the exception of the Cocos and Digitaria could be found, The gene 
ral aspect of the veg’etation indicated a lengthy dry spell prior 
to our visit. 
The Vegetation of Caroline Atoll 
The earliest recorded observation of the vegetation of Caro¬ 
line Atoll is found in Bennett (Bennett, I 84 O), In this volume the 
author related that Cocos were found on the South Island only (see 
map page 366) — a grove on the northeast side of South Island, The 
area covered by this grove is approximately one-fifth of the land 
area of South Island, Bennett refers to the island as ”covered with 
verdure" and goes on to say that ” the interior of the island is a 
sxirface of sand, mingled with coral debris as well as with decayed 
vegetable and animal matters, which give it increasing fertility". 
He also mentions trees "attaining the hei^t of twenty feet". This 
may indicate that the Cordia and or Pisonia gloves covered a more 
extensive area in the last century, Bennett mentions that ” the 
woodlands are chiefly composed of two species of Toxrmefortia ", It 
is difficult to interpret this statement other than by supposing 
Bennett was referring to Cordia as well as Tournefortia . Three spe¬ 
cies were introduced by Bennett: Ipomoea batatas, Taoca leontopeta- 
loides and Inocarpus edulis Porst, Ho siga of any of these species 
other than the Tacca was found in I 965 . The early observation of 
only this smallish grove of trees on South Island bears witness to 
an early aboriginal population or visiting population of Polynesians 
who planted groves extensive enou^ to cover the needs of the people. 
Later under the atispices of Brown and Bros. (Salmon in Holden, Mem. 
Hat, Acad, Soi, 2: 2—22) more Cocos were planted on the island and, 
presvunably on Hake Island. In I 872 guano operations were begun (Arun¬ 
del in Holden, Mem. Hat, Acad. Sci, 2: 2-22), It is not clear whe¬ 
ther the guano was mined on South Island or Hake Island or both from 
these early reports. However since the settlement was on South Island 
it would be plausible to assume that at least some of the mining was 
