174 PROCEEDINGS OP THE AMERICAN ACADEMY 



up the residue of sodium nitrate, which by its insolubility might effect 

 to some extent the protection of the boric acid from the action of the 

 alcohol, 2 cm.^ of water were introduced and evaporated between the 

 second and third, and again between the fourth and fifth distillations. 



When acetic acid was made use of to free the boric acid, the six 

 distillations with methyl alcohol were made as before ; but, sodium 

 acetate being soluble in methyl alcohol, the intermediate treatments 

 with water were unnecessary. With the fourth portion of methyl 

 alcohol a few drops of acetic acid were added to preserve the acidity 

 of the residue, which, as has been pointed out, tends to become alkaline 

 under the treatment. 



The residues of both processes of treatment were found to be free 

 from boric acid by the exceedingly delicate test with turmeric, care 

 being taken in the series of experiments in which nitric acid was used 

 to oxidize nitrites by means of bromine (expelling the latter before 

 making the test), and in the acetic acid series to acidify with hydro- 

 chloric acid sufficiently to counteract the tendency of the acetate by 

 itself to brown the turmeric on evaporation. 



The lime to retain the boric acid in the distillate was ignited in the 

 crucible in which the evaporation of the distillate was to be made sub- 

 sequently, and then transferred to the receiving flask attached to the 

 condenser, so that the boric acid might be fixed during the distillation. 

 To prevent the caking of the lime by the action of the alcohol, it was 

 slaked with a little water before the distillation was begun. 



In experiments (13) to (16) nitric acid was employed, and in (17) 

 to (20) acetic acid was used, with the precaution noted, to liberate the 

 boric acid. 



In experiments (13) to (16) the mean error amounts to 0.0012+ 

 grm. ; in experiments (17) to (18) the mean error is a little more 



