If you inserted that thermometer in a liquid warmer than 200 degrees, the mercury column would be as high as it could go (200), but you know that is not high enough, so you decide to extrapolate. You look at the mercury column and estimate that it could go 3″ higher, if there were room on the device. The length of the mercury column from 150 degrees to 180 degrees is about 3 “, so you assume that 3” is equivalent t0 30 degrees of temperature’ and are thus led to estimate the water temperature to be 230 (200 + 30). And of course you would be completely wrong, because the water would have turned to steam at 212 degrees. Extrapolating can be very misleading.

===============================

You can buy the working paper direct from Cato.org/publications