May, 1923] 
REED-SPECIES CONCEPT 
235 
i.e., B. typhosus. The maximum dilution in which it will act depends upon 
the degree of immunization. The serum, moreover, will not react in these 
high dilutions even with closely related organisms (as the other members 
of the typhoid-co/^ group), and will not react at all with organisms outside 
of this group. If in a particular sample the serum reacts with B. typhosus 
organisms in a dilution of 1-5000, it will react with the closely related 
B. paratyphosus in a dilution of about 1-200, and with B. coli in one of about 
1-20. If typhoid organisms are added to such a serum in excess and later 
removed by centrifuging, both the specific typhoid agglutinins and the 
group agglutinins will be removed with the organisms. Similar treatment 
with the other bacteria of the related group will serve to remove their 
small content of agglutinins but will not have any appreciable effect upon 
the specific agglutinin. 
Such results as this are confirmed by other immune reactions, and 
pathogenic bacteriology is a series of such examples. The recent application 
of this method to the study of the Pneumococci is the basis of one of the 
most illuminating chapters in modern medicine. 
The significance of these results for the present discussion rests upon 
the fundamental chemical nature of the immune reaction, as previously 
noted. On this basis it must be concluded that the typhoid organisms 
react because of chemical identity with the antigenic substances of the 
individuals used in the immunization; closely related organisms react to a 
certain degree, because of chemical similarity, and less closely related forms 
fail to react because of chemical dissimilarity with the original antigen. 
II 
Although not utilized in practical taxonomy in other groups, we have 
ample evidence of the existence of similar distinctions. The extensive 
studies of Nuttall on the immune reactions of vertebrate blood are most 
significant. Nuttall found that when an immune serum containing pre- 
cipitins for human blood serum was mixed with the latter in a high dilution, 
a definite volume of precipitate was formed; but when mixed with other 
sera the precipitate was quantitatively less according to the degree of 
relationship of the animals supplying the sera. 
The work of Uhlenhuth, Wells, Osborne, Gohlke, and many others has 
brought forward an enormous amount of evidence to show that chemical 
differentiation exists between homologous proteins in both plants and 
animals. 
Reichert and Brown have given us some more direct evidences of chem¬ 
ical specificity. Ten years ago they demonstrated that the haemoglobins 
of vertebrates could be distinguished by their crystallography. The results 
of these extensive studies indicate that there is a common structure of the 
haemoglobin molecule, whatever the source of the haemoglobin; that the 
crystals of a genus belong to a crystallographic group which represents a 
