146 CONTRIBUTIONS TO OUR KNOWLEDGE OF SOIL-FERTILITY, xvi., 



plained; they certainly indicate a degree of toxicity which is not 

 confirmed by the dilution-numbers. 



The numbers Avith Penicillium cladosporioides are peculiar. A 

 flask wdth new moss had been sterilised and allowed to stand for 

 several w^eeks before being infected with Bac. T.P.2. At the 

 end of the incubation-period, the mould was plainly seen growing 

 as a floating mass upon the surface; it had ousted the bacteria, 

 for few bacterial cells were obtained from the fluid. The rise in 

 the numbers upon dilution is typical of a toxin, but the reduction 

 upon boiling is not. Again, the dilution-effect was not obtained 

 in tests 7 and 8. 



On the whole the experiment was unsatisfactory, and con- 

 sequently other media were tried. 



Experiments were made with dilute solutions of nitrogenous 

 salts such as ammonium sulphate, ammonium phosphate, and 

 potassium nitrate, as well as with alkaline salts as potassium 

 phosphate, but there was little sign of any probable formation of 

 toxin with any of them. 



The USE of Collodion-Capsules. 



In an endeavour to improve the experimental method, use was 

 made of collodion-capsules, as it w^as considered that by growing 

 the bacteria outside and the test-organism inside the capsule, the 

 production of toxin might be rapidly determined. Accordingly, 

 capsules were prepared by coating the insides of 3/4 in. test-tubes 

 with 4% agar in w-ater. After drying at 37", they received one 

 or two coats of thick collodion, and a narrow paper- scale was 

 fixed inside near the middle, and a short length of thin tubing 

 with a thread attachment near the top. The separation of the 

 capsule from the tube was effected by filling the tube with water 

 and slowly raising it to near the boiling-point, when the collodion 

 separated easily from the glass. The capsules were washed in 

 changes of water, steamed to get rid of all traces of alcohol, and 

 finally inserted in wide test-tubes, which were plugged. The 

 medium was added, 20 c.c. in the outer tube, and 5 c.c. in the 

 capsule, and the whole was sterilised. The outer liquid was 



