Scientific journals often summarize the contents of significant reports found in other journals as a service to their readers. This is especially useful when the report being summarized is written in another language or, as is the case for this problem, the report is in a journal on the other side of the Atlantic Ocean and had to be transported between continents by sailing ship. A group of scientists in Paris did what is considered the definitively best early measurement of the effect of pressure on the boiling point of water back in 1829. They published their findings in French in a 42 page report in 1830. The report was republished as a 2 page summary in English by a journal in the United States in 1831. Somewhere along the line, the data in the US summary got fouled up. Six values were incorrectly reported. Three of these are minor and hard to detect, but three of them are rather obvious — especially if the data were graphed. The reason these data probably weren't graphed is certainly because graphs of data weren't popular at the time. Also, who has the time to check the results of every single report? Also, also, I have to wonder — am I the first person to realize that this data set has transcription mistakes in it? Why don't you go and figure this out.
Construct a graph of pressure in atmospheres vs. boiling point in degree Celsius. (In the old-timey language of 1831, the degree Celsius was called "centigrade" and the saturated vapor pressure of water was called the "elastic force of steam" or the "elasticity of steam".) Copy the numbers from this text file if you would rather not have to type them yourself when you do this analysis.
Identify the 3 most obviously wrong data points and correct them.