BY JOHN JICLUCKIE. 327 



for nitrate by means of diphenylamme sulphate — no reaction for nitrate was 

 obtained. The bacteria did not flourish as freely as on the serum-agar medium. 

 I came to the conclusion that the solution was not of a sufficiently nutritive 

 value to the bacteria and subsequently employed the following: — Cane sugar 10 

 grams, Amm. sulphate 1 gram, Potassium phosphate 1 gTam, distilled water 1000 

 c.c, Calcium carbonate 1 gram. Three flasks, each containing 100 c.c. of this 

 solution which bad previously been sterilised in the autoclave and cooled, were 

 inoculated from a pure culture of the bacteria, and kept at laboratory tempera- 

 ture for 14 days. Three other flasks were similarly prepared and incubated at a 

 temperature of 26° C. for 14 days. 



Nitrate tests were again applied to a few c.c. of the solution, but there was 

 no evidence of any free nitrate being present in the solution. 



The experiments which I have just described were undertaken in order to 

 demonstrate whether or not the bacteria of Macrosajmo-tubercles are nitrate- 

 forming bacteria; if so, then the cultural solution ought to contain increasing 

 quantities of nitrate. The absence of the blue-colouration on the addition of 

 diphenylamine sulphate demonstrates the absence of nitrate. 



But the bacteria may be able to utilize the atmospheric nitrogen, and although, 

 nitrogen may be absent from the solution at the beginning, there may be a con- 

 siderable amount present at the end. The following solution was then pre- 

 pared: — Cane sugar 10 grams, K. phosphate 1 gram, Mg. sulphate .5 gram, 

 CaCOs 1 gTam, H2O distilled 1000 c.c. 



Three flasks (A), each containing 100 c.c. of solution were inoculated with 

 a pure culture of the bacteria and sterilised. Another three flasks (Bl, B2, B3), 

 each containing 100 c.c. of solution, were sterilised, cooled, and inoculated with 

 a pure culture of bacteria. All flasks were incubated at 30° C. for 14 days; the 

 nitrogen content of each was then determined by the Kjeldahl method with the 

 following results : — 



In the series A only the slightest trace of nitrogen was found to be present. 



In B 1. the nitrogen content was 5.47 mg^s. 

 B 2. ,. „ „ „ „ „ 5.30 mgs. 

 B 3. ,, „ „ „ „ „ 5.23 mgs. 



As there was only a trace of nitrogen in the flasks Al, A2, and A3, there 

 had obviously been a considerable fixation of nitrogen by the bacteria introduced 

 into the flasks Bl, B2, and B3. As there was no combined nitrogen in the 

 culture solution at first, the bacteria must have "fixed" the free nitrogen of the 

 air. 



The bacteria of the root tubercles of Macrozamia spiralis are therefore nitro- 

 gen-fixing forms which live in a symbiotic association with the roots of the host, 

 fix the atmospheric nitrogen, and render it available, in some combined form, 

 for the metabolic processes of the host. 



Summary. 



"Root-tubercles," somewhat similar to those of Cycas revoluta, occur upon 

 many of the seedlings and older plants of Macroi:amia spiralis, particularly 

 about the soil-level. 



The tubercles are seldom present upon the more deeply situated secondary 

 roots, but may be induced to develop by artificial inoculation. 



The tubercles are always negatively-geotropic ; uninfected secondary roots 

 are plagio-geotropic, but after inoculation become negatively geotropic. 



The tubercles occasionally branch dichotomously. 



