FOG-SIGNALLING 91 



applicability of a method of alternating out-blast and suction for producing 

 economically sounds of high intensity. The original idea did not develop 

 satisfactorily ; but some experiments in conjunction with Crum Brown led 

 him in 1878 to the construction of a new form of siren suitable for fog- 

 signalling. In these experiments an organ note was made discontinuous by 

 being sounded through a partition and a revolving disk cut into separate 

 sectors. Unfortunately the siren effect superposed on the effects which were 

 being studied disturbed somewhat the quality of tone. Tait found that when 

 there was no organ note being sounded the mere rotation of the perforated 

 disk produced a sound whose intensity could be greatly increased by soldering 

 plates perpendicularly to the revolving disk so as to increase the thickness of 

 the back edges of the apertures. When rapid rotation was set up the sound 

 emitted was almost terrifying in its intensity. It shrieked out through the 

 open window of the Natural Philosophy Class Room into the quadrangle to 

 such a degree as to interfere with the lectures in the neighbouring class rooms. 

 Tait was accordingly obliged to conduct his experiments at hours when no 

 classes met. 



In February, 1880, Tait communicated a short note to the Royal Society 

 of Edinburgh describing his unsuccessful attempt to measure the velocity of 

 the particles which constitute the cathode rays in a Crookes' tube, by means 

 of observations of the spectrum made in directions perpendicular and parallel 

 to the lines of motion of the charged particles. One cause of the failure 

 was the loss of light by multiplied reflections when a powerful spectroscope 

 was used. This led him to construct a rotatory polariscope whose principle 

 depended upon the rotation by quartz of the plane of polarisation, combined with 

 sufficient prism dispersion just to separate the various bright lines of the source 

 from one another. The final form of the apparatus was described in Nature, 

 Vol. xxii (Sci. Pap. Vol. i, p. 423). When a plane polarised ray of 

 light is subjected to rotatory dispersion by transmission through quartz, 

 and, after further transmission through a double image prism, is dispersed 

 prismatically by means of a direct vision spectroscope, there appear side by 

 side two spectra of the original ray crossed by one or more dark bands 

 according to the thickness of the quartz plate used. The dark bands in the 

 one spectrum correspond in position with the bright bands on the other. 

 When the polarising nicol is rotated the bands move along the spectra. 

 Tail's idea was to use this form of apparatus for studying the bright line 

 spectra of faintly luminous objects, such as nebulae and comets. By employing 



12 a 



