With its correct amount of bitumen of, say, 11.5% it would be quite mobile under the rollers during construction. Its resistance to shear (the distorting force of traffic) is low compared with modern mixtures.

On the other hand, Platte River sands, which are productive of low voids, may be combined in such manner as to produce in combination with 20% of limestone filler, voidage at least as low as 12.5% which in the mixture would mean only 5.4% bitumen to completely fill the voids. This mixture is far from plastic, in fact, it is so resistant to compression that under a load of 12,000 lbs. per square inch the highest density obtainable was only 93% of the theoretical maximum, whereas, with more plastic mixtures 100% maximum density was readily obtained with only one-half this load. The grading of the mineral aggregate in this mixture was:

The source of its resistance to compression was unquestionably due to the low percentage of material between the 200 and 10-mesh sieves. Such mixtures are of value from the scientific standpoint only, under present methods of laying, since they are too resistant to compression to be practical.

These two extremes show that the plasticity, or workability of the mixture is an important factor and must be considered in designing. Improved methods of laying will steadily decrease the degree of mobility required during construction. Traffic requirements of today are demanding increasingly resistant mixtures, and these requirements are being readily satisfied in every instance where special study is being given the problem.

In order to briefly describe the procedure of designing a mixture from a known premise rather than by the old trial and error method, Plate CT, is introduced. This shows in tabular and graphical form the study of local materials in combination, and enables the designer to quickly select the best combination of sand and filler and establishes without argument the maximum amount of bitumen permissible, by determining all the mixture can possibly hold at air temperature with the mixture compressed to maximum density.

as would produce original voids in the sand aggregate of 32.4 or torpedo sand (ordinary concrete sand), in such combination asphalt and other asphaltic types, in combination with coarse

Lake Michigan sand, ordinarily available in Chicago, for sheet script. We used in this particular instance a combination of (Turns to chart.) This is the curve. I will forget the manu

per cent. The Lake sand itself has voids usually between 35% and 37%, so that by the addition of a given amount of coarse material we could reduce the voids to 32.4 per cent. In the course of our study, to find out the amount of filler to be added to reduce the voids to the minimum, we began to add filler in definite increments, using 5 or 10 per cent increments, as the case may be. In this particular case I think we skipped the 10 because we had worked that out before and knew about where it was going, and knew what shape the curve would take. We jumped to 10, 20 and 30 and so on up the line. The last point on the curve is 100 per cent filler. The voids in the 100 per cent filler were 26 per cent. That is ordinary commercial limestone dust in Chicago, passing 85 per cent through the 200mesh screen. We found 30 per cent filler gave the minimum voids on the curve of 18 per cent. By this method, which might be called a rational method, or more or less exact method of designing, we find the maximum permissible amount of bitumen. In actually laying the mixtures we do not use quite the maximum because of the difference in temperatures. This would be the maximum at air temperature where we have to lay it at temperatures around 300° or 350° F. and the asphalt expands at a greater rate than mineral, so we can not use quite the maximum. We usually stay two or three-tenths per cent below the maximum amount. With this as a basis it is very easy to calculate the percentage of bitumen which would be required to exactly fill those voids at the minimum point or at any other point in here (indicating chart) which might be selected as the practical mixture, simply by the volume basis.

(Returns to paper.) Our Lake Michigan fine sand that is available for asphalt work is not ideally graded. It is relatively high in material passing the 80-mesh and retained on the 200-mesh sieve, and is quite high in material, passing the 40 and retained on the 80, with very little material retained on the 40 and practically none passing the 200. For many years it has been the general custom among most of the contractors laying asphalt pavements in and around Chicago to use this sand without admixing a coarser material to raise the 10-mesh fraction. This practice is not considered good, as excellent coarse sand is readily available and can be mixed with the regular run lake sand so that a low voidage aggregate will re

to 18.5%. (30% filler in this aggregate would mean approxi

mately 27% of raw filler in terms of the whole mixture.) Numerous pavements have been laid this season in Chicago carrying 25% of raw filler. Such mixtures are easily handled

sult.

Without the addition of coarse material, the voids in the lake sand frequently run as high as 35% to 37%, and this may be easily reduced to about 30%.

The initial sand shown in the curve is a mixture of local coarse or torpedo and lake sand, showing 32.4% of voids. By the addition of 30% of limestone filler the voids are reduced

and laid as long as the bitumen is kept down where it belongs, but if it is raised only a fraction of a per cent (in terms of the mixture). the mixture stiffens up so that it is difficult to handle.

(Leaves manuscript.) Right here I would like to mention that one of the earliest high-filler mixtures of which I knew anything is right here in Des Moines, being a section of Locust Street, beginning at the Chamberlain Hotel and coming this way about two or three blocks. You can very easily tell where the limits of the job are because there is creosote block pavement on this end; it runs up to the creosote block pavement. It is in its 5th year, laid early in the summer four years ago. When that mixture was laid we did not know nearly as much about voidage as we do now. We were still working on the old surface area theory to a large extent, which is, briefly, that the mixture will require more bitumen as the percentage of fine material increases. In other words, if we added this high filler we would have to increase our bitumen to coat the particles. At that time we felt we were called upon to give them a mixture which would stand, not only heavy traffic-one of the heaviest traffic streets in Des Moines-but also the rutting along the curbs from parking during the summer temperatures (at which time the pavement is subject to parking all day and practically all night). My first thought was that we would increase the filler because we had run into the fact-and I guess it was pretty generally known-that an increase of filler would ordinarily stiffen the mixture somewhat. So we had the idea that we would run more filler than was ordinarily used, and I thought, although I could not quite reconcile it in my mind, that we would have to increase the bitumen. We set out to use 10% to 11 per cent of bitumen with about 20 per cent of filler. The sand carried a small percentage of material passing the 200-mesh. I think the hopper sand usually showed about 4 or 5 per cent passing. We began working with our mixture, and the stuff was so fat we could hardly do anything with it, and the rakers complained about it, it would ball up under the rakes, and they could hardly get the rakes through it. So I went back to the plant and commenced cutting the bitumen and maintained 9.5% as the correct amount of bitumen all through the work. I think the pavement shows for itself that the mixture is pretty well designed, because while today it is