21. The method of calculating the strength of timber under the different kinds of strains to which it may be subjected will be found fully explained in the "Architects' and Builders' Pocket Book," and will not be considered here. There are, however, certain variable conditions in timber which it is impossible to recognize in a formula, but which often need to be taken into account when it is desired to utilize the maximum strength of the wood. These conditions are not generally explained in the handbooks, and may appropriately be considered in a work of this character.

Effects of Moisture. - In making the "United States Timber Tests" on long-leaf Southern pine an actual moisture determination was made for every test of strength. Most of the tests were also made either in pairs or sets of three, the wood for each set being taken from the same tree and being as nearly identical as possible, with the exception that one piece was tested while green and another was seasoned before testing. These tests were so numerous and thorough that Professor Johnson considers that it may be stated as probably a universal fact "that all kinds of timber are about twice as strong when thoroughly seasoned as they are in a green state."

"It has also been shown that the maximum strength corresponds to about 5 or 6 per cent, moisture, the strength at absolute dryness being somwhat reduced."

The moisture in fairly well-seasoned timber is about 15 per cent.

## 22. Weight

For any given species of timber, and for any given degree of dryness, the strength is almost directly proportionate to the weight. This is considered to be due to the fact that the weight (for the same percentage of moisture) indicates the density, and consequently the proportionate number of fibres - that wood which has the most fibres would naturally be supposed to have the greatest strength.

The strongest portion of a young tree is the heart. The strongest portion of a very old tree is about midway between the heart and sap. The upper portion of a tree is also slightly weaker than the lower portion.

Knots. - The weakening effect of a knot is about as much in compression as in tension. Large knots should be regarded, therefore, as sufficient cause for the rejection of timber columns as much as for the rejection of timber beams.

"The best single test of timber is the test of a short column or the crushing endwise test."

"There is no evidence that timber loses its strength from age or use alone," * although it is well known that little more than one-half the breaking load of a beam, if left on the beam continuously, will ultimately break it.

## 23. Selection Of Timber For Special Purposes

In selecting wood or timber for a special use that species should be chosen which appears to meet most fully the particular requirements of the case. Thus, for framing timbers, woods that are abundant and consequently cheap, and which can be obtained in large dimensions, are generally chosen, although in some instances extra strength and durability are more important considerations.

For outside finishing, ease of working and freedom from warping and checking are the principal requirements.

For wood that is to be buried in the ground in whole or in part, or to be used for piling, durability should be the chief consideration, although the question of cost must also often be considered.

For floors the wearing quality is the main consideration, while for fine interior finishing and cabinet work the color and grain of the wood generally control the choice.

The following list indicates those woods which are usually considered as best adapted to the particular requirements met with in building construction and finishing:

* Professor J. B. Johnson, in Proceedings of the Twenty-third Annual Convention, A. I. A.

For light framing, for dwellings, tenement houses, etc.: Spruce, white pine, Northern yellow pine and hemlock give good satisfaction and are generally used on account of their lightness and cheapness.

For posts, girders, truss timbers and heavy framing: Georgia pine, Oregon pine or white oak are to be preferred. Next to these are the short-leaved Southern pine, Canadian red pine, or Norway pine, as it is often called, and the best qualities of spruce.

For very long truss timbers and for flagstaffs, etc.: Oregon pine and Georgia pine* are about the only available woods; good-sized timbers of these woods can be obtained up to 60 feet in length.

For outside finishing: White pine, redwood or cypress should be used.

For shingles: Redwood, cypress, cedar and white pine, in the order named.

For siding and clapboards: Redwood, cypress and white pine are the best w6ods; spruce and Oregon pine are also used, but are not as satisfactory as the other woods.

For posts and sleepers set in the ground: White cedar, chestnut, redwood, cypress and black locust.

For piles and cribbage: Oak, elm, Southern hard pine, Oregon pine, Norway pine, cypress, spruce, white pine and hemlock, in the order given; only the first three should be used in salt water.

For sash, solid doors, as a base for veneers and for all joiners' work that is to be painted: Clear white pine gives the best satisfaction, although poplar (whitewood) is often used on account of economy; cypress is also well adapted for sash and solid doors.

For thresholds and floors, or wherever hardness and resistance to wearing is required: White oak, maple, Georgia pine, all quarter-sawed.

For linen chests and closets: Florida or Alabama red cedar.

For interior finish: Redwood, cypress and white pine and any of the hard woods are suitable. They are generally selected to please the especial taste of the owner, and all are sufficiently durable. Every hard wood needs to be thoroughly seasoned and kiln-dried, and all hard wood doors or sash should have a core of pine, covered with a 3/16-inch veneer of hard wood.