[Illustration: 20. Test No. 8, on 1906 Fence. Surface Checking Only] [Illustration: 21. Combination Cracking and Checking on Section 69, on 1908 Fence] [Illustration: 22. Cracks in Paint Coating, Caused by Cracking of Hard Pine Wood] [Illustration: 23. Section 65 on 1908 Fence. Showing Early Breakdown of Corroded White Lead] CHAPTER XIII TENNESSEE PAINT TESTS =Location and Object of Tests.= On September 15, 1910, the erection of a wooden test fence was completed on the State Fair Grounds at Nashville, Tenn. Upon this fence were exposed forty-two samples of white paint, the object of the test being to determine whether the combination type of formula is superior to the single pigment type in the southern plateau, of which Nashville is the centre. =Construction of Tests.= The construction and outline of these tests differ somewhat from those conducted at Atlantic City and elsewhere by the Scientific Section. The fence frame is 150 feet long, being made of 6-inch bevelled girders supported three feet from the ground by 4-inch posts set six feet apart. Upon this girder were placed a series of forty-two test panels supported at top and bottom with weather strips and braces. The test panels used were 40 inches high, 30 inches wide, and one inch thick, being made of the highest grade white pine, tongued and grooved together, and protected on the edges by weather strips projecting from the surface of the panels. Each panel was painted on both sides with the same paint, thus giving an eastern and western exposure, the fence running north and south. The formulas used in the test vary in their percentage composition, being made up in some cases of single pigments, and again with combinations of the opaque white pigments, with and without certain percentages of the crystalline or inert pigments. The paints were applied under the supervision of prominent master painters and a committee representing the Scientific Section and other technical organizations. Other field tests have shown that the sap and knots in hard-grained woods, such as yellow pine, cypress, etc., have been the cause of the failure of even the best paints, and that all tests should be conducted upon soft woods, such as white pine and poplar, if definite results are to be obtained. Paints tinted with ochre, chrome yellow, lampblack, iron oxide, etc., have shown on the other field tests which have been conducted at Atlantic City, Pittsburg, and Fargo the value of these pigments in giving to the paints increased wearing properties. On the Southern Test Fence, therefore, all the formulas were ground in white only and placed upon white pine so as to make the test primarily one to determine the value of the various white pigments upon good wood. [Illustration: Tennessee Test Fences]