difficulties and they are thought to be amply large to cover any probable method of construction.

By increasing the height of the Phillipsburg dam, thereby raising the water surface to 422 feet above sea level, the flooded area will be 58 square miles, and the available storage capacity of the reservoir, when drawn down to elevation 402, will be 219,000,000,000 gallons; with this available storage the minimum yield in dry years will be about 417,000,000 gallons daily. When the reservoir is full the total amount of water impounded will be 387,000,000,000 gallons, and when drawn down to its lowest level to yield 417,000,000 gallons daily the water will stand twelve feet deep over the entire flat portion of the valley.

This reservoir would be the largest artificial reservoir in the world, and would take, after closing the gates, about 500 days to fill.

By diverting the waters of Shawangunk Creek into the reservoir through a tunnel 18 feet in diameter, leading from Bloomingburg to Mechanicstown, a distance of a little over six miles, 47 sq. miles may be added to the watershed and the Wallkill reservoir may be made to yield about 460,000,000 gallons daily in dryest years, and over 500,000,000 gallons in ordinary years.

In developing this Wallkill project to its full capacity, it is assumed that the filter plants, aqueduct lines and distributing reservoirs at New York are duplicated. It will then also become necessary to purify the sewage of Goshen and Middletown, as the entire flow of the stream will be diverted to maintain the daily draught of 460,000,000 gallons, and the discharge of sewage into the open dry channel of the river would soon create offense. The costs of these works are all included in the estimates.

It may be said, against this project, that the construction of the Phillipsburg dam will cause the flooding of lands in New Jersey, and that objections might be raised in that State. It may be answered, however, that the Wallkill, although rising in New Jersey, flows into New York State, and that New York is entitled to the flow of the river, in the same manner that New Jersey claims the right to the waters of the Ramapo, Mahwah and other streams rising in New York, but flowing into New Jersey. It would seem, in view of the importance of the streams in the Ramapo Mountains as possible future water supplies for the

metropolitan district of New Jersey, that an amicable agreement could be reached between the two States that would make it possible for each to avail itself of the needed watersheds in the other State, instead of both being deprived of these benefits by reason of inability to exercise the right of eminent domain in the acquisition and holding of property for reservoir sites.

In such a transfer of rights, New Jersey would certainly not be the loser, as the lands desired for the Wallkill reservoir are now largely valueless for farming purposes.

The reservoir would be a lake varying from one to five miles in width, having an area of nearly 60 square miles, dotted with numerous small islands. The shores would be steep and it could be surrounded by fine, level roads, the length, around the water's edge, being about 70 miles. The lake would be surrounded on all sides by high hills, and could, at slight expense, be made a beautiful metropolitan park district.

A concrete idea of the enormous size of this reservoir, a reservoir which could be created by the construction of one small dam, may be gained by a study of the following table, which exhibits, by the lengths of the lines opposite the names of the reservoirs, their relative capacities. The largest existing reservoir in the world is the Periyar, in India, used for storing water for the purposes of irrigation. Its total capacity is only about onefourth the total capacity of the proposed Wallkill reservoir:

VII.

ADDITIONAL SUPPLY FROM LONG ISLAND.

In January, 1896, Mr. I. M. de Varona, Engineer of Water Supply of Brooklyn, submitted a report on the extension of the present Brooklyn Waterworks to secure an additional supply of 100,000,000 gallons per day. Mr. de Varona's investigation was carefully and thoroughly made, and covered this field of inquiry so completely that an abstract of his report, here given, will show the possibilities of this project.

It is proposed to use the surface flow of eleven streams flowing into Long Island Sound, draining about 200 square miles of land in Suffolk County, 30 to 55 miles from the present Ridgewood Reservoir.

The summer of 1894 is shown by the Brooklyn rainfall record to have been the dryest in 70 years. The minimum yield of the eleven streams above-mentioned during that summer was 80,000,ooo gallons per day. Mr. de Varona proposed to obtain 20,000,000 gallons daily additional by means of driven-well plants, four of which were to be placed about equal distances apart along the line of the proposed conduit in the territory to be developed. The two systems together would represent a development of the territory of 500,000 gallons per square mile per day.

In 1894, the actual yield of the original Brooklyn watershed was 730,000 gallons per square mile per day. Mr. de Varona's assumption of a possible draught of 500,000 gallons per square mile per day would, therefore, seem to be not excessive.

Judging by the published gauging of the streams in the original watersheds, it seems to me that to get 500,000 gallons per square mile per day it might be necessary to develop the ground water supplies more highly than was proposed.

Mr. de Varona's plan is to allow the water to flow by gravity conduits to the nearest one of three pumping stations, located at Babylon, Connetquot and Connecticut, respectively. At each of these stations the water is to be pumped to tanks placed on the tops of high trestles, whereby it may flow by gravity through riveted steel pipes to Ridgewood reservoir. From Babylon to Ridgewood is 30.4 miles, and the water is to be conveyed between these points in two 81-inch steel pipes.

The above estimates are based on 1895 prices, and on actual costs of similar works in the Brooklyn system.

A reserve boiler and pump power of more than 50 per cent. was provided for in the estimates, and, based on these considerations, the price of the water per million gallons would be $39.03.

The estimates do not include the cost of a new distributing reservoir.

If it should be found necessary to develop the ground-water supply more highly than supposed, the cost of construction would thereby be considerably increased.

In this connection a few notes with regard to the extension of the present ground-water plants will be pertinent.