322 
DR. E. H. GRIFFITHS AND MR. EZER GRIFFITHS ON THE 
temperature of the walls of this enclosure were rendered visible by the movements of 
a galvanometer spot. By controlling the flow of air the oscillations of the spot could 
be kept within narrow limits and, under normal conditions, the oscillations did not 
exceed a hundredth of a degree. 
The interior of the wooden vessel, M, was packed with slag wool, the passage for 
the withdrawal of the copper enclosure being kept clear by a cylindrical tube of card¬ 
board, N. 
The space between the top of the enclosure and the outer lid was filled by wrap¬ 
ping felt matting around the glass tubes and leads. 
The block of metal, G, was suspended within the enclosure by a single glass tube, H. 
The central hole contained the heating coil, 0, of manganin wire wound on a mica 
rack and immersed in. a light paraffin, usually petrol. 
The beating coil was fixed to a short taper plug of copper, K, which closed the 
central hole. Many such coils and plugs were used in the course of the work, since 
it was rarely possible to withdraw a plug, after it had once been firmly inserted, 
without doing considerable damage to the heating coil. The usual resistance for the 
heating coil was about 26 ohms. 
In our preliminary experiments we used many modifications of the above 
arrangements : e.g ., the central hole was but one half the diameter and was closed by 
a rubber washer pressed down by the lid ; the heating coil was of 12 ohms resistance 
with leads either of manganin or of copper. The defect of this arrangement was that 
the rubber disintegrated and, with such a volatile liquid as pentane, it was necessary 
to ensure that no evaporation could take place during the interval (prior to an 
experiment) that the block was at room temperature. 
The platinum thermometer was inserted in the cylindrical hole, T, the annular gap 
between the stem and the walls being closed by a packing of asbestos thread. The 
thermometer was connected to a Callendar and Griffiths resistance bridge which was 
carefully recalibrated during the course of the work. 
It may be remarked that, in the preliminary experiments referred to above, a 
thermometer with a much larger size coil was used. The results obtained were in 
complete agreement with the values subsequently found by the modified arrange¬ 
ments, using a smaller sized thermometer (Q) the coil of which was constructed of 
platinum wire of a higher degree of purity. 
The same samples of metals were used in this as in the previous work, their 
dimensions being reduced to the required size by machining. 
In the case of sodium, this procedure could not be followed, so another copper 
sheath was constructed and filled in the manner described below. 
(2) Construction of the Sodium Block. 
The block of sodium was constructed on a different plan* from that described in the 
* We are indebted to Mr. Edgar A. Griffiths for this design. 
