W. A. Davis and G-. C. Sawyer 375 



per cent, on the t.v.d.m.), but because the quantity and rotatory power 

 of the hexoses is exceedingly small (0-15 to 1-27 per cent.). At 8 a.m., 



for instance, if the pentoses are taken as arabinose, y = 0-34, but if 



they are assumed to be xylose y becomes 0-85. 



That the results are vitiated by the presence of a /aeyo-rotatory 

 impurity appears clearly in the data for 6 a.m., 10 a.m., 2 p.m., 4 p.m. 

 and 10 p.m., in all of which cases the amount of dextrose appears to 

 be nil. If in these cases the whole of the reducing sugar is assumed 

 to be laevulose, the negative rotation calculated does not account, on 

 the assumption that the pentose is xylose, for the negative rotation 

 actually observed. The following table shows the differences : 



* After allowing for the pentoses (as xylose) and saccharose present. 



If the results are calculated on the assumption that the pentose is 

 arabinose, the negative rotation not accounted for becomes even greater ; 

 thus at 6 a.m. it becomes - 0-1.38° instead of - 0-083°. The number 

 of cases in which dextrose appears to be entirely absent is increased 

 on this assumption. 



It is clear therefore that the apparent predominance of laevulose 

 in the potato leaf is due to the presence of relatively large quantities 

 of a laevo-rotatory impurity (? asparagine), and it is probable that the 

 dextrose and laevulose, as in the mangold leaf, are really present in 

 equal proportions, that is as invert sugar, and are formed from sac- 

 charose. It is interesting that the dextrose appears in largest amount 

 at 6 p.m. and 8 p.m., that is at the time when the starch content reaches 

 a maximum and is subsequently put under contribution. As we show 

 on p. 357, the starch is broken down by the leaf enzymes completely 

 to dextrose. Fig. 3 shows the variation of the apparent dextrose and 

 laevulose during the 24 hours (pentoses assumed to be xylose) ; as an 

 index of the real fluctuation of the hexoses the curves have, of course, 

 no value, but they are interesting as showing that the laevo-rotatory 

 impurity varies regularly during the 24 hours. The laevulose always 



25—5 



