EXPEEIMENT ON HEATING EAETH. 183 



phrase, we could '* bum anything." Towards the end of May 

 the greater part of the plants were put in, and othei^ at intervals 

 through the months of June and July. A Casella's ground-heat 

 thermometer plunged 6 inches near the side, and consequently 

 immediately over the pipe, stood at 75 degrees without varia- 

 tion during the whole of the warmer part of the year. In the 

 centre of the bed, or midway between the pipes, it stood as 

 constantly, at 70 degrees. This relation between the two 

 indications held good when the general temperature fell, with 

 the close of the year ; the fire was lit or revived daily as for a 

 vinery at work. My border, being a horizontal platform cut 

 into the face of a rising ground a few yards from the bottom, 

 enabled me to avoid sinking the fireplace, which I purpose con- 

 cealing with rough rock-work. The flue was carried up the 

 rise and served to heat another shorter bed, at right angles to 

 the first. Finally, a chimney was placed where it was concealed 

 by some shrubs, and the whole was accessible by a few steps cut 

 into the short slope in front. 



The plant arrangements were carried out very roughly, and the 

 plants themselves put to the severest tests th^t they would ever 

 be likely to be subjected to. In the outset of any investigation, 

 it is well known to experimenters how much truth may be elimi- 

 nated by working in extremes. Thus most of the plants were taken 

 at once from the stove, — some few after a week's probation in a cool 

 house. They were all planted in the same soil, and no protection 

 or shading of any kind given to them. The first observation 

 worthy of remark was that they appeared to receive no check 

 from planting out ; their roots, doubtless, taking at once to the 

 warm earth, and they began at once to ** grow away," while the 

 bedding plants in the garden were still in the chilled and inactive 

 state which they generally assume for the first weary week or two 

 after turning out. Two or three Cannas, including the curious C 

 flaccida, with Hedychinm coronanum^ were placed so as to hide my 

 rude cistern, which they soon did most effectually. <7, iridiflora 

 I failed in procuring in time, and I much regret not being able 

 to try this fine and still rare plant. C fiaccida produced its 

 beautiful yellow Hedychium-like blossoms, but so frail in texture 

 that they were annihilated by the first puff of wind or shower of 

 rain. The Hedychhim flowered well, upon such shoots as had 

 been made previously, in the stove. It then, together with the 

 Cannas^ made a huge tTiicket of shoots. The Cannas neither 



