]84 A. J. Ewart : 



Glucose equivalent Per cent. 



Lo.s3 of ol: reducing polymerization > 



(CHoO).2H.,0 Alkali weiijht sugar in terms of glucose- 



2.09 gram SrH ,0.^ 0.8 gram 0.63 gram 0.35 17% 



1.53 „ BaH.,02=l „ 0.55 „ 0.34 22% 



1.69 „ NaHO 0.5 ,. 0.74 „ 0.08 5% 



(CHaOa 



1.25 „ BaH^Oa 1 „ 0.41 ., 0.33 26^^ 



In each case the amount of alkali was sufficient to remove alh 

 the formaldehyde. When sodium hydrate and the polyhydrate are- 

 pounded together, the reaction starts spontaneously, and is very 

 violent. There is a heavy loss of formaldehyde vapour, and much: 

 of the sugar is caramelised. For this reason the percentage poly- 

 merisation is less than when a boiling dilute solution of formalde- 

 hyde is neutralised with dilute sodiimi hydrate. 



If tlie reaction is started by locally warming several grams of 

 the mixture in a small distilling flask, a large amount of distillate 

 is obtained witliout applying further heat. This consists in the 

 case of the polyhydrate (and crystalline Ba H.^Oo) of water, methyl" 

 alcohol and formaldehyde. 



If a slight excess of the polydrate is used, the whole of the 

 BaH202 is converted into barium formate. Tliis can be obtained 

 by dissolving the residue in a little water, and adding an equal 

 bulk of alcohol. A large part of the sugar slowly settles out with 

 some barium formate. On filtering after 48 hours standing, and 

 doubling the bulk in the alcohol, a second ppt., mainly of barium 

 formate, is produced, which, can <'asily be washed and purified. 



A point worth noting is that if the finely ground polyhydrate is 

 mixed thoroughly Avith finely ground crystalline barium hydrate, 

 the mixture smells strongly of formaldehyde, and its temperature 

 falls 15°C. for some time. The temj^erature then slowly rises, Vmt 

 not appreciably above that of the room. The mixture slowly 

 develops traces of reducing sugar, but does not undergo any com- 

 plete reaction even after days in contact, until this is started by 

 heating one j^oint of the mixture. 



A similar preliminary fall of temperature is show^n with di- 

 and tri-ninlecular formaldehyde. If, however, these are mixed 

 witli freshly slaked dry quicklime, or with powdered calcined 

 barium or strontium hydrates, the mixtuie remains dry, the fall" 

 of temperature is hardly noticeable, and on heating the mixture 

 locally the reaction does not spread. Only the parts heated turn 

 brown, and methyl alcohol distils over (yielding methyl iodide with 

 Iodine and red phosphorus). The residue contains an abundance 



