"In the thermometer which is graduated from 212 deg. (the boiling-point) to 180 deg., similarly to those intended for the purpose of measuring heights, there must have been a starting point, or zero, from which to begin the graduation. I have asked an optician in London how he fixed that zero, the boiling-point. 'By boiling water at my house,' he replied. 'Where is your house?' In such a part of the town, he answered. I said: 'What height is it above the sea?' to which he replied, 'I do not know;' and when I asked the state of the barometer when he boiled the water, whether the mercury was high or low, he said that he had not looked at it! Now, as this instrument is intended to measure heights and to decide differences of some hundred, if not thousand feet upwards, at least one should endeavour to ascertain a reliable starting point. From inquiries made, I believe that the determination of the boiling-point of ordinary thermometers has been very vague, not only from the extreme difficulties of the process itself (which are well known to opticians), but from the radical errors of not allowing for the pressure of the atmosphere at the time of graduation--which may be much, even an inch higher or lower, than the mean, or any _given height_--while the elevation of the place above the level of the sea is also unnoticed. Then there is another source of error, a minor one, perhaps: the inner limit, the 180 deg. point, is fixed only by comparison with another thermometer; it may be right, or it may be very much out, as may be the intermediate divisions; for the difficulty of ascertaining degree by degree is great: and it must be remembered that the measurement of a very high mountain depends upon those inner degrees from 200 deg. down to 180 deg., thereabouts. Hence, the difficulty of making a reliable observation by boiling water seems to be greater than has been generally admitted." =95. Method of Calculating Heights from Observations with the Mountain Thermometer.=--Having considered how to make observations with the proper care and accuracy, it becomes necessary to know how to deduce the height by calculation. That a constant intimate relation exists between the boiling temperature of water and the pressure of the air, we have already learned. This knowledge is the result of elaborate experiments made by several scientific experimentalists, who have likewise constructed formulae and tables for the conversion of the boiling temperatures into the corresponding pressures of vapour, or, which is equivalent, of the atmosphere, when the operation is performed in the open air. As might be expected, there is not a perfect accord in the results arrived at by different persons. Regnault is the most recent, and his experiments are considered the most reliable.