Mr. Sabin.

Mr. Aiken.

Mr. Greenman.

Mr. Larned.

consistency he is using when manipulating dryer mixtures, yet so far as the report goes, it refers to the wet mixtures which are apt to be used in making other tests.

MR. W. A. AIKEN.-As far as the Vicat apparatus is concerned, I think it is a good thing for inexperienced testers to use in testing a great many unknown brands of cement at one time. I do not know of anything better; but where there is a set of men handling a single brand of cement right along, they learn the peculiarities of that material, and in such a case I certainly would not use it, because it is too slow. As far as the percentage of water is concerned, I think it may be used to determine the normal consistency satisfactorily. What I call normal consistency is rather wetter, perhaps, than most people would usc.

I think, however, that we should not knock a thing down, and say it is worthless, unless we have something better to put in its place.

MR. GREENMAN.-It is not my intention to knock down something that has been put up, except in regard to this one particular point. I should like to state a particular case. We had a cement come in during the last two or three weeks which required, to get a penetration of 5 m.m., 3 per cent. more water than usual. Now, the water actually flowed out of the bottom of the mold and ran all over the glass. That cement was altogether too wet, yet you could not get a penetration of more than 5 m.m. That's not right. Now, in such cases as that, should the experience of the operator determine what per cent. of water is to be used, or should we say positively that the Vicat apparatus shall be used, no matter how wet the cement might be?

MR. E. S. LARNED. In referring to tests for normal consistency, I think the Committee in the recommendations made have plainly recognized the difficulties and variables connected with this test, and it does not appear to me to be reasonable to expect absolutely uniform results. The age of the cement, its degree of fineness, moisture contained and the operator himself, all have important influences upon the results. If the cement powder previous to this test was uniformly dried at the required temperature, using the cement that passed the 200 mesh sieve only, I believe results could be maintained within a narrower range of variation. The method and time of troweling, however, is a most important factor in this test, as it is in the tests for tension.

I believe the experienced and skilled operator can determine Mr. Larned. normal consistency without the assistance of the Vicat needle, but unfortunately we have too few operators of this character testing

cement.

MR. HUMPHREY.-Mr. Greenman states that one of the Mr. Humphrey. samples of Portland cement that he tested generally required 24 per cent. of water for the normal consistency, and that with the Vicat apparatus it required 28 per cent. in order to get the proper penetration. It seems to me that this was a very abnormal cement and that 28 per cent. for a Portland cement would ordinarily make a very sloppy mixture. A normal Portland cement should not take more than 22 per cent. in order to secure a penetration of 10 m.m. with the Vicat needle apparatus, and it would seem to me that if a cement required from 24 to 28 per cent., an investigation should at once be made to determine whether it was not adulterated or in other respects abnormal.

When the Committee on Standard Specifications for Cement began the formulation of their rules, their first action was to adopt the Methods which had been reported by the Committee on Uniform methods of Test of Cement, and in order that they might test the value of these methods they sent five brands of Portland and four brands of natural cement to some thirty testing laboratories covering the most prominent in this country, but of varying degrees of experience. These samples were accompanied by the report of the Committee on Uniform Methods of Test of Cement, together with instructions as to the tests to be made.

A study of the results shows in general that while there was considerable variation in the results of the tests for tensile strength, time of setting and some of the other tests, the results as to the proper percentage of water required for each cement were concordant, the variation being hardly more than one-half per cent. These laboratories were scattered all over the country, and in most cases the operators had had no previous experience with the Vicat apparatus.

MR. GREENMAN.-I wish to state that the views I expressed Mr. Greenman. are not my own views only, but represent the views of certain institutions of learning in our immediate vicinity, and also the views of certain cement manufacturers.

ECONOMICAL MOLD FOR FORMING COMPRESSION

TEST PIECES FOR CONCRETE.

BY CLIFFORD RICHARDSON AND C. N. FORREST.

In the course of the initiation of an elaborate study of the effect of variation in the proportions and character of the materials entering into the composition of Portland cement concretes upon their strength, it was found that the purchase of a sufficient number of molds for the formation of cubes for the compression tests would involve a very large expense. One firm quoted $11, another $7 each for 12 six-inch cube molds with bottom plates. As these prices were practically prohibitive, recourse was made to the ingenuity of our machinist to supply the want and he has designed a mold which is not only cheap but more satisfactory than a mold for cubes for several reasons. The mold may be described as follows:

Tank steel in. in thickness is cut in strips 6 in. wide and 18.85 in. long. It is then rolled into a circular form until the two edges are brought together within in., as thown in illustration No. I. An iron ring is then cast 1 in. wide, in. thick and 6 in. in inside diameter. This is provided with two screws passing through it and of sufficient length, when the ring is slipped over the tank steel cylinder, to compress it at such points as to bring together and close the opening which exists owing to the spring in the steel, in this way forming a closed mold which is ready for use. In order that the collar may always be applied longitudinally and the pressure produced evenly a cylinder of slightly larger diameter, as shown in illustration No II, is provided which shall hold the collar during compression parallel with the base. It has been suggested that two collars might serve more satisfactorily than one. The cost of such molds need not exceed $16 to $20 per dozen, ten made for us costing $16.

The advantage of the mold is that the test pieces can be much more uniformly compressed than in the case of cubes where it is difficult to bring about thorough compression at the edges. That

this is the case is shown by the fracture of the cylinders. There is no wedging out of the sides of the test pieces and the fracture is a perpendicular one in contradistinction to the hour glass form which generally appears with cubes.

Mr. W. Purves Taylor has made some comparisons of the strength of cylinders as compared to cubes, on the same mortar, which shows that neat cement at 28 days gave an ultimate strength per square inch of 5,026 pounds for the former and 4,920 pounds for the latter, or a ratio of 1.022 in favor of the cylinder. In the case of a test piece made with a mortar consisting of 1 of cement

to 3 of sand the ratio was 1.133 in favor of the cylinder. On the addition of stone in various proportions the results obtained with the cylinder were not always as high as with the cube, but the average of all tests showed a ratio of .976 for the cylinder as compared with 1.0 for the cube. Mr. Taylor notes in many of these tests that the mortar was crumbly or soft. As the tests with neat cement and sand mortar without stone show a superiority in the strength of cylindrical test pieces the conclusion may be drawn that the cylindrical form is a satisfactory one. They also possess the great advantage that the test pieces can be used in a smaller machine than is required to break cubes six inches square.

LOW-PULLING EARLY-STAGE PORTLAND CEMENT US. THE ORDINARY EARLY-STRENGTH

DEVELOPING PRODUCT.

By W. A. AIKEN.

The detailed strength requirements of the Standard Specification for Cement adopted by this society emphasized its recognition of the growing demand for a product of lower strength at early stage than has generally in the past been specified. The revolution in the engineering profession's views on this subject is in the author's opinion largely due to the general knowledge that the Board of Rapid Transit Railroad Commissioners of the City of New York was requiring such cement, which from the records of its Department of Inspection of Material continually showed its greater value from long time tests: the positive worth of such records being in the multitude of tests on such a large amount of material.

The following tables show average strengths at various periods up to 3 years of several classes of cement, involving over 1,250,000 barrels, inspected for the New York Subway construction. The classification fully and clearly impresses the value of our specification cement by showing test results on: (1) All accepted cement, this being arbitrarily divided into two classes-the first showing a neat strength less than 700 lbs. at 7 days and the second showing a neat strength more than 700 lbs. at 7 days; (2) Such cement as has been "turned over" to the manufacturer simply because the tensile strength was so high at 7 days as to render unlikely a satisfactory approximation to our specification gain requirement; (3) Such material as was absolutely "rejected" at 28-day period, solely for non-compliance with the gain requirement, which for the benefit of those not conversant therewith, is now stated to be 15 per cent. in neat briquettes and 25 per cent. in 2:1 sand mixtures, between 7 and 28-day periods. (4) Lastly for comparison, strikingly illustrating the age value of low early strength material, I give in Table I similar period averages for "accepted" natural cement used in certain parts of our work.