward in the 
soils differed between the kipukas. The current 
soil water content increased considerably in the 
bottom part of the profile in Kipuka Ki and 
was higher than in the soils of Kipuka Puaulu, 
whereas the upper part of the profile was drier 
than that of the soils in Kipuka Puaulu. This 
indicates a different rain shower pattern be- 
tween the kipukas. This may be a random pat- 
tern, however, which then would have no bear- 
ing on the total amount received. Except for 
the lower profile parts (B3 and C2), there was 
little difference in the amount of organic carbon 
in the two forest soils, indicating a similar pro- 
ductivity in both kipukas. Thus, the differences 
in number of species cannot be related to dif- 
ferences in productivity. 
2. The larger number of introduced weed 
species in Kipuka Puaulu may be caused by (a) 
its greater exposure to man and cattle, and (b) 
its larger sun-exposed area, which favors the 
establishment of shade-intolerant weeds. It is 
interesting that the fewer weed species in Ki- 
puka Ki occupy more ground. Some of them 
have formed dominant communities. 
C. V egetation 
Several obvious plant communities occur un- 
der the forest cover. They are represented by 
native and introduced plants as follows: 
NATIVE 
Microlepia association 
Nephrolepis association 
Peperomia patches 
Pipturus shrub strata 
Coprosma thickets 
INTRODUCED 
Commel'ma association 
Rubus penetrans association 
Solanum association 
Dactylis patches 
Com m,eUna-N ep h rolepis mixed community 
Nephrolepis communities and Dactylis 
patches are common also in open areas. Cop- 
rosma thickets are characteristic only for Kipuka 
Puaulu. The Rubus penetrans and Solanum as- 
sociations are characteristic for Kipuka Ki. Only 
one small Solanum patch was observed in Ki- 
puka Puaulu. All other associations occur in both 
kipukas. Peperomia patches seem to be estab- 
lished on ground that has been rather recently 
scarified by pigs, and form there a pioneer com- 
munity in shaded habitats. Similarly, Coprosma 
thickets are associated with pig scarification, 
