264 The Carbohi/drates of the Mangold Leaf 



extraction of the sugars. Had any change, such as inversion, occurred, 

 it would have been impossible to obtain, for example, the regular 

 series of results actually found in the case of the potato leaf (see p. 367), 

 where the cane sugar rises and falls between very narrow limits (2-1 to 

 3-5 per cent.) along curves which are practically straight lines. 



Methods of Analysis. If work such as we have been engaged on is 

 to have any permanent value, it is necessary to ensure that the analytical 

 methods give trustworthy results with the class of material actually 

 dealt with. Much time was therefore devoted in the beginning to 

 testing these methods, and we have shown in several earlier papers 

 (Davis and Daish [191:5 and 1911]) that very grave errors may arise 

 in investigations of this kind. Thus the important work of Brown and 

 Morris in 1893, so far as it refers to the sugars, suffers from the fact that 

 the only method then available to estimate maltose gave entirely 

 incorrect results owing to the destruction of laevulose brought about 

 by the hydrochloric acid used in the hydrolysis ; as the accuracy of the 

 values for the other reducing sugars, dextrose and laevulose, depends 

 upon correct values being taken for the maltose, it is clear that the 

 proportions found for these sugars are equally incorrect. We have 

 shown also (Davis and Daish [1914]) that, in estimating starch by the 

 diastase method, large errors may be caused by the loss of dextrin; 

 we have introduced therefore a new method based on the use of taka- 

 diastase, the enzyme of AsjKrgiUus oryzae, which converts starch 

 into a mixture of maltose and dextrose only. The method of hydro- 

 lysing cane sugar which was used by Campbell [1911] may, we have 

 shown, give rise to entirely false results when applied to plant extracts, 

 owing to the cane sugar being only partly inverted by the 2 per cent, 

 citric acid which suffices to invert completely fure solutions of the 

 sugar. In Campbell's analyses the cane sugar would therefore be 

 underestimated; this would lead to very high values being returned 

 for maltose, even though considerable destruction of laevidose occurred 

 during the hydrolysis with hydrochloric acid. As the values for sac- 

 charose and maltose were incorrect, the data for dextrose and laevulose 

 are equally invalid. Campbell's work on the carbohydrates of the 

 mangold leaf must therefore be regarded as merely preliminary in a very 

 difficult field and the data and conclusions entirely withdrawn. 



Parkin's recent work [1912] was, fortunately, carried out with a 

 plant in the leaves of which starch and maltose do not occur. The 

 analvses were therefore not coin])licated by the necessity of estimating 

 these substances. Parkin carefully tested many points of the analytical 



