542 Dr. E. W. A. Walker. [Feb. 18, 
(at 37° C.), and the power of growth in gelatin at 20° C. with or without 
the liquefaction of that medium. It would appear that they regard their 
method as enabling them to subdivide the morphologically indivisible type, 
Streptococcus, into a number of distinct and independent groups or “ pro- 
visional species.” ’ 
It is clear that so simple a method of distinguishing definite varieties 
(if such exist) among the streptococci would possess the greatest value, not 
only from a theoretical and taxonomic standpoint, but also from its enormous 
practical utility and most important bearing both on the diagnosis, the 
prognosis, and the treatment of streptococcal infections.. But while a method 
of such great potential value cannot be lightly set aside or disregarded in 
dealing with the organisms in question, neither can it be accepted with 
reliance without the most thorough and complete investigation. | 
The problem of the identity or diversity of the streptococci which invade 
the human body is one whose importance is almost equalled by its difficulty. 
Many observers have attacked the problem without conclusive result, and 
Marmorek’s (4) contention for the unity of human streptococci still held the 
field successfully at the period when Gordon’s tests were introduced. 
The question had been prominently before my own mind, owing to the 
necessity of endeavouring to determine whether the micrococcus which 
(following Triboulet and Paine and Poynton) I found in a number of cases of 
acute rheumatism, and spoke of as the Mterococcus rhewmaticus, was or was 
not distinguishable from other members of the group Streptococcus. Such 
definite distinction it was eventually found impossible to establish, since 
even the remarkable production of formic acid by the streptococcus from 
rheumatism which Ryffel and I (5) observed proved to be only quantitative 
and not qualitative in character. I was therefore anxious to determine 
whether Gordon’s tests would afford a reliable means of differentiation. 
Clearly the first requisite for test reactions such as those proposed is that 
they shall yield constant and definite results with any given strain of 
streptococcus. Gordon himself fully recognised that “stability of the 
reactions” was important, and gave some attention to the question, but 
regarded it as sufficiently decided when he had ascertained that the 
reactions remained unchanged in observations extending over a fortnight, 
and were only modified in two cases out of eleven (and then only in respect 
of one or two of the reactions) by one passage of the micro-organisms through 
the mouse. 
In their preliminary criticism of Gordon’s tests, Andrewes and Horder 
also emphasised the fact that “constancy of the tests is clearly a cardinal 
point,” and they went on to state that as the result of their own observa- 
