On this page
How do you tell whether a clear, fresh-looking lake has sewage in it? In the 1880s there was no simple test for germs in water. A team in Massachusetts found an answer in ordinary salt, and the laboratory that measured it was run by Ellen Swallow Richards. This post covers her path to science, what the survey did, a worked example using the survey's own numbers, and what students can learn from her.
Who was Ellen Swallow Richards?
The main early source is The Life of Ellen H. Richards, a 1912 biography by Caroline Hunt. It gives her birth as 3 December 1842 on a farm near Dunstable, Massachusetts. She graduated from Vassar College in 1870.
Hunt describes how MIT's faculty discussed her application in December 1870 and recommended her admission "as a special student in Chemistry", adding that admitting women was "as yet in the nature of an experiment". She entered in January 1871. In 1873 she received a Bachelor of Science in chemistry, becoming MIT's first woman graduate, and the same year a master's degree from Vassar. MIT's own Infinite History profile also describes her as the first woman to graduate from MIT and notes that she taught sanitary chemistry. She died in 1911.
What was the Massachusetts water survey?
Hunt calls the survey, which began in 1887, probably the greatest direct contribution of her life to public health. It lasted nearly two years, with monthly analyses of samples from all parts of the state, representing the water supply of 82 percent of the population. Thomas Drown, chemist to the State Board of Health, whose analyses were made in MIT's laboratories, planned it and put Richards in charge of the laboratory and its assistant chemists.
The pace was hard. In March 1888 she wrote to a friend, as Hunt quotes: "We are testing all the public supplies once a month and we are up to 2,500 samples already." Samples had to be analysed within hours of arriving or they were useless. Hunt says more than forty thousand samples were analysed wholly or partly by her; MIT's profile gives a figure of about twenty thousand. The counts differ, but either way it was an enormous body of careful data.
The state's report, Examinations of the water supplies and inland waters of Massachusetts, 1887–1890, records Drown's view of her role: "The accuracy of the work and the no less important accuracy of the records are mainly due to her great zeal and vigilance."
What is "normal chlorine"?
The report explains the idea. Chlorine in water analysis means chloride from common salt. In waters far from people, the chemists found a regular decrease in chlorine going from the coast westward. By testing sources with almost no population around them, the chemists fixed a "normal chlorine" for each region and drew it on a map of the state.
Household waste water carries salt as well. So any chlorine above the regional normal is a sign of past or present sewage, and roughly of how many people live on the land that drains into the water. Unlike organic matter, the report notes, chloride does not change as waste water filters through soil, so it stays as evidence of "previous pollution".
Worked example: reading the Boston numbers
The report gives these values in parts per 100,000:
Normal for the region around Boston's sources (Snake Brook and a reservoir in Ashland): 0.23
Water as supplied to consumers in Boston: 0.41
Mystic Lake, supplying Charlestown and nearby towns: 1.93, where the regional normal was under 0.50
Step 1. Excess chlorine in Boston's water. 0.41 − 0.23 = 0.18 parts per 100,000. The report says this difference represents the effect of the population on the collecting area.
Step 2. Convert to milligrams per litre. One litre of water has a mass of about 1,000,000 mg. One part per 100,000 is 1,000,000 ÷ 100,000 = 10 mg per litre. So 0.18 parts per 100,000 is 1.8 mg of chloride per litre.
Step 3. Compare Mystic Lake. The excess is more than 1.93 − 0.50 = 1.43 parts per 100,000, or over 14 mg per litre. That is about eight times Boston's excess. The report blames tanneries and a large population on that drainage area.
Common mistakes, several of which the report itself warns about:
Trusting one measurement. The report stresses that a single determination cannot tell you a water's real condition, and that chlorine in clean waters varied by 300 or 400 percent with rainfall.
Using the wrong normal. A coastal town and an inland town have different baselines. Always subtract the local normal.
Mixing up chloride and chlorine gas. Here "chlorine" means chloride from salt, not the disinfectant.
For more practice with units like these, see our guide to concentration, molarity and molality.
What can a student take from how she worked?
Richards' great strength was reliability: samples tested on time, results recorded without gaps, methods improved as the work went on. A clever idea like normal chlorine is only as good as the measurements behind it. She also took a path no one had cleared for her and used it to answer a question that mattered to every household.
When a calculation like this one has several unit steps, the Fermi Problem Library has worked problems with a step-by-step solver that explains the reasoning, so you can check each conversion.
Try it now: convert 0.23 parts per 100,000 into milligrams per litre, then into parts per million.
