( 968 ) 
tube or by pouring over the content of it the substance was brought 
into contact with the water and the rise in temperature read off. 
The form of the curve indicating how IW depends on ¢ has been 
determined with six substances, of which | only communicate those 
of cellulose (ash-free filters of Scuimrcner & SeHürr) and casein (pur. 
Honcustrr FARBWERKE)!). Of the curves published by other investigators, 
only those of Roprwarp and Karrni for artificially prepared starch 
granules *) and of Vorsrnr for prepared woody jibres*) (as in the 
determination of crude fibre) seem to me fit for Comparison with 
liquids. The other substances investigated by Roprwarp and Karrrin 
were species of amylum which, owing to their stratiform structure, 
form more intricate objects; but there also curves were found which 
in form agree completely with those of artificial starch granules *). 
On the annexed illustration the curves of these four substances, 
carefully drawn to scale, are shown. All four make the impression 
of a hyperbola; this is particularly striking in the case of cellulose 
and casein *). 
In the ease of the other two curves and the cellulose I have 
questioned whether they may, indeed, be represented by the formula 
of the hyperbola 
The subjoined tables show that the agreement is satisfactory *). 
1) [ beg to offer my sincere thanks to the directors of the Hoechster Farbwerke 
vorm. Meister, Lucius and Briinine for kindiy placing at my disposal the casein 
required for these experiments. 
2) Zeitschr. f. physikal. Chem. 33, p. 551. 
3) Untersuchungen über die Quellung der Holztaser, Inaugural-Dissertation, Kiel 
(1896) p. 32. VorBenR had also observed that the line determined by him had 
the character of a hyperbola. 
4) Amylum solubile (Mercx) is, for reasons to be stated afterwards, not {it to 
be compared with liquids; for the rest it presents a line of the same form. 
5) In the case of the other four substances which [ have investigated, lines were 
also obtained which make the impression of hyperbolas. 
6) The deviations are not much greater than the presumable experimental errors. 
