TABLE 9.
Precipitation in percentages of means within 150 miles of Philadelphia, during two-year periods of minimum precipitation,
the numbers being the average per year.

It seems that the fall at the hospital may be taken at 112, if that recorded by the Signal Service is 100. Then, as the minimum rainfall at the former was assumed to be 33.6 inches, the minimum fall at the latter should be taken at 30 inches per annum.

TABLE 10.
Comparison of Rainfall, recorded by the U.S. Signal Service and the Pennsylvania Hospital, in Philadelphia.

The next step which became necessary was to establish the minimum precipitation upon the watersheds to be investigated. The quantity falling upon them since the beginning [PAGE 278] of the surveys has been reported every year, but not until last year was it practicable to make a comparison, because of the short period over which the observations had extended (see Table, Report, 1885). The rainfall at the various stations was expressed in percentages of the fall recorded at the Government station in Philadelphia. Unfortunately some of the totals were incorrectly printed. They should have been as follows: (See Table 7.)

[Table of Corrected Rainfall Figures]

[PAGE 279] A longer time will be required to arrive at percentages which will represent the true mean values. For the present I have assumed the above results as being the best available data for the purpose. In order to determine from them the quantity of rain representing the average fall upon each of the water-sheds which could be used for storing water, it was necessary to carefully compare their topographical features, their mean elevation, the wooded areas, the relative amount of exposure to the rain-bringing winds and the elevation of the gauges above the surface of the ground. This comparison indicated the results as given by Table 11.

The minimum precipitation on the different water-sheds is given both in inches and in percentages of the rainfall recorded at the Signal Service Station in Philadelphia. The average minimum monthly falls are given in inches. Their relation to the annual fall was obtained from the ratios of the mean monthly to the mean annual fall at Philadelphia, as given in the first column of the table. It will be evident that the figures for each separate month cannot represent the absolute minimum for said month, but only the average minimum, and that therefore the mean monthly stream-flows, which were estimated from these figures, also do not represent an absolute minimum flow during the month, but an average. Inasmuch as the storage reservoirs hold and equalize the flow for over half a year, the latter quantity is the proper one to use in calculation.

It is to be hoped that the rainfall observations on the watersheds in question can be continued, so that these quantities may be established with a greater degree of precision, in order to make it possible to better adjust the size of the storage basins. As the stream-flow is less than the consumption from about May 1 to December 1, the minimum rainfall of this interval should be deduced from the records of the Pennsylvania Hospital gauge, before the sizes of the storage basins are finally determined.

I have appended to this report, as Plate VIII., a specimen sheet of the rainfall charts showing how the records have been [PAGE 281] plotted. They admit of ready comparison, and indicate at a glance the depth of fall by the blue -lines representing each separate fall, and also the rate by the degree of their inclination.

TABLE 11.
Deduced average minimum rainfall on sundry watersheds

I have also appended, as Plates VI. and VII., drawings of the ordinary and of the automatic gauges. The former were made by Messrs. Schultzbach & Co., in Washington, the latter by Messrs. Black & Pfister, in New York.

STREAM-FLOW.

Not until during the present year has it been possible to present the results of the stream gaugings since the beginning of this investigation, because we had been unable to obtain certain necessary measurements of high flows until last winter, which it was essential to have before the high flows of previous years could be computed. In former reports the methods of gauging streams have been described. It therefore remains now only to state the results thereof. The daily flow where gauge stations had been established has been tabulated, and the records are on file in the department. Tables 12 and 13 of this report show the monthly and annual yields of the streams. On Table 13 the first column gives the area, the second the average rainfall, and the last the average flow per second per square mile. I have added for comparison similar data concerning the streams supplying New York and Boston with water, viz., the Croton and Sudbury rivers. For reasons already mentioned the years have been reckoned from October 1st to September 30th. The results of only two such years could be embodied in the table. The first year shows a flow above the average and the second a flow below it. The results of the Sudbury and Croton rivers are derived from observations extending over six years. It is interesting to note that the Tohickon creek gives the greatest average yield per square mile, and the Perkiomen creek above Green Lane the next greatest, while the North East Branch of the Perkiomen gives the least.

TABLE 13.
Annual yield of sundry streams.

[PAGE 283] Table 16 gives the maximum and minimum daily flows that have been observed in our watersheds and in those of the Sudbury and Croton rivers. It will be noticed that the least summer flow is generally found in those streams that also have the greatest winter flow. The Perkiomen at Frederick has the largest summer flow per square mile of any of the streams observed, which no doubt is due to the mountainous and wooded region near its head. The variation in the flow during the different months is .very great. It is apparently even greater here than in the Croton and Sudbury rivers. The Perkiomen has a smaller maximum and larger minimum flow than the Tohickon and Neshaminy creeks, due to the somewhat greater rainfall in the higher altitudes of the Perkiomen watershed, and partly to the greater area of wooded territory, which tends to retain the water and deliver it into the streams more gradually.

Tables 14 and 15 are summaries of the flow in different streams for each month. In one table the flow has been reduced to cubic feet per minute per square mile; in the other, to ratios ?f the average monthly quantities. In these forms the results of our gauges have been compared with the quantities given by Fanning in his treatise on "Water Supply" for the Cochituate and Sudbury rivers in Massachusetts, and for the Croton and West Croton rivers in New York. An examination shows again that our creeks have a larger proportion of flow in the winter months and are dryer in summer than the Massachusetts and New York rivers. This is due partly to the lower latitudes of the former, permitting the accumulation of less snow, but mainly to the larger proportion of cultivated and open ground in our watersheds, which allows the rainwater to run off more rapidly.

Plate IV is appended as a specimen of the stream-flow charts, showing the discharge for every day in the year and having for comparison both the rainfall and the daily temperature plotted. Plates XI, XII and XIII show the various stream gauges which have been previously described.

TABLE 16.
Table showing minimum and maximum daily flows in sundry streams

[PAGE 285] From the facts recited it is clear not only that our streams will require very large storage reservoirs to equalize the flow for a uniform daily delivery throughout the year, but that there will be required a greater proportionate storage capacity than on the Sudbury and Croton watersheds. In order to calculate the required amount of storage the minimum stream flow must first be determined. This quantity, as already said, is obtained from the minimum rainfall upon the watersheds. As this has been given above, it remains to discover the relation between the rainfall and stream flows, or, in other words, the proportion of the rain-water reaching the creeks.

Tables 17, 18, and 19 contain the percentages of rain flowing off the watersheds in question. For comparison I have added the same percentages recorded for the Croton, Cochituate, and Sudbury watersheds. The last line of Table 17 gives the monthly percentages that were finally assumed for our cases. Generally they are practically the same as those found by observation during the last two and a half years, and given in the first line of the table. It was thought well to decrease them somewhat for January, February, and March in the winter, as it seemed probable that our observed results were greater than they would be for a longer term of years. The figures from June to October were likewise decreased for the same reason.

As these percentages were to be applied to the minimum and not to the average rainfall, the question arose whether they would hold good for minimum years. Inasmuch as the difference would be small, and as further time is required to establish a number of points, for instance, the actual minimum rainfall upon the watersheds, which might affect the result in a greater degree, it was deemed sufficient for the present to assume the percentages to be the same. It is evident, however, that they cannot be alike for all of the watersheds under consideration owing to the different topographical and other conditions. Yet with the limited time at our disposal and [PAGE 287] with the rainfall and stream-flow, observations extending over 80 short a time, it was decided to use them only as an average for the whole territory, leaving it for the future to discover the percentage for each special watershed.

TABLE 17.
Average monthly percentages of rain flowing off sundry watersheds

TABLE 18.
Annual Precipitation on Sundry Watersheds, and Percentages of same reaching the Streams.

Table 20 contains the results as derived from the above data. It gives the supposed minimum yield for each stream for each month and the daily average for the year, It is not understood that during anyone month the computed flow represents the least flow that can be counted on during the same, for Table 12 shows that this is not the case, But it is understood that for the whole minimum year there will be a general distribution resembling that which is given. As it will be [PAGE 288] necessary during such a year to draw water from large storage reservoirs capable of holding more than half a year's supply, the particular flow during one month is of no importance as compared with the average flow for several months.

The computed average minimum daily yield in million gallons from each watershed is as follows; Tohickon, 86.5; Neshaminy, 112.1; Perkiomen at Frederick, 129.3; Perkiomen at Green Lane, 60.1; East Swamp creek, 40.3; West Swamp creek, 43.4; North East Branch, 49.7.

From what was said above it will be evident that these values must be too great for some areas and too small for others. This will not be a serious matter, however, for the present purpose, as the average topographical conditions for each of the two gravity schemes remains nearly the same. The Neshaminy resembles the North-east ranch and the West Swamp creek of the Lower Perkiomen, while the Tohickon resembles the Upper Perkiomen. The error that is made will tend to give a greater quantity than would actually be available for the lower and open watersheds and a smaller quantity for the Upper Perkiomen and Tohickon. We can see that this is really the case by comparing the flows for 1885, which closely approach a minimum year, with the flows deduced by means of the assumed percentages from the minimum rainfall. Expressed in million gallons per annum the following table gives the yield of the different watersheds for 1885, the computed minimum yield and the resulting differences.

The Neshaminy, West Swamp creek and North East Branch show that the estimated minimum flows are too great, while those of the Tohickon and Upper Perkiomen are probably not large enough. Rounding off the figures and taking this point . into consideration, we may designate the probable average minimum daily yield in million gallons to be as follows: Tohickon, 90; Neshaminy, 110; Perkiomen at Frederick, 130; Perkiomen at Green Lane, 61; East Swamp creek, 41; West Swamp creek, 41, and North East Branch, 46.

TABLE 20.
Average minimum flow deduced from assumed minimum rainfall

TABLE 21.
Minimum stream flows in million gallons

[PAGE 291] It was found that a uniform delivery of 25,000,000 gallons a day from the Neshaminy dam would supply all the mills below it and obviate the necessity of paying damages to riparian owners. This quantity of water has been provided for in the estimates, and it may be added that the 25,000,000 gallons can, if properly utilized, develop two hundred and fifty horsepower at the dam in descending to the stream. But while it is possible to spare enough water for compensation from the Neshaminy creek, owing to the practicability of pumping any deficiency from the Delaware river at Point Pleasant, it is not possible to spare it in the Perkiomen valley without curtailing the available supply for the city. During the dry seasons the entire quantity would be needed by it. I have therefore included in the estimate for the Lower Perkiomen scheme the value of all the mill privileges between Schwenksville and the Schuylkill river, amounting to $130,000, and in the estimate for the Upper Perkiomen scheme between Green lane and the Schuylkill, amounting to $160,000. These figures are included in the amount given in Table 35 under the heading "Cost of Storage." If a compensation of 25,000,000 gallons daily must be given to the riparian owners, the available amount for the city from the Perkiomen would be reduced to 169,000,000 gallons during the years of minimum rainfall.

The Delaware river below Point Pleasant (see Table, page 351, Report of 1885), having a minimum flow at this point of some 1,500,000,000 gallons daily, would not be damaged by the extraction of 200,000,000 gallons.

STORAGE RESERVOIRS.

From an inspection of the territory and with the assistance of the topographical maps all the available sites for storage or impounding reservoirs were noted. Tables 22, 23, and 24 give a list of the same, with their principal features indicated, together with the cost of storage in each case. Time did not permit of making as detailed a study of the more important [PAGE 292] ones as would be desirable for an accurate estimate of the cost. The capacities were computed from the contour lines as taken from the general maps, and the profiles of the sites for dams were taken in most cases likewise from the contour maps, special surveys having been made only in a few cases. For preliminary estimates, however, the results are sufficiently close. Plate I. is a general map of the entire water-shed investigated, showing all the reservoir sites that .were considered. The total storage capacity in the Tohickon watershed is over 25,000,000,000 gallons, and in the Neshaminy about 23,400,000,000 gallons. In the Perkiomen watershed there is a capacity beyond what could be used, and a selection of the best and least expensive sites was possible. In the Lehigh watershed rough approximations had to be made because there was no time for more detailed work, nor aid the necessities of the case absolutely demand it. The natural facilities for impounding water in most of the valleys are quite good, and the expense is therefore not excessive. The cost of the principal reservoirs, for instance, is as follows:

Tohickon valley at Hancock, about 18,000 mill. gal. at $82.53 per mill:
Perkiomen valley at Green Lane, about 12,000 mill. gal. at $93.88 per mill.
E. Swamp creek valley at Millville, about 8,000 mill. gal., at $103.13 per mill.
W. Swamp creek valley at Zieglersville, about 12,000 mill. gal. at $76.21 per mill.
N. E. branch at Lederachville, about 15,000 mill. gal. at $100.20 per mill.

The average cost of the storage basins in the Croton valley is given at $200 per million gallons, and the estimated cost of the large Croton reservoir, about to be built, as $125 per million gallons.

In order to select the reservoirs that are required from the list contained in Table 22, it is necessary first to ascertain the amount of storage which must be provided in each valley to equalize the flow.

[PAGE 293]

TABLE 25.
Relative consumption of water in Philadelphia.

With the data contained in the previous pages, it would now be possible to estimate the same for the constant delivery of a uniform daily supply. As the supply, however, is not quite uniform, the summer months showing a greater consumption per day than those of the winter, it remains to ascertain what quantity of water is likely to be needed during each month. Table 25 has been prepared for this purpose from the experience gained in Philadelphia during the last six years.The values are percentages of the average daily supply for the year, or the actual number of million gallons per day, if the average daily supply for the year is 100,000,000 gallons. It will be seen that while in August the consumption is 15 per cent. greater in January and February it is 15 per cent. less than the average.

The quantity of water which must be impounded in a given watershed increases in a greater ratio than the supply to be daily furnished. As the latter becomes greater in proportion, not only a larger quantity of stored water must be drawn, but it must be drawn for a longer time, because the period when the stream carries a deficient amount becomes longer. In the Sudbury watershed, in order to furnish 70,000,000 gallons daily, a storage capacity is required of 2,909,000,000 cubic feet, and for 40,000,000 gallons daily, a capacity of only 450,000,000 cubic feet is needed. The reservoir capacity is [PAGE 294] in a ratio of 6 1/2 to 1, while the daily supplies are in a ratio of 1 3/4 to 1.

In the Croton watershed, in order to furnish 100,000,000 gallons daily, a storage capacity of 1,200,000,000 cubic feet is required; for 200,000,000 gallons a capacity of 4,000,000,000 cubic feet, and for 300,000,000 gallons daily, a capacity of 7,300,000,000 cubic feet. The reservoir capacity is in a ratio of 6 to 3-1/2 to 1, while the daily supply is in a ratio of 3 to 2 to 1. It is evident that the expense of storage becomes comparatively great when the amount of water used approaches the total flow of the streams. From the average minimum daily yield of the creeks that we are considering, it will be seen that the entire flow during years of minimum rainfall must be impounded in order to furnish the required supply.

With the above data it is now possible to compute the necessary storage capacity for each valley. Tables 26 to 31 inclusive, give the results for the Tohickon, Neshaminy, Perkiomen at Green Lane, East Swamp, Perkiomen above Schwenksville, and the North. East Branch valleys. Column 1 in each table contains the monthly stream-flow; column 2 the loss by evaporation and percolation from the reservoirs; column 3 the water consumption for each month; column 4 the water added to or drawn from the reservoirs, and column 5 the water stored in the reservoir at the end of each month. Inasmuch as May is usually the first month in which the consumption exceeds the stream-flow, it has been assumed that the reservoir shall be full at the end of April. At the end of November, on the other hand, the reservoirs would be drawn down to the lowest point. To provide for the contingency of an extremely dry summer and fall, and also to prevent the necessity of drawing all the water from the reservoir at any time, it has further been assumed that the amount of water in the reservoir at the end of November should be equal to two months' consumption. [PAGE 295] In order to study this question further, the lowest possible rainfall for the seven months, from May 1 to November 30, should be carefully considered with the aid of the long record at Philadelphia given in Table 5.

With these conditions the tables give the following requisite storage capacities and the greatest mean daily supply for the different watersheds:

Tohickon, 2454 mil. cu. ft., 80 mil. gals. daily.
Neshaminy, 3181 mil. cu. ft., 101.3 mil. gals. daily.
Perkiomen, at Green Lane, 1705 mil. cu. ft., 52.8 mil. gals. daily.
East Swamp creek, 1144 mil. cu. ft., 36.2 mil. gals. daily.
Perkiomen, above Schwenksville, 4829 mil. cu. ft., 151.2 mil. gals. daily.
North East Branch, 1402 mil. cu. ft.,. 43.5 mil. gals. daily.

It will be seen that the Tohickon and Neshaminy, embodying together one project, could not furnish more than 181.3 million gallons daily; the entire Perkiomen, above Schwenksville, together with the North East Branch, not more than 194.7 million gallons; and the Perkiomen, above Green Lane, with the East Swamp creek, only 89,000,000 gallons.

At the beginning of the investigation it appeared probable that a very close discrimination might be required between the different watersheds, because their general character was quite similar. Besides making a careful topographical survey and gauging of the rainfall and stream-flow at as many points as possible, it was thought desirable also to have at hand whatever data might otherwise throw light on the relation between the rain and stream-flow from the separate areas. It was therefore concluded to abstract the following data from the topographical maps, which would assist in this direction. The areas were divided into vertical sections; the first comprising all the territory between 0 and 200 feet elevation; the second that between 200 and 400 feet elevation, and so on, each section being bounded by a 200 feet contour line. This division would facilitate the making of a mean profile of the areas and of their respective surface characteristics, with which a better interpretation of the above relation might be obtained. [PAGE 296] The surface characteristics noted were the areas of the ground slope less than 2 feet per 100, between 2 feet and 20 feet per 100 and over 20 feet per 100; also the areas of the roads, of the cultivated soil, of the wooded and untillable ground and of the swamps and meadows. Tables 2 and 3 contain these data.

In addition to this compilation some field work was undertaken, which in connection with other work, could be done without much expense. Certain areas of, different surface characteristics were staked out, a rain gauge set up in the middle of each, and a meter placed at the lowest point to measure the water which ran off during each rain. A comparison of the general rain and stream-flow with the data in the above compilation would have been very much facilitated by these observations.

Should the progress of the investigation have made it certain that only stored water from the Perkiomen and neighboring watersheds could be used for a future supply, it would have been necessary to enter into the question of storage and available quantity more fully, and these data would have become useful As, however, the economy of procuring the Delaware water at Point Pleasant and the superior quality of the water in the Tohickon ,watershed as compared with the Neshaminy, and particularly of the Lower Perkiomen, became evident, it was not considered essential to spend the necessary time for the comparison outlined above. The deductions which have been made above and the results reached therefrom were considered sufficiently close under the circumstances.

After having determined the amount of storage required, the most suitable reservoirs from among those given in Table 22 were chosen. It was evident that certain reservoirs were absolutely necessary, although their selection incurred either a heavy expense or other disadvantages. For instance, Reservoir No. 7 at Schwenksville had to be selected, although flooding several villages, because the water required a delivery [PAGE 297] at a certain elevation, and Reservoir No. 1 at Sumneytown required a long conduit to deliver the water into the main aqueduct.

Among determining elements also the following should be considered. The larger the reservoirs the better will be the quality of the water. A large surface facilitates wave action, and thereby a better aeration of the water, which is quite essential where the creek water to be stored comes from agricultural areas. Large and long reservoirs act also as excellent settling basins, because the slow velocity of the water passing through them allows the suspended particles to settle. Deep reservoirs, further, keep the water cooler, cause less evaporation, and retard the growth of organic matter. Steep banks allow a minimum amount of surface to be alternately wet and dry, consequently to develop low vegetation which is injurious to health. Table 23 gives the flooded areas for each 10 feet elevation, and permits this point be to readily considered. The lower down the reservoirs are in the valley the more rapidly will rains fill them after having been drawn down.

The geological structure of the valley sometimes has a great effect on its ability to store water. If the stratification across the valley is synclinal, it will favor the retention of the water, while if it is anticlinal it will facilitate leakage. Fissured trap rock which forms the dyke at Schwenksville through which the Perkiomen has worn its path, would allow water to escape more readily than compact rocks. The question of percolation has, however, not been considered a serious one. The water from all the creeks is more or less muddy after rains, and the fine silt will in a short time close the pores of the porous materials and practically make them water-tight. Want of funds precluded a geological survey of the proposed reservoir sites.

On the Neshaminy watershed everyone of the available sites is needed to furnish the required supply. In the Tohickon valley Reservoirs No. 1 and No. 2 were selected as [PAGE 298] being the best. For the Upper Perkiomen scheme it was necessary to consider the proper storage in each of the several valleys. In that of the East Swamp Creek Reservoirs No. 1 and No. 2 were chosen, although not particularly favorable sites. Reservoir No. 5, at Green Lane, was the best one for storing the water of the Perkiomen, and Reservoir No. 8, at Dale Forge, for storing the water of the West Branch. Reservoirs No. 1 and No. 8 were the best in the valleys of the North East Branch and the West Swamp creek. Reservoir' No. 7, finally, was necessary to store the water above Schwenksville.

Table 32 gives the list of these reservoirs, with their capacity and cost. It will be seen that the average cost per million gallons is $122.70 for the Tohickon, $165.89 for the Neshaminy, $133.61 for the Upper Perkiomen, $135.31 for the Perkiomen above Schwenksville, and $100.20 for the North East Branch. A brief description of the selected reservoirs follows:

Reservoir No. 1 of the Tohickon watershed was located at a point about one mile north of Point pleasant, where the valley is quite narrow, and separated from the Delaware river by a distance of only 1700 feet, which makes its location favorable for an extension of the aqueduct further up the Delaware river. The dam will also serve the secondary purpose of a crossing for the aqueduct when extended, thus saving the expense of syphoning [sic]. The height of the dam at the deepest point is about 150 feet, including the foundation, and the extreme length is 946 feet. The flooded territory covers an area of 316 acres, which is about one-half covered with timber, and is of little use for cultivation, owing to the steep rocky nature of the ground. There would be flooded: 1 grist-mill, 1 grist and saw-mill, 6 dwellings, and 2 barns. There is also 1 grist and saw-mill below the dam, which would have to be abandoned.

Reservoir No. 2 is located on the Tohickon creek, just below the mouth of Haycock run, and forms a very large reservoir, being capable of storing about 18,000 million gallons, [PAGE 299] and extending back 7 1/2 miles. For the most part the valley is favorable for a reservoir, the slopes drop off quickly and the valley widens above the dam to large proportions. The territory to be flooded is mostly under cultivation, the wooded area not being more than about one-third of the whole. The dam is 100 feet high above the creek, and 1,510 feet long. It floods an area of 1,829 acres, with 7 grist and saw-mills, 2 creameries, 1 tannery, 35 dwellings, and 27 barns.

It was found that while there was an abundance of storage capacity on the Little Neshaminy, only a portion of the flow could be stored in the valley of the Big Neshaminy without going to a great expense. By a fortunate circumstance it is practicable to store the water of the latter in the reservoir of the former by connecting the two valleys with a short tunnel 12 feet in diameter. Reservoir No. 1 is situated on the Little Neshaminy, three-fourth or a mile above its mouth. It has a favorable site for a dam, as the sides of the valley approach each other sufficiently to make its extreme length 1,550 feet, and extreme height 98 feet, while the valley opens out to three-fourths of a mile in width. The slopes of the reservoir vary from steep to nearly level, but for the most part they are steep. The dam backs water over 2,531 acres, of which 189 acres are wooded, and it floods 88 buildings, as follows: 4 grist-mills, 2 saw-mills, 2 school-houses, 2 chapels, 1 creamery, 41 dwellings, and 36 barns. Owing to the great width between the banks, and the bays formed by tributary creeks, the cost of changing the location of the roads and of bridging is very large.

The Big Neshaminy is such a wide and open valley throughout that it was difficult to decide on a suitable location for a dam. The point selected is 1 3/4 miles above the forks. The right bank of the valley rises up almost perpendicularly over 100 feet, but the left bank rises gradually at a grade of about 7 feet per hundred to the proposed height of the top of the aqueduct, from where it continues nearly level for a distance of over 4,000 feet, thus requiring a very long dam. It [PAGE 300] is proposed to build 1,200 feet of masonry across the valley, and the remaining 4,725 feet of earth. The greatest height of the dam, including foundation, is 89 feet, and the area flooded covers 2,273. acres, of which 203 are wooded. The territory flooded has a long, irregular shape, its length is about 11 miles, and it reaches as far back as New Britain. The slopes of the valley average from 5 to 12 feet par 100 over a surface that will alternately be covered with water and again exposed, except at the extreme upper end, where level and shallow areas occur that will have to be kept flooded by subsidiary dams. The storage capacity of the reservoir is not more than one-half of the size required to store the minimum flow of the stream, the remainder being provided for in Reservoir No. 1 on the Little Neshaminy. Seventy-two buildings will be flooded, viz.: 7 grist and saw mills, 1 store, 1 schoolhouse, 38 dwellings, and 25 barns.

Reservoir No. 3, on the north branch of the Neshaminy, floods an area of 369 acres, which is nearly all cultivated. The average slopes of the sides of the valley are about 5 feet per 100. The height of the dam is 47 feet, and its length 1,420 feet. The sites of 1 mill, 7 dwellings, and 2 barns will be submerged.

Reservoir No. 1 of the Upper Perkiomen is located on the East Swamp creek, above Sumneytown. The dam is 71 feet high and 1,030 feet long, and floods a valley of 196 acres heavily wooded. The sides of the valley are steep and in some places precipitous, forming a deep and narrow reservoir. The country is of little value for farming purposes. The dam floods 2 mills, 8 dwelling-houses, and 2 barns. If not used in connection with the Lower Perkiomen scheme it will be necessary to connect this reservoir and Rich Valley creek with the main aqueduct at Green Lane or Perkiomenville by constructing a branch conduit.

Reservoir No. 2 is located on the East Swamp Creek below Millville, and covers over 1,648 acres, of which 232 are wooded. The valley at the lower end has steep slopes, but the upper [PAGE 301] portion is comparatively level, so that about two-thirds of the reservoir has a shallow depth varying from 10 to 20 feet, which is exposed when the water is drawn down. The dam is 55 feet high, including foundation, and 800 feet long. It floods 5 grist and saw mills, 1 hotel, 24 dwellings, and 7 barns.

Reservoir No. 5 is located on the Perkiomen creek just above Green Lane. The site is a very favorable one for a dam, the valley being narrow and steep at this point, and widening out above it to large proportions. The territory flooded covers an area of 1,705 acres, of which 209 are wooded. The reservoir slopes are steep for the most part as far up as Red Hill. Above this point they begin to flatten, and in some cases become nearly level, forming shallow areas from 10 to 15 feet deep, exposed during low water. Seventy-nine buildings will be flooded, viz., 6 grist and saw mills, 36 dwellings, and 37 barns. The height of the dam, including foundations, is 95 feet, and the length is 634 feet. About one-third of the territory flooded is good farming country.

Reservoir No. 8 is situated on the west branch of the Perkiomen creek near Dale Forge. It is the highest of all the reservoirs proposed, being 609 feet above tide-water, and floods only a few important buildings; The height of the dam is 78 feet, and its length 384 feet. The flooded area is 226 acres.

Reservoir No. 1 of the Lower Perkiomen scheme is located on the North East Branch west of Lederachville. The dam is 100 feet high and 4,025 feet long, more than half of which has an average depth of only 8 feet. The dam floods an area of 1,928 acres, of which 117 are wooded. The slopes of the reservoir are generally good, and there are but few shallow places except at the extreme upper end. The area flooded is generally good farming land, with 99 buildings, as follows: 5 grist and saw mills, 1 meeting house, 1 hotel and store, 58 dwellings, and 34 barns. This reservoir is connected with the main aqueduct by an auxiliary conduit nearly one mile long.

[PAGE 302] Reservoir No. 7 is located on Perkiomen creek, above Schwenksville. The dam is to be built in the gorge at Zieglersville station, and floods 2,307 acres. Its capacity is limited only on account of the villages Green Lane and Sumneytown, which are situated on the proposed banks of the same; The available depth of water is only 12 feet. The dam is 99 feet high and 1,430 feet long. Two small villages, viz., Frederick Station and Perkiomenville, are flooded out entirely, and of Sumneytown and Zieglersville the lower buildings.

The slopes of the valley are good, and about one-seventh of the flooded area is wooded. Two hundred and fifty-four buildings would be submerged, viz., 15 grist and saw mills, 1 planing-mill, 1 powder-mill, 4 hotels, 1 tannery, 2 creameries, and the remainder are dwelling-houses, barns, and ice- houses. From the proposed dam to near Green Lane the present line of the Perkiomen Railroad would also be flooded, and an estimate was made for the following new location: Leaving the present line south of Schwenksville, and extending up the valley to the left at a maximum grade of 50 feet per 100, it crosses the proposed reservoir south of Zieglersville, and then extends due north until it again reaches the present line just below Green Lane. In this reservoir, and also in the following one, the change of roads, bridges, etc., will be very costly on account of the large area and configuration of the territory flooded.

Reservoir No. 8 is located on the West Swamp creek, a little over three miles above its mouth. The location for the dam is the most economical one of any that have been proposed. The valley at this point is a narrow gorge, and immediately above it widens out into a large basin of 2,301 acres. The reservoir thus formed covers a flat and nearly level country, so that over a large portion the water is very shallow. The slope of the ground averages not more than 2 feet per 100, which leaves a large area exposed at low stages of the water. The country is a good farming district with very little woods, the latter covering about one-tenth of the area. The dam is [PAGE 303] 85 feet high and 498 feet long. Five grist and saw mills, 1 tannery, 3 stores, 51 dwelling-houses, and 28 barns would be submerged.

WATER-POWER AT POINT PLEASANT.

The economical feature of the project for obtaining water at Point Pleasant lies in the existence of an undeveloped waterpower sufficient to raise into the aqueduct a daily quantity of Delaware water equal to 120 million gallons during the low-water stage. A close examination with a view of utilizing the same was made last spring. The site of the proposed dam is above the bridge; its elevation is assumed at 85 feet above tide-water, which backs the water to the head of Wharford's First Rift, or about one and one-half miles; and gives an available head of 15 feet. The flood waters, based on the freshet of 1862, would raise the level of the pool 20 feet. It would therefore be necessary to raise the tracks of the Belvidere Division of the Pennsylvania Railroad about 10 feet, and to protect the canal at the proposed dam with double gates, to be used in times of extreme high water.

The minimum flow of the river was assumed at 1,500 million gallons per day (see Table, page 351, Report of 1885). Deducting the quantity to be raised into the aqueduct, there will remain enough water to supply power equivalent to 3,640 horse-power. Assuming that the motors employed will utilize 80 per cent. of the theoretical power, there will remain 2,912 actual horse-power. The aqueduct at Point Pleasant is 217 feet above tide-water. Adding for friction, etc., the lift of the pumps would be 137 feet. The pumping mains are 30 - inches in diameter, and the distance to the aqueduct is 600 feet, The velocity in the same is assumed to be 3t feet per second. Computing the loss by friction of the pumps at 3 per cent., it is found that 117,463,000 gallons can be raised into the aqueduct every twenty-four hours during the lowest stages of the river. As it is practicable to supply a much larger quantity of water during ordinary stages of the river, and at [PAGE 304] favorable times to pump into the lower storage reservoir of the Tohickon valley, I have assumed the available capacity of the Delaware river to be 120 million gallons per day with a slight increase of cost.

Table 33 gives the cost of the water-power in detail as prepared by Mr. Harvey Linton, assistant.

TABLE 33.
Cost of Water-power at Point Pleasant.

[PAGE 305]

PERSONNEL.

The following persons have been engaged on the work:

Special Work.
Dana C. Barber, Sanitary Surveyor.
R. H. Sanders, Geologist.
Murray Rush, Appraiser.

Department Observers.
J. Kirk, Forks of Neshaminy, January 1 to May 14, 1884.
J. Wisler, Schwenksville, January 1 to May 1, 1884.
N. S. Renninger, Green Lane, July 24,1883, to April 1, 1884.
G. H. Hart, Pennsburg, September 9, 1883, to June 1, 1884.
G. W. Roth, Ottsville, September 1, 1883, to August 1, 1884.
Thomas H. Walton, Doylestown, October 5,1883, to December 31, 1885.
Dr. J. A. Roth, Seisholtzville, June 1, 1884, to date.
J. H.. Steltz, Green Lane, December 1, 1884, to December 31, 1885.
Edwin F. Heavner, Ottsville, August 1, 1884, to date.
[PAGE 307]
Dr. C. D. Fretz, Sellersville, May 24 to December 31, 1885.
H. L. Shull, Lansdale, May 1, 1884, to date.
George Lowder, Smith's Corner, January 15, 1886, to date.
Albert Stover, Point Pleasant, October 23 to December 31, 1885.

The Department is indebted to the following parties who have kindly furnished rainfall records:

General W. B. Hazen, Chief Signal Officer, Washington.
Serg. T. F. Townsend, U. S. Signal Service, Philadelphia.
Serg .C. H. Kitchel, U. S. Signal Service, Philadelphia.
Serg. L. M. Dey, U. S. Signal Service, Philadelphia.
Mr. E. F. Smith, Chief Engineer Canals, Reading, Pa.
Mr. Thomas Meehan, Germantown, Pa.
Pennsylvania Hospital, Philadelphia.
Mr. Thoman [sic] J. Beans, Moorestown, N. J.
Dr. Charles .Moore, Pottstown, Pa.
Mr. S. B. Lehman, Lebanon, Pa.
Milnor Gillingham, Fallsington, Pa.
Mr. M. McNeill, Princeton, N. J.
Mr. J. L. Heacock, Quakertown, Pa.
Miss Emily Kent, Phillipsburg, N. J.
Prof. S. J. Coffin, Easton, Pa.
Dr. J. C. Green, West Chester, Pa.

The Department is also indebted to the following gentlemen and corporations
for assistance rendered in lending maps, furnishing reports, etc.:

Prof. J. P. Leslie, Geologist, Pennsylvania.
Col. H. M. Robert, Corps of Engineers, U. S. A.
Pennsylvania Railroad Company.
Philadelphia and Reading Railroad Company.
Lehigh Valley Railroad Company.
Joseph S. Harris,. President Lehigh Coal and Navigation Company.
Prof. James Hall, Geologist, New York

[PAGE 308]

It is due to the members of the corps, and particularly to Mr. F. L. Paddock, Principal Assistant, to state that they displayed praiseworthy industry and skill, without which it would not have been possible to complete the investigation within the given time, nor for the available funds.