March 17, 1899.] 



SCIENCE. 



391 



S. Compressional Space Waves. — Either of 

 the adjustmenbs, Figs. 4 and 5, is adapted to 

 actuate sound waves, as will be shown below, 

 §25. 



9. Botary Polarization. — Figure 6 shows 

 the apparatus adapted to compound two 

 equal and opposite circular motions, Fig. 

 10 being a detail relative to it. Both the 

 front and rear series of eccentrics have the 

 same diameter and swing, but there is one 

 turn in front to two in the rear, respectively 

 left and right. The riders are gutters 

 about 4" long, joined at right angles with 

 the concave sides toward the eccentrics. 

 The extensible levers '(tubular as above) 

 are soldered in the prolongation of the bi- 

 sectrix of the riders, and project from the 

 salient side of the right angle obliquely up- 

 ward, each passing through a perforation in 

 the horizontal laths of folded tin plate shown 

 at U U' and V V. The levers are effect- 

 ively about 18" long, and are held down 

 upon the cams by springs * (like the above) 

 one end of each of which engages the lever 

 while the other is revolubly attached to the 

 axle, between the cams (see Fig. 6). If 

 U V and F V (adjustable) are symmetric- 

 ally placed with reference to the two effect- 

 ive ends of the levers the upper ends will 

 trace a circle-like figure, corresponding to 

 the circular motion of the lower ends. 

 With the pulleys cross-belted as shown, the 

 pin eyelets X Y (3" long, soldered axially 

 to the upper ends) may then be adjusted to 

 the counter circular motion indicated in 

 Fig. 10. 



Two methods of compounding were tried. 

 In the first the ends of two silk threads, 

 Fig. 10, carrying the cork W (vibrating 

 particle) between them were fastened to 

 delicate helical springs surrounding the 

 upper ends of the levers. This method con- 

 structs the wave very well, but in motion 

 the friction at the eyelets (one of which is 

 often high and the other vertically below 



* These springs are seen on the helix in Fig. 2. 



it) is apt to be too unequal to keep the 

 particle in the symmetrical position neces- 

 sary. Better results are obtained by stretch- 

 ing a very thin India rubber band, dd ee, 

 between the eyelets, carrying the particle 

 as before. Springs were similarly tested. 

 Parallelogram motion is hardly appreciable 

 here without elaborate construction. 



The vertical vibration is in this way very 

 well obtained(of course, in semi-amplitude). 

 The horizontal vibration is noticeably curvi- 

 linear, seeing that the two motions com- 

 pounded are not quite uniformly circular. 

 Even in this case, however, the connectors 

 dd ee move parallel to themselves. 



The helical characters of the wave ob- 

 tained is well shown in Fig. 6, calling to 

 mind that each ball vibrates normally to 

 the strings by which it is suspended. 



The laths U V are supported by uprights 

 Z Z', which fit in flat sockets (seen at /, in 

 Figs. 4 and 5). With these the whole 

 superstructure of laths, levers and riders is 

 removed from the machine at once in a 

 manner easily suggested. The bed plate 

 then returns to the appearance of Fig. 1. 



The method of obtaining similar results 

 in compounding circular motion for the case 

 of Fig. 4 is given below §24. A special cam 

 axle carrying two screws (alternate cams 

 differing 180° in phase) is here needed. If 

 rotary polarization is wanted the wave- 

 lengths of the front and rear axle must 

 differ. 



EXPERIMENTS. 



10. Method of Designating Phases. — Before 

 describing the consecutive experiments to 

 be performed with the machine, it is well 

 to come to an understanding as to the phases 

 in which the two component disturbances 

 meet. These are conveniently determined 

 by the long axes of the first eccentrics on 

 each axle, which (axes) may, therefore, be 

 called pointers. Since the waves for clock- 

 wise rotation at the crank travel from left 

 to right, along the axle, and since a rise of 



