70 



experience in tank-heating, has in great measure cor- 

 roborated these views ; and as his corroboration of the 

 plan I have recommended, embodies some useful hints, 

 I will quote the substance of his remarks : "I had a 

 pit erected, thirty-eight feet long, seven and a half 

 wide, divided into four compartments, for growing 

 melons and cucumbers, with a tank extending the whole 

 length of the pit, six feet wide and six inches deep. 

 Across this I put larch spars, and upon them turves, 

 with the grassy side downwards, and on them the soil 

 for the melons and cucumbers. The plants grew and 

 did well for a time, but they were of short duration in 

 comparison with the dung-bed. Instead of the mois- 

 ture ascending through the soil as I expected, I found 

 that the heat from the tank dried the turves and soil 

 next to them as dry as dust, and that there was no 

 such thing as obtaining a moist heat from hot water 

 without the soil was in contact with it. Next year I 

 put broken stones upon the spars, and turves upon 

 them, and made my arrangements so that I could occa- 

 sionally run water in the tank to wet the turves and 

 the soil next them. This was an improvement ; and I 

 went on prosperously for some years, till the spars be- 

 gan to decay. I then had iron bars put across, and 

 two of the compartments covered with squares, a foot 

 in diameter, and one inch thick ; the other two with 

 slates; both slates and squares jointed with Roman ce- 

 ment, to prevent the soil from getting into the tank, 

 as I had found the inconvenience of it when using the 

 spars. I put some broken stones upon the covers, 

 and turves upon them, and then the soil. Here my 

 original difficulty occurred; the soil next the covers 

 got too dry, and to moisten it from above was imprac- 

 ticable, without making the soil a complete puddle, 

 which would have stopped the healthy growth of the 



