368 Professor Sir James Dewar [Jan. l.s, 



which may have equal or different vohimes as desired, and may be 

 further controlled so as to present themselves either as a vertical 

 column or as a cluster. 



The shape of the cluster obtained depends on the dimensions 

 of the nozzle, the speed of the air current, and the relation between 

 the mass of the drop and the size of the constriction. If the drop 

 be too large or the air current too slow, extra films result from the 

 accumulation of superfluous liquid ; the same thing happens with an 

 elongated constriction. On the other hand, too rapid a currant may 

 simply spray the liquid over the interior of the nozzle. The cleanest 

 working is obtained from a short neck blown out above and below 

 in a spherical shape as shown in the figure at K. 



A succession of equal bubbles may either form a regular cluster 

 on the mouth of the nozzle, or a vertical chain or column of spherical 

 segments united by plane circular films. A cluster will usually be 

 formed when the bubbles are less in diameter than the nozzle, and a 

 chain or column when the bubbles are larger. For the manipulation 

 of these complexes an air-tight, cubical, plate-glass chamber was 

 made (Fig. 5), on a frame of aluminium alloy (edge, 50 cm.). The 

 glass plate for the top was 1 cm. thick ; holes were cut in it for the 

 rubber corks that held the nozzles and supported the movable glass 

 rods which carried the different small pieces of apparatus used to 

 catch and control the chains while being formed. 



For the manipulation of the longer columns the glass cylinder 

 shown in Fig. 6 was used. It was 3 feet long and 6 inches in 

 diameter; its ends made air-tight joints with two plates, the upper 

 of which held the blowing nozzle, while the lower held a support 

 ring. Suitable exit stopcock and soda-lime guard tube were also 

 fitted. The support ring was sealed to the top of a vertical rod that 

 could slide up and down in an air-tight joint in the bottom plate. 

 The cylinder was fixed on a shelf in the middle of a massive wooden 

 stand, and a hole was cut in the shelf to allow free movement to the 

 sliding rod. When a to-and-fro motion was given to the support 

 rod, tlie flexibility of the column was readily seen ; transverse waves 

 appeared to pass up the Avhole length, each segment moving on the 

 one adjacent as if attached by a very free universal joint. By 

 drawing the rod down a considerable extension could l)e obtained; 

 contrariwise by raising it a bulging or spiral formation would result. 

 If this be carried further and the surface l)e sufiiciently moist, the 

 column will sag out into contact with the glass walls and form a 

 chain of oval segments. In a Avidcr vessel, however, the segments 

 can in this way be linked up successively to form regular clusters. 



A beautiful exhibition of multiple scintillating coloured reflections 

 is obtained when a beam from an arc lamp l)elow is directed up 

 tlirough the column while the support rod is moved. An even more 

 complete illumination is secured by using a white lined hood above, 

 in which is hidden a 200-watt lamp, with two white wings of stiff 



