35 
same as in the pot experiments; No. 4 is a very productive silt, hi^h 
in lime (3.80 per cent) and humus; Nos. 5 and 6 are brown-clay soils 
belonging to the class of clays which contain a larger amount of iron 
than aluminum; Nos. 7 and 8 are clay soils belonging to the class of 
clays which show a higher content of aluminum than iron; No. 9 is 
a soil containing 20 per cent titanium, about 40 per cent iron oxid, 
and 8 per cent aluminum; No. 10 is a soil which is principally coral 
sand (about 90 per cent calcium carbonate); No. 11 is a sandy soil 
from a humid district and is high in both lime (6.3 per cent) and mag- 
nesia (5.8 per cent). These 9 types of soils include all the important 
ones of the islands, for which reason the data should be of wide 
application in drawing conclusions regarding the locking up of the 
phosphate. 
The data given in the table show that there are included soils 
which possess all the conditions generally considered essential for the 
fixation of phosphates. There are the normal conditions, such as 
high clay content, colloidal clay, and organic compounds, which pro- 
mote physico-chemical absorption; the humic conditions which pro- 
mote biological absorption; and, finally, the chemical conditions, 
such as high content of lime, magnesium, iron, aluminum, and tita- 
nium, which, either through an actual combination or a reversion to 
a less soluble form, influence chemical fixation. At least one or 
possibly all three of the above factors may influence the maintenance 
of a favorable medium in the soil for plant growth, in so far as plant 
growth is affected by the presence of a readily available source of 
phosphoric acid. A relation may be established from data given in 
Table XII between the chemical and physical composition of the soil, 
the solubility of phosphoric acid in various solvents, and its availa- 
bility as measured by plant growth. 
Table XII shows the relative solubility of phosphoric acid in all the 
important soil types of the Hawaiian Islands. Since three of these 
types, Nos. 1, 2, and 3, were used in the pot experiments, a com- 
parison of the data will indicate the relation between the solubility 
and the availability as measured by the growth of millet. Soil No. 1 
did not respond to phosphate fertilizer, thus showing the high avail- 
ability of its phosphoric acid; soil No. 2 was greatly in need of avail- 
able phosphates, as indicated by the marked increase in plant growth 
following the application of phosphate fertilizers in all forms; soil 
No. 3 was loss in need of phosphate than No. 2, as indicated by a 
smaller increase resulting from the addition of phosphate. The 
other soils have not been used in pot experiments. The results of 
previous experiments ' with soils of the same type as certain of those 
included in Table XII are summarized in Table XIII. 
I W. P. Keller, Jour. Indus, and Kn-in. (hem., 2 (1910), p. 277. 
