228 The Philippine Journal of Science 1919 
storage were arranged. The cement was spread out in thin 
layers and thoroughly mixed from time to time to expose new 
surfaces. At intervals, portions were removed and tested for 
setting time. The results are shown in Table IV. Sample 13, 
with 0.62 per cent sulphuric anhydride, changed from a flash 
set to a normal set; sample 15, 1.72 per cent, changed from 
normal to quick setting; No. 1, 2.15 per cent, remained slow 
setting; No. 4, 1.92 per cent, became quick setting in nine days, 
but had returned to normal at the end of eighty-five days. In 
this series of experiments, it is evident that the effect of ex- 
posure is practically independent of the sulphuric anhydride 
content. 
Hydrated lime is said to retard the set of cement.® Some tests 
were made to determine its effect in the presence of various 
amounts of gypsum. Cement having a flash set and containing 
1.50 per cent sulphuric anhydride was used. Gypsum ground 
to equal fineness and containing 35.81 per cent sulphuric anhy- 
dride was mixed with this cement in proportions to give cements 
of 1.65, 1.75, and 2.00 per cent sulphuric anhydride, respectively. 
To the original cement and to each of the mixtures were then 
added 0.5, 1, and 2 per cent hydrated lime, respectively. The 
setting time of each mixture was then determined, and the 
results are shown in Table V. The effect of the lime is appa- 
rently independent of the sulphate present. The normal con- 
sistency increases with the lime present. 
POSSIBLE CAUSES OF IRREGULARITY 
The peculiar irregularity of this cement is difficult to explain. 
A number of quick-setting samples have been analyzed and the 
analyses compared with those of slow-setting samples having 
a lower sulphuric anhydride content. Some differences were 
found, but they were not sufficiently large to account for the 
irregularity. The cement is somewhat higher in iron and in 
alkalies than the average cement, but othervrise there is nothing 
unusual in its composition. High iron content should not make 
cement quick setting.® The various alkalies are present in con- 
stant amounts and therefore should have a uniform influence, 
unless their effect is caused to vary by changes taking place 
during the manufacturing process. 
Since the irregularity has not been accounted for on the basis 
'Carpenter, R. C., Eng. Rec. (1904), 50, 769; Meade, R. K., Portland 
Cement. Easton, Pa., The Chemical Publishing Co. (1911), 424. 
* Meade, op. cit., 35. 
