109 



As it was desired to send out samples containing considerable carbonate, and such 

 urn- not at hand they were prepared as follows: The sample bottle was weighed and 

 into this was \\ ei-hed 2 grams dry sodium carbonate C. P. then, as rapidly as possible, 

 18 grains of < -ommerc -ial sodium hydrate. The bottles were then stoppered and sealed. 



The analyst \va.- directed to dissolve the entire content of the bottle in carbon 

 dioxid-free water, make up to 2,000 cc and use 50 cc portions for the titrations (0.5 

 gram sample). The n -suits submitted have been multiplied by two and reported in 

 percent in the following table: 



Soda lye. 



The nvulis on sodium hydroxid arc very good. As expected, Method II gives 



1\ higher re-nit- i'-r hydroxid and lower on carbonate than. Method I. The 



rence, ho\ mail. The referee determined carbon dioxid in a portion of 



i In- -ample gravimetrically and found 11.62 per cent and 11.71 per cent calculated as 



sodium carbon.; 



I '.-ing these two indicator* in the same determination, as is done in this method, the 

 :icv would always be to high results on sodium carbonate. Phenolphthalein, . 

 more -en-it ivo to acids, becomes colorless immediately when the solution is 

 neutral, while with methyl-orange the acid must be in slight excess to develop the 

 pink color, tlu> excess required depending on the amount of indicator used and the 

 depih .i olor titrated to. A blank should be made, using the same amount of water 

 and indicator, and deducted in each case when methyl-orange is used. For the 

 determination- in tin- --cnnd report in the table the analyst used normal acid. This 

 may account for the results in -odium carbonate being high, as 0.1 cc normal acid is 

 equivalent to ..\er I per ci-nt sodium carbonate, when operating on 0.5 gram of sub- 

 stain-. 



1'oKMAI IH.HYHK. 



x samples were sent out for analysis, No. 1, a strong solution to be worked by the 

 modified hydrogen peroxid method, and No. '_', a dilute solution to be worked by the 

 nd method. lx>th found in Bulletin 107, revised, page 33. 



Formaldehyde. 



