B.—CHEMISTRY 37 
demanded by the X-ray diagram, but this diagram fulfils every particular 
dimension of the repeating pattern of the cellobiose formula : 
This mental picture of the constitution of cellulose will remain incom- 
plete if we can gain no knowledge of the number of @-glucose units 
constituting the chain length of cellulose. Here an endeavour has been 
made to reach an approximation of the size of the chemical molecule 
by a gravimetric assay of the end group of methylated cellulose. A 
specimen of the latter material, prepared under very carefully controlled 
conditions to avoid chemical rupture, enables us to gain an insight into 
this problem by investigation of its products of hydrolysis. An estimation 
of the weight of tetramethyl glucose obtained from these hydrolysis 
products has yielded under proper technique a value for the average 
length of the cellulose chain, which is thus found to consist approximately 
of 200 glucose units. In this connection, and especially when we come 
to consider the same problem in reference to starch, it must be recognised 
that native cellulose is most probably a molecular aggregate consisting of 
a much larger physical molecule than is here represented. A molecular 
aggregate of such chains, joined by physical links or by co-ordination 
end to end, is calculated to give a much enhanced value for the determina- 
tion of molecular weight by viscosity methods as interpreted by the 
Staudinger constants. Similarly a determination of particle size by the 
sedimentation method of the ultra-centrifuge is now found to give a 
much higher value than that of 200 glucose units. Here again the physical 
molecule or molecular aggregate may be expected to take account not 
only of an aggregation which increases the length of the chain but also 
of the forces which affect molecular aggregation laterally between adjacent 
chains. I think these factors must be recognised in any comparison of 
the molecular weights of cellulose determined by physical and chemical 
