tions to the arms, are locked. The entire works are housed by an iron shell, and the brake just above C throws a shoe against crank wheel, stopping the mill when thrown out of gear.

In operation, this mill, in a strong wind, allows the wheel to swing around to assume a position more or less parallel to the vane, presenting less effective wind surface, and reducing the power of the wheel in accordance with the variation in velocity or force of the wind.

A chain passes from the governor hinge, or lever, over a pulley and down through the hollow pitman, extending to lever close to the ground, by which the mill may be thrown in or out of the wind as desired. The short tubular pitman at the top, working through the turn-table, is con nected with the pump-rod by a swivel joint, as is required in all mills.

The pump used on this mill is what is known as the Syphon, and is manufactured by Rumsey & Co., Seneca Falls, N. Y. It consists of a brass cylinder three inches in diameter, and allowing for a stroke of four and five eighth inches. This cylinder is suspended in an iron cylinder, forming a water space around it, into the upper part of which the water is supplied, and the water is sucked up through a valve at the bottom, and lifted by the piston, passing into an air-chamber over the cylinder, from which it flows to the discharge pipe. The piston-rod passes through the top of the air-chamber and is always well lubricated with water, keeping a tight joint even when the packing lacks oil. The pump is single acting.

The derrick, upon which the engine and pump are supported, is formed of four timbers converging toward the top and terminating in a square platform suitably ornamented, and suitably braced by the usual batter-board and cross-ties. The mast, which supports the engine proper, is placed upon this derrick, passing through the platform and being supported upon cross frames.

The price of this mill, without derrick and pamp, is eighty-five dollars.

LOCATION OF MILLS.

The five mills in the contest were located in the north-westerly portion of the fair grounds, and arranged fifty feet apart; the Halladay, Leffel. Perkins, and Kennedy mills being arranged in a direct line south-west and north-east, and the Manvel mill at right angles to and south-east of the Kennedy mili.

All of the mills had a clear unobstructed wind space, being raised forty feet above the ground and entirely clear of all buildings, trees, or structures of any height. The main building was east of the Halladay mill, but was located several hundred feet therefrom, and, during the test, only twice did the wind come from that quarter-during part of Monday, September 23, and Friday, September 26, (see table B.) When from that direction it was with the small velocities of seven and two miles respectively, and the density of the air was light. Thus it is seen that the obstruction, if it can be so called, was insignificant.

The ground was as level as could be expected; the different levels, considering the Halladay mill as zero, were as follows:

Halladay, 0; Leffel, +12 inches; Perkins, -24 inches; Kennedy, -30 inches; Manvel, --3 inches.

The direction of the wind most of the time was such that the air-currents were not required to pass one mill before reaching the next, thus enabling each and all to receive the full effect of whatever air was in motion during the period of trial.

TESTING APPARATUS.

The data it was desired to obtain consisted in the number of revolutions of each mill, and the quantity of water pumped under a given pressure, both in light and heavy winds, and for short and protracted periods, and under varying barometrical changes of the atmosphere.

The speed or number of revolutions of the mill (strokes of the pump) was obtained by speed indicators, worked by a lever pivoted to the reciprocating pump-rod.

The amount of water pumped was obtained by use of the Keystone water meter, manufactured by the Keystone Water Meter Manufacturing Company, (Limited,) corner Front and Laurel streets, Philadelphia.

The meters were carefully tested, before being applied. to insure the accuracy of their registration. They were placed in the return pipe, which returned the water to the wells, so as not to increase the resistance of the pump.

The water was pumped from two large barrels, united by a syphon to increase the water surface, and into one of which barrels the ball and cock connecting with the city mains were placed to compensate for any water lost by evaporation or otherwise, and from the other of which barrels the water was pumped and then returned, after passing through the meter, so that no excessive duty would be demanded of the syphon during a strong wird or gale. The water was forced up a stand-pipe for a distance of twenty-five feet from the fixed level in the tanks, and from a goose-neck it was discharged into a large vessel provided with a float-actuated valve and a screen surface of one square foot of No. 60 wire gauze, to insure a filtration of all water passing to the meter. The float-valve was designed to automatically close the inlet to the pipe communicating with the meter when the level of the water in the tank came within three inches of the valve-seat, so as to prevent the passage of air down said pipe, and the registration of the same by the meter.

By this means, no matter how irregular the pump may be worked, the water will pass through the meter in a solid column, to the exclusion of the air, and thus cause the meter to work under a constant pressure.

The lower portions of the mills were housed in and kept under lock and key, so as to exclude all persons from tampering with the pump or testingapparatus.

The mills were oiled at the beginning of the test and were not allowed to be oiled or touched until the close thereof, they being closely watched, both day and night, by attendants. The direction of the wind, its velocity, and barometric condition (density) of the air, were accurately taken by the United States Signal officer, with the Government instruments.

TEST.

The tables A and B give all the data taken, with the exception of a few stoppages, in which slight interruptions to the regular working of the apparatus took place, and these were as follows:

The ball float-valves, arranged in the tank at top of return-pipe, were made to insure the prevention of air passing through the meters, and, when designed, it was not thought that they would be subjected to such a strain as practice has proved them to have been. The breakages noted were where the brass arm holding the float broke, allowing the valve to shut down tightly, and the tank to overflow, until the valve was raised and the ball-float once more attached. The same weakness showed itself in nearly all the mills.

The breakages in the speed recorders were due to the fact that the American Meter Company, who made them, used narrow brass pawls to work the ratchet wheels in use on the mills. Where much vibration in the pumprod was experienced, the strains and friction were so great as to wear the pawls down, and render them unfit for further duty. The recorders being securely inclosed within a soldered case, it was impossible to repair them. The enormous pump used on the Halladay mill caused so great a vibration that it broke two recorders before a run of ten hours could be had, hence, in table A, it will be noted that no speeds are given for this mill.

Referring to table A, we see that when the wind ranged between twelve and sixteen miles' velocity, and the barometer stood at 29.973, Thursday night, September 18, that the speed of the mills was not nearly so great as when, with an average velocity of wind of about nine miles an hour, the barometer stood at 30.132, showing that the density of the air, as well as the velocity, plays an important part in the available duty of a mill. The relative speeds of the mills may be seen by the record of September 20, at 5, P. M., table A, in which we see that during a run from Thursday, at 3.45, P. M., the Leffel made 35,200 revolutions; the Perkins, 35,700; the Kennedy, 20,520, and Manvel, 24,550, showing a great difference between speeds.

Comparing three of these mills on September 23, at 5.30, P. M., we find them as follows: Leffel, 77,720; Kennedy, 51,190, and Manvel, 58,710; showing the Kennedy and Manvel wheels to be gradually losing ground in velocity. Again, comparing the Leffel and Manvel velocities, after the full test, we have: Leffel, 180,750, and Manvel, 174,770, showing a gain by the