records is due in some measure to the influence of wind, which is particularly active on skating speeds.

Column XII of Table IX indicates that the mean percentage deviation of the series of computed times against record times is 7.6 per cent between 500 meters and 100 miles (160.9 kilometers). If the same test is applied to the entire series between 50 yards and 100 miles, the mean percentage of deviation, without regard to sign, is 13.3 per cent.

The following formulas pertain to the straight line of Figure 13 :

The data for bicycling have been taken from pages 267 and 268 of "The World Almanac" for 1906, and are embodied in columns I and II of Table X. The first five events are professional paced records against time. The next series of events, from 2 miles to 100 miles, inclusive, are professional motor-paced records, in competition. The remainder are stated to be American competition, professional, paced, hour records. The average speeds over the distances are set down in column V. It will be seen that the speed between 2 miles (3.22 kilometers) and 30 miles (48.28 kilometers) was almost precisely uniform at 23.5 meters per second. In fact, it appears from the table in "The World Almanac" that all of these records were made by one and the same individual, on one and the same day (May 31, 1904), at Charles River Park, Mass.. Again, from 31 miles (49.89 kilometers) to 50 miles (59.59 kilometers), the speeds are nearly uniform at 22.8 meters per second (51 miles per hour), and these appear to have been likewise made at Charles River Park by the same rider on the same day (September 1, 1903). There is very little fall of speed between the 50-mile event and the 2-mile event, or apparently but little fatigue as far as 50 miles. Beyond 50 miles, however, the speeds fall off and fatigue is indicated.

Figure 14 shows log 7 plotted against log L, as in preceding cases. The observations fall near to a straight line, A B, B C, as far as 50 miles. This line makes an angle of 45° with the axis of distances. For the purposes of comparison a broken line ab, be is drawn through the 2-mile record point at an angle of 48° 22' to correspond with the lines in all of the other figures. The contrast is very noticeable. Beyond 100 miles (160.9 kilometers) the points tend to follow the direction of the broken line, but not with any degree of precision.

The inference to be drawn from Table X and Figure 14 is that