W. A. Davis, A. J. Daish and G. C. Sawyer 265 



procedure used iu estimating the cane sugar and reducing sugars; as 

 will be seen later, many of our results confirm those obtained by Parkin 

 (except as regards the dextrose : laevulose ratio, a subject which is dealt 

 with separately, see Paper II), so that considerable confidence may be 

 placed in these observations. The principal points open to criticism 

 are : (1) How far the results were affected by the process of drying the 

 leaves adopted by Parkin ; (2) how far the cane sugar results were 

 lowered owing to incomplete inversion occurring on account of the 

 presence of lead acetate in the solution interfering with the ordinary 

 Clerget process. From the experiments Parkin actually made to test 

 these points it would appear that in the case of the snowdrop the error 

 arising from either cause was but small. 



In earlier papers (Davis and Daish [1913 and 1914]) we have given 

 an outline of the methods of analysis we have adopted ; we need there- 

 fore only add a few details which were formerly omitted. 



Cane sugar has always been estimated by two distinct methods: 

 by inversion with 10 per cent, citric acid [1913, p. 466] and by inversion 

 with invertase (autolysed yeast). This gives a means of checking the 

 results. 



Maltose was estimated by the use of maltase-free yeasts [1913, 

 p. 464], such as S. marxianus and S. exiguus, duplicate fermentations 

 being carried out with ordinary baker's or distiller's yeast so as to make 

 allowance for the pentoses present which remain unfermented. 



Starch was estimated by taka-diastase [1914, p. 159] in the dry leaf 

 material from which the sugars had been completely extracted by 

 alcohol. Special details are given of this method, as apphed to estimate 

 "soluble starch" or "dextrin" when present, in our experiments on 

 the potato leaf (see p. 361). 



Pentoses. As shown in a previous paper (Davis and Sawyer [1914]) 

 appreciable quantities of pentoses are invariably present in the 

 alcoholic extracts of leaf material ; these have been estimated by 

 distilling a known volume (50 cc.) of the original purified solution used 

 in the sugar estimations with hydrochloric acid and weighing the 

 furfural produced as phloroglucide according to the Krober-Tollens 

 method. 



Pentosans. A suitable quantity (1-5 grm. of the oven-dried leaf 

 material, from which the sugars have been extracted, or about 1 grm. 

 of extracted stalk) is distilled with hydrochloric acid under Krober- 

 Tollens conditions (Allen's Commercial Organic Analysis, i. 402); the 

 furfural evolved is precipitated and weighed as phloroglucide. 



