more traction will be required with wide tires on a hard surface that is not smooth.

The committee of the Society of Arts previously referred to experimented on the streets of London in 1875 to ascertain the force required to draw loads over different roadways at varying rates of speed. Table No. 49 gives results.

The report added that the asphalt experimented on was not in good condition, and for that reason the force shown for asphalt was undoubtedly higher than it otherwise would have been. These figures, however, are valuable as they give the effect caused on the draft of increased speed.

Table No. 50 is made up from the results of different experi menters, the figures representing the force in pounds to draw one ton at a speed of approximately 3 miles per hour.

From all these figures it is estimated, taking into consideration the varying conditions under which all tests were made, as well as the improved character of pavements at present, that the force

expressed in pounds to draw one ton over the different pavements herein considered would be: granite A 34, granite B 40, asphalt 16, brick 20, Belgian 40, macadam 40, and cobblestone 65. This will give a percentage to granite A of 7, granite B 6, asphalt 15, brick 12, Belgian 6, macadam 6, and cobble 4.

The general opinion among engineers is that the tractive force varies inversely as the diameter of the wheels, but some say inversely as the square root of the diameter. Mr. W. Hewitt in a paper before the Surveyors' Institution of England says: "From experiments made with Eastren and Anderson's horse-dynamometer at the Royal Agricultural Show, 1874, a slightly greater ratio than inversely as the diameter was given, and I am inclined to think that inversely as the diameter is the more correct view of the two."

Slipperiness.-A great many conditions affect this property: conditions of the street, temperature, whether wet, damp, or dry,

etc.

Mr. Wm. Haywood, Engineer to the Sewer Commissions of London, made some very extended observations in the London streets in 1873 to determine the liability of horses slipping on asphalt, granite, and wood pavements.

The asphalt observed was the ordinary rock asphalt of that time, 2 inches thick on a 9-inch concrete base, with the surface in good condition. The grades varied from 1 in 58 to 1 in 550.

The granite pavement consisted of Aberdeen blocks 3 inches wide, 9 inches deep, and from 9 to 15 inches long, laid stone to stone, the joints being filled with stone-lime grout. The pavement as a whole was not in good condition. The grade varied from 1 in 30 to 1 in 1000.

Two wood pavements were experimented with. One was formed. of fir blocks 3 inches wide, 5 inches deep, and 9 inches long. The blocks were laid touching each other at the ends, but crosswise of the street; the joints were inch wide, filled in with thin gravel and grouted in with a bituminous composition. The other consisted of beech blocks 3 inches wide, 4 inches deep, and 6 inches long, with 1-inch joints at side and ends, filled in with cement grout. The grades varied from 1 in 30 to 1 in 260.

The asphalt was sprinkled slightly with sand, and the wood four times with gravel. The wood and granite were watered to lay

the dust, but the asphalt was not treated. All the pavements were kept as clean as their nature and respective surfaces admitted with the usual amount of labor. All observations were taken between 8 A.M. and 9 P.M.

was:

The mean number of horses passing daily in March and April

Table No. 51 shows the total number of horses that fell on the different streets during the fifty days on which observations were taken, as well as the daily mean.

The mean being: Asphalt 191 miles travelled for each accident.

Accidents" in this connection mean falls on knees, falls on haunches, and complete falls. No account was taken of horses slipping simply. During the last thirty-two days record of these different occurrences was kept, and the percentage of each is shown in Table No. 52.

Table No. 53 shows the distance in miles horses travelled without accident under three different conditions of surface moisture.

Mr. Haywood thinks that the accidents on the beech pavements should be eliminated, as they were not typical pavements, when the true order of slipperiness would be:

Granite
Asphalt

Wood

Miles.

132

191

446

It must be borne in mind that the asphalt experimented on was the European natural rock asphalt, which is admittedly much more slippery than American asphalt.

In summing up Mr. Haywood says:

Taking the whole group of conditions into account, the asphalt was the most advantageously placed, the wood was the next so, and the granite was the worst placed.

On the average of the whole fifty days' observations, the granite was found to be the most slippery, the asphalt the next so, and the wood the least.

Separating the accidents under three conditions of surface as regards moisture, it appears that asphalt was most slippery when merely damp, and safest when dry; that granite was most slippery when dry, and safest when wet; that wood was most slippery when damp, and safest when dry."

In 1885 Capt. F. V. Greene had a series of observations made in ten of the principal cities of the United States to determine the relative slipperiness of the same kinds of pavements as laid in this country. From his results he decided that on pavements in American cities a horse would travel 272 miles on wood, 413 on granite, and 583 on asphalt without an accident. His accidents were divided also into falls upon the knees, falls upon the haunches, and complete falls. On a rough pavement falls upon the knees should not be wholly charged to slipperiness, as a great many must