338 The Bcxtrose-Laevulose Ratio in the MuikjoUI 



be inferred that the hexoses are formed more rapidly than the saccharose, 

 and consequently precede it; instead, the quantity of each of the 

 reducing sugars is roughly proportional to the cane sugar present, a 

 fact which points to the formation of the hexoses from this sugar. 

 Series II. In Series II (Table II) the sugar percentages are far 



higher than in Series I and the fluctuations in the value y are far less 



marked in consequence. With few exceptions (e.g. at 6 p.m. and 



8 p.m.) the value of y does not differ much from 1, the percentages 



of dextrose and laevulose being as nearly equal as one could expect 

 bearing in mind the errors to which the calculations are subject. In 



general, the values of y obtained by assuming the pentose to be ara- 



binose are slightly lower than unity, whilst by assuming it to be xylose, 



they become sUghtly higher than unity. It is probable that j would 



be almost exactly 1 were it not for the presence of small quantities of 

 optically active impurities, which in some cases increase the value, in 

 others lower it. The following table shows that the departure of the 



value Y from 1 goes hand in hand with the divergence A between the 



Ij 



values for cane sugar found by the reduction and double polarisation 

 methods ; this divergence is no doubt also caused by the presence of 

 these optically active impurities (see p. 344). 



* A represents the difference between tlie reduction and polarimtion valuer for cane 

 sugar, expressed as a percentage of the areragr value found by reduction by the invertaso 

 and eitric acid methods; thus, e.g., +10-0 per cent, shows that the average value found 

 by the double polarisation method is 10 per cent, higher than the average value found by 

 reduction. In this particular case (6 a.m.) the cane sugar found by reduction was 4-24 

 per cent, on t.v.d.m , and by double polarisation 4-65 per cent. 



