certain Dilute Aqueous Solutions. 377 



To protect the liquid a swell as possible from dust and other 

 impurities in the air, the tray (D) and the fork (F 2 ) were 

 placed in the galvanized iron box ((!), which was closed by a 

 tightly-fitting lid provided with a window of good, plane, 

 parallel plate-glass, through which all readings were taken. 



The fork F x is an ordinary electrically driven tuning-fork 

 provided with a mercury-platinum contact-maker. F 2 is also 

 an electrically driven tuning-fork ; it is tuned in unison with 

 Fj, and is driven by it by means of a shunt circuit. 



By means of L^ the light from S is focussed on a point 

 midway between the two holes in S 2 ; it then passes 

 through L 2 which renders it parallel, then it is reflected by 

 Mj to the surface of the liquid, from this it is reflected to M 2 , 

 and from M 2 into the telescope T. If now the forks are 

 started and the telescope is focussed for the light reflected from 

 the crests of the waves, the entire field of the telescope, which 

 is dimly illuminated, will be crossed by a series of parallel 

 bright lines which correspond to the crests of the waves. 

 These lines appear stationary, and the distance between them 

 is approximately one-half the wave-length, since each wave 

 is seen in two positions (once when the prongs of F x pass 

 through their positions of equilibrium and are approaching 

 one another, and again when they are separating). The 

 crests seen in one position do not lie exactly midway between 

 those seen in the other. 



The plan of operation is to set the spider-line of the tele- 

 scope on one of these lines, read the position of the carriage, 

 then move it along several centimetres, and set on another 

 line of the same series. The difference of these two readings 

 divided by the number of waves passed over gives the wave- 

 length. 



Sources of Error, and Adjustments. 



The first objection that presents itself is that, as the carriage 

 is moved, the mirrors will not remain parallel to their original 

 positions. Owing to the length of the arms supporting the 

 mirrors (30 cm.) and to the distance of the liquid below 

 them (45 cm.), any slight displacement of the mirrors will 

 produce a great error in the determination of the wave-length. 

 These errors may be introduced either by irregularities in the 

 ways of the engine, or by a motion of the arms with respect 

 to the carriage. The latter is simply a matter of rigid con- 

 nexions and can be easily remedied, so we shall first consider 

 the former. 



To test the accuracy of the ways, the rotation of the carriage 

 was tested about three rectangular axes — one along the screw, 



