whereas the cost per pound does vary considerably, and is therefore not a good unit of comparison. I have assumed a uniform rate of wages for old and new mills, and the same cost of power per horse-power (although of course both have changed considerably), and also the same first cost of machinery and steam-plant, so as to simplify the comparison. For first cost I have taken mules at $3.00 per spindle, with all their appurtenances, shafting, belting and floor space; frames at $3.60 per spindle; steam-plant at $40 per horse-power for the plant itself, and at $50 including foundations, settings, chimney, buildings, etc. I have taken interest at 5 per cent., depreciation on machinery at 7.5 per cent., and on buildings at 4 per cent. Cost of a horse-power with compound engines, $10 a year for fuel and $19 a year for total cost. I do not pretend that these figures are ideal or absolutely accurate. They of course will vary in different mills, but I think they are fair, and any reasonable variations that can be made in them will not materially affect the point the table is intended to illustrate.

COST PER YEAR OF OPERATING ONE SPINDLE ON PRINT CLOTH
NUMBERS.

Comparing high-speed mules with slow, the cost for power is from 16 to 70 per cent. more for the high speed, and yet the total cost of production is 10 to 13 per cent. less. Even with 70 per cent. more power per spindle, the high speed saves 10 per cent. in cost. Taking the old slow-speed warp frames and the modern high-speed spindles, the power per spindle is about the same, and the saving in total cost about 32 per cent. And, comparing the high-speed spindles with Sawyer, at 7,500 revolutions, the power costs 41 per cent. more for the former, but the net saving is 13 per cent.

The old slow-speed mules and frames are of course now nothing but reminiscences. But it is not such a very great while ago that 7,500 was considered high speed for warp spindles. The above table will show easily that a mill to-day cannot afford to run spindles at that speed in competition with modern spindles at 9,200 revolutions, even if it could have a free gift of all the fuel required to drive them. Thus the allowance in the table for fuel for power to drive the spindles at 7,500 revolutions is 9 cents per spindle per year. If this is wiped out, it leaves the total cost of production $1.10 per spindle per year, and $1.96 per 100 pounds of yarn, or over 6 per cent. more than the cost with high speed. Even if we deduct all the expenses of the steam power (17 cents per spindle per year), it would still leave the total cost $1.02 per spindle and the cost of product $1.82 per 100 pounds, against $1.84 with high speed. In other words, the free gift of a modern steam-plant large enough to run the slower spindles, and also of all the fuel and labor and expenses of all kinds to operate it, would barely enable them to compete in cost of production with the high-speed spindles, the latter being charged with the cost and all expenses of a steam-plant 41 per cent. larger, and consuming 41 per cent. more fuel than the slow ones. And this is not all. For, when the profits of business are taken into account, the high-speed spindles, with their increased product, would easily leave the slow ones "out of sight" even under the above disadvantages in regard to power.

If it should be thought that the item of depreciation ought to have been made larger for the high-speed spindles than for the slow ones, there is ample margin for this item to be modified to suit any reasonable opinions in this regard, without materially affecting the point of comparison which is the object of the table. The item of profits of business, if taken into the account, would offset any possible modifications of this sort several times over in favor of the high speed. Besides, the item of depreciation is intended to cover the cost of the continual changes and improvements being made in machinery, involving sometimes the discarding of the machines entire, and this item is more than half of the whole amount I have allowed; so that the percentage of increase on account of wear and tear simply would be so small that I did not consider it worth while. to calculate it for the purposes of this comparison.

A word in regard to the variation of 5 cents per spindle in the table, in the cost of power for modern mules. The existence of this variation is well known, and of course this difference in cost of power cannot be ignored, providing other things are equal, — quality of work, repairs, waste, wear and tear, etc., as it then represents a positive and considerable loss, without any compensating advantages. But, if it were a question of cost of power only, as between a slow and a high speed mule, other things being equal, the results of a comparison would be the same as with frames, only to less marked extent. The high speed would leave the slow ones "out of sight" as to economy of manufacturing.

Taking now a general view of the entire mill, a comparison between the old and modern mills is no less striking. Going back to the time when both warp and filling were spun on mules (1869), my data show that what was then a good average mill of 44,000 spindles, all mules, was producing 37,700 pounds per week, and required 677 horse-power to drive it, a product of .85 pounds per spindle per week and of 55 pounds per horse-power per week.

Another, with 28,000 spindles, all mules, produced 24,300 pounds per week, with 430 horse-power; or .87 pounds per spindle per week and 56 pounds per horse-power.

Another mill (1874), with old ring warp and mule filling, with 90,000 spindles, produced 78,000 pounds per week, and required 1,583 horse-power to drive it, a product of .87 pounds per spindle per week and of 49 pounds per horse-power per week. These were all on print cloth numbers.

A mill with Sawyer warp spinning and fairly high-speed mules produced .93 pounds of cloth per spindle per week, and 1 horse-power produced 46.5 pounds of cloth.

And to-day a modern mill with all frames, at high speed, produces 1.17 pounds per spindle per week, and 1 horse-power produces 46.75 pounds. In round numbers, the product of a 30,000-spindle modern mill is equal to that of a 40,000-spindle mill of twenty years ago.

From these figures we find that the Sawyer warp mill required 19 per cent. more power to turn off a pound of cloth than the old all-mule mills. But this fact did not weigh a feather against the adoption of Sawyer spindles.

Since that period the product per horse-power has not changed materially. But the improvements made in steamplants meantime have reduced the actual cost of the power per pound of cloth, so that it is less than it was in the old slowspeed"all-mule" mill, the total cost of fuel for power alone being now about .41 cents per pound of cloth, while then it was .66 cents (taking the same price of coal in both).

The deduction to be made from the above table is, that, in regard to cost of production alone, any increase of speed and product will be in the line of economy, so far as cost of power is concerned, even if the latter should increase four times as fast as the production. And this is so unlikely a supposition that practically the question of power is not to be considered for a moment as against speed.

One of the items of power in a cotton mill, and not a small

one, is the friction load of the shafting. Power expended for this is in a sense wasted. It produces nothing and costs a great deal. In the best mills it will be not less than 22 per cent. and often 25 per cent. of the total power. (This includes the friction due to the belts on the loose pulleys of all the machines, as this is the usual method of weighing this load, so that it does not of course represent the mere friction due to weight of the shafting.) In a mill requiring 1,000 horse-power, therefore, 220 to 250 horse-power will be expended in this manner, costing, at $19 per horse-power per year, about $4,200. Various methods have been tried from time to time to reduce this loss. One way has been by reducing the diameter of shafting, sometimes to extremes, and increasing the speed; but not much has been accomplished in results. The percentage remains about the same. In the course of my work I have had occasion to test the power of a large number of mills of all descriptions, old and new, large and small, with excessively heavy and excessively light shafting, at extreme slow and high speeds, and medium heavy at medium speeds, and with all sorts of bearings, and in all sorts of conditions. I have found the friction load to run from 22 up to 39 per cent. The lowest I have ever found was 21.25 per cent., and this was in a very old mill, requiring 1,055 horse-power, with rather heavy shafting, but all at slow speeds, from 210 to 250 revolutions. The friction load was 224 horse-power. I have never found this result equalled in a modern mill. Several years ago I tested two mills in the same yard: one a very old one, with extremely, even ridiculously, heavy shafting, but at very slow speeds; and the other a new mill just completed at that time, with very light shafting at high speeds, and with bearings about 5 feet apart. I remember that I expected to find the friction of the older mill so much more than the new one that it would pay to change the shafting. The test showed so little difference between the two that it was not worth considering. This result was a surprise to me at that time, but would not be so now.