662 PROFESSOR THOMSON ON THE ELECTRO-DYNAMIC QUALITIES OF xMETALS. 
the metal collars round the india-rubber, were stopped with red lead, and after some 
trouble were made water-tight. Thus the heater and coolers, without any metallic 
communication with the conductor, served the purpose of keeping the required sup- 
plies of hot and cold water round it in the proper places. The spaces for the ther- 
mometers were firmly stopped below with corks fitted to support the lower ends of 
the bulbs in perfectly fixed positions (fig. 3 ). Little collars of cork were put round 
the tubes just above the bulbs, and pushed down into the upper end of the hollow so 
as to hold the thermometers firmly and prevent all motion of their bulbs. 
32 . Various methods of heating the central part of the conductor were tried. 
First, as in previous experiments, water in the central vessel was kept boiling either 
by a gas-lamp under it, or by steam blown into it from a separate boiler ; then a com- 
plex system with a boiling fountain, by which I attempted to get a perfectly uniform 
stream of water at a constant temperature, as little short of boiling as possible, to 
flowthrough the open spaces between the different slips within the central vessel, 
was used during several experiments. Lastly, water filling the eentral vessel was 
kept at a very constant temperature, near the boiling-point, by a gas-lamp below it, 
regulated by a person watching the indications of a thermometer with its bulb fixed 
in the middle space between the slips, as nearly as possible in the centre of the 
compound conductor. This last I found to be by far the best plan, and I used it in 
all subsequent experiments in which any external application of heat to the conductor 
was required. 
33 . Each cooler was divided into four compartments by partitions of tin-plate, 
stopping all communication from one to another, exeept through the spaces between 
the different slips composing the conductor. A constant stream of cold water (§ 22), 
introduced by the eompartment nearest to the middle of the conductor, and drawn 
off by an overflow pipe, from a compartment next the end, was thus forced to flow 
all through among the different slips, and, as I found by placing thermometers in 
various positions in each compartment, gave a very satisfactory effect in fixing the 
temperature of the whole section of the conductor. 
34 . The experiments were made in other respects exactly as described above 
28 and 29 ) ; the electric current, however, not being often again kept up for a 
longer time than ninety-six minutes, since the fumes, which always began to rise from 
the battery after the current had been flowing for about an hour, began after half an 
hour more to occasion great irregularities and inconvenience by causing the liquid 
(which sometimes became very hot,) to foam and overflow in some of the iron cells. 
The atmosphere had been in previous experiments sometimes rendered intolerable for 
the observers, by the acid vapour ; but this evil was done away by covering the battery 
with cloths kept moist with ammonia and water, and by moistening other surfaces 
in the neighbourhood in the same way, so that the fumes never got far without 
meeting vapour of ammonia and combining into white clouds, which were perfectly 
innocuous. 
