EXPERIMENTS WITH THE FLAME-MANOMETER. 399 



of the note, of which, the sonorous vibrations are acting 

 upon the membrane in the capsule of the apparatus. The 

 quicker the apparatus is turned the greater is the space 

 through which the image travels in a given time, and 

 the greater is therefore the distance between the images 

 of the successive flames which correspond to the succes- 

 sive vibrations. Again, if the box is turned constantly 

 with the same velocity, there will be less images of 

 flames within a given space if there are fewer vibrations 

 in a given time; in other words, the images are the 

 farther apart from one another the deeper the note 

 which acts upon the issuing gas. Fig. 236, B, shows 

 the series of images which corresponds to a note which 

 is one octave lower than the note which produces the 

 series of images at A. 



The later in time a flame is produced the more 

 its image appears to have moved laterally; since the 

 upper portions of the flame are produced earlier than 

 the lower, the image of the upper portions moves on- 

 ward while that of those later produced appears to lag 

 behind; hence the upper portions of the flame appear 

 curved, with points which are turned towards the direc- 

 tion in which the images move. If the successive im- 

 pulses which cause the interruption of the flow of gas 

 are not very rapid nor strong, the efflux of gas never 

 ceases altogether, but is only diminished, and as a con- 

 sequence the images of the successive flames form a 

 connected band at their lowest portions, as shown in 

 fig. 236, B. 



The two saucer-shaped halves of the wooden capsule should be 

 made by a turner of hard wood; their section may be seen from fig. 

 234, A. After boring the holes for the tubes, a thin layer of glue is 

 spread over the rim of one of the saucers, and gold-beater's skin is 



