GASTRO-INTESTINAL FLORA OF NORMAL INFANTS 071 
flora in health is the chemical composition of the ingested food.^ 
Escherich,'- as far back as 1887, clearly showed that a very charac- 
teristic change in the intestinal flora of dogs could })e brought about 
by feeding protein, during which bacteria that li(iuefy gelatin become 
abundant in the feces. 
Assuming that food is an important factor in determining the more 
common types of bacteria found respectively in the intestinal tracts 
of nurslings, artificially-fed children and adults, it would be reas()nal)le 
to expect that the same or similar bacteria should develop in the 
intestinal tracts of experimental animals, provided they were fed upon 
the same foods as nurslings or adults. A prolonged series of experi- 
ments upon monkeys,^ dogs and cats have shown that alternations in 
diet do influence the prevailing types of bacteria in the intestinal 
tract to a marked degree. The essential features of these experiments 
were that monkeys, dogs and cats fed upon cow's milk containing 
sufficient lactose solution to bring the percentage of protein and 
carbohydrate approximately to that of human breast milk excreted 
feces which, in appearance and in bacterial content, approached very 
closely those of the normal hiunan nursling. The acid reaction, 
practical absence of obligately proteolytic bacteria, the dominance of 
B. bifidus and acidophilus and the appearance of Micrococcus ovalis 
in numbers similar to corresponding types in normal nurslings' feces 
were in striking contrast to the feces of the same animal after a pro- 
longed feeding with a purely protein diet. In the latter event large 
numbers of proteolytic bacteria were present in the feces, which were 
alkaline in reaction and rich in indol, phenols, hydrogen sulphide, 
ammonia and other products indicative of intense proteolytic decom- 
position. Obligately fermentative bacteria of the bifidus-acidophilus 
type were few in number, or practically absent. Rettger^ has made 
somewhat parallel observations in mice and rats.^ 
The nature of the dominant organisms which develop in diets rich 
in carbohydrate varies with the carbohydrate itself. B. bifidus is 
more commonly predominant when lactose is the sugar fed, without 
an excess of protein; if dextrin, maltose or glucose is substituted for 
lactose under the same conditions, B. acidophilus is very frequently 
the more prominent. In like manner, the nature of the protein 
influences the types of proteolytic bacteria to a very marked degree; 
in general, animal proteins other than casein appear to encourage a 
somewhat more active proteolytic flora than vegetable proteins. These 
1 Kendall: Jour. Med. Res., 1911, 25, 117; Am. Jour. Med. Sci., 1918, 156, 157. 
2 Darmbakterien, etc., p. 111. 
3 Kendall: Jour. Biol. Chem., 1909, 6, 499. Herter and Kendall: Ibid., 1910, 7, 
203. Kendall: Jour. Med. Res., 1910, 22, 153; ibid., 1911, 24, 411; 1911, 25, 117. 
Torrey: Jour. Med. Res., 1919, 39, 415. 
4 Centralbl. f. Bakteriol., orig., 1914, 73, 362. 
5 It is rather more difficult to replace a proteolytic flora in adult animals by a fer- 
mentative flora than it is in young animals of the same species; the explanation of this 
relative refractoriness to substitution of obligately fermentative types of bacteria for 
the facultative organisms commonly found in the intestinal tracts of the older animal 
is by no means clear. 
