XIII, A, 1 Witt: Cement and Concrete 35 



of boric acid, barium chloride,- calcium oxychloride, calcium 

 thiosulphate, and lithium chloride. When a substance has been 

 investigated by a number of persons, it often happens that 

 conflicting statements are published. According to Chandlot and 

 Rohland,'' sodium chloride has no efi'ect on cement, but Dobrzyn- 

 ski and Passov/ say that it increases the tensile strength. Roh- 

 land classes sodium carbonate as an accelerator of the set of 

 cement, while in a previous paper he stated that it might either 

 retard or accelerate it. Similar disagreements on the effect of 

 calcium chloride mav be found, though its general effect on the 

 set has been established. One interesting example of dis- 

 agreement is that of ferrous compounds. Desch ° is of the opi- 

 nion that if any ferrous iron is present in cement it may be 

 converted into ferrous sulphide by calcium sulphide, which is 

 produced by the reducing action of the fuel. Ferrous sulphide is 

 considered objectionable because of its ability to become oxidized, 

 which would result in a change in volume. However, there is a 

 French patent for the right to add from 1 to 3 per cent dry fer- 

 rous sulphate to cement. It is claimed that this will accelerate 

 the set and increase the strength. The general effect of calcium 

 sulphate is perhaps better known than that of any other sub- 

 stance, because so much research has resulted from its use in the 

 manufacture of cement. However, there are still many conflict- 

 ing views on the role of the substance in controlling the set and 

 the amount that can be added without endangering the quality.^ 



The causes that have thus far prevented many of the investi- 

 gations carried on by various workers from being comparable 

 may be divided into two classes : 



1. Details that either cannot be controlled by the investigator 

 or which can be controlled only with considerable difficulty. 

 These include the individual characteristics of cements, resulting 

 from the composition of the raw materials and the methods of 

 manufacture; the personal equation of the investigator; varia- 

 tions in temperature and humidity due to change of season and 

 to location ; and differences in apparatus and methods of pro- 

 cedure in various countries. Most of these points do not require 

 explanation. It is well known that cements of the same chemical 

 analysis may be entirely different in their physical properties. 

 While it would be possible to control the temperature and 

 humidity of laboratories so that cement could always be tested 



"See Table I. 



" Desch, C. H., op. cit., 75. 



' Cf. Witt, J. C, and Reyes, F. D., This Journal, Sec. A (1917), 12, 133. 



