It is the policy to present only the results of the tests, eliminating all statements of opinion or theories concerning them. Occasionally in bulletins on a particular phase of a subject curves will be drawn. showing graphically certain relations, such as the variation of strength with age or the variation of deformation with increasing stress. Attention will also be called to points of interest or some unusual feature in the results.

CONSTITUENT-MATERIALS SECTION.

OUTLINE OF INVESTIGATIONS.

Nature of the work. All the sand, stone, gravel, and other materials which are used in the investigations being carried on by the division are tested for their quality in this section.

In addition a series of investigations are under way covering the determination of the relative value of the various sands, gravels, cinders, crushed stone, and other constituent materials used in cement mortars and concretes.

For these investigations there is used a typical Portland cement, obtained by thoroughly intermixing a large number of standard brands of Portland cement. In the earlier stages of these investigations, it was the practice to test each brand separately, and also each mixture of these brands. It was found that the difference between the result obtained for the mixture and the average of the results of the individual brands was so slight as to be negligible for all practical purposes. It is the practice now, however, to test the various brands. in order to ascertain whether the cement meets the requirements for a typical cement contained in the programme for these investigations (p. 17). The cement obtained by mixing several brands, here referred to as a typical cement, is mixed in large quantities at a time and sealed in air-tight galvanized cans of about 800 pounds capacity.

Laboratories.-Interior views of the constituent-materials laboratories are shown in Pl. II. A briquet is shown in the clips of the long-lever testing machine at the left in A, and the soapstone immersion tanks are shown in the background. The glass-top mixing tables are shown in the foreground of B. This view also shows the 200,000-pound Olsen testing machine at the left, the long-lever cement-testing machine and the hydraulic hand-operated compression-testing machine in the center, and the moist closet at the right. Register numbers.-Each sample of material as it is received at the laboratories is given a register number, a record of which is made and filed in a card index. The sample is subsequently known by this register number, which is so chosen that it indicates the nature of the material. For example, each material is designated

15767-Bull. 329-08-3

by the first and last letters of its name; thus, Ct. for cement, Sd. for sand, Gl. for gravel, Cr. for cinder, Sg. for slag, and Se. for stone. The first sample of the cement was called Ct. 1, the next Ct. 2, etc. When the mix is made it is given a single number and each sample taken from that mix is given a subsequent number; thus the second mix was called Ct. 133, and the first sample from that mix Ct. 133-1, etc.

PHYSICAL TESTS OF CEMENT.

All the cement used at the laboratories is subjected to the following physical tests: (1) Specific gravity. (2) Fineness. (3) Time of setting. (4) Tensile, compressive, and transverse strength, neat and with three parts Ottawa sand. (5) Percentage of water for normal consistency. (6) Soundness.

The methods recommended by the special committee on uniform tests of cement of the American Society of Civil Engineers are used in making these tests.

The specific gravity is determined upon untreated, dried, and ignited samples. A large series of special tests have been made to determine the relative merits of the different forms of apparatus; the Le Chatelier apparatus, however, is the standard for routine work.

The time of both initial and hard set is determined both by means of the Gilmore wires and by the Vicat needle apparatus. The latter apparatus, however, is the standard for all tests. The temperatures of the air and of the water are recorded on Draper automatic recording thermometers, which record covers a period of one week.

The accelerated testing apparatus for making tests for soundness is illustrated in Pl. III, B. The test consists in maintaining the pats in an atmosphere of steam over boiling water for five hours. The percentage of water for normal consistency is determined with the Vicat needle apparatus. The test for fineness is made by sifting the cement through the No. 100 and No. 200 sieves, using a special sifting apparatus made by Howard & Morse, which is capable of holding nine sieves at one time.

STRENGTH TESTS OF NEAT CEMENT AND OTTAWA SAND

MORTAR.

The neat cement is tested for tensile, compressive, and transverse strength at ages of 1, 7, 28, 90, 180, and 360 days, and a parallel series. of tests is ma 'e of 1:3 mortar, using standard Ottawa sand for all ages except one day. Three test pieces are broken for each age.

Tension briquets are approximately 1 inch square at the smallest section; in order to get accurate results the least dimensions are carefully measured at the time of testing and the actual unit stress

is computed. The briquet molds are of the 5- and 3-gang type and are made of brass (Pl. II, A).

Compression test pieces are 2-inch cubes, the molds being of the 3-gang type. The mortar cubes are tested in the 40,000-pound hydraulic hand-operated compression machine shown in Pl. III, A, in which a is a ball-and-socket bearing and b and c are handles that give fast and slow motions, respectively. A cube is shown in the machine just above a. The stronger neat cubes are tested in the 200,000-pound testing machine.

Transverse test pieces are 1 inch square in cross section and 13 inches in length; they are tested on a span of 12 inches. The transverse molds are of the individual type. The transverse specimens are usually tested on the machine shown at the left of the center in Pl. II, A, but can be tested in the briquet machine by means of the apparatus illustrated in Pl. III, C.

The test pieces are stored in the moist closet for twenty-four hours, and are then placed under water, in the immersion tanks, until tested. The water in the immersion tanks is kept constantly running and heated when necessary in order to maintain a constant temperature of about 70° F.

The order of testing is regulated by a filing system with cards 6 by 4 inches in size which have blanks for data, as shown in Form A.

When a sample of neat cement, for example, is molded into test pieces, a card is made out for each group of three, and the date upon which the group will be tested is placed at the top of the card. The cards are then filed in the order of the dates. Calendar guide cards are used, dividing the months into three parts. Each morning the cards bearing the date of that day are removed from the front of the file, and the test pieces indicated on them are tested that day. In this way the cards are kept moving toward the front of the file and reach the front on the day when the test pieces whose register numbers they bear are to be tested.

The results of these physical determinations are entered upon Form B for filing at the laboratory, and are subsequently copied into tables for publication.

a Forms B to R, inclusive, are condensed in the pages of this bulletin, in connection with the subject-matter to which they severally relate, for the purpose of showing in detail how the results of the various tests are recorded. In practice, each form is printed on a sheet of paper 8 by 10 inches in size.