ESTIMATION OF CARBOHYDRATES 227 



II. (a) The Gravimetric Estimation. 



No attempt seems to have been made to estimate sugars by determining 

 the weight of cuprous oxide precipitated until 1878 when experiments were 

 made by Maercker, who found that accurate results were obtained if the re- 

 duction were carried out under definite conditions. His procedure was to 

 boil for 20 minutes an excess of Fehling's solution with a sugar solution of 

 about 0*1 per cent., the total volume of the solution being kept constant ; the 

 cuprous oxide was filtered off rapidly through filter paper, washed free from 

 alkali with boiling water and reduced in a current of hydrogen to metallic 

 copper which was weighed. The possible errors were unreduced cuprous 

 oxide in the filter paper and absorption of hydrogen by the reduced copper. 



Allihn repeated these experiments in 1880; instead of filter paper he 

 used an asbestos pad in a glass tube, which was designed by Soxhlet, and 

 boiled the Fehling solution for 2 minutes. The accuracy of Maercker's results 

 was confirmed and the method has since been known as Allihn's method. 

 Salomon in 1881 again stated that the method was accurate if the solution 

 contained about 0*1 per cent, of sugar. Kjeldahl in 1895 studied the 

 method and stated that the main error was introduced by reoxidation on the 

 surface of the liquid during the boiling rather than by reoxidation during 

 filtration. He recommended passing a current of hydrogen or coal gas through 

 the liquid during the period in which it was heated. 



H. T. Brown, Morris and Millar published in 1897 the results of a very 

 extended study of the methods of estimating sugars in which they pointed out 

 that the chief considerations to be attended to were : (i) the use of Fehling's 

 solution of constant composition, (2) the maintenance of the same degree of 

 dilution in all experiments, (3) the precipitation of an amount of copper which 

 shall fall between certain limits, and (4) an invariable method of determination, 

 both as regards mode and time of heating. 



Their process is carried out as follows : 



50 c.c. of freshly mixed Fehling's solution are placed in a beaker of about 

 250 c.c. capacity and having a diameter of 7*5 cm. This is placed in a boil- 

 ing water-bath ; when the solution has attained the temperature of the 

 boiling water, an accurately weighed or measured volume of the sugar solution 

 is added and the volume made up to 100 c.c. with boiling d stilled water. 

 The beaker is covered with a clock glass and the heating continued for 

 exactly 1 2 minutes. The precipitated cuprous oxide is rapidly filtered through 

 a Soxhlet tube with asbestos pad, washed first by decantation with hot water, 

 then with water, alcohol and ether, and finally dried. The dry cuprous 

 oxide is reduced to metallic copper by gentle heating in a stream of hydrogen 

 and weighed ; or it is oxidised by heating to cupric oxide and wei hed. 



A correction must be made for the reduction which takes place when 

 the Fehling's solution is heated alone. The amount of sugar corresponding 

 to the copper is given in the table on p. 605. 



These figures were confirmed by Davis and Daish in 1913 who also drew 

 attention to other particulars, (a) The asbestos used in filtering. This should 

 be treated for 30 minutes with boiling 20 per cent, sodium hydroxide and 

 washed thoroughly with water, (b) The filtering through a Gooch crucible 

 instead of a Soxhlet tube, (c) The oxidation of the cuprous oxide to cupric 

 oxide by heating the crucible in a protecting crucible for 30 minutes with 

 a inch Teclu burner until the weight is constant. The conversion into 

 cupric oxide seems particularly favourable, if sugar estimations in plant ex- 

 tracts be required. 



Pfliiger has also examined the gravimetric estimation of glucose and has 

 found that the principal error in Allihn's procedure is the time of boiling j 



