Building the Filter

Bringing Water to Brookland, Pt. 2

Panoramic view looking northeast over filter beds (c. 1910) Courtesy of the Archives of the Washington Aqueduct
Washington Times July 9, 1899

As I mentioned in my last post on the Lydecker Tunnel, Congress had been looking into a filtration system to clean up Washington’s muddy water supply since the 1880s. Until then, the Washington Aqueduct, which brought Potomac River water to the city from Great Falls, had relied on sedimentation to let impurities settle. But typhoid fever fears at the end of the 19th century brought a new urgency to the search for pure, clean water.

Poughkeepsie, New York, was the first city in the U.S. to filter its water, beginning in 1872. They used what is called slow sand filtration, where untreated water is slowly passed through fine sand and gravel, which traps impurities. First developed in the UK, it was called the English system. Other medium-sized cities followed. Albany had the largest slow sand filtration system, capable of cleaning 15 million gallons per day. A second method of filtration, known as mechanical or rapid sand filtration, was gaining ground in the United States and was called the American system. Mechanical filters rely on the addition of a chemical, alum (aluminum sulfate), to the untreated water to coagulate the impurities for easier filtering. The two competing systems each had fervent advocates, who conferred with the District Engineer, Colonel Alexander Miller in 1898.

Col. Miller listened to the engineers espouse the benefits of the two systems, then took a year and a half to consider the options and do his own testing. He set up two small tanks near his office — one slow sand, one rapid sand — filtered river water through them, then had chemical and bacteriological experts test the results. He also considered the overall costs, with mechanical filtration significantly cheaper and less land-intensive than slow sand. Finally, in March of 1900, he filed his report to Congress. He recommended a rapid sand filtration plant, capable of cleaning 60 million gallons of water per day.

But Miller had only spoken with engineers, and now the medical community weighed in, and they weren’t happy. The response was reported by the Washington Post:

“A vast majority of the medical profession of the District, in fact, almost the entire fraternity, condemns the American system as an experiment, and declares that it would be risking the health of the entire population in the future…”

Using alum to clean water was not a new concept, the Egyptians discovered it in 1500 BCE, but its use on such a large scale was unnerving to both the scientific community and the general public. The Senate’s McMillan Committee on the District of Columbia held hearings and listened carefully to the passionate arguments from both sides. In February 1901 they decided in favor of a slow sand system, with the proviso that a coagulant could be added after further study. The plant would be the largest ever built and was to be designed by engineer Allen Hazen, who had designed Albany’s slow sand plant. Money was appropriated and land on the east side of First Street was acquired. Plans and preparation took two years, and construction was finally begun in May of 1903.

First came the excavation and leveling of the filter site. They had to lower the north end (Michigan Avenue) 10 feet while raising the southern end (Channing Street) 16 feet. This was no easy task. Permission was granted to build a spur rail line from University Station in Brookland, along Michigan Avenue to the filtration site. This allowed builders to more easily bring supplies to the site and remove excavated soil, some of which was used for regrading the Union Station site, then under construction. In the photo below from 1904, workers on the far left are laying track for the spur. The two men in the center are standing atop the site’s concrete mixing plant. Trinity College’s Main Hall, still under construction, can be seen in the background. On the right is the north end of the excavation, with a steam shovel hard at work.

Excavation for the filtration plant in 1904. Click to enlarge. Archives of the Washington Aqueduct

The timeline for construction was tight, so even before excavation was complete, the concrete foundations of the filters were being laid. Altogether there would be 29 one acre filters, 9 on the west side of First Street, and 20 on the east side. This 1904 photo shows some of the filter’s concrete flooring in the foreground, and in the background are numerous columns that would support the vaulted roofs.

Concrete construction of filter beds. Click to enlarge. Courtesy of the Archives of the Washington Aqueduct

The diagram below shows what a pair of filter beds would look like when complete. Each covers an acre of land and contain numerous vaulted bays. Water would be piped from the reservoir to the top of the filters, then percolate through the sand and gravel. A thin layer of biofilm formed on the surface of the sand, and the microorganisms that developed within it did the actual filtering of organic matter. The cleaned water would then exit through a drain at the bottom of the vault. The filters were covered with 2 feet of topsoil, and each bay had a manhole where clean sand could be added.

Source: Robert Sillman Associates, Existing Conditions Assessment & Feasibility Evaluation

All the real filtering was done underground, out of sight. The concrete cylindrical bins that are the most visible part of the filtration plant were for storing clean sand. Here is a view of court #3, showing a sand washer in the foreground with the sand bins lined up behind it. The gates on both sides of the court lead down to the filter beds.

Court #3, facing east. (1904) Click to enlarge. Courtesy of the Archives of the Washington Aqueduct

The sand in the filter beds would need to be cleaned after the impurities built up. For that, workers would use a portable sand ejector that brought the dirty sand to above-ground sand washers. From there the sand would be propelled into the clean sand bin while the dirty water was sent to the sewer. When needed, clean sand would be loaded onto a horse or donkey-drawn wagon that would deposit the sand through the manholes to the filter beds. See the illustration below:

1903 illustration, modified from the report “Purification of the Washington Water Supply,” prepared for the McMillan Commission

Here is a photo of workers shoveling the dirty sand into an ejector for washing. Working in the filter beds was not the most pleasant experience, but leaving the manholes open allowed for light and ventilation.

Evening Star photo from June 12, 1938. Click to enlarge.

Once the water had been filtered, it was piped into the underground Filtered Water Reservoir, which could hold 14 million gallons. Here is a plan for the filtration plant that shows the location of the underground basin.

Source: “Works for the Purification of the Water Supply of Washington, D.C.” American Society of Civil Engineers (1906) Click to enlarge

From there about half of the clean water was sent to the city’s lower elevations through gravity-powered mains, while the other half was sent to the Bryant Street Pumping Station, where it was pumped to higher elevation areas, including Brookland.

The Bryant Street Pumping Station. This beautiful building was designed by Henry Brauns and built by
the George A. Fuller Company in 1904. It is still in service. Carol M. Highsmith Archives, Library of Congress.
Evening Star, October 1, 1905

After nearly three years of planning and construction, the filtration plant was put into operation in the autumn of 1905. The effect on the water was immediately noticeable with a substantial increase in clarity and a drop in colon bacilli. Typhoid fever and malaria epidemics were soon dramatically reduced. As a work of civil engineering, the filtration plant was a great success and played a hugely important role in the health of Washingtonians for the next 80 years.

But even before the plant was finished, more plans were being made, plans that would call for the beautification of this industrial area into something resembling a park. That’s the next chapter of the story.


Back in 2009, I was fortunate enough to go on a tour of the filtration plant remains. It was a fascinating experience, I learned a great deal, and was able to take a few photos. Click to enlarge.

Court #3, looking east
Brick regulator houses were designed and built in 1905 to control the flow of filtered water from the sand filters to the clear water basin. There were originally seven of them, built in line with the sand bins.
The gates to the filter beds had some impressive hinges.
Beyond the gates is a ramp leading down
The interior of the vaults is undoubtedly eerie, with beams of light coming in through the manholes.
The openings to the sand bins were designed to be large enough for a small wagon to fit. Sand would be released from the valve above.
After the wagon was filled with sand, a horse would pull it up this ramp to deposit its load through the manholes.
Court #2, looking east.


Hazen, Allen and E.D. Hardy. “Works for the Purification of the Water Supply of Washington, D.C.” Transactions of the American Society of Civil Engineers 57, 1906

Kockritz, Justin, The Bryant Street Pumping Station and the McMillan Park Reservoir Historic District: A Question of Boundaries, School of Architecture, Planning, and Preservation, University of Maryland, 2009

Scott, Pamela, Capital Engineers, The U.S. Army Corps of Engineers in the Development of Washington D.C.  1790-2004, U.S. Army Corps of Engineers, 2011

Ways, Harry C., The Washington Aqueduct: 1852-1992, Baltimore, MD: U.S. Army Corps of Engineers, Baltimore District, 1996

Existing Conditions Assessment & Feasibility Evaluation, McMillan Slow Sand Filtration Site, Robert Silman Associates, Washington DC, 2014

The History of Drinking Water Treatment, United States Environmental Protection Agency, 2000

History of the Washington Aqueduct (1852-1952), Washington District Corps of Engineers, 1953

McMillan Park Reservoir Historic District, National Register of Historic Places nomination, 2008

McMillan Slow Sand Filtration Plant, Historic Preservation Report for the Proposed Redevelopment of the McMillan Slow Sand Filtration Plant, EHT Traceries, 2014

Purification of the Washington Water Supply: An Inquiry Held by Direction of the United States Senate Committee On the District of Columbia, Government Printing Office, Washington DC, 1903

The Washington Aqueduct, Water Supply District of Columbia, War Department, United States Engineer Office, Washington DC, 1939

Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s