JYewton's Colors by Transmitted Light. 225 



trough of such dimensions as just to admit the set facilitates 

 the dipping. 



Daring the course of this experiment, before the films have 

 become thin enough to show colors, certain other phenomena 

 of a circulatory character are very noticeable. These may be 

 studied most easily in the projected image of a single film, and 

 for observations on conditions immediately after formation, it 

 is convenient to support a ring on the shorter, branch of a J- 

 shaped " handle " and to fix it permanently in the focus of the 

 projecting lens, a narrow trough being raised so as to submerge 

 the ring when necessary. The moving images are of two 

 kinds, one corresponding to pear-shaped air-bubbles running 

 up at the sides of the film, the other to spheroidal drops running 

 down through the middle. These sets of images are always of 

 opposite colors, and if, under given circumstances, the air- 

 bubbles show white and the drops black, then moving the film 

 some 4 cm away from the lens will reverse the colors, the bub- 

 bles showing dark and the drops white. From this it is evident 

 that the latter act as convex lenses with real foci (bright spots) 

 some 2 cm in front of the film, while the bubbles are concave 

 lenses with virtual foci 2 cm behind it. 



These air-bubbles require but little explanation, but the 

 drops are more interesting. The happenings in a typical film 

 (dipped at a time t—0) are as follows: At first its image is 

 clear, but almost immediately a few large drops appear at vari- 

 ous parts of the film and begin to descend. These " wanderers " 

 are much larger than their successors (d~0'5 or 0*6 mm ) and their 

 number varies from one to a dozen, being largest when a pre- 

 ceding film has dried on the ring and has not been removed. 

 At t=l sec. a "skirmish-line" has formed about a third of the 

 way down, above which line the film is closely dotted with 

 drops. This travels down in the film about as fast as do the 

 color-bands later, reaching the bottom at t— about 8 sec. The 

 drops in its front rank are usually small, but a half a centimeter 

 behind come the largest of the "ordinaries" (d-0'2 to 0*3 mm ) 

 and after these a succession of smaller ones until (at £=60 to 

 100 sec.) the film begins to show colors and great viscosity is 

 noticeable. The size of the average drop is some # 2 mm and, 

 roughly speaking, the size of those passing a given point 

 decreases as t increases. The velocity of descent varies from 

 to 7 mm per sec. (that of the wave front averaging 5-7 ram per 

 sec.) and is greater (1) for large drops, (2) for drops near the bot- 

 tom of the ring and (3) at times corresponding to small values of 

 t, but, if other things are equal, is nearly the same at all points 

 in the same horizontal line. A number of observations gave 

 an empiric formula ?; 2 = 2j? (y-q) (y being the distance from the 

 top of the ring, andj? and ^functions of £, the former decreas- 



