5. Rubber hose for wrapping around the finger during the taking of the blood sample.

6. Tannic-pyrogallic acid mixture (0.04 gm. of a 1:1 mixture), for producing the color suspension in the diluted specimen of blood. 7. Small bottle of water for diluting the blood.

PROCEDURE FOR TESTING FOR CARBON MONOXIDE

SAMPLING THE GAS

Use any type of narrow-mouth bottle for taking the samples to be tested; preferably the bottles should hold at least 200 c. c., and the neck should be small enough to be closed with a cork or rubber stopper. Take a sample by aspirating the gas to be tested into the bottle in a manner as described under sampling of mine gases. Squeeze the bulb at least fifty times to make sure that the original air in the bottle has been completely removed and replaced with a sample of the gas to be tested. Then close the bottle with a cork or rubber stopper.

ANALYSIS

Make a small puncture wound at the tip of the finger with the hæmospast. If the blood does not flow freely, wrap the finger with the rubber hose, beginning at the base and progressing toward the tip. When a large drop of blood appears, catch it in a spot plate or draw it directly into the dilution pipette to the 0.1-c. c. mark. Dip the tip of the pipette into the container of distilled water, and draw water into the pipette until the solution reaches the 2-c. c. mark, then force it into the bottle containing the sample of gas to be tested. Replace the stopper, and rotate the bottle constantly for 15 to 20 minutes, so that the blood will absorb the CO in the sample of gas. Pour the blood solution into a test tube, add a small portion of pyrotannic acid (an amount which will remain on the point of a penknife blade is enough), close the end of the test tube, and invert the tube three or four times to insure thorough mixing. Place the tube in the rack and let it stand 15 minutes at room temperature, then compare it with the standards by interposing it between them and finding the standard which it most nearly matches. The percentage of CO-hæmoglobin is estimated from the value of that standard. Make observations by daylight, if possible, but not in direct sunlight. Calculate the percentage of CO present from the curves shown in Figure 24.

CORRECTIONS

Be sure that the sample of gas is so large (200 c. c. or more) that the amount of CO absorbed by the blood solution does not materially

alter the percentage of CO in the sample, otherwise a correction must be made.

Temperature and percentage of oxygen in the air also affect the results, but corrections for these are seldom necessary. However, as shown in Figure 24, whenever the oxygen is not below 19 per cent and temperature is between 65° and 72° C., no corrections are necessary. If corrections for size of sample, reduction of oxygen percentage, and change of temperature are necessary, the correct percentage of CO from the saturation of the blood used in the test can be found directly from the curves in the other three quadrants of Figure 24. A typical example of the application of the curves is given at the bottom of this figure.

DETECTION OF CARBON MONOXIDE IN MINES

The Bureau of Mines has already called attention to the futility of attempting to detect CO in the mine by the ordinary test 34 with the flame safety lamp, but some persons still believe the contrary. The bureau has found that the cap or any other change in the flame of the safety lamp from the action of CO is not distinguishable from changes produced by methane (CH). It will be granted that the cap or other change produced by 1 per cent of CH, in the mine air can be seen only with difficulty when the ordinary testing flame is used. Such a proportion of CO is enough to cause unconsciousness in less than 5 minutes and death in 5 to 10 minutes. The bureau made detailed tests to determine the exact effect of CO and CH, on the flame of a lamp. Although chemical tests, such as a blood test or absorption of the CO in cuprous chloride solution in a gas-analysis apparatus, are reliable, in the authors' opinion the use of birds (canaries) is much superior in that a test is quickly made, requires no technical experience, and is exact enough. Mice were formerly used as well as birds, but it has been found, both in the mine and laboratory, that mice are unsuitable subjects and they are now rarely employed in the detection of carbon monoxide. The Bureau of Mines uses canaries only, in mines after explosions or after and during fires, and in the laboratory. Some observations as to their behavior follow: 34a

USE OF BIRDS IN MINES AND THE LABORATORY

Because mice may be slow in responding to the presence in mine air of such small percentages of CO as would cause distress to a man at work, experiments similar to those performed with mice were tried with birds. Canaries were confined in a bell jar in atmospheres containing the following percentages of CO: 0.09, 0.12, 0.15, 0.2, and 0.29 per cent.

Burrell, G. A., The use of birds and mice for detecting carbon monoxide after mine fires and explosions: Tech. Paper 11, Bureau of Mines, 1912, 15 pp.

34a Burrell, G. A. Work cited, pp. 10-11.

After an exposure of 1 hour to an atmosphere containing 0.09 per cent of CO, a bird was not affected to such an extent that it would, if carried into a mine, indicate by its actions the presence of that proportion of CO. Only by close observation could one detect that the bird at the end of an hour felt slightly distressed.

With 0.12 per cent of CO in the atmosphere of the bell jar a bird did not show clearly symptoms of being affected. In about 15 minutes it had lost its liveliness and thenceforth remained comparatively quiet. The bird did not fall from the perch, but close observation showed that it was decidedly weaker at the end of the hour than was the bird placed in air containing 0.09 per cent of carbon monoxide.

In air containing 0.15 per cent CO a bird evinced symptoms of slight distress in 3 minutes. It gasped, gradually became weaker, swayed, and at the end of 18 minutes fluttered from the perch. At the end of an hour it had not lost all muscular power, but showed symptoms of extreme weakness.

In air containing 0.2 per cent of CO a bird showed pronounced symptoms of distress in 11⁄2 minutes; it became very unsteady in 3 minutes and fell from the perch in 5 minutes. After it was taken from the jar it regained its feet in 2 minutes and appeared to be in normal condition in 5 minutes.

In air containing 0.29 per cent of CO a bird fell from the perch in 21⁄2 minutes. When placed in fresh air again it had almost revived in 5 minutes.

Canaries are good indicators of the presence of noxious gases in mine atmospheres, as they quickly show signs of distress when small quantities of CO are present. A bird sways noticeably on its perch before falling, and its fall is a better indication of danger than the squatting, extended posture some mice assume without struggles, attempts to walk, or other preliminary symptoms of poisoning. Consequently birds not only give more timely warning of the presence of small quantities of CO, but exhibit symptoms that are more easily noticed by exploration parties.

The Bureau of Mines has made an extensive series of tests, which show that after canaries and other small animals have apparently recovered from exposure to CO, they may be used repeatedly without losing their efficiency in detecting gas. In these tests the animals were subjected to the action of the gas until they collapsed and then were placed in fresh air. As far as the eye could see they recovered as quickly after they had been exposed many times as after the first time. In the presence of 0.25 per cent of CO, birds and mice collapsed much more quickly than men, but recovered much more quickly. Of all the small animals tested, the canaries seemed most suitable for use in mines.

DETERMINATION OF OXIDES OF NITROGEN IN MINE AIR The authors have performed some experiments to determine the adaptability of Busch and Gutbier's nitron method 35 for the de

Gutbier, A. [The quantitative analysis of nitric acid and nitrates by means of nitrons"] Ztschr. angew. Chem., Jahrg. 18, Mar. 31, 1905, p. 494. Busch, Max, [Oxidation of nitrous acid by hydrogen peroxide. Estimation of nitrates in the presence of nitrates]: Ber. Deutsch. chem. Gesell., Jahrg. 39, March, 1906, pp. 1401-1403.

termination of small quantities of oxides of nitrogen in mine air. Blasting sometimes produces these oxides in harmful quantities. Any nitric oxide when first produced combines with oxygen to form red fumes consisting chiefly of nitric peroxide.

Nitron (diphenylendoanilodihydrotriazol) has the formula C2H1N. The nitrate of this base has the formula C2H1N,HNO,. According to Busch, nitron can be used qualitatively to determine 1 part of HNO3, or nitrate, in 60,000 or 80,000 parts of a solution. The writers are indebted to C. G. Storm, former explosives chemist of the Bureau of Mines, for suggesting an adaptation of Busch and Gutbier's method for the estimation of small quantities of nitrous fumes in mine air. Before the authors made their experiments, Storm obtained satisfactory results in using this method for the analysis of a gas sample containing a large amount (11.1 per cent) of oxides of nitrogen. The sample represented the gases that resulted from burning a gelatin dynamite in an atmosphere deficient in oxygen.

NITRON METHOD

In the authors' experiments measured amounts of nitric oxide prepared in a Lunge nitrometer from pure nitrates were added to bottles filled with air, and the following procedure used by Storm was adopted for estimating the nitric oxide.

By means of a separatory funnel, 10 c. c. of a 5 per cent solution of potassium hydroxide and 10 c. c. of a 3 per cent solution of hydrogen peroxide were poured into each bottle, the hydrogen peroxide to oxidize the nitrous fumes to nitric acid, and the potassium hydroxide to form potassium nitrate. The bottles were shaken frequently to insure complete oxidation and absorption. Next the solutions were washed into 200-c. c. beakers, wash water being used sparingly so that the total volume of solution was not greater than 30 or 40 c. c.. The solutions were then made slightly acid with 10 per cent sulphuric acid solution, a few drops of which at a time were allowed to flow down the sides of the beaker; finally 0.5 c. c. of acid was added in excess of that required for neutralization. These solutions were heated to boiling, and an excess of nitron solution (10 per cent by weight of nitron in a 5 per cent solution of acetic acid) was added. Theoretically, 1 c. c. of nitric oxide requires 0.01344 gram of nitron, or 0.1344 c. c. of a 10 per cent nitron solution. Then the burners were removed and the solutions allowed to cool to room temperature. Frequent stirring of the solutions during cooling hastened the precipitation of the nitron nitrate. After cooling, the beakers were put in an ice bath for about 2 hours, when precipitation was complete.

The nitrate usually appeared as long, fine needles, but when the larger amounts of nitrate were present it had much the appearance

of fine particles of asbestos. When concentration of the solution was carried too far before precipitation, the solution became semisolid, but this condition did not interfere with the formation of the precipitate. When 3 or 4 c. c. of ice water was added and the mixture shaken, the solution again liquefied, the nitron nitrate only remaining as a precipitate. Satisfactory results were not obtained with four samples whose volume of solution was much more than 30 or 40 c. c. before the addition of nitron solution. The precipitates of nitron nitrate in ice-cold solution were decanted into a weighed Gooch crucible with an asbestos mat and subjected to slight suction. The filtrates were used to wash the precipitates from the beakers. After the precipitates had been completely transferred to the filter, they were washed with four or five successive small portions of ice water, about 2 c. c. to a portion, and the precipitate was allowed to suck dry after each addition of ice water. The Gooch crucibles were then dried at 105° C. to constant weight.

The authors obtained the following results when they followed closely the procedure described above:

Results of gas analysis with the nitron method

These results show that with a large enough sample of mine air (5 to 10 liters) detection of quantities of nitric oxide much below that required to produce harmful effects on men is possible by this method. To calculate the volume of nitric oxide corresponding to the weight of nitron nitrate, the following data are required:

Molecular weight of nitron, 312; molecular weight of nitron nitrate, 375; molecular weight of nitric oxide, 30; weight of 1 c. c. of nitric oxide, at 0°C. and 760 mm. pressure, 0.001341 gram; 0.001368 gram of nitron nitrate equals 1 c. c. NO.

PHENOL-DISULPHONIC ACID METHOD

PROCEDURE

The nitron method is not accurate enough for determining the very low percentage of oxides of nitrogen usually found in mine air; therefore, more sensitive methods were investigated. This study resulted in the development of the phenol-disulphonic acid method,36

36 Allison, V. C., Parker, W. L., and Jones, G. W., The determination of oxides of nitrogen: Tech. Paper 249, Bureau of Mines, 1921, 16 pp.