Desilication of Halloysite and Its Relation to Gibbsite Formation 1 
Goro Uehara, Haruyoshi Ikawa, and G. Donald Sherman 2 
ABSTRACT: The evidence presented points to the alteration of halloysite to gibb- 
site. Mineralogical data, as determined by X-ray and differential analysis, verify 
identification of halloysite and gibbsite. Chemical data confirm the expected lower 
silica and higher alumina content for samples which are predominantly gibbsitic. It 
is reasonable to assume from petrographic evidence that gibbsite develops by desilica- 
tion of halloysite. Halloysite amygdules undergo desilication along the outer peri- 
pheries, where acid silica-deficient waters pass, attacking the halloysite by dissolving 
silica. Halloysite is stable only if it is protected from such solutions, or if the solution 
passing by is saturated with silica. 
Whereas alteration of feldspar to halloysite involves a gain in volume, a loss in 
volume follows desilication of halloysite. This loss in volume is exemplified by the 
surface cracks clearly visible in the desilicated halloysite. 
In AN earlier PAPER Sherman and Ikawa 
(1959') described the occurrence of gibbsite 
amygdules in the Haiku bauxite area of Maui, 
Hawaii. At that time it was proposed that 
gibbsite amygdules formed in rock vesicles by 
rhythmic precipitation of hydrated alumina. Sub- 
sequent and more thorough investigation of the 
same area and elsewhere in the Hawaiian Islands 
suggests a more complex sequence of events in 
the development of these gibbsite amygdules. In 
this paper the results of field and laboratory 
studies are presented to explain halloysite forma- 
tion and its subsequent desilication to gibbsite. 
Bates (1962) suggested that gibbsite is pro- 
duced in Hawaii by ( 1 ) removal of silica from 
halloysite, (2) dehydration of aluminum gel, 
and (3) precipitation from solution. The first 
mode of formation is of primary interest in this 
study. 
FIELD DESCRIPTION 
All data presented in this work are for samples 
collected from the Haiku bauxite area described 
1 Published with the approval of the Director of the 
Hawaii Agricultural Experiment Station, University of 
Hawaii, Honolulu, as Technical Paper No. 660. Manu- 
script received September 4, 1964. 
2 Assistant Soil Scientist, Junior Soil Scientist, and 
Senior Soil Scientist, respectively, Hawaii Agricultural 
Experiment Station, University of Hawaii, Honolulu, 
Hawaii. 
by Sherman and Ikawa ( 1959). Similar features 
have been observed on the islands of Hawaii, 
Molokai, Oahu, and Kauai, but the Haiku site 
represents the most extensive halloysite-gibbsite 
deposit of this kind. This site is located ap- 
proximately 2 miles east of Pauwela Village, 
Maui, and can best be studied along deep road- 
cuts on the highway leading to Hana. 
The soil occurring in this area Is mapped as 
the Pauwela series of the Humic Ferruginous 
Latosol. Soils of the Pauwela series are found on 
the island of Maui from sea level to 1,500 ft 
elevations in areas receiving an annual rainfall of 
80-150 inches. 
Highly vesicular andesitic pahoehoe probably 
constituted the parent rock. Pahoehoe lavas are 
often vesicular and have a rather regular distri- 
bution of vesicles; some are cellular to a degree 
that makes them nearly pumiceous, so that when 
these vesicles are filled with secondary minerals 
a larger volume of the rock may be occupied by 
them. 
A description of the Pauwela soil can be 
found in Cline et al. (1939). Approximately 
3-5 ft below the surface there appears a yel- 
lowish-brown weathered rock ( saproly te ) re- 
taining much of the original rock structure. In 
the area investigated this weathered zone per- 
sists even in the deepest portion exposed by 
the roadcut, although occasional unweathered 
cores also may be seen. 
119 
