Illinois Division of Highways for use in cold weather, both as an integral compound and as a surface application for curing concrete roads. In the former case the amount is limited to 2% by weight of the cement. The specification for road work requires 21⁄2 lb of the flaked, or granular material, uniformly distributed over each square yard of pavement, in lieu of the customary requirement that the concrete be kept moist during the curing period. Laboratory tests also indicate that concrete samples cured in dry air show great benefit at all ages from the use of calcium-chloride. This material, however, should not, in any sense, be considered as insurance against the danger of frost action, although it undoubtedly reduces the critical time during which freezing is injurious. Furthermore, its value as a curing agent seems to depend very largely upon the existing atmospheric conditions.*

In the light of our present knowledge calcium-chloride should not be employed as an admixture in reinforced-concrete work except where thorough embedment of the reinforcement can be assured. Up to about 4% of the commercial material by weight of cement (2% chlorine by weight of cement), its use as an accelerator may often be desirable. The following statement, the result of a large number of tests, is abstracted from Bulletin 13 of the Lewis Institute.

"In the use of calcium-chloride no advantage was gained for percentages of the commercial product greater than 2 or 3% of the weight of cement (chlorine content 1 to 12%). This amount when used in mixes of about 1:5 and in consistencies suitable for building construction showed an increase in compressive strength of from 100 to 200 lb per sq in which increase was practically constant at ages of two days to three years. For richer mixes and drier consistencies the strength increase was greater and for leaner mixes and wetter concrete it was less."

As various brands of standard cements hydrate four or five times as fast as others when incorporated with a constant solution of calcium-chloride, tests should be made with the particular brand of cement intended for use before determining upon the specification. There is no reason to believe that the use of calcium-chloride increases the tendency toward efflorescence. Calciumchloride in liquid form is dissolved in the mixing water before the latter is introduced into the mixer.

CAL, which is obtained by pulverizing the dry, or undried, product resulting from a mixture of either quick-lime or hydrated-lime, calcium-chloride, and water is similar to calcium-chloride in its general application as an admixture. Like the former, its chief value is to accelerate the hardening process, giving considerably greater strength at the two and seven-day periods, and practically the same strength at the 28-day period. The effectiveness of Cal as an accelerator differs with different brands of cement, and the same precautions should be observed as in the case of calcium-chloride. Tests upon mortar treated with Cal show an even greater relative increase in strength for the two and seven-day periods than for the concrete samples, and an increase in strength is apparent at even the 28-day period. In regard to the use of Cal in reinforced* Bulletin No. 15, Structural Materials Research Laboratory, Lewis Institute.

concrete there is as yet no authoritative data relative to its possible effect upon the reinforcement; it is therefore recommended that its use be avoided where there is likelihood of moisture being present. Cal is particularly valuable as a curing agent for concrete subjected to premature drying. It apparently does not cause increased efflorescence. As to the relative value of Cal compared with calcium-chloride it should be noted that practically every advantage of the former, except perhaps a slightly greater workability of the plastic concrete, may be obtained by the use of calcium-chloride, and where an accelerator, or curing agent, is desired the latter is usually the cheaper material.

COMMON SALT, Sodium-chloride, often used in the past as insurance against injury from frost-action, has only a slight value in lowering the freezing point of water, and when used in any appreciable quantity results in a material reduction in the strength of the concrete. Its use is not recommended and should be particularly avoided in reinforced-concrete work, as it is likely to cause the reinforcement to rust. It is also objectionable from the standpoint of efflo

rescence.

HYDRATED-LIME is generally considered to increase the workability of concrete and mortar. Small percentages added to the concrete mixtures suitable for building construction do not appreciably affect the strength, in fact the leaner mixtures are benefited. Richer mixtures, however, such as 1:12:3, suffer a decided decrease in strength when tested at an age of 28 days. In the Author's opinion, not over 8% of hydrated lime, by weight of cement, should be used in structural concrete. In contradistinction with the chloride compounds, specimens containing hydrated-lime show a proportionately greater reduction in strength under dry curing conditions than when kept in damp storage; in fact hydrated-lime does not prevent loss of water from the concrete by evaporation. It is interesting to note that the voids in concrete are affected only slightly by the addition of lime, apparently reflecting the effect of the admixture upon the amount of water required for a given workability. Hydrated-lime and similar powdered admixtures are added to the cement at the time of mixing.

KAOLIN AND CELITE also increase the workability of concrete. For a nominal 1:2:4 mixture the maximum percentages which can be used without endangering the strength of the concrete are 8% for the former and 4% for the latter, each by weight of the cement. As in the case of hydrated-lime, they show a proportionately greater reduction of strength under dry curing conditions, and apparently their chief value is to reduce somewhat the amount of water needed to give the required workability.

In general it may be said that substances such as CALCIUM-CHLORIDE AND CAL have a distinct function as accelerators in shortening the curing period of concrete and giving increased strength at the early stages. They are, furthermore, of use as curing agents and add very appreciably to the strength of concrete for which it is impossible to supply the moisture necessary for normal hydration. HYDRATED-LIME, KAOLIN, AND CELITE represent a class of powdered admixtures which are probably of some value in increasing the workability of concrete of the leaner mixtures, and their judicious use has no ill effect upon the

concrete, and in fact may be of distinct help in producing a more uniform mixture. It is believed, however, that, all things considered, an increase in the amount of cement is more desirable for reinforced-concrete work of a structural nature than the use of admixtures of this type. *

3. PROPORTIONING THE INGREDIENTS OF CONCRETE

7. Methods of Proportioning Concrete. The problem of proportioning concrete-mixtures is to determine the amounts of cement and water, together with the amounts and grading of the aggregates, TO PRODUCE THE DESIRED STRENGTH MOST ECONOMICALLY. During the last decade numerous theories have been advocated and various systems proposed to effect this result. The problem, however, involves many variables, and none of the systems offered up to the present time are entirely satisfactory when subject to the necessary limitations of field-practice. The following paragraphs give a brief summary of the more important of the different methods proposed.

(a) Proportioning by Arbitrary Volumes. This method consists of mixing the dry materials in certain proportions, determined by assuming an average void-content in each of the aggregates, usually without any reference to the water content. The relative gradation of the aggregates is ordinarily covered, in a general way, by inadequate and occasionally pernicious stipulations. This is usually the limit of the specification as affecting the question of proportioning the ingredients, and when the ratio of these volumes is expressed in the formula 12:4, it is considered ample justification for basing the design upon concrete to test 2 000 lb per sq in at an age of 28 days. That such a practice is unscientific might be considered a rather academic criticism, as the concrete in most cases undoubtedly fulfills its mission, owing occasionally to over design, and more often to the factor of safety: BUT IT IS NOT ECONOMICAL. CEMENT CAN BE

SAVED BY PROPER CHOICE AND GRADING OF THE AGGREGATES AND A WORKABLE

MIXTURE OBTAINED WITHOUT FLOATING THE CONCRETE INTO PLACE. EQUAL STRENGTH MAY BE HAD FOR LESS MONEY, OR GREATER STRENGTH FOR THE SAME MONEY, BY APPLYING EVEN THE CRUDEST METHODS OF SCIENTIFIC PROPORTIONMENT.

(b) Proportioning by Voids. By this method the percentage of voids in the two aggregates is either assumed or approximately determined by immersion. The volume of the mortar is then computed as somewhat more than equal to the voids of the coarse aggregate. Similarly, the cement quantity is determined so as to fill the voids of the sand, and have a little to spare. This method is also inaccurate and consequently wasteful. The attempt to fill the voids of the dry aggregates was of course with the intention of producing a mixture of the maximum density, but took no cognizance of the bulking effect of the sand, the space occupied by the water, or the changed void-content after the materials were combined.

* Compare the Economic Value of Admixtures by J. C. Pearson and Frank A. Hitchcock, Proceedings Amer. Conc. Institute, 1924, page 312, with Bulletin No. 8, Structural Materials Research Laboratory, Lewis Institute.

(c) Proportioning by Mechanical Analysis.-Fuller's Theory. This method is based upon the assumption that an aggregate so graded as to have a maximum density shows maximum strength with any given quantity of cement. A sieve analysis is made of each aggregate and the points of a curve determined by plotting the sieve-openings as abscissae, and the percentages passing the sieves as the corresponding ordinates. Having drawn a curve representing maximum density, the screened aggregate is then recombined with a grading to approximate, as closely as possible, the IDEAL CURVE. This method, although somewhat widely used in laboratory work, usually results in a rather harsh mixture which is not easy to handle and place. Furthermore, it assumes that density of the dry mixture is the only criterion of strength, which hypothesis can hardly be accepted in the light of recent tests, which show approximately the same strength in mixtures of widely different gradings. It is also true that the strength of a concrete is rather a function of the sum of the air and water-voids in the mixture, than of the air-voids alone in the dry materials.

(d) Proportioning by Surface Area Method-Edwards' Theory. This theory assumes that the strength of the concrete depends upon the ratio of cement by weight to the nominal surface-area of the aggregate. Although this method of proportionment, and the resulting discussions, have brought to light many interesting relations between the various factors involved, the system in its practical application has been largely disproved as the surface areas of different aggregates often show wide variation without there being any appreciable difference in the strength of the concrete.

(e) Proportioning by Voids-Cement Strength Relation.* This method of proportionment is based upon several fundamental conceptions, among which are the following:

(1) The percentage of air and water-voids in the concrete, under certain conditions, is an index of the concrete-strength. That is, for any given aggregate combined in fixed proportion with a particular cement, that gradation which results in the lowest percentage of voids (air and water-voids) in the concrete, other conditions remaining similar, produces the strongest mixture.

(2) If a given aggregate is combined with different quantities of cement in such a way that the percentage of voids is the same in all mixtures, other conditions remaining similar, the amount of cement is an index of the strength.

(3) The strength of a concrete mixture in which both the gradation of aggregates and the cement-content are varied, other conditions remaining the same, is a function of the ratio of the amount of cement used in a unit of volume of concrete to the voids (air and water) in this volume.

Considering that the voids in concrete are made up of the air and water-voids in the mortar, these latter are investigated by plotting CHARACTERISTIC MORTARVOIDS CURVES for a given fine aggregate combined with different quantities of cement. Such curves show the relation at various water-contents, between the mixture, expressed as a ratio of the fine aggregate to the cement, and the voids in the mortar. These curves are used as a basis for the computation of the *Bulletin No. 137 of the University of Illinois,