the tunnel, I find that the slope required to produce this velocity is 0.0041, or 0.41 of a foot in 100 feet.

The open channel from Mill Creek to the tunnel will have to carry the same quantity of water, viz, 418 cubic feet per second. In determining the slope of this channel I have thought that it would be better to provide for a velocity too great than one too small. If it be too great and erosions of the bed should occur at points where the soil is less resisting, these places can be paved as successively may be found necessary, whereas if it be too small there would be required an annual expenditure from the appropriation for maintenance and repair of the aqueduct for removing deposits from the channel. I therefore propose to provide for a mean velocity of about 4 feet per second. This would be slightly excessive if the channel should run full, but as the calculation of 418 cubic feet per second was under the extreme supposition of a rainfall of 13 inches per hour over the entire watershed of the reservoir in a storm of several hours duration, and as the highest rate of rainfall recorded at the signal office in Washington between June, 1876, and November, 1892, was only 1.20 inches per hour for one hour, the probabilities are that the channel will never run full.

When it does not run full or nearly full (during the major part of the year there will be but a few inches of water in the channel) the mean velocities will be less and there will be a danger, that can not be avoided in so changeable a stream, of deposits in the channel.

For a mean velocity of 4 feet per second the waterway would require to have a cross section of not less than 104.5 square feet. If the channel be made 6 feet deep and 9 feet wide at bottom, with side slopes of one vertical to one and a half horizontal, its cross section would contain 108 square feet, and by the Kutter formula, assuming 0.03 as the value ofn, I find that the slope required to enable this channel to carry 418 cubic feet per second is 0.0012, or say 1.2 feet in 1,000 feet.

I therefore propose for the tunnel an interior diameter of 7 feet and a slope of 0.0041; to make the open channel between Mill Creek and the tunnel 6 feet deep and 9 feet wide at bottom with the side slopes just mentioned, and that the channel shall have a slope of 0.0012.

Under the same supposition of 14 inches per hour rainfall, the clannel from East Creek to Mill Creek will have to carry (see also my last annual report) 110 cubic feet of water per second. For a mean velocity of 4 feet per second the cross section of the channel must therefore have an area of not less than 27 square feet. A channel 3 feet deep and 5 feet wide at bottom with the same side slopes as before would have an area of 28 square feet. The proper slope or inclination in this case would be 0.003, or 3 feet in 1,000 feet, and I propose to make the channel accordingly.

Estimates for completion of the works of improvement of the Dalecarlia receiving reservoir and purchase of land authorized by the act of March 3, 1893, for lowering the height of the cross dam at the distributing reservoir, and for cleaning out the distributing reservoir, will be found in the list of estimates appended hereto, and explanations of the same will be found further on in this report under the title "Expla. nations of estimates."

THE CONDUIT AND THE CONDUIT ROAD.

For want of funds nothing has been done during the last fiscal year in the work of removal of deposits in the conduit, which my inspection of its interior in September, 1891, found to amount to about 15,000 cubic

yards. The deposits interfere with the full flow of the conduit, and, with the want of height of the dam at Great Falls, although not to so great a degree as the latter, they are a cause of a deficiency of water in the distributing reservoir, which in turn give rise in summer tɔ complaints to the District Commissioners from householders in the city. Their removal requires the emptying of the conduit throughout its length and the digging up and loosening of the deposits and sluicing them out through the waste gates and valves in the conduit, and is very expensive by reason of the necessity of employment of night labor, which costs more than day labor, and by reason, also, of its frequent interruptions during the refillings of the distributing reservoir required to keep up the supply of the city. The work can not be done by means of the small annual appropriations for repairs of the aqueduct, and in my estimates of 1892, and again in 1893, I asked for an appropriation of $14,000 for this purpose, but it has not yet been granted by Congress.

It is a most important work, and I again include the item in my annual estimates.

The deposits in the 7-foot by-conduit at the distributing reservoir were removed in July. There was a depth of about 2 feet at the influent gatehouse, and it decreased to about 6 inches at the auxiliary gatehouse. The trouble heretofore had in opening the waste gate in the dam of wasteweir No. 3 was ended by the making of an iron ratchet for maneuvering the gate.

Seven hundred and eighty-four cubic yards of flint rock, purchased in February, were crushed in April and piled on the side of the Conduit road above the distributing reservoir, for use in the repair of the road from this reservoir to culvert No. 24 during the next winter. The stone cost 93 and 95 cents and $1 per cubic yard, and the cost of setting up the steam crusher, crushing the stone, and piling it was 41 cents per cubic yard.

Sixteen boundary stones were planted between Cabin John bridge and Griffiths Park bridge (bridge No. 3), and six were planted between the distributing reservoir and the Dalecarlia receiving reservoir.

The Conduit road, from the intersection of the Foxhall road to the upper end of the distributing reservoir, was repaired in February with 1,088 cubic yards of crushed bluestone, from the quarries on the Virginia side of the Potomac, instead of the white flint rock heretofore used on this road. The distance is about 5,000 feet, or about 1 mile. The stone was put on about 4 inches deep, and it was thoroughly rolled with the 15-ton steam roller kindly loaned me by the District government. I was induced to use bluestone for this repair of the road by the exorbitant demands of the owners of flint rock in the vicinity as to prices and by an experiment I made respecting the comparative resistance to abrasion of bluestone and flint rock. This was made at a foundry in a large cylinder termed a "rumbler," used for cleaning castings. The cylinder, partially filled with 300 pounds of broken stone and 100 pounds of broken iron castings, was revolved at the rate of 30 revolutions a minute, the fine material as fast as it was worn from the stone falling out through interstices in the cylinder. The loss of weight of the bluestone by this process was found to be considerably less than from the flint rock, but experience has since proved that the latter makes by far the better pavement. It is not so dusty in summer or muddy in winter, and this is doubtless due to the fact that the particles worn from the flint rock are in the form of sand, while those ENG 94--201

from the bluestone make a fine powder. This trial of the relative values of bluestone and flint rock for macadam was conclusive, and I will use no more of the former for repairs of the Conduit road.

The only hill on the Conduit road this side of the hills near Great Falls is Dalecarlia Hill, near the upper reservoir. It had for several years been in a bad condition, especially on its western side, and it was very thoroughly repaired in December by covering it with a macadam pavement over a length of about 650 feet. About 500 cubic yards of flint rock were used in the work.

Estimates for the removal of deposits in the conduit, for raising the masonry casings of the manholes along the line of the conduit, and for commencing the work of widening the macadam pavement of the Conduit road by widening the pavement of the road between the two reservoirs, will be found in the list of estimates appended hereto, and explanations of the same will be found farther on in this report under the title "Explanations of estimates."

THE MAINS.

The trunk mains that lead from the distributing reservoir and supply the distributing system of street mains were laid by the United States, and are under the care of this office. The aggregate length of these mains is about 21 miles. The distributing mains were laid by the District of Columbia and are under the care of the Commissioners of the District.

There have been no breaks in the United States mains during the last fiscal year, and the only labor expended upon them has been in the driving up of the lead in a few of the joints where it was found necessary, and in providing against the bursting in freezing weather of the small valves designed for admitting air while emptying the 48-inch and 30-inch mains that were laid under the provisions of the act of March 2, 1889. These valves are at the following places: 2 valves at Thirtyfourth and M streets, 2 valves at Thirty-second and M streets, 2 valves at Twenty-fourth and M streets, 2 valves north of Dupont Circle, 2 valves at Fourteenth and R streets, 1 valve at Third and East Capitol streets.

Our experience in the winter of 1892-'93, when the frost penetrated the ground to a depth of 4 feet, showed the necessity of this precaution. The lines of the trunk mains have been carefully inspected and flushed monthly, and the valves have been regularly oiled and cleaned during the year.

An estimate for inserting efficient air valves and blow-off valves in the old 30-inch and 36-inch mains will be found in the list of estimates, and explanations of the same will be found farther on in this report under the title "Explanations of estimates."

THE AQUEDUCT LANDS.

I have been unable during the year to extend the surveys of the aqueduct lands beyond Griffiths Park bridge, or bridge No. 3, but it is my intention to recommence these surveys early in the next fiscal year and to carry them as far toward Great Falls as other necessary work and the funds available will allow.

On April 25 notices were served, by direction of the Secretary of War, on the owners of encroachments on the aqueduct lands developed

by previous surveys, to terminate these encroachments within specified times and under specified penalties, as follows: Heirs of the late Mrs. R. Bobinger, E. & E. Baltzley, Ignatius Belt, Benjamin Newman, Thomas Tuohy.

On the application of the first of these encroaching owners, the Secretary of War granted on May 7, 1894, a revocable license to William Bobinger to occupy so much of the land of the United States as is covered by the Cabin John bridge hotel, under the conditions that early steps be taken by the owner to secure the passage by Congress of an act to authorize the sale of the land so occupied, and that if he shall fail to acquire title to said land by April 25, 1895, he shall terminate his encroachment within three months of the said date. If the encroachment be not terminated within said period, the removal of the encroachment may be effected by the United States at the expense of the owners without any right to damages by him on account thereof, and any sum that may be expended by the United States for this purpose shall be repaid by the owner on demand. A bill (S. 2118 and H. R. 7502) has been introduced at the current session to authorize the Secretary of War to sell the land in question, and the bill is now pending.

THE BRIDGES.

The wrought-iron riveted girder bridge north of the M street highway bridge, which carries the 48-inch main across Rock Creek, was thoroughly tightened up, and it and the Pennsylvania avenue aqueduct bridge were painted. Both of the bridges are in excellent condition.

The pavements of Griffiths Park bridge (bridge No. 3) and Cabin John bridge (bridge No. 4) remain in the same condition as previously reported, but it is expected that an appropriation will be made at this session of Congress for the repavement of these bridges, there being an item of $5,000 for this purpose in the District bill as it has passed the House of Representatives.

Estimate for replacing the wooden bridge over the spillway at the Dalecarlia receiving reservoir by a stone structure, commensurate in durability and appearance with the other bridges on the line of the aqueduct, will be found in the list of estimates appended hereto, and explanations of the same will be found farther on in this report under the title "Explanations of estimates."

FILTRATION.

Every year, mainly in the latter part of winter, when heavy rains falling on the watershed of the Potomac and its tributaries, especially the Shenandoah, have found the ground loosened up by alternate thawing and freezing, there are complaints in the press and elsewhere of the tur bidity of the Potomac water, and the same is true in the spring when the fields are plowed. They are not very urgent or very prolonged, for the water soon clears; but as they are sometimes accompanied by demands for the filtration of the water, it may be well in this, my last annual report before retirement from active service, to describe in a general way the two systems of filtration now in use, and to state what would be approximately the cost of each if applied to the water furnished to the District of Columbia.

I should first remark that it seems to be commonly believed that, although our Potomac water is often so offensive to the eye as to make

it appear to be unfit even for bathing purposes, it does not contain germs of disease and is not unhealthful. I think this belief is well founded.

Dr. Busey, president of the Medical Society of the District of Columbia, in an address before the Appropriation Committee of the House of Representatives on June 14 last, urging an appropriation for improving the sewerage of the city, stated that no germs of typhoid fever (one of the most dreaded of all disease germs) have been found in Potomac River water; and Dr. Wales, recently director in charge of the Museum of Hygiene of the Navy Department, at which daily analyses of the water have been made for some years, in a communication to this office giving the results of a chemical examination of Potomac water when it was in a turbid condition,* stated that although there are found in Potomac water five forms of micro-organisms, they are all innocuous and are generally present in all river waters, and added that he regarded this water, "after a careful study for three years bacteriologically, as good as any river water in the world."

That the Potomac water should compare favorably with the best river waters might be inferred from the character of that part of its watershed that is above Great Falls. From the source of its north branch, in western Maryland, about 200 miles above Great Falls, following the course of the stream; from the source of the south branch in West Virginia, about the same distance, and from the source of its principal tributary, the Shenandoah, about 175 miles above the falls, the country is mostly wooded and mountainous. There are but a few large towns on these streams, and I understand that none of them are sewered. The following table gives the names of the principal towns above Great Falls on the Potomac and its tributaries, with their population and distances above the falls, stated approximately:

The foregoing table, when considered in connection with the wellestablished fact that rivers, especially those that flow over rocks aud dams and those that have wide surfaces exposed to the sun and air, tend to purify themselves, gives assurance that under ordinary conditions we have little to fear from our Potomac water.

As an example of the self purification of rivers, I may mention the river Limmat in Switzerland, which, receiving at its upper end the