22 BULLETIN 348, TJ. S. DEPARTMENT OF AGRICULTURE. 
an average cementing value of 24, with decomposed sandstones 
showing an average binding value of 73 (Table 6, No. 7), and to note 
that the secondary products of the latter rocks consist almost 
entirely of colloids in the form of ferruginous clay (kaolin and limo- 
nite), while in the former crystalline epidote and chlorite are present 
in proportionately greater amounts. 
The effects of alteration products on cementing properties of rock 
powders are shown in figure 3, where varying percentages of secondary 
minerals in rocks and gravels have been plotted against their respec- 
tive cementing values. The great difference in binding properties 
between quartzose plutonic rocks (granite and diorite) and quartz- 
free types of the same class (syenite and gabbro), as well as between 
plutonic and volcanic rocks of similar composition, is clearly demon- 
strated. It will be observed also that where more complete curves 
could be drawn, as in the case of basalt, diabase, and sandstone, a 
critical point is reached at about 14 per cent of secondary minerals up 
to which the rate of increase in cementing value is almost directly pro- 
portional to that of these minerals, but beyond this point a marked 
deflection or break in the curves occurs in accordance with decreased 
cementing values resulting from largely increased proportions of 
highly crystalline products of alteration (epidote, zeolites, and cal- 
cite). The curve representing gravels from which the secondary 
minerals or clay content have been obtained by sifting through a 
200-mesh screen and washing, indicates a relatively high cementing 
value up to samples containing 2 per cent clay, and shows that 
beyond this the relative increase corresponds in general to that of 
sandstone, reaching a maximum of 120 in samples having an average 
of 16.8 per cent of clay. 
The property to slake in water is also common to many rock types, 
and as shown in Table 8 is largely dependent upon the physical 
character of the rock-forming constituents. The tests have been 
made by placing thoroughly desiccated briquettes of rock powders 
in water and noting the time required for their complete disintegra- 
tion. Briquettes remaining intact for 24 hours in water are con- 
sidered nonslaking. 
It will be observed that the samples tested have been separated 
into acid and basic crystalline rocks (Nos. 1-4) and siliceous sedi- 
mentary rocks (Nos. 5 and 6), and the average mineral composition, 
time of slake, and cementing value given in each case. In comparing 
the results it will be noted that the nonslaking acid rocks (No. 1), 
while somewhat richer in secondary minerals, especially chlorite and 
epidote, contain a large excess of primary hornblende and consider- 
ably less quartz and mica than the slaking varieties (No. 2), while 
their cementing values are about the same. Nonslaking basic and 
