760 REPORT—1896. 
hemoglobin solution into the bottle. The bottle is then recorked, removed from 
the basin, and shaken for ten minutes, so that a maximum saturation with carbonic 
oxide may be attained. During the shaking the bottle must be covered, as bright 
daylight alters the result very markedly. The blood solution in the bottle is then 
poured out into one of three narrow test-tubes of equal diameter. Into another 
of these test-tubes 5 c.c. of the original solution of blood are measured out with a 
pipette. The third is filled with the same blood solution after the hemoglobin has 
been completely converted into carboxyhsemoglobin by shaking for about a minute 
with coal gas. 
Carmine is now added from a burette to the 5 c.c. of oxyhemoglobin until 
first the tint of the solution from the bottle of air, and afterwards the tint of the 
solution saturated with coal gas attained. Water may also be added if the car- 
mine solution alters the depth of colour of the liquid. From the readings of the 
burette the proportion of carboxyhiemoglobin to oxyhsmoglobin may easily be 
calculated. In practice the estimation may be 2 per cent. too low or too high, 
but this is about the limit of error. 
When 0:09 per cent. of carbonic oxide is present in the air the hemoglobin is 
shared equally between the oxygen and carbonic oxide. The affinity of carbonic 
oxide for hemoglobin is thus about 230 times as great as that of oxygen when 
twice 0:09 or 0:18 per cent. of carbonic oxide is present ; two-thirds of the heemo- 
globin go to the carbon oxide and one-third to the oxygen, and so on. Roughly 
speaking, the percentage of carbonic oxide in the air can be calculated by multi- 
plying the number of parts of carboxyhemoglobin to one part of oxyhzmoglobin 
by 0:09. Thus, if the hemoglobin were found to be 10 per cent. saturated with 
carbonic oxide, then, as there would be to each part of oxyhzemoglobin one-ninth 
of carboxyhemoglobin, one-ninth of 0:09 or 0:01 per cent. of carbonic oxide would 
be present in the air. 
It is evident that the method cannot be used directly when high percentages 
of carbonic oxide are present. The sample in such a case must be diluted. Coal 
gas, for instance, requires dilution to about ;45th with air when this method is 
employed. When the oxygen percentage in the air is much diminished the sample 
must also be largely diluted with air, or a correction made in calculating the result. 
Blood solution was originally suggested by Vogel as a qualitative test for CO in 
air. He used the spectroscopic test, and found he could detect 0:2 per cent. of the gas. 
2. The Detection and Estimation of Carbon Monoxide in the Air by the 
Flame-cap Test. By Professor Frank Crowes, D.Sc. 
The detection of carbon monoxide in the air is mainly of importance on account 
of its poisonous nature when inhaled. It would rarely happen that serious ex- 
plosions arise from its being fired in admixture with air, since a carbon monoxide 
explosion is of a comparatively mild character, and further air only commences to 
be feebly explosive when the carbon monoxide is present in the proportion of at 
least 13 per cent, ; this is an amount which would render the air rapidly fatal to life. 
The introduction of carbon monoxide into the air may arise from leakage of 
many forms of gaseous fuel, such as coal-gas, producer-gas, Dowson-gas, water-gas, 
and flue-gas from smelting works, whether the metal is smelted by the old reducing 
methods, or by the newer method more recently applied by Mr. Mond to the 
smelting of nickel ores. This gas is also produced by the detonation of the nitro- 
cotton explosives, and by the imperfect combustion of any ordinary fuel which may 
occur either slowly or explosively. Hence cases of poisoning by this gas have 
mainly arisen from the ‘gas’ taken from the iron blast-furnace, from water-gas 
either used alone or in the enrichment of coal-gas, from coal-gas leakage, and from 
the ‘after-damp’ of the colliery explosion or ‘ gob-fire.’ It will be seen that this 
insidious poison is, therefore, of not infrequent occurrence in the air. The author 
finds that 0-25 per cent. of the gas can be detected in the air by a ‘cap’ 0°5 inch 
in height over the standard hydrogen flame. ‘This test is, therefore, sufficiently 
sensitive for practical application, and furnishes the most rapid means of detecting 
