Soil-Vegetation Relationships in Kipukas — Mueller-Dombois and Lamoureux 
291 
up much of this ash and redeposited it as a dune 
at a place where the smooth lava was inter- 
cepted by a rough aa flow. The ash-dune now 
represents an island supporting pioneer vegeta- 
tion. This process is accumulative since the vege- 
tation, once established, catches more eolian de- 
posits and in turn contributes organic matter, 
soon forming a moisture- and nutrient-improved 
habitat that also differs in elevation from its 
surroundings. It is quite conceivable that such 
elevated dune ecosystems can be surrounded by 
subsequent lava flows. Such an occurrence, on a 
much larger scale, could account for the origin 
of the kipukas discussed here, although addi- 
tional evidence to support this hypothesis must 
be obtained. Also, many, if not most, of the 
Hawaiian kipukas, such as Kipuka Nene, un- 
doubtedly have developed merely by the dis- 
integration in situ of older lavas, the kipuka 
area being subsequently surrounded by newer 
flows. 
The upper ash deposits in the kipuka soils 
are more or less horizontal with respect to the 
present soil surface, a form of deposition which 
Powers (1948) calls "blanket deposits." The 
"blanket deposits" in the savannah soil show 
that there have been at least 9 ash deposits in 
Kipuka Puaulu since establishment of the thick 
black horizon (Alb). Not all of these may have 
been derived from different explosions, but 
Powers has discovered ash from at least 26 
eruptions in the area that occurred later than 
the big Kilauea ash explosion of 1790. The 
latest recorded near Kipuka Puaulu was from 
the 1924 eruption. This shows that the soil is 
not of one (old) age, but is of several ages 
from older to younger, and the surface soil may 
even be much younger than the surrounding 
rough aa flow, rather than older as indicated by 
Rock. The surrounding flow is prehistoric, thus 
at least pre-1778. 
Fragments of charcoal were found in both 
kipukas in the forest soils. They occurred at 70 
cm depth in Kipuka Ki and at 80 cm in Kipuka 
Puaulu. This indicates two facts. First, there 
was fire in both kipukas at an earlier date in 
their development; and second, both had woody 
vegetation growing on them at that time. Al- 
though charcoal was not found in the savannah 
soil, fire may explain its origin. It is interesting 
that the savannah soil looks quite different from 
the forest soils which, in spite of being in two 
separate kipukas, show much similarity in ap- 
pearance. Both forest soils are deeply melanized, 
dark brown in color, and are rather uniformly 
enriched with organic carbon (Table 1). The 
savannah soil shows more clearly the parent 
material, because of less uniform melanization. 
Here organic carbon content fluctuates greatly 
between soil horizons. These two patterns, that 
is, the more uniform color and organic carbon 
distribution in the forest soils and the greater 
variation in color and organic carbon distribu- 
tion in the savannah soil, are undoubtedly asso- 
ciated with past rooting zones. One may assume 
that a mixed, well-stocked forest occupies the 
soil volume more uniformly than does a domi- 
nantly grass-covered savannah. The grass and 
ground vegetation roots may have been more 
restricted to the black horizon zones. Such a 
concentration in rooting depth was also found 
at present at the soil surface of the savannah 
soil. This pattern supports the assumption that 
the savannah originated after a fire. It is prob- 
able that the fire occurred when the 20 cm-thick, 
black layer, the buried A1 horizon (Alb), was 
at the surface supporting actively growing veg- 
etation, because the charcoal was found right at 
the lower boundary of this layer in both forest 
soils (Fig. 2). Therefore, the savannah may be 
quite old. The C-14 date of the 70 cm -deep 
charcoal in Kipuka Ki came to 2,170 dz 200 
years, i.e. 9 about 220 years b.c . 3 
Analyses of potentially available water, or- 
ganic carbon, and pH show no significant dif- 
ferences between savannah soil and forest 
soil in Kipuka Puaulu, so that neither soil water 
nor nutrient differences can be assumed to be 
responsible for the difference in vegetation. 
Moreover, there is no distinctive topographic 
pattern associated with either type, so that the 
savannah’s origin is not attributable to environ- 
mental differences related to topography. 
B. Flora 
Between November 1963 and March 1965 
botanical surveys were made of both kipukas. 
Voucher specimens have been deposited in the 
herbarium of the Department of Botany, Uni- 
versity of Hawaii, and duplicate specimens in 
3 Sample GX0394, Geochron Laboratories, Inc. 
