W. A. Davis, A. J. Daish and G. C. Sawyer 285 



simple. As in the earlier stage of growth (see Fig. 4) the curve more or 

 less closely follows the temperature curve during the day and the 

 greater part of the night. As the temperature rises from 10 a.m. to 

 1 p.m., the invert sugar increases faster than the saccharose, but when 

 the temperature falls, the ratio of invert sugar to cane sugar falls off 

 along practically a straight line, exactly as was the case in Series I 

 (Fig. 4). From 2 a.m. to 6 a.m., the saccharose is disappearing faster 



than the hexoses, so that the ratio — - -' rises, again along nearly a straight 



line, until jiist after sunrise, when the formation of cane sugar begins 

 more rapidly than that of hexose sugar. 



It is very difficult to explain the night maxima, B and B' , which 

 form such a striking feature of the sugar curves at this stage of growth 

 and also in the later and final period in October (see Fig. 6, p. 289). 

 The maxima reached at night are in the case of both sugars considerably 

 higher than those attained during actual insolation. Both sugars 

 increase together and fall together, so that interconversion cannot explain 

 the result. The sum of the sugars at the night maximum (16-6 per cent.) 

 is slightly higher than at 6 p.m. (15-3 per cent.) and far higher than at 

 1 p.m. (12-1 per cent.), when the direct formation of the sugars under the 

 influence of light reaches a maximum. It is improbable that, at night, 

 a reverse current of sugar sets in from the roots to the leaves and our 

 actual analyses as well as a careful microscopic examination of all the 

 samples have shown the entire absence of starch from the leaf during the 

 day and night. Had starch been present, the increase in the amount of the 

 sugars at night might be due to the transformation of starch into these. 

 In the absence of starch, any explanation of the large increase in the 

 proportion of soluble sugars which is observed to attain a maximum in 

 the neighbourhood of 2 a.m. (at about 3 a.m. in October, in both cases 

 about 3 hours before sunrise) must be more or less conjectural. Both 

 in September (Series II) and October (Series III) the maximum con- 

 centration of the sugars is reached at nearly the same time, whilst the 

 proportion of cane sugar is practically identical (about 8 per cent.) in 

 both cases, in spite of the day values for saccharose being far lower in 

 Series II than in Series III. As starch and the product of its hydrolysis, 

 maltose, are entirely absent from the mangold leaf it seems probable 

 that some other substance acts as a reserve at this period of growth, 

 and, during the night, is broken down to cane sugar and invert sugar, 

 thus causing the rise which is observed. The mangold leaf undoubtedly 

 contains a large amount of gummy substance which is soluble in water 



