220 REIBLING AND REYES. 



As none of the above experimental data threw any light on the varia- 

 tions in the set brought about by exposure to the air, we had recourse 

 to the microscopic tests for the lime described in Part I of this paper *® 

 and we thus were convinced that changes in the rate of set were brought 

 about largely by the alterations in the condition of the free lime. 



It soon became evident that the field of investigation would have to 

 be carried beyond the study of the commercial material. Ordinarily, 

 commercial Portland cements consist of a mixture of variovis kiln 

 products the different parts of which have not seasoned in the same 

 manner, or to the same extent. They also contain various amounts of 

 calcium oxide, hydroxide and carbonate. Consequently, further season- 

 ing does not produce as marked an influence on the set as when condi- 

 tions are more uniformly regulated. Therefore, we again foimd it 

 necessary to supplement our work on ordinary commercial cements with 

 a study of known and selected materials which we ground, plastered 

 and seasoned ourselves. 



For this purpose we secured large qiiantities of fresh, non-seasoned 

 clinkers (brands 0, K, and L) from three manufacturers who were 

 selected in order to obtain standard cements with characteristically 

 different chemical and physical properties. The normal setting prop- 

 erties of these three brands of cement as previously ascertained by our 

 routine and research work are given in Table XXIII. 



^^Loc. cit. 



