FORMS OF THE COLON BACILLUS 127 



diluted in sterile tap water for plating. Plates were made for the 

 most part in triplicate, though on a few occasions five plates were 

 poured for each dilution. The counts recorded are the mean of all 

 plates of the same dilution, that dilution being chosen which gave the 

 nearest to one hundred colonies per plate. Another 1 c.c. sample 

 was mixed with Congo red solution, and slides were prepared for 

 counting and photographing preparatory to measuring their size 

 and form. The photographs were projected to give a final magnifi- 

 cation of 1 9,000 X; the image of the cells was traced on cards as 

 previously used for the cholera vibrio. In addition to the direct 

 microscopic counts the proportion of cells stained by Congo red 

 was also recorded, and this yielded information of considerable sig- 

 nificance. 



The measurement of the rate of death in a culture, and the 

 determination of the limits of the death phase present some diffi- 

 culties. The death phase has generally been defined as beginning 

 at the point where the number of living cells determined by plate 

 counts begins to decrease, but undoubtedly the factor which causes 

 the cells to die, whatever it may be, actually begins to act upon them 

 earlier than this, at least at the point of inflection in the growth 

 curve from positive to negative acceleration in rate of growth, and 

 possibly earlier. This has been claimed by Wilson, who finds at 

 all stages of growth a discrepancy between the microscopic count 

 and the plate count, which can only be explained by the fact that 

 some of the cells are dying, while the majority of them are still 

 growing. By the time the culture has reached the beginning of the 

 death phase, as defined above, there must be already a considerable 

 number of dead cells present, and this is clearly indicated in my 

 own observations here reported by the appearance in every case 

 of a considerable number of cells stained by Congo red before there 

 had occurred any decrease in the plate counts. On the other hand, 

 when the number of living cells shows a decrease, and the culture 

 may definitely be said to have entered the death phase, some of the 

 cells are still multiplying, as indicated by the increase in microscopic 

 counts, which generally continues for some time longer. 



In these experiments we have three sets of data from which to 

 compute the death rate; the plate counts which indicate the decrease 



