294 
PACIFIC SCIENCE, VoL XXI, April 1967 
which is particularly pronounced under larger 
Sapindus trees where the pigs seem to search 
for their fruits. 
SUMMARY AND CONCLUSIONS 
1. The soil of both kipukas is derived from 
several ash deposits. The lower, sloping ones in 
Kipuka Puaulu differ from the upper ones, 
which are stratified horizontally. 
2. Charcoal was found in both kipukas un- 
der forest in association with a buried, black 
surface horizon, at 70 cm depth in Kipuka Ki 
and at 80 cm in Puaulu. 
3. The C-14 analysis of the 70 cm-deep 
charcoal in Kipuka Ki indicates that a fire 
occurred at about 220 B.c. 
4. The forest soils of both kipukas are uni- 
formly melanized, showing considerable mega- 
scopic similarity, and differ markedly from the 
savannah soil, which showed melanization re- 
stricted to narrow layers and which exposed a 
clear parent material stratification. 
5. The soil parameters tested indicated no 
significant differences between the forest soil 
and the savannah soil of Kipuka Puaulu in 
terms of soil water, organic carbon, and pH. 
6. The forest soils of the kipukas differ only 
in current soil moisture distribution and organic 
carbon content of the lower horizons (B3 and 
C2). 
The work so far is only an introduction to 
the plant ecology of Hawaiian kipukas and 
points to the need for the following further 
research : 
1. Analysis of photographs. It would be 
profitable to examine all photographs Rock 
made of Kipuka Puaulu and, if possible, to 
identify some spots for rephotographing. This 
could reveal certain interesting successional 
changes over the last 50 years. 
2. Current observation indicates reoccupa- 
tion of the savannah by forest. This appears to 
be accomplished by sucker growth of Acacia 
koa. Invasion of trees by seed seems practically 
impossible. It would be interesting to study the 
rate of reinvasion, now, when there is no more 
interference by cattle. 
3. Studies of cloud interception. Differences 
in soil water supply as a result of fog drip 
should be investigated, to determine the role 
this environmental factor plays in influencing 
the rate of reinvasion of forest into the savan- 
nah. 
4. Measuring pig damage. Current observa- 
tion indicates that pigs affect the forest vegeta- 
tion in two ways, (a) By scarifying the surface, 
they eliminate ground vegetation and provide 
ideal seed beds for tree seed germination of that 
which is left. Many formerly pig-scarified areas 
seem to come back in thickets of tree seedlings 
of Sapindus and Coprosma. (b) During periods 
of food scarcity or over-population pigs seem 
to gnaw away the bark of trees, particularly of 
Coprosma, thus damaging them severely, e.g., 
by providing entrance avenues for pathogens. 
The food habits of pigs should be studied in 
connection with population counts to explain 
their influence on vegetation patterns. 
5. Quadrat studies of vegetation patterns. 
These should be done in particular with Pepe- 
romia , as a probable native pioneer on pig- 
scarified ground; with Commelina and Nephro- 
lepis mixed associations, to determine whether 
Commelina takes over the habitats occupied by 
the native fern, Nephrolepis ; and with the two 
weed communities formed by Rubus penetrans 
and Solanum pseudocapsicum , to determine 
their effect on the native Microlepia association. 
6. An ecological survey of all kipukas and 
their surroundings should be made in an at- 
tempt to assess their development in succession 
and their influence on the vegetation of the sur- 
rounding more recent volcanic material. 
CHECK LIST OF PLANTS IN THE KIPUKAS 
This check list includes all of the vascular 
plant species of Kipuka Puaulu and Kipuka Ki 
as of May 1965. The symbols used are: * = 
native Hawaiian species ; # — native tree ; 
-f- = growing, apparently naturally, in kipuka 
in 1963-65; a = growing in kipuka in 1963- 
65 only as individuals recently planted by Na- 
tional Park Service ; b = specimens, collected 
between 1930 and I960, in herbaria at Hawaii 
Volcanoes National Park or B. P. Bishop 
Museum, but species not found growing in 
kipuka in 1963-65 ; c = reported by Fagerlund 
and Mitchell (1944), but no specimens avail- 
able; d = reported by Rock (undated, and 
1913), but no more recent specimens available. 
