814 Mr. C. W. Heaps on the Effect of 



The apparatus used was the same as described in the former 

 paper, but as the non-uniformity of the field of the electro- 

 magnet has been examined quantitatively there, by comparison 

 with the field of a solenoid, no attempt has been made to 

 allow for this error in the present experiments. In any case 

 it is too small to cause very serious trouble. The electro- 

 magnet with poles four centimetres in diameter and two 

 centimetres apart has been used throughout for producing the 

 magnetic fields. Using a bismuth spiral in the first instance 

 to measure the fields, a curve wKs plotted between current and 

 field strength, so that knowing the current through the electro- 

 magnet at any time, the corresponding field could be read off 

 from the curve. 



The tellurium (from Kahlbaum) used in the experiment 

 was cast in cylindrical form, so that for this substance the 

 crystalline structure was especially pronounced. The small 

 needle-like crystals could be seen radiating like spokes from 

 the axis of the cylinder. Zinc (Kahlbaum) and gold 

 (Johnson, Matthey & Sons) were used in the form of wire. 



The natural crystals examined had in general a fairly high 

 specific resistance, so that small bars cut out with a diamond 

 saw could be tested very easily. The chief difficulty lay in 

 making electrical connexions with the bars. Streintz and 

 Wellik* conclude from their experiments that because of 

 imperfect contact it is extremely difficult to measure the 

 resistance of crystals. Konigsberger f , on the other hand, 

 says that if proper precautions are taken this contact re- 

 sistance can be made exceedingly small — errors caused by it 

 should be no greater than one to five per cent. He used 

 pressure contacts and connexions soldered on copper plating. 

 The latter method was tried and in most cases found to be 

 the better. For bars of lead sulphide, soft tinfoil held firmly 

 by springs against the end made a fairly good contact, but 

 the following method was hit upon as most satisfactory. A 

 small pit was made in a piece of copper and filled with soft 

 bismuth solder. The end of the galenite rod was immersed 

 in the melted solder, and when cool the expanded solder 

 clamped the rod very securely. The presence of contact 

 resistance was easily discovered from the behaviour of the 

 galvanometer when the Wheatstone bridge was in balance. 

 Ordinarily, temperature changes caused a slow and steady 

 drift of the galvanometer, but when a poor contact was made 

 with the specimen the drift was not steady. Waverings and 

 sudden jumps occurred, and often when a lead was touched 



* Phys. Zeitschr. xii. p. 845 (1911). 

 t Phys. Zeitschr, xii. p. 1139 (1911). 



