292 
PACIFIC SCIENCE, Vol. XXI, April 1967 
the herbarium of Hawaii Volcanoes National 
Park. At the end of this article will be found 
a check list of the plants found in the two ki- 
pukas. It includes also records from Rock (un- 
dated, and 1913), and Fagerlund and Mitchell 
(1944), as well as specimens in the herbaria of 
Hawaii Volcanoes National Park and the Ber- 
nice P. Bishop Museum, Honolulu, Hawaii. 
Table 2 summarizes the information provided 
in the check list. It shows that Kipuka Puaulu 
now contains, and has contained, significantly 
more species of vascular plants than has Kipuka 
Ki. Table 3 provides an analysis of the numbers 
of species common to both kipukas and of those 
found only in Kipuka Puaulu or Kipuka Ki. 
This indicates that, while each kipuka contains 
species which the other does not, Kipuka Puaulu 
has a significantly greater number of unique 
species than does Kipuka Ki. 
Thus, the observations recorded in Tables 2 
and 3 and in the check list agree with Rock’s 
(1913) observation that there are more species 
TABLE 2 
Numbers of Species, Varieties, and Forms of 
Vascular Plants Recorded from Kipuka 
Puaulu and Kipuka Ki 
PLANTS 
ALL 
DATA 
KIPUKA 
PUAULU 
KIPUKA 
KI 
Total number 
of spp. 
Native spp. 
Native trees 
Introduced spp. 
158 ( 104) * 
86 (52) 
42 (21) 
72 (52) 
145 (92) 
81 (48) 
42 (21) 
64 (44) 
73 (63) 
36 (30) 
15 (11) 
37 (33) 
* Figures outside parentheses include all spp. ever 
Figures within parentheses include all spp. growing 
in 1963-65. 
recorded. 
naturally 
TABLE 3 
Distribution of Species, Varieties, and Forms 
of Vascular Plants Between Kipuka 
Puaulu and Kipuka Ki 
plants 
COMMON 
TO BOTH 
KIPUKAS 
KIPUKA 
PUAULU 
ONLY 
KIPUKA 
KI 
ONLY 
Total number 
of spp. 
Native spp. 
Native trees 
Introduced spp. 
60 (51)* 
31 (26) 
15 (11) 
29 (25) 
85 (41) 
50 (22) 
27 (10) 
35 (19) 
13 (12) 
5 (4) 
-(-) 
8 (8) 
* Figures outside parentheses include all spp. ever recorded. 
Figures within parentheses include all spp. growing naturally 
in 1963-65. 
in Kipuka Puaulu. The number of native trees 
now growing in Kipuka Puaulu (21) is almost 
twice as large as that in Kipuka Ki (11). How- 
ever, Rock reported in 1913 that there were at 
least 40 native tree species growing in Kipuka 
Puaulu. Even allowing for differences of taxo- 
nomic opinion, the decrease in number of tree 
species during the last 50 years appears quite 
remarkable. In his book, "Indigenous Trees of 
the Hawaiian Islands,” Rock (1913) included 
19 photographs of tree species in Kipuka Pu- 
aulu. Fifteen of these photographs show bits 
of landscapes and ground vegetation, which at 
that time appeared badly abused by cattle graz- 
ing. Many areas appear barren or show trampled 
ground vegetation, and several pictures show 
broken-down trees. From the photographic rec- 
ord one could assume that the present savannah 
formation is caused entirely by cattle grazing. 
However, two photographs show what appear 
to be sections of the present savannah forma- 
tion. One of these shows a dense cover of 
Pteridium, which today is also well established 
in the savannah. Inasmuch as fire has definitely 
occurred in both kipukas, it is believed that fire 
may have created openings in the forest that 
were aggravated and maintained by subsequent 
cattle grazing. It seems probable that cattle were 
guided in their grazing habits by this fire- 
conditioned vegetation pattern, since a denser 
ground vegetation would be found in the open, 
coupled with fewer obstacles for the movement 
of cattle. Increased cloud interception and fog 
drip in the forest (Ekern, 1964) may also have 
contributed to maintaining the pattern. This is 
indicated by the greater current moisture con- 
tent in the lower profile of the forest soil 
(Table 1) and the location of the kipukas in 
a zone of common cloud occurrence (Krajina, 
1963). 
There are several possible explanations for 
the larger number of both native and introduced 
species in Kipuka Puaulu. 
1. The larger number of native species in 
Kipuka Puaulu may be related to: 
(a) larger area. Both kipukas are so much 
larger than the "minimal” area-size of forest 
stand communities cited in the literature (Ellen- 
berg [1956] gives 500 m 2 ; Cain and Castro 
[1959], < 20,000 m 2 for tropical rain forest), 
that one may think that size is not a factor. 
