ARNOLD H. EGGERTH AND MARGARET BELLOWS 673 



Cataphoresis experiments with a micro apparatus similar to that 

 of Putter (1921) show that the bacteria in this zone carry a positive 

 charge. 



In Table II, the experiment is repeated in the absence of salt. The 

 general results are similar to those found in Table I, except that the 

 zone of fiocculation of untreated Bacterium coli is found to lie between 

 pH 1.6 and 3.0. It will be noted that when isoelectric gelatin is 

 added to a suspension of Bacterimn coli, the mixture is more alkaline 

 than pH 4.7; this is because the bacterial suspension, though carefully 

 washed and in pure water, maintains a pH of 6.8 to 7.0 even though 

 unprotected from the CO2 of the air. 



The use of indicators in determining the H ion concentration of 

 unbuffered or poorly buffered solutions calls for a word of explanation. 

 Tizard (1910) has shown with methyl orange and methyl red that 

 serious errors can be made. In a series of unpublished experiments 

 made in collaboration with C. B. Coulter, in which all results were 

 checked electrometrically, it was found that a carefully purified 

 methyl red indicator could be used without serious error in ranges 

 from pH 4.6 to 5.4. With the various sulfonephthalein indicators, 

 which can be made up in aqueous solution, it was found that if the 

 dye be adjusted so that its pH is not far from the pH of the solution 

 tested, quite reliable results could be obtained. Thus, brom thymol 

 blue indicator, adjusted so that when viewed in thin layers it matched 

 the pH 6.8 standard, can be used over the range pH 6.4 to 7.2, without 

 an error greater than 0.2 pH. Without adjustment of the indicator 

 for particular ranges, the error may be much greater than this. 



The results obtained in these experiments are strikingly similar to 

 those found by Michaelis and Davidsohn (1912) in the precipitation 

 of nucleic acid (isoelectric at about pH 0.7) by serum albumin, which 

 is isoelectric at pH 4.8. Maintaining a constant concentration of 

 nucleic acid, these authors found a concentration of albumin that 

 produced a precipitate whose optimum occurred at pH 4.1 to 4.4. On 

 lowering the concentration of albumin, the optimum now shifted to 

 the acid side, and the optimal zone was greatly broadened. This is 

 essentially like what we have described in Tables I and II. Michaelis 

 and Davidsohn explained their results on purely electrical grounds. 

 Nucleic acid is negatively charged at all reactions alkaline to pH 



