318 
Transactions of the Royal Society of South Africa. 
A number of cultural tests had been commonly employed which were 
shown by him to be of little differential value, and he further established the 
importance of a selected though comprehensive series of biochemical 
reactions as type criteria. 
MacConkey (1905), in his first paper dealing with this subject, arbitrarily 
divided the lactose-fernienters into four sub-groups according as they did or 
did not decompose saccharose and dulcite. The first group, represented by 
the classical B. acidi lactici, included those which fermented neither dulcite 
nor saccharose ; the second included those which fermented dulcite but 
not saccharose, e. g. B. coli communis ; the third compi-ised types such as B. 
iieapolitamis which fermented both dulcite and saccharose ; and the fourth 
consisted of strains which fermented saccharose, but not dulcite. 
This classification was of course entirely arbitrary and incomplete. 
In 1909 MacConkey reviewed the whole subject, and indicated that if, in 
addition to the fermentation of dulcite and saccharose, further tests were 
added — effect on adonite and inulin, presence of motility, indol production, 
the Voges and Proskauer reaction — theoretically 128 possible varieties could 
be differentiated. At that time he had examined 497 strains from human 
and animal faeces, water, etc., and of the 128 possible types had met with 36 
varieties differentiated according to their action on (1) gelatin, (2) dulcite, 
(3) saccharose, (4) adonite, (5) inulin, in some instances, (6) inosite, and by 
(7) the presence or absence of motility, (8) indol production and (9) the 
Voges and Proskauer reaction. 
MacConkey had at the same time tested a number of other fermentable 
substances which had been commonly employed, but indicated that no 
further information was to be obtained by the use of sugars, etc., such as 
galactose and laevulose, on which the various sub-groups had all the same 
effect; and in the case of quercite and erythrite found that practically none 
of his strains had any fermentative action. Out of 497 strains examined 
178 were from human faeces, and of the various types noted the most 
prevalent were type No. 71, B. coli communis, and B. vesiculosus (see 
Table I). 
MacConkey's system of classification has been supported and adopted by 
Bergey and Deeham, Clemesha and others, and modified by Jackson, who 
employed the reactions in mannite and raffinose as further differential 
characters. 
Howe, on the other hand, claimed that motility, indol formation, 
mannite and dulcite fermentation were of little value for classification owing 
to the fact that, from the statistical point of view, these reactions showed no 
correlation with one another or with other criteria. Prescott and Winslow 
have also urged the value of the statistical basis for a biological classification, 
and that the characters of these organisms should be considered not inde- 
pendently, but in relationship to one another. 
