Manhattan. The system was designed for a population of 450,000 people using 20 gallons per day, which would give the requirement of 9 million gallons to be delivered by the Aqueduct. It is interesting to note that this appeared to be perfectly adequate in the minds of the designers who had no idea that the New York met- ropolitan area would exceed 10 million In population in our day and use perhaps five times as much water per capita. After the engineers Canvass White and Major Douglass studied a number of potential water sources in the New York City area, it was decided to build a gravity system utilizing water from the Croton River more than 40 miles away. The advantage of using a gravity system is that it functions without employing any pumping machinery. The quantity delivered in the system is a function of the slope of the conduit, the roughness of the walls and bottom of the aqueduct, and the cross sectional area of the water in the channel. As early as 1735. Daniel Bernoulli, a member of the famous family of mathematicians, dehved the basic equations of the flow in a gravity channel. From the equation and diagram (figure 20) it can ^ -■ eneg.^) HeAD, feet hf ' HMt> jLffSi , feet figure 20: Bernoulli's Equation for Open Channel Row B Cj ■7777777V777777777 \_ iyettsd. / perimeter Cross Section of Channel be readily seen that the Bernoulli equation involves the elevations of various points along the channel, the height of the water flowing in the channel, and the velocity of the water. Later in the century, the French engineer and mathematician Chezy presented his well- known equation for determining the velocity of water in an open channel (figure 21). In the equation, the velocity is equal to a con- stant which is a function of the roughness of the conduit multiplied by the square root of the hydraulic radius and the slope. The hydraulic radius is the ratio of the cross-sectional area of water divided by the wetted perimeter. This relationship is shown in figure 22. Subsequent investigators derived similar formulas for de- termining the hydraulic radius and velocity in the conduit. After the completion of the Aqueduct, Jervis compared the measured flow to the design parameters based on formulas of Prony, Eytelwein and Robison. Baron Riche de Prony was a famous 18th-century hydraulic engineer and director-general of the Ecole Nationale des Ponts et Chaussees, and Johann Albert Eytelwein was a German army engineer who prepared a well-known book on hydraulic engi- neering published in 1801. It should be noted that Robison was an academic colleague of James Watt, the famous steam engine inventor and John Rennie, the British bridge builder. Robison's work was published early in the 19th century and was available to people like Jervis for design purposes.^ The quantity of water flowing in an open channel is the product of the cross section of the water and its velocity. Therefore, in order to provide the 9 million gallons per day, the engineer must establish an approximate cross section of the aqueduct, the rough- ness of the inside surface of the conduit, and the slope. To deter- mine the grade line or slope and its most suitable location, it was essential to survey the entire route. The survey work was undertak- en in 1833-34 on the Croton line under the direction of Major D. B. Douglass and he was joined in 1834 by John Martineau. It was Major Douglass who determined that a gravity aqueduct could be built from a dam on the Croton River to the north bank of the Harlem River at a constant slope of 13V4 inches per mile. Having set the slope, trial designs of various cross sections were evaluated so that figure 21, above left: Chezy's Equation for Row Velocity figure 22, above right: Hydraulic Radius for an Open Channel figure 23, right: T. E. Sickels, Viaduct near Sawmill River, c.1837. watercolor and ink on paper Courtesy Jervis Public Library, drawing #245. Photo: G. R. Farley the conduit selected would yield the quantity of water required. Jervis recorded Douglass' decision: He [Douglass] made the location of the line of the aqueduct from the Croton river to the north bank of the Harlem River 33 miles, and determined the grade of the aqueduct at about 13' /4 inches to the mile. It was, in the main, well located. In regard to plans of work, he proposed a cross section of the masonry of the conduit with which some modification was adopted. So far as I have known, he prepared no specifications for the work that were approved by the commissioners.'^ Jervis noted that Douglass did not proceed with the design of any of the structures for the Aqueduct. Jervis described