182 CONTRIBUTIONS TO OUR KNOWLEDGE OF SOIL-FERTILITY, Xvi., 



Taking the figures as a whole, there is not sufficient difference 

 between them to justify the consideration that any toxin had 

 been produced by the amcebse, for rises in the dilution-curve are 

 obtained in their absence. In view of later experience, the rises 

 might well have been caused by the reaction of the culture-fluid. 



About this time, it became evident that the method of de- 

 termining the reaction by the use of phenolphthalein, as the 

 indicator, might be faulty when hay-infusion was used, and that 

 results based upon the reaction of a medium such as extract of 

 vegetable-mould did not hold for another medium, such as hay- 

 infusion. The method had been to add a few drops of phenol- 

 phthalein to a portion of the extract, and boil for ten minutes. 

 If the solution became purple, another portion was boiled with 

 acid, and titrated back. If it only became slightly reddened or 

 tinted, it was titrated without boiling with acid. When methyl- 

 orange was employed as the indicator, the reaction-numbers were 

 much higl^er. 



It became necessary to examine the influence of various 

 strengths of hay-infusion, as a direct toxic action had been found 

 in a 01% solution, and had not again been obtained in 0-2% solu- 

 tions. The bottles received 700 c.c. of infusion, 60 grams of 

 sterile soil, and 20 cc. of a starter containing 112,000 mobile 

 amoebae. 



Experiment xxvi. 



