— 230 - 



Directly, the error can be estimated by comparing the apparatus 

 with another of greater accuracy. Thus I estimated the respiration 

 for leaves of Sambucus to 0.405 mg G0 2 pr. hour pr. 100 cm 2 . At 

 the same time an estimation was carried out with the apparatus, 

 described in my paper (1912 p. 21). The respiration was estimated 

 at 0.426 mg G0 2 . The error in this last apparatus is about + 5 %; 

 the right value for the respiration in the leaves of Sambucus is there- 

 fore 0.46 mg C0 2 and the error in the assimilation apparatus is 

 about 13 %. Further I tried to estimate the C0 2 content of pure 

 anhydrous Na 2 C0 3 . 0.5 g of this substance was dissolved in 1 1. of 



water. 5 cm 3 of this solution + 5 cm 3 ^ HCl was poured into a 

 tube, fitted in place of the receiver, and an air current, free 

 from C0 2 , was passed through the apparatus. In the 5 cm 3 of the 

 Na 2 G0 3 solution was 2.5 mg Na 2 C0 3 , corresponding with 1.04 mg 

 C0 2 . By titration of the content of the absorption tubes I found 

 in two experiments 1.04 and 1.06 mg C0 2 . Here a deficiency of 

 absorption could not be detected. 



2. The accidental errors of the apparatus arise from the titra- 

 tions. Every estimation of the G0 2 assimilation is obtained as a 

 difference between 2 titrations. The standard deviation (<r) on each 

 titration is about 0.07 cm 3 and on the difference 0.1 cm 3 . When 



HCl is used, 0.1 cm 3 corresponds with 0.02 mg C0 2 . The experi- 

 mentation time is in this case 1 hour, and the leaf area should be 

 so great, that about 0.4 — 0.5 mg C0 2 is assimilated. 



Supposing that the error arising from the deficiency of ab- 

 sorption is about 10 %, and the value of the standard deviation 

 0.02, the corrected value for the G0 2 assimilation is (a + 0.1 a) i 

 0.02, when a is the amount of C0 2 , assimilated during the experi- 

 mentation time. The corrected value is then calculated per hour 

 per 100 cm 2 leaf area. (In the tables I — VIII the column "C0 2 assi- 

 milated" indicates the values directly found, increased by 10 %). 



I am inclined to believe that the error on the corrected values 

 will not surpass about 5 % of the value ; at any rate the accuracy 

 of the apparatus can not be said to be very high ; but it should be 

 remembered, that the apparatus from the beginning was adapted 

 for field-work and short experimentation times. The last point is 

 of importance. Frequently experiments of lesser accuracy with un- 

 altered plants are to be preferred to experiments of greater accuracy 

 with plants, that on account of a long experimentation time are 

 altered in uncontrollable ways. On the other hand the accuracy of 

 the apparatus can be increased to any desired degree by using for 



