THE ANALYST. 97 BACTERIOLOGICAL, PHYSIOLOGICAL, ETC. Quantitative Estimation of Pentoses in Urine. A. Jolles. (Xeits. a d . them., 1907, 46, 764-771.)-1n a previous paper (Biochewt. Zeits., 2, 243) the author described a method for the detection of pentoses in urine which consisted in forming the osazone from 15 C.C. of the urine, washing the precipitate, and distilling it with a small quantity of hydrochloric acid. The distillate was then' tested with Bial's reagent, which would give a distinct green colour if only 0.05 per cent. of pentose were present in the sample. The author has now applied his volumetric method for the estimation of furfural by means of sodium bisulphite (ANALYST, 1906, 31, 116) to the estimation of pentoses in urine. One hundred C.C. of urine are placed in a flask of about 1-5 litre capacity together with 150 C.C.of hydrochloric acid of 1-06 specific gravity. In a second flask are placed 900 C.C. of water, and the steam generated by boiling this water is paased through the liquid in the first flask. The distillate con- taining the furfural is collected, an aliquot portion of it is carefully neutralised with 20 per cent. sodium hydroxide solution in presence of methyl orange, and 10 C.C. of standardised bisulphite solution are added, the excess of bisulphite being titrated back after two hours with -& iodine solution. The author has proved by a large number of analyses that, in the absence of pentoses, neither healthy nor patho- logical urines yield any appreciable quantity of substances capable of combining with bisulphite when treated in this manner. The highest; error due to substances other than pentoses amounted to 0.032 per cent. (calculated aa pentose), and was caused by the presence of large quantities of glycuronic acid derivatives. I n cases of pentosuria, however, the quantities of pentose found range from 0.1 per cent, upwards. The error mtty be still further reduced by previously boiling the urine with dilute acetic acid. J. F. B.