i7, i Rustia: Autoinfection during Puerperium 121 
temperature out of the nine with hemolytic streptococcus post- 
partum had streptococci antepartum; one of these was found 
distinctly hemolytic antepartum while the other was not tested 
for hemolysis. Similarly, they found two out of four cases 
with morbid temperature due to nonhemolytic streptococci 
postpartum which had the same microorganisms antepartum. 
Permar (1916) , writing on the bacterial flora in later pregnancy, 
apparently recognized puerperal infection carriers. He states: 
The presence in the vagina of streptococci giving the carbohydrate reac- 
tion of virulent organisms as well as those of less virulent character, but 
corresponding to forms recognized as having definitely invasive qualities 
is comparable to that recognized in the other cavities in which virulent or 
apparently virulent organisms are constantly present without giving rise 
to disease processes. 
Bumm claimed to have found streptococcus in 50 per cent of 
normal, pregnant women, and that 75 per cent of healthy puer- 
peral women were carriers of streptococcus. In his communi- 
cation to the Tenth German Gynecological meeting, at Wurzburg 
in 1903, he says: 
The great problem is this: Are the streptococci which we find regularly 
in the majority of cases of all pregnant and parturient women identical 
with the streptococcus found in septic wound infection only existing in a 
temporarily avirulent stage; and have they nothing at all in common with 
the streptococci causing septic infection except only in outside similarity 
in form belonging altogether to another harmless type of organism? I be- 
lieve that at this stage all the future bacteriologic researches will have 
to be applied towards the differentiation of the various types of streptococci 
found. That streptococci are present in the genitalia can not be doubted 
to-day. The question to be studied now is the relation of such to the 
quality of the streptococci derived from sepsis. 
This problem laid down by Bumm has been the subject of many 
investigations. 
Various classifications of streptococcus were made and several 
varieties of streptococci were observed in the numerous 
pathologic processes; but, while the bacteriologist has largely 
used the carbohydrate fermentation tests, the medical man 
has observed the hemolytic character, so that the results 
of Ihese scientists cannot be compared. In recent years, the 
majority of authors who have used the blood agar plate method 
as a means of primary classification of streptococci into hemo- 
lytic and nonhemolytic strains have further extended the class- 
ification of streptococci already grouped on the basis of 
fermentation tests. In fact, the consensus to-day is that the most 
satisfactory classification of streptococci for practical purposes 
is that of Holman, which consists in a primary differentiation of 
