BY H. 8. HALCRO WARDLAW. 519 



the amounts of blood delivered by them. Care was taken, by 

 thorough mixing, to ensure a uniform distribution of the cor- 

 puscles in a sample of blood before withdrawing a portion for 

 analysis . 



The percentages of oxyhemoglobin were determined in two 

 ways: (1) By calculation from the oxygen capacity, taking as. 

 100 per cent, the amount of haemoglobin in blood which combines 

 with 18.5 volumes of oxygen (at normal temperature and pres- 

 sure) per 100 cc. ; (2) colorimetrically, by comparison of solu- 

 tions of known dilution with standard tinted glass in the 

 Miescher — von Fleischl haemoglobinometer. The scale of the 

 heinoglobinorneter was calibrated by determination of the oxygen 

 capacity of samples of blood corresponding to different readings. 

 Five series of determinations showed a maximum range of 

 variation of 5 per cent, from the values obtained by the chemical 

 method for a dilution of blood of 1 in 200. 



Attempts were made to detect the presence of pigments other 

 than oxyhemoglobin by examination of the absorption spectrum 

 of the blood. Metheinoglobin was especially looked for. It 

 was not found possible, however, to detect the small proportions 

 of this pigment in the presence of a large proportion of oxyhemo- 

 globin by means of the spectroscope, as already stated. The 

 effect of various proportions of metheinoglobin on the colori- 

 metric estimation of hemoglobin in the blood was then deter- 

 mined. It was found that the presence of 5 per cent, of 

 methemoglobin could be detected with certainty by this method. 

 For a particular sample of blood, this proportion of methemo- 

 globin altered the reading of the hemoglobinometer from 40.8 

 to 30.3, that is, caused an error of 25 per cent in the determina- 

 tion of hemoglobin, and gave rise to an appreciable change in 

 the tint of the diluted blood. Percentages of methemoglobin 

 well within the range of variation between the results of the 

 chemical and colorimetric methods of determining hemoglobin, 

 would therefore be detected by this means. 



The acidity and reactivity of the samples of blood were deter- 

 mined by the method described by Cullen (loc. cit.). The cor- 

 puscles of the blood were separated from the plasma in a cen- 

 trifuge. The dissolved carbon dioxide was then removed from 

 the plasma by exposing it in a thin layer for several minutes 

 to a pressure of about 6 cm. .of Hg. In this way, uncertainties 

 due to the escape of varying amounts of carbon dioxide during 



