The copper oxid should be heated for a considerable time to redness in a muffle with free access of air before using and the copper gauze be reduced to pure metallic copper in a current of hydrogen at a low red heat. The anterior layer of copper oxid serves to oxidize any hydrogen that may have been occluded by the copper. When a sample is burned containing all or a considerable part of the nitrogen as nitrates, the longer piece of copper gauze is used.

160. The Combustion.—The tube having been charged and connected with the pump it is first freed from air by running the pump until the mercury no longer rises in the manometer. The end of the tube containing the sodium bicarbonate is then gently heated so that the evolution of carbon dioxid will be at such a rate as to slowly depress the mercury in the manometer, but never fast enough to exceed the capacity of the pump to remove it. The lamp is extinguished under the sodium carbonate and the carbon dioxid completely removed by means of the pump. The delivery-tube is then connected with the azotometer, and the combustion tube carefully heated from the front end backwards, the copper gauze and coarse copper oxid being raised to a red heat before the part containing the sample is reached. When the nitrogen begins to come off, its flow should be so regulated by means of the lamps under the tube, as to be regular and not too rapid. From half an hour to an hour should be employed in completing the combustion. Since most samples of fertilizer contain organic matter, the nitrogen will be mixed with aqueous vapor and carbon dioxid. The former is condensed before reaching the azotometer, and the latter is absorbed by the potassium hydroxid. When the sample is wholly of a mineral nature it should be mixed with some pure sugar, about half a gram, before being placed in the tube. When bubbles of gas no longer come over, the heat should be carried back until there is a gradual evolution of carbon dioxid under the conditions above noted. Finally the gas is turned off and the pump kept in operation until the manometer again shows a perfect vacuum when the operation may be considered finished. In the manipulation our chief variation from the official method consists in connecting the combustion apparatus with the measuring tube before the heat is applied to the front end of the combustion tube. Any particles of the sample which may have stuck to the sides of the tube on filling will thus be subject to combustion and the gases produced measured. Where it is certain that no such adhesion has taken place it is somewhat safer on account of the possible presence of occluded gases to heat the front end of the tube before connecting the combustion apparatus with the azotometer.

161. Method of Johnson and Jenkins.—In the method of Johnson and Jenkins the principal variation from the process described consists in introducing into the combustion tube a source of oxygen whereby any difficultly combustible carbon may be easily oxidized and all the nitrogen be more certainly set free.[135] The potassium chlorate used for this purpose is placed in the posterior part of the tube, which is bent at slight angle to receive it. The sodium bicarbonate is placed in the anterior end of the tube. The combustion goes on as already described, and at its close the potassium chlorate is heated to evolve the oxygen. The free oxygen is absorbed by the reduced copper oxid, or consumed by the unburned carbon. Any excess of oxygen is recognized at once by its action on the copper spiral. As soon as this shows signs of oxidation the evolution of the gas is stopped. Care must be taken not to allow the oxygen to come off so rapidly as to escape entire absorption by the contents of the combustion tube. In such a case the nitrogen in the measuring tube would be contaminated.

It is rarely necessary in fertilizer analysis to have need of more oxygen than is contained in the copper oxid powder in contact with the sample during the progress of combustion.

162. Calculation of Results.—The nitrogen originally present in a definite weight of any substance having been obtained in a gaseous form its volume is read directly in the burette in which it is collected. This instrument may be of many forms but the essential feature of its construction is that it should be accurately calibrated; and the divisions so graduated as to permit of the reading of the volume accurately to a tenth of a cubic centimeter. For this purpose it is best that the internal diameter of the measuring tube be rather small so that at least each ten cubic centimeters occupies a space ten centimeters long. The volume occupied by any gas varies directly with the temperature and inversely with the pressure to which it is subjected. The quantity of aqueous vapor which a moist gas may contain is also a factor to be considered. Inasmuch as the nitrogen in the above process of analysis is collected over a strong solution of potassium hydroxid capable of practically keeping the gas in a dry state the tension of the aqueous vapor may be neglected.

163. Reading the Barometer.—Nearly all the barometers in use in this country have the scale divided in inches and the thermometers thereunto attached are graduated in Fahrenheit degrees. This is especially true of the barometers of the Weather Bureau which are the most reliable and most easy of access to analysts. It is not necessary to correct the reading of the barometer for altitude, but it is important to take account of the temperature at the time of observation. There is not space here to give minute directions for using a barometer. Such directions have been prepared by the Weather Bureau and those desiring it can get copies of the circular.[136]

The temperature of a barometer affects its accuracy in two ways. First the metal scale expands and contracts with changing temperatures: Second, the mercury expands and contracts also at a much greater rate than the scale. If a barometer tube hold thirty cubic inches of mercury the contents will be one ounce lighter at 80° F. than at 32° F. The true pressure of the air is therefore not shown by the observed height of the mercurial column unless the temperature of the scale and of the mercurial column be considered.

Tables of correction for temperature are computed by simple formulas based on the known coefficients of expansion of mercury and brass. For barometers with brass scales the following formula is used for making the correction:

C = -h t - 28.63 .
1.113t + 10978

In this formula t = temperature in degrees Fahrenheit and h = observed reading of the barometer in inches.