480 
PACIFIC SCIENCE, Vol XV, July 1961 
mediate stage in the weathering sequence of 
halloysite to gibbsite. Bates also reported the 
formation of gibbsite crystals upon dehydration 
of amorphous Fe-Al gels. This observation sup- 
ports Sherman (1957), who reported that crys- 
talline gibbsite aggregates formed when the 
soils of the Hydrol Humic Latosol group were 
air dried. These soils have a high content of 
amorphous mineral colloids which contain a sub- 
stantial amount of gel material. 
Some of the Japanese workers (Sudo, 1954; 
Sudo and Ossaka, 1952) conclude that allo- 
phane precedes halloysite in the weathering 
sequence from ash to allophane to halloysite. 
Aomine and Yoshinaga (1955) have also em- 
phasized that the clay fraction of the volcanic 
ash soils of Kyushu and Hokkaido formed un- 
der similar well-drained conditions is pre- 
dominantly allophane, regardless of differences 
in temperature, weathering time, vegetation, 
and ash origin. 
The New Zealanders have also tried to pro- 
perly position the amorphous materials in the 
weathering sequence. Fieldes and S win dale 
(1954) have prepared a flow sheet tracing the 
mechanism of silicate minerals weathering. 
They have proposed that the nature of the clay 
constituents of any soil can be predicted if its 
parent material and weathering stage are known. 
The amorphous materials occupy a great role 
in this flow sheet in that it is thought that the 
primary silicates (aside from the micas) can- 
not form layer silicates without first passing 
through an amorphous stage. Clays derived from 
rhyolitic and andesitic ash pass through the 
weathering sequence from amorphous hydrous 
oxides through allophane to meta-halloysite and 
kaolin. It is believed that many of the Hawaiian 
soils derived from andesitic ash follow this same 
sequence. 
In an earlier paper Fieldes et al. (1952) em- 
phasized that the amorphous hydrous oxides 
played more than a brief transitory role. They 
reported some soils of the lower Cook Islands 
which showed high cation exchange capacity 
values. These soils were all low in silica, and 
also allophane was not found to be a constitu- 
ent in them. They attributed the cation ex- 
change capacity mainly to the amorphous hy- 
drous oxides. 
In later papers Fieldes (1955, 1956, 1957) 
reported enough fundamental differences in allo- 
phane to warrant recognizing three types: allo- 
phane A, allophane B, and the intermediate 
type, allophane AB. Based mainly on infrared 
absorption data, it was found that silica is linked 
with alumina to form allophane A while some 
silica is discrete as amorphous hydrous silica in 
allophane B. Fieldes could offer no satisfactory 
explanation as to why co-precipitation and link- 
ing of alumina and silica occur to only a limited 
extent in allophane B. He did not want to state 
that allophane B is simply a mixture of amor- 
phous alumina and silica. Differential thermal 
analysis shows that a high temperature exo- 
therm between 850° and 1000° C. is strong in 
allophane A, not present in allophane B, and 
weakly developed in the intermediate form, 
allophane AB. Fieldes (1955) has presented a 
weathering sequence of clays derived from rhy- 
olitic and andesitic ash: allophane B — allophane 
AB- — allophane A — meta-halloysite — -kaolinite. 
He has stated that in this sequence the stable 
form is meta-halloysite and progress towards 
this stable form through allophane A is consist- 
ent with the mechanism proposed by Tamura 
and Jackson (1953). The structure consisting 
of hydrous alumina octahedra randomly cross- 
linked by silica tetrahedra and called allophane 
by Tamura and Jackson would hence correspond 
to allophane A as proposed by Fieldes. 
There is a growing consensus among investi- 
gators in this field that amorphous colloids may 
play a very important role in soil formation and 
in establishing properties of many soils of the 
continental United States as well as in the Pacific 
islands. Because of their noncrystalline nature, 
identification of allophane and other amorphous 
constituents is at present very difficult, and at 
best very unreliable, by standard methods of 
analysis unless they be present as predominant 
components of their system. In view of the fact 
that transition from primary silicates (with the 
exception of micas) to the secondary layer sil- 
icates must include some solution and reprecip- 
itation, it is reasonable to suspect that amor- 
phous colloids exist, at least as a transition stage, 
in most of our soils. The extent of their pres- 
ence is masked in most of our mineralogical 
studies by the crystalline components present. 
