374 Dr. H. Garrett on the Viscositjj and 



[)laye{l directly against the ijottom of the tube, raising it to 

 a l)right red heat. 



In order to see Avhether an increase in the length of the 

 absorbing column o£ the vapour produced the same effect as 

 increasing the density o£ a short column, a steel tube fiv«' feet 

 in length was used. 



In this were placed eight lumps of sodium of the same 

 size at inter^-als of about 6 inches. Eight Bunsen burners 

 were regulated so that when one of them was placed under 

 a lump of sodium the D-lines were strongly reversed, but no 

 trace of the fluted spectrum appeared. On adding Ijurner 

 after burner exactly the same sequence of events was 

 obserA'ed as in the previous experiments, where the density 

 of the vapour was increased by raising the temperature of 

 the tube. Though this was to be expected it seemed worth 

 while to try the experiment. 



AYe have measured the wave-lengths of about -160 of the 

 strongest lines between wave-lengths 4616 and 5738. 



The tables of these Avave-lengths appear in the A stro- 

 ])hysical Journal for September of this year, together with a 

 complete set of the photographs of the fluted absorption- 

 spectrum. 



Johns Hopkins University, 

 Biiltimore, June 1908. ' 



XLVII. The Viscosity and Composition of some Colloidal 

 Solutions. By Hexry Garrett. B.Sc, Land., Fh.B., 

 Heidelbera* . 



THE object of this research was to study the changes in 

 the viscosity (t;) of typical colloidal solutions, gelatine, 

 silicic acid, and albumen, when the temperature and concen- 

 tration are varied, and thereby to arrive at a clearer under- 

 standing of the constitution of such substances. 



Two methods were used : — {a) 0. E. Meyer''s method, in 

 which the viscosity is deduced from the damping of the oscil- 

 lations of a disk vibrating in the liquid ; [0) the Hagen- 

 Poiseuille method of flow through a capillary tube. 



The results obtained were quite inexplicable on the older 

 theories of colloids, but agreed well with Quincke^s '' foam- 

 cell ^^ theory, which can be shortly outlined. A colloidal 

 solution, on this theory, is not a homogeneous fluid like water, 

 but consists of a mixture of two solution?, havino- surface- 

 tension at the surface of contact, one rich in colloid, the other 



* Communicated by the Author, being an abstract of the author's 

 Inaugural-Dissertation zur Erlangung der Doktorv:'urde. Heidelbers: 

 Universitat (1903). 



