CONSTITUTION OF GLYCOCOLL AND ITS DERIVATIVES. 93 



assume that methyl chloride, supposed to be liberated during the first 

 stage of the reaction, goes to interact with dimethyl-glycocoll, produc- 

 ing betaïn and hydrogen chloride : 



I-LC-N(CH3).JH + CliCHa ILC-K(CH,)3 



I I = I I +HC1. 



OC-0 OC-0 



This assumption, which tins no merit as an explanation over that 

 ah'eady given, is, inoreover, hardly warranted by facts. For, even if 

 w^e suppose that dimethyl-glycocoll is first formed and then changed 

 into trimethyl-glycocoll by interaction ^^ith metliyl chloride, it is 

 difficult to imagine that this interaction would be so complete that 

 none of the interacting bodies should be left unaltered. The fact, 

 then, that by the action of ti-imethy lamine upon chloracetic acid, betaïn 

 is produced unmixed with dimethyl-glycocoll goes to deprive the 

 ground of the assumption necessitated above of any probability and, 

 thei'efore, to corroborate the \dew here advocated as to the mode of 

 formation of o'lycocoll and its trimethvl derivative. 



The consideration of the mode of formation of snrkosi ne (raono- 

 methyl-glyrocoll), a bodv which A ol hard synthetically obtained by the 

 action <^f methylamine upon ethyl chloracetate ÇLieh. Ann., 123, 2G1 

 [1<S62]), leads to the same conclusion. The reaction is expressed as 

 follows, according to the view liere urged : 



H,C-C1 H,C-NH,(CH,)C1 



+ 2 NHo(CH3) + H.O - I \+ CoH^.OH 



oc-ocj-is ' oc-oxri,(CH3) — ~- 



= H„C-XH.,(CH,) 



^^^_I^ + Nri,(CH3)CI . 



If, however, we attempted as before to represent tlie aboA'e change in 



