200 HEMOGLOBIN. 
Were we to admit the accuracy of the work of all the observers, 
whose results are exhibited on the table on p. 199, we should be forced 
to the conclusion that haemoglobin is a body which does not only vary 
considerably in composition in different animals, but does not possess a 
constant composition even in different individuals of the same species. 
Thus, whilst Kossel found the percentage of carbon in the oxy haemoglobin 
of the horse to be 54 - 87, and the mean of a large number of analyses 
by Kossel, Otto, and Biicheler gave 54 - 68, Zinoffsky, as a result of his 
analyses (only two in number, so far as the carbon and hydrogen are 
concerned !), found the percentage of carbon in the hemoglobin of the 
horse to be 5115 (!!). A body in which the carbon differs by 3-72 per 
cent, in different specimens cannot, it will be argued, be a chemical indivi- 
dual. But to draw this conclusion in reference to hemoglobin from the 
facts in our possession would certainly be an error. The discrepancies 
between the results of the analyses of the hemoglobin of the same 
animals are doubtless due to differences in the purity of the substance 
analysed, and to errors of analysis. The preparation of perfectly pure 
oxyhemoglobin, entirely free from contamination with other con- 
stituents of the blood corpuscles and from products of decomposition, is 
much more difficult than has, until very recently, been supposed. In 
the attempt to purify the substance by crystallising it as frequently as 
practicable, nearly all observers have in all probability decomposed it, 
and have afterwards analysed a mixture of oxyhemoglobin and products 
of its decomposition. How far this is the source of the above discrep- 
ancies must now, in the light of recent spectrophotometry work, be 
carefully enquired into. Moreover, assuming that perfectly pure crystal- 
lised oxyhemoglobin is at the disposal of the analyst, the task of drying 
without decomposing it is one of peculiar difficulty, concerning the 
method of execution of which the chemists who have carried out the 
researches under discussion have been by no means agreed. Thus, whilst 
some (following Hoppe-Seyler's directions) have dried the oxyhemoglobin 
intended for analysis, in the first instance in vacuo at 0° C, and only 
afterwards at higher temperatures, others (Zinoffsky, Hiifner, Jaquet) 
have dried the substance in vacuo at ordinary temperatures (15° to 
18° C), and subsequently at 110° to 115° C. 
It is conceivable, nay probable, that some of the differences in the results 
of different observers may have depended upon the above-mentioned differ- 
ence in the treatment of the substance analysed. But, unquestionably, some 
of the best marked differences must depend upon differences in the 
method of analysis employed (e.g. where one observer determines the X hi 
oxy haemoglobin by Will and Yarrentrapp's method, -whilst another employs 
Dumas' method), and upon accidental errors of analysis, which can easily be 
rendered obvious, by making a considerable number of analyses. 
For instance, it appears to me that the percentage of carbon given by 
Zinoffsky, 1 as representing the proportion of this element in the haemoglobin 
of the horse, must be due to imperfect combustion. "Whilst this observer 
carried out the determinations of iron and sulphur hi the haemoglobin of the 
horse in the most elaborate and perfect manner, making many analyses of 
each of three separately prepared specimens of crystallised haemoglobin, he 
rested satisfied with only two determinations of carbon and hydrogen, and 
two determinations of nitrogen (the latter by the method of Will and 
1 "Ueber die Grosse des Hamoglobinmoleculs," Ztschr. f. physiol. Clara., Strassburg 
1SS6, Bd. x. 
