W. A. Davis, A. J. Daish and tl. C. 8awvkii 291 



per cent.) between these hours. This is very different from the August 

 and September pickings when a rapid formation of the pentosans (gums 

 or ligneous tissue) was clearly obvious in spite of the simultaneous 

 increase in the sugars. A certain amount of pentosan formation does 

 apparently occur, however, even at this later stage of growth, and is 

 visible in the rise between 11 a.m. and 1 p.m. The fluctuations at night 

 seem to be principally relative — thus the increase in the matter insoluble 

 in alcohol which occurs between 7 p.m. and 11 p.m. is due to the rapid 

 fall of sugars from 21 to 16 per cent., but as there is a fall of the pentose 

 curve between the same points, it appears that some real pentosan 

 formation occurs from these sugars. The fall of pentosan and insoluble 

 matter from 11 p.m. to 3 a.m. is also a relative effect due to the increase 

 of the sugars and the rise from 3 a.m. to 7 a.m. is also relative, owing 

 to the faUing oft' of these. 



The free pentoses increase during the day as in the earlier pickings 

 and follow very largely the invert sugar curve ; thus there is a continuous 

 increase from 7 a.m. to 1 p.m., after which the pentose falls, apparently 

 giving rise to pentosan. At night the pentose falls between 7 and 11 

 p.m. and the pentosan rises ; the sudden rise of pentose between 11 p.m. 

 and 1 a.m. occurs simultaneously with the sudden increase of hexoses. 

 The fall of pentoses subsequently to this, from 1 a.m. to 3 a.m., is partly 

 a relative effect, due to the rise in the reducing sugars. 



The total variation of the pentoses during the day is only very small, 

 viz. 0-82 to 0-92 per cent., but at night the fluctuations (probably 

 largely relative) are from 0-92 to 0-61. During the day the fluctuations 

 are considerably less than during the daytime in Series II ; but at 

 night the changes are greater, mainly representing a falling off in the 

 pentoses. 



Comparison of Series III with Series II. 



1. The actual proportions of saccharose and hexoses and the range 

 of variation of these sugars are considerably greater in Series III (October 

 llth-12th) than in Series II (September lOth-llth) and therefore far 

 greater than in Series I (August 26th-27th). 



September lOth-llth. Cane sugar varied from 8-27 to 4-24 per 

 cent. ; hexoses, 8-9 to 5-4 per cent. 



October llth-12th. Cane sugar varied from 9-52 to 4-98 per cent.; 

 hexoses, 12-41 to 9-39 per cent. 



In both series the cane sugar varies between wider limits than the 

 hexoses, and the difference is most marked in Series III. 



