85% of inside unbalanced area; and its weight per square inch of inside unbalanced area = 1.31 lbs.; its weight per square inch of outside unbalanced area = .71 lbs.

The above valve was of brass, and one of iron with wooden seats was substituted. Its seat area was 94% of inside unbalanced area, and its weight per square inch of inside unbalanced area = 1.16 lbs. ; its weight per square inch of outside unbalanced area = .60 lbs. A 25-inch valve had a seat area of 75% of inside unbalanced area, and its weight per square inch of inside unbalanced area lbs.; its weight per square inch of outside unbalanced area

1.41 .80 lbs.

An 8-inch valve had a seat area of 75% of inside unbalanced area, and its weight per square inch of inside unbalanced area = .96 lbs. ; its weight per square inch of outside unbalanced area = .55 lbs.

Water pressure 65 lbs.

All these valves worked poorly and noisily. To improve their working, spiral springs were added to their already overburdened weight, and then follows the story of the large quantities of air pumped into the pump chambers in order to make the valves work well.

Page 189.-" To keep these valves from lifting too high, and to make them seat quickly, spiral springs are used. They work all right as long as they are new; but as soon as they lose their elasticity, or break, the valves cause an unnecessary thumping noise in the pumps, and a trembling of the floor."

Page 190.-" A few weeks ago one of these large center valves broke the spiral springs and stripped the nut off the bolt which holds the valves." Speaking of the 25-inch valve, the report says: "Thus it has .845 lbs. of weight per square inch of floating area (what I have termed the outside unbalanced area) to bring it down to its seat. To accelerate this the four spiral springs are required."

Of the 8-inch valve it says: "They have only .56 lbs. per square inch to seat them. These valves work noisily, as soon as the spiral springs are weakened or broken, which is immediately noticed in the working of the pump."

Page 219." One of the new engines stripped the nut from the bolt which holds the large double-beat suction valve in place.'

It is difficult to distinguish between the noises in a pump caused by the closing of one valve, and the opening of the other. I have no reason for the noisy action of these valves other than has been pointed out in this paper, namely the excessive seat area. Spiral

springs served to soften the blow upon the nut, or stop, in the rising of the valve, and when they were broken by the violent concussion in opening, the nut was also soon broken.

I do not think the springs were needed to help close the valvesthey were already excessively heavy. It is very doubtful in my mind whether these broad seats act at all in the nature of a "floating" surface. The real surface, or area, against which the water impinges to lift the valve, is on the inside, or between the seats, and to that area the weight of the valve should be referred. The outflowing current moves parallel with the seats, and hence has no vertical component acting to lift the valve.

We have already seen that the pressure due to the velocity of outflow is reduced to nearly one-third of its force in its vertical effect to lift the valve, by the oblique direction of the water passing through the valve; hence I do not think the seat surface, which is parallel to the flow, should be added to the lifting area.

The report proceeds, after speaking of the constant difficulties with these valves, to speak of the great improvement effected by substituting for them numerous small (4") rubber valves. These valves are substantially the same as the Reynolds pattern, illus trated in Fig. 128.

CCCXXXVI.

EXPANSION OF TIMBER DUE TO THE ABSORPTION OF WATER,

BY DE VOLSON WOOD, HOBOKEN, N. J.

(Member of the Society.)

HAVING had occasion to use the facts in regard to the expansion of timber in the direction of its fibers, and failing to find, after a limited search, definite figures, I prepared a specification for the Department of Tests in Stevens Institute, according to which the results in the following table were reported.

The specification directed that two specimens each of pine, oak, and chestnut, each about three feet long and six inches wide, should

be selected, of good quality, fairly straight grained, free from knots and fairly seasoned. The specimens were then to be kept in a dry warm room for about three weeks. Eight brass pins were to be inserted, one near each of the four corners, two near the middle of the length and near the edges, and two near the middle of the width and near the ends. The arrangement of the pins and general dimensions of the specimens are shown in the annexed figure (Fig. 129), drawn of full size.

A very small mark was made on each pin with a center punch to facilitate accuracy of measurements. These were on one side only of the specimen, and the readings were to the nearest halfhundredth of an inch.

The report states that the specimens were, according to the specification, kept in the office for three weeks, when the initial measurements were made, after which they were immersed in water

June 7, 1888, and removed from the water July 14, having been in the water thirty-seven days, and the final measurements made on the latter date.

It will be seen that chestnut expanded more than the pine or oak both transversely and in the direction of the fiber; that the mean elongation of the chestnut was 2.6 times that of the pine, and the lateral expansion was 1.4 times as great as that of the pine.

DISCUSSION.

Mr. F. H. Ball.-I happen to have had a little experience, which I will explain; it may be of interest. In our pattern shop we had occasion to make an engine frame. The side of the engine frame was made by gluing up pine, piling the pieces on top of one another. The two sides were also made in that way, then there was a flat piece put across. That part had to be

and

worked out into shape, and so far as we observed the grain of the wood was straight, and very much like the other wood. They were both pine. First let me say, the round end of the pattern was put on for convenience with a dove-tail, so that it lifted off from its end of the frame. We noticed very soon after we began to use the pattern that this last piece appeared to be shorter than the others. We did not suppose that between two different kinds of pine there would be any material difference in the shrinkage lengthwise, if there was any shrinkage in the length; but we noticed that that piece was shorter, and kept getting shorter. I won't say exactly what the difference is; but my impression, just from looking at it, was that this piece was an eighth of an inch shorter than the other, or else the other parts were an eighth of an inch longer. So finally we had to take this piece off and put on another. We did not notice any appreciable shrinkage in other directions. This piece I do not think was over three feet long, and it was screwed to the rest of the frame. The rest beyond it was glued on. There was no opening of the joint, but the joint was perfect. We thought of course this piece shrunk. Our pattern-makers were slow to believe that pine would shrink nearly one-eighth of an inch in 36 inches.

Prof. Denton.-I would like to ask Prof. Wood what his experiments show-whether it is possible for a piece to have shrunk three-eighths of an inch by the experiments.

Prof. Wood.-By no means. The pieces that we experimented with were longer, almost double the length of the others, and the measured elongation was two one-hundredths of an inch.

Prof. Denton.-That is the direction of the grain.

Prof. Wood. The direction of the grain.

Mr. Ball.-I do not want to be put on record as saying that our shrinkage was an eighth of an inch, but it was a very appreciable amount. Our pattern-makers called it one-eighth of an inch without measuring it. I know it was a good deal more than onesixteenth of an inch.

Mr. H. R. Towne.-It is pertinent to remember in this connection that strips of wood are commonly used to form a hygrometer, and I believe afford the best method of constructing an instrument of that kind for measuring the humidity of the atmosphere. They are usually made, I believe, of two strips of wood of different kinds glued together, the action of it depending on the absorption of moisture by the two pieces.