328 Transactions of the Royal Society of South Africa. 
of the urinary tract, and are found either alone or associated with the 
pyogenic cocci. In order to ascertain what types were characteristic of 
these pathological conditions a number of coliform bacilli isolated from 
urinary cases were carefully investigated. 
The specimens of urine were mainly from cases of cystitis and pyelitis. 
The samples were centrifugalised in sterile tubes, and from the sediment, 
plates of Endo-agar or MacConkey's bile-salt-ueutral-red-lactose-agar were 
inoculated by successive strokes of the platinum loop. Separate colonies 
were thus obtained, and to ensure the purity of the ultimate culture 
investigated further successive strokes were generally made (from single 
colonies) on another plate, and the final culture made from a single colony 
on the second plate. In some cases agar slope cultures were made from 
single colonies on the primary plate ; this culture was then replated and the 
final culture obtained by subinoculating a single colony. 
One hundred and seven strains of coliform bacilli from 90 cases of urinary 
sepsis were examined. In the majority of instances pure cultures of individual 
types were isolated. It might be expected in dealing with a bacterial group 
such as B. coli that separate colonies similar in appearance might on examin- 
ation prove to belong to different types. I have on several occasions selected 
two or three similar colonies at random from the same plate and found on 
examination that they represented the same cultural type. Not infrequently, 
however, mixed cultures are met with, but it is remarkable that in such 
instances one notices some difference in the colonies, e.g. difference in size, 
thickness, depth of colour on Endo or MacConkey's agar, or complete absence 
of pink coloration. Of course non-lactose- fermenters produce "pale" 
colonies on these differential media, but in primary culture the absence of 
red coloration is no proof that the organism is a non -lactose fermenter. 
The fact that mixed cultures of different coliform species may occur in 
urinary sepsis is of special and obvious interest as regards the preparation 
of autogenous vaccines. 
As in the case of faecal strains there is a distinct tendency towards the 
prevalence of certain types in urinary conditions, and the commonest intestinal 
species are also the most frequent types met with in pathological urines. 
Type 1, Sub-group A, corresponds to B. MacGonkey No. 71 and is the 
commonest coliform species met with (y. Tables I, II, VI). It proved the most 
frequent type (15 strains) in urinary sepsis. The other prevalent varieties 
were types A2 [B. coli communis), A3 {B. vesiculosus), A4 {B. Gri'mthal) A5 
{B. Schafferi), BI {B. MacConkey No. 74). 
Of the 107 strains— 
66 belonged to the sub-group A. 
28 belonged to the sub-group B. 
5 belonged to the sub-group C. 
8 belonged to the sub-group D. 
