As for the second fixed point, it is marked during an experiment with boiling water, performed as follows:--You employ a vessel of tin plate sufficiently high to enclose the whole thermometer; you pour into this vessel distilled water, till it is about an inch deep, and then you heat it. The vessel is surmounted by a cover pierced with two holes, one of which is intended to receive the stalk of the thermometer, the other to allow the steam to escape. When, on continuing the ebullition, you observe that the mercury ceases to rise in the tube, you mark the point at which it has stopped, just as you marked the first point. The last mark indicates the _boiling point_; the one hundredth degree of the centigrade scale, the two hundred and twelfth degree of Fahrenheit’s scale. You transfer to paper the distance which is found between the first point and the second point determined, and you divide this distance into one hundred equal parts, or degrees, for the centigrade thermometer, into eighty parts for the thermometer of Réaumur, and into one hundred and eighty for that of Fahrenheit. If the tube of the instrument is very regular in the bore, the degrees should be equal in length; if, on the contrary, you have been obliged to divide it into parts of equal capacity, you find how many of these parts or little spaces it is necessary to take to constitute one of the above degrees. You find this by dividing their whole number by 100, or 80, or 180, according to the degrees of the scale which you intend to make use of. Thus, if you find between the two points fixed by melting ice and boiling water, three hundred divisions of equal capacity, it is necessary to include _three_ of these divisions in every _degree_ of the centigrade scale. The vessel employed to take the boiling point must be of metal, and its surface should be perfectly clean and well polished, and have no rough points. If sand, or other matters, were permitted to repose on the vessel, and to form asperities, the water would enter into ebullition at an inferior temperature. This operation should, moreover, be performed under an atmospherical pressure, which is indicated by the barometer when the mercury stands at twenty-nine inches and a half. But as this pressure is different according to the elevation of the place of operation, and, indeed, suffers continual variations even in the same place, it follows that the temperature of boiling water is subject to continual changes, and that, in the graduation of the thermometer, it is indispensably necessary to take notice of the height of the barometer at the very moment that the point denoting the degree of boiling-water is fixed upon. You succeed in making the necessary corrections by the help of the following table, which is founded on the experiments of Sir G. Shuckburg and of the Committee of the Royal Society. [See the Table on the opposite page.]