The Circular says that Table 1 "represents the returns from about 1,200 counties " ; while Table 2 shows that replies were re ceived from 1,061 counties. The first is equivalent to about one county in twenty-five, and the second to about one in thirty. In either case, the number of answers was entirely inadequate to secure data representative of the entire country. It is stated by the Director of the Road Inquiry Office that " nothing is known as to the number of replies received from each county. " Average. Haul.The value of the reply as to the average haul will depend upon the manner of determining it. "The Road Inquiry Office has no copy of the letter of inquiry." Apparently the letter merely asked for the average haul, and gave no instruc tions as to the method to be used in deducing it. A man on receiv ing such an inquiry and knowing that farm products were hauled to a certain town from all distances up to ten miles, would probably reply that the average haul was one half of 0+10, or 5 miles. By many trials, the writer has found that in the great majority of cases this answer is accepted as correct; while in fact it is erroneous, and in many cases greatly so. This method of determining the average haul does not take into account the number of loads hauled each distance. The average of the distances hauled is ordinarily con siderably more than the average haul; and for this reason, it is probable that the values in the first column of Table 2 are consider ably too great. The distance hauled will vary greatly with the locality. Farm products will be hauled much farther to a large city than to a small village; certain kinds of products will be hauled much farther than others; and the distance hauled will vary greatly with the kind of roads. The result for the average haul to large cities, where a considerable part of the freight traffic on the public highways is hauling vegetables, fruit, etc., over good stone-roads to market, even though correctly determined, is of but little value in finding the cost of hauling the average farm product to market. Gardeners, dairymen, etc., live chiefly near large cities and are more interested in good roads than the average farmer; therefore it is possible that the replies from near large cities were in undue pro portion. According to the U. S. Census for 1890, gardeners, dairy men, florists, nurserymen, and vine-growers constitute only one fifty-seventh of the so-called farming class; and the "area devoted to raising fruit and vegetables for market is 534,440 acres," or one six hundred and sixty-ninth of the area of the cultivated land.
The value of the replies in determining an average will depend upon their distribution with reference to productiveness. One part of the state furnishes more traffic and is better supplied with railroads than another. The replies should be distributed pro portionately to the amount of traffic.
An examination of a large-scale railroad map of the United States indicates that probably the average haul as given in Table 2 is considerably too great—at least for the states that furnish the bulk of the traffic. Portions of a state may be found which are relatively at a considerable distance from a railroad station, but in nearly every case it will be found that throughout that area there is but little traffic on the public highways.
The reliability of the value of the average haul as given in Table 2 can be approximately tested in another way. For example,
Illinois has an area of 56,600 square miles, and in 1895 had 10,752 miles of railways exclusive of sidings and second tracks; or 1 mile of railroad for each 5.2 square miles of area. Investigations also show that the distance between railroad stations averages a trifle under 4.5 miles. Therefore, if we consider a strip of land 5.2 X 1 miles laid transversely across the railroad half way between railroad stations, the maximum haul will then be approximately of 5.2 + of 4.5 = 4.8 miles. This may be regarded as the average maximum distance of haul in Illinois. The average haul is prob ably approximately half this, or say 2.4 miles. There is a slight error in the above computation, since no account is taken of the fact that the railroads cross each other or that they converge toward railroad centers; and for this reason the above result is slightly too small.
On the other hand, the above method is slightly in error since no account was taken of water transportation; and for this reason the above result is slightly too great. Again, the above result is slightly too small, since produce is hauled on wagon roads consider able distances to large cities; but the amount of these products is small as compared with the total agricultural products (see § 23). Therefore we may conclude that the above result is not much too small, although it is less than half that in Table 2.
A similar investigation for Iowa, Eastern Kansas, Eastern Nebraska, Ohio, and Indiana, leads to the conclusion that the values for the average haul in Table 2 are about twice too great. The above method of testing the data in Table 2 is applicable to states that have considerable areas of untilled land, since under these circumstances the distribution of railroads and railroad stations will not be even approximately uniform.
The reliability of the data in Table 2 may be tested in an other way. The Illinois Agricultural Experiment Station, in Bulletin No. 50, published the data received in 316 replies from 76 counties in Illinois concerning the cost of producing and marketing corn and oats. According to these data the average distance hauled was 3.2 miles—about six tenths of the value in Table 2 for the aver age haul in Illinois.
Average Weight of Load. The weight of the average load varies chiefly with the grade of the road and the condition of its sur face; and in most localities the latter varies greatly with the season, and is not the same for any two successive years. Further, with earth roads (and they constitute the great majority of agricultural roads in this country) the most of the freight is hauled when the roads are in their best condition. For these reasons, it is a matter of considerable difficulty to determine the weight of the average load for any particular place, much less for an average of several states. However, as these data are not directly used in determining the supposed cost of bad roads, this phase of the subject will be dis cussed only briefly.
In the Bulletin just referred to the average load in marketing 311.845 bushels of corn was 62 bushels. If it were all shelled. this is equivalent to I 74 tons per load; and if it were unshelled. it is equivalent at least to 2.2 tons per load. As part of it was shelled and part not, the average load in this case was somewhere between 11 and 2} tons.