1659verted to new type of operation; low temperature regeneration allows use of ordinary steel and keeps corrosion at a minimum; and odor nuisance during regeneration is eliminated, since disulfides are formed as a by-product. CONCLUSIONSThe use of mercaptan oxidation catalysts, soluble in alkaline treating reagents, offers an attractive, economical means of regenerating solutions used in the removal of mercaptans from gasoline. The ease of regeneration possible through the use of oxidation catalysts materially improves operations of mercaptan extraction systems. ACKNOWLEDGMENTThe authors wish to express their appreciation to J. B. Rather of Soconv-Vacuum Oil Company, Inc., for permission to publish this paper. Since the paper was originally offered for publication, arrangements have been made with the Shell Development Company for the joint licensing of Shell's solutizer process incorporating the described tannin-catalyzed air regeneration technique, the combination being known as tannin solutizer process. The authors also wish to acknowledge the assistance of members of that organization as regards information on the more recent commercial applications.

presents research results, test methods, and performance criteria related to the structural and environmental functions and the durability and safety characteristics of building elements and systems. studies of sulfate expansion, heat of hydration, autoclave expansion, compressive strength after normal and steam curing, shrinkage, absorption, freeze-thaw durability, as well as the effect of outdoor exposure on concrete specimens.

The relationship between cement properties and drying shrinkage of partially hydrated neat cement pastes of normal consistency made of a large number of portland cements of different types and composition were studied by fitting multivariable regression equations with the aid of a digital computer. The time of cracking of annular specimens restrained from shrinking by a steel core varied from 0.1 to more than 40 hours. The principal variables which appeared to have the greatest effect were the fineness and K2O and C4AF contents of the cements. The fineness and K2O contents of the cements appeared to have the greatest effect on the nonrestrained shrinkage of prisms at the time of cracking of the annular specimens. The shrinkage of neat cement bars after 1, 6, 27 days, and 6 months in laboratory air differed greatly with different cements. The principal variables appeared to be the C3A, SO3, and C3A/SO3 ratios. Other commonly determined variables were associated with each of the above test values. The trace elements, other than Na20 and K2O, were not generally associated with the shrinkage or cracking characteristics. Keywords: Chemical composition of portland cements; cracking of portland cements; cracking resistance of cements; portland cement; shrinkage of Portland cements; trace elements. Page 1. * Included among those notations etc. are the use of C3A, C3S, CzS, and C4AF for the calculated potential compound composition of the cements, viz, tricalcium aluminate, tricalcium silicate, diealcium silicate and tetraealcium aluminof errite. Also used are Loss for loss on ignition, Insol for insoluble residue, and APF for airpermeability fineness. Also used are AE + NAE to designate airentraining plus non-air-entraining cements and NAE for the nonair-entraining cements. The compressive strength in psi of 1:2.75 (cement to sand) mortars cured for 1, 3, 7, 28 days, and 1 year are designated by STOl, ST03, ST07, ST28, and STIY respectively. * Note (1) Ratio of number of plotted points in pairs of diametrically opposite quadrants. Note (2) General trend of lines drawn through plotted points. Note (3) Apparent nature of relationship. L = linear; NL =: nonlinear; NO = no apparent relationship; and L? = nature of relationship not determinable. (Note (4) Quadrant sum in corner test (see reference 5). * Note (1) Ratio of number of plotted points in pairs of diametrically opposite quadrants. Note (2) General trend of lines drawn through plotted points. Note (3) Apparent nature of relationship. L = linear: NL = nonlinear; NO = no apparent relationship; and L? = nature of relar tionship not determinable. Note (4) Quadrant sum in corner test (see reference 5).

Brunauer-Emmett-Teller (BET) surface areas of hydrated portland cement have been calculated from water-vapor and nitrogen-adsorption data. The adsorption and desorption of wat~r vapor caused. measurable decreases in specific surface areas by both water-vapor and mtrogen adsorptIOn. Changes were !tlso produced by wetting and drying. The adsorption and desorption of nitrogen at the boiling point of liquid nitrogen did not produce similar effects.The specific surface and nonevaporable-water content of a hydrated cement are also somewhat dependent on the initial drying treat ment. The effect of measuring water-vapor surface a nd nonevaporable water with specimens dried for different periods of time is also considered.

The relationships between cement characteristics and compressive strength of 1:2.75 (cement to graded Ottawa sand) mortars of standard consistency at ages of 24 hours to 10 years, and made with 199 cements of different types, were studied by fitting multivariable regression equations with the aid of a digital computer. The dominant variables associated with the differences of compressive strength, strength gain, and strength ratios wore different at the various test ages, and after different curing conditions. The additional use of certain trace elements with commonly determined independent variables resulted in a significantly better fit between the equations and the observed data. Interactions between the fineness values and othei;-independent variables were noted. Certain parallelisms and differences were noted with respect to variables associated with compressive strength differences at various ages and the heat of hydration of the cements at these ages.