286 The Botanical Gazette. [July, 
mustard were used. In his last paper,? he brings together 
the results of several experiments made during the last few 
years, in which is shown a gain in N, both in crop and soil, 
over what was in the seeds and soil at the beginning. Frank's 
methods are not given in sufficient detail to enable one to 
judge of his results critically; in fact this charge has been 
made repeatedly against many of his observations.® He uses 
mainly the indirect method of nitrogen determination, plant- 
ing the seeds in a soil containing a known quantity of fixed 
nitrogen and then determines by analysis the content of the 
soil and crop. If the sum at the end exceeds the total 
amount available at the beginning, he reasons that the plant 
has assimilated gaseous nitrogen. A control pot with un- 
planted soil is usually analyzed to see if there is any change 
in the fixed nitrogen of bare soil. As his experiments are 
usually carried out on unsterilized soil and his unplanted 
check soils often show a gain in fixed nitrogen, there 1s 
hardly any doubt that the N-increase in his experiments with 
non-leguminous plants is in part due to fixation by lower ot- 
ganisms, alge, fungi, or bacteria that are common to the 
oil 
Kreusler® points. out a serious objection to his methods of 
analysis as not sufficiently accurate to discriminate in the case 
of non-leguminous plants where such small increments are 
to be noted. 
In the résumé above referred to two experiments with ys 
leguminous phanerogams are given which were made in 4 
solutely N-free land, the results of which are as follows: 
Sinapis alba (4 plants): grams of N in seed, 0.0012; in crop, 0.0043. 0.2186. 
Solanum tuberosum (4 pieces): grams of N in seed, 0.022; in crop, 9-*" h 
He also describes a still more recent experiment, made wit 
Sinapis alba. In this case he used large bell jars and 
though the plants did not develop normally (they were 
able to unfold their flower buds in this closed space), he fo 
a certain amount of nitrogen fixed. 
The N content at beginning was as follows: 
Three seeds, 0.0009%"; soil, 0. 162%; 
at close of experiment, 
Crop, 0.0507; soil of pot, 0.215%; 
0.195%. a 
7 Bot. Ztg. 51: 150. 1893. 
® Journ. f. Landw. 41: 144. 1893. 
* Bied. Cent. 21: 257. 
un- 
und 
soil of control, 
