412 DR. F. HORTON ON THE NfODULUS OF TORSION AT, RIOIDTTV OF 



necessary to make a special determination of this, especially seeing that the subject 

 has been recently investigated by several observers.* The value used for a in the 

 above factor is that given by HOLBORN and HENNING, viz., '00000054. The method 

 by which y was determined forms Part II. of this paper. In some cases the 

 temperature coefficient was actually found for the fibre in question, in others a mean 

 value was taken. 



The logarithmic decrements of the torsional oscillations of the fibres were found at 

 the same time as the periods of vibration. The correction to the observed period on 

 account of the damping of the vibrations was applied when necessary. 



The Measurement of the Radii of the Fibres. 



When the observations of the torsional vibrations were completed, the fibre was 

 removed from the jacket, and its mean radius was determined. For this purpose the 

 apparatus shown in fig. 3 was used. The method consists in measuring the circum- 

 ference of the fibre by rolling it between two fine glass tubes and counting the 

 number of revolutions it makes in travelling a distance of 5 millims. 



The glass tubes A and B are capillary tubes of uniform external diameter. B is 

 fixed by means of wax on to a small adjustable table D, shown on a larger scale in 

 fig. 4. The height of this can be carefully adjusted by means of the milled head E, 

 and the table can be set horizontally by levelling screws. A was drawn down from 

 a wider tube which is firmly attached to the brass sliding table F. This moves along 

 two slides, one of which is V-grooved and the other plane, after the manner usually 

 adopted in travelling microscopes. The table F can be pushed forward by turning 

 the milled head H, which is attached to a steel screw with a |~millim. pitch thread, 

 so that a slow motion of the table can easily be made. The distance through which 

 the table moves is measured by means of a fine Zeiss scale on glass, which lies above 

 a circular aperture in the centre of the table and is illuminated from below. The 

 scale is viewed from above by the microscope M, fitted with a rocking-plate micro- 

 meter,f by means of which measurements to '00002 centim. can be made. 



In order that the fibre shall roll, it has to be perfectly free from dirt or grease. 

 The tubes A and B were generally renewed for each fibre ; the glass, when freshly 

 drawn out and afterwards protected from dust, being found to work better than when 

 cleaned in the ordinary way. 



A length of about 1 centim. of the fibre is experimented on at a time. This is 

 prepared in the following manner : The end of the fibre to be measured is first well 

 cleaned in chromic acid and afterwards washed in water. It is then placed transversely 



* SCHEEL, ' Ber. d. Deut. Phys. Ges.,' 5, 1903; HOLBORN and HENNING, 'Ann. d. Phys.,' 10, 2, 1903; 

 P. CHAPPIUS, 'Verb. Naturf. Ges. Basel,' 16, 1903. 



t For description see POYNTING, 'Phil. Trans.,' A, vol. 182, p. 589, 1891, or 'The Mean Density of the 

 Earth,' p. 95. 



