Intelligence and Miscellaneous Articles. 547 



serve as measure may easily be too large, and hence the accuracy 

 of the observation be materially restricted. In such cases it is best 

 to place from the outset the stronger source of light at double or 

 treble the distance, and to multiply the corresponding angles by 

 4 or by 9. For use in the determination of gas-flames, which have 

 usually from 5 or 6 to 10 or 14 times the intensity of the normal 

 candle, it is advisable, owing to the easiness of the multiplication, to 

 replace the distances 2 and 3 by 1 s/ 5 and 1 a/ 10, — thus, for instance, 

 with a distauce of the normal candle of m, 3, to place the light 

 to be measured at the distances 



m '3x V5 =0 m '6708, 

 and0 m '3x V 10=0*9486; 

 and then multiply the tangent of the angle read off by y> or 10. 



The instrument I use, as constructed by the skilful maker, Hugo 

 Schickert, in Dresden, is arranged as follows. 



In a cylindrical brass box, 1 decimetre in height and diameter, on 

 a stand, on which it can be adjusted at any height, four tubes 

 3 centimetres in width are inserted, two of which have conical pro- 

 jecting pieces, and two have weak lenses. In this box is the screen 

 moveable about a vertical axis, and provided at the top with an ali- 

 dade, which moves over the horizontal divided circle, the null -point 

 of which is exactly over the axis of two tubes opposite one another. 

 The null-point of the alidade is in the plane of the paper screen ; 

 the angle of the screen, with the vertical plane through the zero- 

 point of the circle, may be read off to ten minutes by a nonius. The 

 conical tubes, which cross at right angles and lead the incident 

 rays against the paper screen, have an arrangement to introduce 

 diaphragms and ground or coloured glass plates, the use of which 

 is advantageous with very strong light, or with flames of unequal 

 brightness. The interior is completely blackened. 



To use the apparatus, the vertical axis is placed just over the 

 apex of a right angle drawn on the table, on the legs of which are 

 the two lights to be compared ; the screen is then turned until the 

 transparent spot quite disappears from the eye of the observer, who is 

 looking through one ocular tube. An observation is then made 

 through the other tube exactly in the same manner ; the mean of the 

 two gives the value to be placed in the calculation of the tangents. 



Blackening the sides of the observing-chamber is superfluous; 

 it is sufficient to place behind the flames dark and non-lustrous 

 screens of such a size that the conical tubes, if continued, would 

 cover them. — PoggendorfFs Annalen, September 1866. 



CONDUCTIVITY OF MERCURY FOR HEAT. BY G. GRIPON. 



This was determined by the method of Despretz and Peclet. Ex- 

 pressed in Peclet's unit it is =1*67 (lead =3-84). Hence the con- 

 ductivity of mercury for heat is =0*407 that of lead. If the conduc- 

 tivity of silver be made equal 100, that of mercury is 3*54, — that is, 

 greater than that of marble, and somewhat greater than that of gas- 

 coke. As the conductivity of mercury for electricity is = 1*80 (silver 

 = 100), it is seen that these two numbers do not coincide as is the 

 case with solid metals. — Comptes Rendus, July 9, 1866. 



2N2 



