meaiifs of tlie diyiiis box shown on Plate i. The temperature of the inflowing air in this drying 

 box wasjcept at about 100'^ 1<\, with suitable precautious against the checking of the wood, and 

 the air was exhausted by means of a fan. The air was, therefore, somewhat rarefied in the box. 

 The temperature was at all times under control. It operated when the fan was running, and this 

 was only daring working Inmrs. 



THE TESTING LABOUATOKY. 



The testing laboi'atory is the basement *tory of the gymnasium building of Washington Tni- 

 versity. Its dimensions are 71 by 10 feet, with one corner partitioned off, as shown on the tioor 

 plan, Plate i. The net area used for hib )ratory purposes is 2,500 square feet. All the apparatus 

 suspended from tlie ceiling, as sliafting, steam pipes, exhaust fan, etc., is shown in dotted lines. 



The apparatus pertinent to the timber tests consists of a 1,0()0,(KK) pound column-testing 

 machine; one IdO.OOO i)ound beam-testing machine, one l()0,(tO()-](oun(l universal testing machine, 

 of Richie's " Harvard" pattern ; (me small portable l)eam machine, one (Miorsc power Prayton coal- 

 oil engine, one 4- horse power steam engine, one planer and one lathe, for ironwork; one planer, 

 one band saw, and one cutting-ott'saw, for shaping and dressing wood specimens; suitably scales, 

 drying ovens, etc., for the moisture and si)ecific gravity tests; the drying box with its steam coils 

 and exhaust fan, and all the necessary appliances, benches, tools, desks, etc., including a Thatcher's 

 slide rule for makint;- the comi)utations. The timber is stored in various parts of the rocmi not 

 otherwise utilized. The broken specimens are stored in the museum building at the Missouri 

 Botanical Gardens. 



UESClllPTlON OE TESTS. 

 The o'iifHihrcal'nKj tests. 



Lar<ifi beams. — The large beams are tested on the large beam testing machine shown on Plate Ti. 

 The base of this machine consists of two long-leaf pine iitii:k» (I'mus palustri.s) inches by IS 

 inches by L' t feet long, with a steel plate three fourths of an inch by 18 inches by 20 feet long, all 

 bolted up as one beam. The power is ai»plied by hydraulic |)res.sure upon a jilungcr below, to flie 

 crosshead of whicli are attached the two side S('rews, on w hich the u])|ier crosshead is moved by 

 sleeve nnls and s[>nv gearing.* The beam to be tested rests on pivots at the ends, ])laced on t()|( 

 of the base beam, and the upper crossheail is moved down by meansof tli<^ gearini; until the central 

 pivot attached to it conies in contact with tlie beam, or rather with the distribution blocks placed 

 on the beam at this point. Theteet then begins, the power originating In a double jilungcr jiunip, 

 operated by hand or by steam jiowcr in another part of the room. 



To prevent the jiivots or "knife-edges" from crushing into the timber, it is necessary to make 

 the contact at liofh ends and center, first upon a cast-iron plate, then through longer wooden 

 blocks to the timber. The center block is curved somewhat on the lower side, to allow for a con- 

 sidei'able deflection in the beam when iiearing its maximum load. 



In the tests of all beams, both large and small, the load is put on at the same uniform rate, so 

 as to eliminate the time eftect, w hich is very great in timber tests. The losid on the small beams 

 is increased at such a rate as to iiroduce an increase in the deflection of one-eighth inch |)er 

 minute without any jiause until rupture occurs. This causes rupture in from ten to fifteen minutes 

 time. The load is read off when it ri^aches certain even amounts, and an observer notes the cor- 

 responding detleittion without stopping the test. The time required for the large beam tests is 

 about the same, the deflection rate being greater when the total dellection is expected to be 

 greater, as is the case with I by S inch sticks 12 feet long. The dellections of the large beams 

 are observed upon a jiolished metallic scale, graduated to inches and tenths, which is tacked to 

 one side of the stick at the center. A tine thread is stretched, by means of a ruliber band, over 

 nails driven into the side of the stick above the end supports on the line of the neutral axis. This 

 string or thread is moved about an inch away from the surface of the timber, and nil parallax, or 

 error of reading from an oblicjue position of the eye, is avoided by keeping the eye where the 

 thread and its image in the metallic mirror coincide and form one and the same line. The read 

 ings are taken to iiu'hes and liundre(ltlis by estimating the tenths of the graduation spaces on the 

 scale. 



