[Vol. 10 



328 



ANNALS OF THE MISSOURI BOTANICAL GARDEN 



HiO). The (NH 4 )iS04 in the acid was then returned to the 

 flask and digestion effected as in the Kjeldahl-Gunning method, 

 5 grams of powdered K 2 SOi being used to raise the temperature of 

 the digesting liquid. Before distillation 25 ml. of 4 per cent 

 K,S solution were added to precipitate the mercury from mer- 

 curammonium compounds. Where the determination did not 

 involve nitrates the Gunning modification of the Kjeldahl 

 method as given on page 7 of 'Official Methods' (Association of 

 Official Agricultural Chemists, '21) was used. Trouble from 

 bumping during distillation was entirely overcome by the addition 

 of approximately a tablespoonful of glass beads and a gram of 

 powdered pumice stone to each flask. In my hands the official 

 Gunning method modified to include the N of nitrates (page 8 

 of 'Official Methods') was entirely unsatisfactory, even with the 

 utmost precautions; particularly if the "hypo" was added all at 

 once a brownish gas could frequently be seen above the liquid, 

 indicating a loss of N. This was to all appearances overcome by 

 adding the thiosulphate slowly and shaking continuously during 

 the addition, but the results on known solutions of nitrate were 

 very variable. The following show the trend of the results ob- 

 tained by the official method. 



TABLE I 



TOTAL NITROGEN INCLUDING NITRATES 



10 ml. 



5 ml. 



5 ml. 

 10 ml. 

 5 ml. of 0.1 



of . 1 M KNO, 8olution 

 of . 1 M KNOi solution 

 of . 1 M KNO, solution 

 of . 1 M NH 4 NO, solution 



M NH,NO, solution 



9.15 

 4.05 

 4.15 

 17.60 

 9 30 



Mgm. N 





13 . 530 



5.986 



6.134 



26.010 



13.750 



Theory 



14.010 



7.005 



7.005 



28.020 



14.010 



The method outlined at the beginning of this paragraph gave 

 dependable results and was much more rapidly carried out than 

 the official method. 



So far as I have been able to determine, there is no satisfactory 

 method for the determination of nitrate N in the presence of am- 

 monium and other forms of N and organic substances. Scales' 

 ('16) method of reduction with Zn-Cu couple gives accurate 

 results, if great precaution is taken with the couple. However, 



