12 
BULLETIN" 1308, U. S. DEPARTMENT OF AGRICULTURE 
in volume of cresol-kerosene mixtures and the solubility of kerosene 
in the alkaline aqueous extracts. Summation of the errors for an 
analysis performed as suggested will indicate that the result is likely 
to be in all cases a little too high, the error ranging from about 0.1 
per cent for a low-boiling sample to nearly 0.1 per cent for a high- 
boiling sample. 
Incidentally two additional sources of error are eliminated in the 
improved method: (1) Because in the Weiss-Hill method one 
measurement in the tar-acid funnel was made of oil plus tar acids 
in glass, whereas the second was made of oil over an aqueous film 
adherent to the glass; (2) because, as demonstrated by experiment, 
about two-thirds of any tar bases present, if not first extracted by 
acid, will pass into the alkaline extracts and will thus be reckoned 
as " tar acids." 
The writer prefers the term " total phenols " for designating the 
sum total of the substances extracted by sodium hydroxide from a 
redistilled commercial cresol, and would restrict the term " tar acid " 
to such molecules as contain the carboxyl grouping. Organic acids 
in general readily react with aqueous solutions of sodium car- 
bonate and even of bicarbonate; phenols, on the contrary, are inert 
and even tend to be salted out of aqueous solution by addition of car- 
bonated alkali. The following group of experiments was designed 
to determine whether a separation between true acids and phenols 
might be effected by distillation over sodium carbonate or bicar- 
bonate. 
EXPERIMENT 5 
The distillations and extractions of the samples were carried out 
according to the codified method later described (p. 15), except for 
the modifications noted in Table 2, wherein are also given the data 
obtained upon the effect of the presence of sodium carbonate or 
bicarbonate in the distillation flask. 
Table 2. — Effect of carbonated alkali upon the distillation of acids and phenols 
in experiment 5 
Nature of sample 
Substances besides kerosene added 
before distillation 
Apparent percentage total phenols 
I 
II 
III 
IV 
Cresol L B 
Per cent 
95.6 
96.0 
96.2 
96.0 
97.0 
96.6 
Per cent 
96.0 
95.6 
96.2 
96.4 
97.0 
96.6 
Per cent 
Per cent 
Do __. 
Solution Naj'C03 saturated at 25° 
C ; (I) 2 c. c; (II) 3 c. c; (III) 1 c. c. 
plus 2 e. c. HjO. 
95.8 
96.6 
Cresol H B 
96.4 
Do . 
DryNaHC0i;(I)lg.;(II)2g 
Air-dried powdered Naj CO3, 2g 
2 c. c. coconut-oil fatty acids, (I) 2 g. 
Na HC0 3 ; (II) 2g. air dried Na 2 C0 3 . 
Do 
Do 
From experiment 5 it is apparent that under conditions ordinarily 
prevailing even sodium carbonate is inert toward phenols during dis- 
tillation, but that both sodium carbonate and sodium bicarbonate 
can probably be trusted to hold back true acids of a nature likely to 
be present. In neither of the two samples of cresol did there appear 
