It was decided to test each formula in three colors, in duplicate, and on each grade of wood, exposing the duplicates on either side of the fence. Thus for one paint formula there were required 18 panels, or 6 painted in each color and on 3 grades of wood.
Paints for Tests. The mixed paints received for the tests were in quart cans, having been especially prepared from the formulas submitted to manufacturers by the technical committee in charge of the work. They were properly labeled with their number and color, in each case. The formulas decided upon for the test are described later. The various white leads and other single pigment paints which were used were received in kegs weighing 121⁄2 pounds each, having been bought in the open market and then given a formula number. The formulas of the paints designed for both the Atlantic City and Pittsburg tests, as well as the numbers of the panels upon which the paints were applied, are shown on pages [131]-[133]-[145]. The analysis of one of the combination paints applied is herewith given, to show the correct method of stating the composition of a paint.
Formula No. 20, Atlantic City Test Fence
| Percentage Composition | |||
| Pigment | Vehicle | Total | |
| Corroded white lead | 67.01 | — | 42.84 |
| Zinc oxide | 19.89 | — | 12.71 |
| Asbestine | 3.86 | — | 2.47 |
| Calcium carbonate | 9.24 | — | 5.91 |
| Raw oil | — | 94.30 | 34.02 |
| Japan drier | — | 3.89 | 1.40 |
| Turpentine | — | 1.81 | 0.65 |
| 100.00 | 100.00 | 100.00 | |
Brushes. Heavy 7-O round bristle brushes were used for the priming coat so that the paint might be well worked into the wood, while for the second and third coats three-inch chisel edge brushes were used. These brushes were, of course, washed several times with turpentine after painting each panel, so that pigments from one paint could not be carried over into a paint containing other pigments.
Cypress Panels
Shellacking Panels. The shellacking of any bad places of minor nature which may have been present on the surfaces of some of the panels, was done with the highest grade orange shellac. It was thought advisable to determine whether shellacking over the priming coat of paint or on the bare wood previous to the application of the priming coat, was the better method. Panels Nos. 1 to 8 in each test were therefore shellacked over the priming coat, while on all other panels the shellacking was done directly on the bare wood previous to the application of the priming coat of paint.
Application of Paints. In order to determine just how much paint was applied to each panel and to reckon the spreading rate therefrom, careful weighings were made during the application of every paint. This was carried out by placing a quart can of paint as received, upon a laboratory balance, the gross weight being taken and recorded. The can was shaken and its contents transferred to a quart-size enameled cup where with the aid of a paddle it was broken up into a mixture of even consistency. A portion of this paint was then transferred to two small sample cans carefully numbered with the formula number, for future reference and analysis. The reduction of the paint was then made. The brush used on the priming coat was placed with the pot and the paint on the balance and the weight taken by the official weigher. The pot was then given to the painter who applied the priming coat to one panel. The brush, pot, and paint were then handed back to the official weigher and the difference in weight recorded. From these data could be reckoned the spreading rate of the formula applied.