4 
AMERICAN JOURNAL OF BOTANY 
[Vol. 8 
O.I gr. in 2,000 cc. of distilled water, 50 cc. of which were used in the prepara- 
tion of each liter of nutrient solution.) The solution thus prepared was 
divided into two equal quantities in large flasks, 1.5 percent agar being added 
to each, and i percent glucose to one portion. The total nitrogen content 
of the medium with any one nitrogen source should therefore be the same 
per unit weight in the two series, with and without glucose, except for traces 
of nitrogen in the glucose or for slight discrepancies in the actual amount or 
composition of agar added. The nitrogen-containing compounds were not 
dried to constant weight nor was the agar purified in any way, all substances 
being added to the solution directly from the stock bottles. The chemicals 
used were Baker's "analyzed" and Merck's ''highest purity"; the agar was 
of the kind known as "Difco bacto." 
The total nitrogen content of each culture medium was determined by 
actual analysis of weighed portions of that medium. It is obvious that by 
this method any nitrogen introduced with the agar or as impurities with the 
glucose would be completely accounted for.^ 
The 1919 experiment was a partial duplication and an extension of that 
of the previous year and included the following series: 
Series 6. Urea (0.375 gr. per liter) — without glucose. 
Series 6A. Same solution, with i percent glucose. 
Series yA. Ammonium sulphate (0.621 gr. per liter), with i percent glucose. 
Series "jB. Ammonium sulphate (as above), with i percent' mannite. 
Series 8. Ammonium nitrate (0.5 gr. per liter) — no glucose or mannite. 
Series 8^. Same solution, with i percent glucose. 
Series 8B. Same solution as series 8, with i percent mannite. 
Series 9. Calcium nitrate [Ca(N03)2.4H20, 1.475 gr. per liter] — no glucose or mannite. 
Series gA. Same solution, with i percent glucose. 
Series gB. Same solution as series 9, with i percent mannite. 
Each complete solution was placed in the autoclav under 15 pounds' 
pressure until the agar was dissolved. The solution was then filtered 
through absorbent cotton; the filtrate was free from sediment. 
Introduction of the Agar. — The Kjeldahl flasks wer-e cleaned in the usual 
way and dried in the hot air oven. On removal from the oven, cotton plugs 
were inserted in the mouths of the flasks to prevent the entrance of dust. 
Each flask was numbered by means of a carborundum point and weighed to 
within 0.05 gram on a Mackenzie one-pan balance, the cotton plug being 
removed only during the weighing. The flasks were then stored in clean, 
dry boxes until required. 
As soon as the medium for a series was prepared the required number of 
flasks (11 in 1917-18, 24 in 1919) were arranged in one of the special wooden 
racks (see fig. i, Plate I) and 150 cc. of the hot agar solution was added to 
1 Numerous analyses of the agar showed the nitrogen content to be about i mg. for 
the amount present in each culture flask. It will be noticed that the analyses for the total 
nitrogen of the media may not have yielded exactly the calculated amounts, because of 
the moisture present in the nitrogen-containing compounds. 
