184 



Dr. R. W. Coppinger on 



[Mar. 14, 



wooden battens (NM), standing 24 inches apart, and connected above by 

 a light crosspiece (00). In the upright battens, at 4 inches above their 

 bases, are apertures 1^ inches square, for the reception of the bar of 

 ice (AA) on which, the experiment is made, while a series of thermo- 

 meters (DDDD) are placed 4 inches apart, secured above to the cross- 

 piece (00), and having their bulbs imbedded in the icebar below. 



For the manufacture of ice-bars I had constructed a strong copper 

 tube, 29 inches long by 1^ inches in sectional area, from which on being 

 filled with water, and exposed to a low temperature, bars of fresh or 

 sea- water ice were obtainable. The removal of these bars from the 

 mould was effected by the application of warm water to the outside of 

 the tube, which, melting the surface of the ice-bar, allowed it to be 

 extracted. 



The ice-bar, being placed in position, projects, when of its full 

 length, 5 inches beyond the upright batten (M), as shown in the figure. 

 Below the projecting end is placed a spirit-lamp (J3), fitted with a 

 sheet-iron plate (C), curved in such a manner as to prevent the bare 

 flame from touching the ice, and to divert the drops of melted ice from 

 the wick of the lamp. This lamp can be so placed that its heat will 

 bear pretty accurately on the under surface of the bar at either 2, 3, 

 or 4 inches as desired from M. The thermometers being placed at 

 intervals of 4 inches apart, their distances from the source of heat are 

 respectively 6, 10, 14, and 18 inches when the lamp is 2 inches from 

 the upright. Their bulbs are sunk in holes bored in the ice, and are 

 packed in with powdered ice, so as to render the connexion with the 

 bar as perfect as possible without incurring the risk of the instruments 

 being broken. 



At all low temperatures the instruments used were Hick's spirit 

 thermometers. Although these instruments were supplied to the ship 

 as of good reputation, their errors at temperatures below —10° differed 

 so much in each instrument for every few degrees of the scale, that I 

 found it best in recording the rate of heat-flow along the ice-bar to take 

 the temperature actually indicated by each thermometer uncorrected, 

 and to note the range of temperature in each, assuming that its error 

 for every two or three degrees, at all events, was uniform. 



In the earlier experiments I noted the temperature every 5 minutes 

 for about 2 hours, beginning 5 minutes after the lamp had been 

 lighted ; but finding that interval too small to show any decided differ- 

 ence in temperature, I had recourse to 10-minutely, and ultimately to 

 15-minutely intervals. The entire duration of each experiment was 

 limited to one hour, as by that time so much of the projecting end of 

 the bar had melted away as to lessen the influence of the lamp upon it. 



In order to determine whether the rise in temperature indicated by 

 the thermometers might be to any extent due to radiation of heat from 

 ' the lamp rather than to conduction through the ice-bar, the thermo- 



