22 BULLETIN 6 2, HAWAII EXPERIMENT STATION 
In the case of the first two soils mentioned, the' determining factor 
for the hygroscopic coefficient is the percentage of clay, the humus 
content being small. However, in the case of the last two soils men- 
tioned, the amount of organic matter is the dominant factor. The 
Clyde clay loam, although possessing only 20.1 per cent clay, shows 
a higher hygroscopic coefficient than the Vergennes clay, the clay 
fraction of which is 74.5 per cent. A comparison of the Dunkirk 
subsoil with the Clyde clay shows that they both possess the same 
amount of clay, yet the Clyde soil has a hygroscopic coefficient over 
three times as large as that of the Dunkirk soil, due to the disparity 
in humus. 
In the above-cited example the magnitude of the humus content 
ranged from 0.2 per cent to 4.34 per cent, small figures, relatively 
speaking, as compared with the percentage of organic matter in 
Hawaii soils. If such disparity is possible with soils containing 
from 0.2 to 4.34 per cent of organic matter, it is logical to assume 
that much larger differences are possible with soils containing from 
two to three times that amount of organic matter. This actually is 
the case with Hawaii soils. The wide limits between which the 
organic-matter content of Hawaii soils varies explains the wide 
limits obtained for the moisture constants themselves. 
The moisture-equivalent figures vary between 21.9 and 71.3 per 
cent, the average of the whole series being 46.7 per cent. Consider- 
able variations were obtained for this moisture constant even within 
the same soil type. For instance, soils Nos. 57 and 9, which are 
classified as clay and clay loam, on the basis of mechanical analysis, 
had moisture equivalents of 27.4 and 67.2 per cent, respectively. In 
this case the large difference in the respective organic-matter content 
(6.27 and 18.52 per cent) may explain the disparity in the moisture 
equivalent. However, in the case of soils Nos. 2 and 49, the moisture 
equivalents of which were 39.6 and 71.3 per cent, respectively, the 
difference in organic-matter content does not explain this disparity 
(organic matter 3.94 and 2.34 per cent). This case is all the more 
remarkable because soil No. 2, with the lower moisture equivalent, is 
a clay with 51.5 per cent colloids, whereas soil No. 49, with a moisture 
equivalent almost twice as great as that of soil No. 2, is a clay with 
only 21.3 per cent colloids. 
These discrepancies may be explained by the considerable differ- 
ences that may exist in the chemical composition of the inorganic 
colloids, and by the amount and nature of the coarser-than-clay frac- 
tions, such as silts and fine silts. As will be pointed out later (p. 25), 
the chemical composition, especially the silica-sesquioxide ratio, may 
vary very considerably in the different fractions of the same soil. 
Similar discrepancies in moisture equivalents are noted in the 
figures published by Bennett and Allison (4, pp. 21, ##, 308) in the 
case of Cuba soils. For instance, Alto Cedro clay No. 32828, with 
a clay content of 52.1 per cent, gave a moisture equivalent of 75.2 
per cent, whereas Nipe clay No. 32830, with a clay content of 44.5 
per cent, gave a moisture equivalent of 23.7 per cent. Both soils 
are very low in organic matter. In this instance the low moisture- 
retaining capacity of the Nipe clay is attributed by the authors to 
certain physical peculiarities, notably to large pore space. 
