"For mortar face work, samples should be taken from each barrel—for general work, on the basis of one sample to ten barrels; all bags sampled should be properly marked, so that they may be resampled if necessary. Cement should not be stored in a damp place. Hard lumps should be thrown out and replaced by good cement.
"Sand. Sand should not contain more than 10 per cent of silt, except where actual tests indicate that the sand proposed is as good as the average sand generally used in the locality where the work is being executed. In this connection I have to say that I recall a case of my own work where sand containing 16 per cent of silt gave a very much better result than a sand containing no silt and which had been originally intended for filtra tion purposes. On waterproofing work, as a rule, a large percentage of silt is desirable up to, say, about 15 per cent, but facing work should not, as a rule, contain over 5 per cent, as the greater amount of silt in the concrete, the greater amount of slime there will be carried to the face work.
"In placing concrete upon a sand foundation con taining a sand without any silt in it, I noticed that the concrete was being robbed of its water. We were putting in soupy concrete. Fifteen minutes after it was placed, it was almost dry. Besides this, I noticed that where a layer was allowed to stand a few minutes without any abutting concrete, the water of the concrete had washed a part of the sand out of the concrete, and also a part of the cement. There was a ring of sand about 5 in. from the end of the layer, and the cement was extended some what beyond that. The sand used was a fine sand in this case. As a result of this observation, I accepted a coarse sand containing about 12 per cent of silt, which held its water satisfactorily.
"Sand Test. On large work I have recently used the following method of testing sand, which has been quite satisfactory: Wash a considerable quantity of sand; separate the silt from it; then make up a special sample from the sand and silt that you have separated, so that this latter sample will contain the maximum silt allowed under the specifications. (If the contractor is furnishing sands from more than one locality, it will be necessary for you to make the test herein described from samples of each of the localities, as will readily be seen from the nature of the test.) Take the specially prepared sample just described, and thoroughly mix the silt with the sand until the distribution is uniform. Then take
a small portion of the proposed sample, about three thimblefuls (on my own work I have given the inspector an ordinary test tube, with a mark on the test tube indicating the height of the final sample of sand re quired). Take this small sample of the prepared sand, drop it into a graduated beaker of clear water about 30 in. high (ou my own work I have used the local hydrant water, which is filtered and reasonably clear at all times of the year) ; shake well until the sample has been thoroughly washed. Then take a domes tic pin on the end of a wire or small stick, and insert it in the top of the beaker, lowering it into the beaker until the pin is no longer visible from the top, and then read the position of the pin on the graduation at the side of the beaker. Make half a dozen similar tests, using the prepared sand, and the average of these samples will be the correct gauge for sand which does not contain over the stipulated amount of silt.
"Of course, it is not necessary to sample every car load or barge of sand that comes in, as the bulk of it can be determined by superficial observation. If you are in doubt, however, take a number of shovelfuls of sand from various parts of the ear or barge, mix thoroughly, and take out the necessary amount of sand for test. Drop same into the beaker, shake, lower the pin, and read the gauge. If the pin is then lower than the gauge, the sand has less silt than the maximum; if it is higher, it has more silt.
"This is a very cumbersome thing to explain, but it is more practical than it sounds. When I first suggested it to the men engaged on the work, it was thought to be too hair-splitting to be of any practical value; but after it was actually put into use, it was entirely satisfactory and enabled the men to make immediate tests of the sand, instead of waiting half a day, which would be necessary under the common method; and it permitted a great many more samples to be taken. If you use it, it will take about an hour to get your first sample pre pared and your gauge determined, after which two or Is three minutes will enable you to make the test. (This system I first saw used by Mr. Allen Hazen.) "Condemn sand for a noticeable quantity of mud balls, regardless of test. In this connection an inspector should use some judgment as to the nature of the mud ball. Sometimes these mud balls are almost solid chunks of mud, and at other times they contain nearly all fine quartz, with a very little mud.