The study of castor oil mixtures was next continued with the idea of ascertaining what if any differences existed in the stabilizing values of certain commercial mineral fillers. This work was described in a paper presented at the January, 1925, annual convention of the American Road Builders' Association, Chicago, Ill. Three types of commercial products were investigated, limestone dust, portland cement and hydrated lime. From the standpoint of size of particle these fillers were classed as relatively coarse, medium and very fine, respectively. The following facts were strongly indicated by test results obtained on mixtures of these fillers with sheet asphalt sand:

1. Extreme fineness (that is the presence of material passing the 350 mesh sieve), of at least a portion of the commercial filler is the most important property tending to produce high stabilizing values in the total mineral aggregate.

2. Graduation of size particles in the mineral filler tending to reduce the percentage and size of voids is also an important stabilizing factor.

3. The stabilizing value of fillers is to a considerable extent a direct function of their compressed volume equivalents. 4. Of the commercial fillers tested on an equivalent weight basis there was practically no difference between the stabilizing value of limestone dust and Portland cement, but hydrated lime produced much higher stability values than the other two fillers.

Having proceeded thus far the authors next turned their attention to developing the test so that it could be successfully used in testing mixtures of fine mineral aggregates with asphalt cement. Here many knotty problems were encountered, the details of which need not be discussed in this paper. It was found necessary to modify the testing apparatus to some extent and to carefully standardize each step of the method, particularly the control of heat treatment during mixing. This required a considerable amount of time and thought, but gradually the test was perfected until it was found possible to obtain apparently consistent results which could be checked usually well within a range of 10 per cent from the average for any given mixture.

This method was described in detail in a recent paper presented before the American Society for Testing Materials.† It is identical in principle with the castor oil test, but all asphalt paving mixtures are tested at approximately the maximum temperature to which a finished pavement under traffic is ordinarily subjected, as displacement is more likely to occur at such temperature than at lower temperatures.

* Proceedings American Road Builders' Association, 1925.

A Practical Method for Determining the Relative Stability of Fine Aggregate Asphalt Paving Mixtures. Proc. Am. Soc. for Test. Mats., 1925

The mixture of hot mineral aggregate and asphalt cement is first made under conditions approaching as closely as possible those obtaining at paving plants. From each mixture, while hot, three clyindrical briquettes are compressed in 2-inch cylindrical molds under a load of 3,000 lbs. per square inch. The briquettes are next ejected from the molds and allowed to cool over night. They are then placed in a water bath and brought to a temperature of 140° F. The test on each briquette is made in the water bath by placing it in a testing mold of slightly larger diameter with a circular bottom orifice 134 inches in

diameter. A plunger is placed on the top of the briquette and a rapidly increasing load applied until the maximum load required to force the mixture through the orifice is reached. The average maximum load developed by the three briquettes is then recorded as the stability value of the mixture. The assembly of bath briquettes and testing mold used in this test is shown in Fig. 2.

Load in Lbs.

With a large number of experimental mixtures tested during the development of the method it was soon demonstrated that with any given sand the stability of the mixture was materially increased by the addition of increasing amounts of mineral filler up to a certain point, and that different fillers possessed different stabilizing values when used in equivalent weight percentages of the total. Certain typical results obtained along this line are illustrated in Fig. 3, showing comparative stability values obtained from the use of a rather coarse commercial limestone filler and a very fine hydrated lime. To this extent the trend of results has been almost identical with results obtained from the castor oil mixtures.

The effect of variations in sand grading has not as yet been investigated or checked against the castor oil mixtures, as there are a number of other important factors upon which it was thought best to first obtain some definite information. Among these were the effect of initial compression, percentage of asphalt, consistency of asphalt and percentage of voids in the mixture. Certain typical results so far obtained in connection with the three first maintained factors are shown in Fig. 4.

It should be understood that these results apply only to certain specific mixtures made with certain constituents and that at the present time it would be extremely unwise to attempt to rely upon the actual test values given in proportioning mixtures from other individual constituents at a paving plant. They are here shown merely for the purpose of illustrating certain points which must be considered in connection with further use of the test.

The upper right-hand curve shows stability values obtained from paving mixtures composed of 80 parts of a given sand and

Fig. Effect of Initial Compression and Consistency
of Asphalt on Stability of Natural Sand Mix.

(Figure 4)

20 parts of a limestone filler when the percentage of a given asphalt in the total mix was increased from 7 to 11 per cent and all mixtures were given the same initial compression. The curve clearly shows that maximum stability is produced by a rather limited range in percentage of asphalt. Starting with a low percentage of asphalt the stability increases rapidly to a maxi

mum with increase in asphalt and then almost as rapidly falls off with further increase in asphalt.

The lower right-hand curve illustrates the effect upon stability of variation in penetration of asphalt for certain specific mixtures identical in every other respect. Here increase in stability with decrease in penetration of asphalt is very definitely indicated and supports the present general tendency toward the use of harder grades of asphalt in the construction of heavy traffic pavements.

The curve at the left illustrates the extremely important part played by the degree of initial compression in securing maximum stability from a given paving mixture. These results were obtained from briquettes of the same mixture which were com

pressed under loads of 1,000, 2,000 and 3,000 lbs. per square inch before they were subjected to test. Within this range it is seen that stability is almost directly proportional to the compressive load to which the mixture was subjected.

Fig. 5 shows certain relations developed between the stability value and percentage of voids of various paving mixtures, subjected to the same initial compression and containing the same constituents, but with varying percentages of asphalt and filler. Each individual curve represents mixtures containing the same