Some five-year tests have just been made on the Northern Pacific Railroad to determine the actual wear of rails under a measured amount of traffic. Apparently one object of the investigation was to determine the effect of variation in the chemical composition of the rails. A pair of each of five types of rails were tested in each of eight situations; six of them being on the Pacific division and two on the Minnesota division. No attempt will be made here to analyze the effect of variation of chemical composition, but, since one of each kind of rail was used in each locality, the average of all rails for each locality will be considered as the typical rail for such an aline-ment and grade. The rails were actually weighed each year for five years, so that their actual loss in weight during each year's wear could be determined. The tonnage passing over these rails was systematically recorded. It varied for these eight localities between 27,021,227 and 29,862,738 tons. For uniformity in each case, the wear was reduced to the uniform basis of 10,000,000 tons. Even though the wear is not strictly proportional to the tonnage, the variation between 27,021,227 and 29,862,738 is not large enough to cause any serious error from this source. The wear of the rails on the tangents in per cent per 10,000,000 tons' duty is given in the following tabular form.

Table XV. Rail Wear On Tangents, Northern Pacific R.R

Five pairs of rails on first tangent, Pacific div.; grade, 0.3%.

Five pairs of rails on second tangent. Pacific div.; grade, 0.525%.

Five pairs of rails on third tangent, Minnesota div.; grade chiefly level at bottom of sag.

Pet. loss in four years.

Pet. loss per 10,000,000 tons' duty.

Pet. loss in four years.

Pet. loss per 10,000,000 tons' duty.

Pet. loss in four years.

Pet. loss per 10,000,000 tons' duty.

1.800

.603

.960

.346

.451

.167

2.336

.783

.900

.324

.825

.305

1.806

.605

1.360

.490

1.135

.420

2.015

.676

1.256

.452

.822

.304

1.706

.572

1.807

.391

.280

.104

.648

..............

.401

..................

.260

It may be at once noticed that the average loss in per cent per 10,000,000 tons' duty on the first tangent was 0.648%; on the second tangent it was only 0.401%. In the endeavor to discover the cause of the uniformly increased wear of the rails on the first tangent over that on the second tangent, the grade of the two tangents was considered, but the grade of the first tangent was 0.3%, while that of the second tangent was 0.525%. As rail wear is usually greater on steep grades than on flat grades, owing to the slipping of the driving-wheels when climbing the grades, or the possible skidding of the wheels when moving down the grade with brakes set, the results are here relatively contrary to what we would expect. The only apparent explanation of the increased rail wear on the 0.3% grade is that it occurs near the bottom of a very long down grade, where a train might have acquired a high velocity and where the wheels might have skidded in the attempt to hold the train, or where engines, hauling a train up grade, are doing their utmost (perhaps by using sand) to obtain a sufficiently high velocity to carry their trains by momentum over the long grade. In the case of the 0.525% grade this tangent occurs at the very upper end of the grade, where the velocity in either direction is probably lower than the average. Whether this is the true explanation, the relative wear on these two tangents for all makes of rails is uniformly as stated above.