220 BACTERIAL ORIGIN OF GUMS OF ARABIN GROUP, 



experiment was weighed and put into the tube, which was shaken 

 until solution had taken place. The tubes were next sterilised 

 by being steamed once for three-quarters of an hour. The tannin 

 medium is not a suitable pabulum for the majority of bacteria, 

 so that one steaming is practically sufficient. Only once or twice 

 have I seen a mould upon the plates, and many hundreds of these 

 have been prepared. After steaming, the tubes were cooled to 

 50° and poured into 9 cm. Petri dishes. The infecting culture 

 was prepared by growing Bad. acacioi upon the sloped surface of 

 saccharose-potato-agar at 30° for 24-48 hours. This produced a 

 loose aggregation of cells without slime. Some of the culture 

 was picked up with the front of a loop of stout platinum wire 

 and smeared completely over the surface of the agar in the Petri 

 dish. After 24 hours the lid of the dish was raised and the slime 

 uniformly distributed with the same large platinum loop, which 

 had an internal diameter of 4 mm. At the end of three days the 

 slime was scraped by means of a small rubber spade into a 

 counterpoised watch glass and weighed. Two days afterwards 

 the slime which had formed during the interval was similarly 

 removed and weighed. The plates were kept under observa- 

 tion for another two days to be sure that slime-production 

 had ceased. The slimes were weighed upon an open coarse 

 balance to the second decimal place, although the balance 

 was sensitive to 5 milligrams. In the tables that follow, the total 

 weights obtained are multiplied by 5 and expressed as the nearest 

 whole number, thus giving the weight of slime in grams from 

 100 c.c. of medium. In one or two tables the actual weighings 

 are recorded to show how the bulk of the slime is formed in the 

 first three days. 



Several experiments were made during the course of the 

 research, when duplicate plates were used, to see if a larger 

 quantity of infecting material had any tendency to increase the 

 amount of slime. The results showed that the quantity of 

 material had no influence upon the weight of slime. In some 

 cases a little more was obtained, in others a trifle less, in others 

 the weights were the same. 



