188 EXPERIMENT ON HEATING EARTH. 



party-wall would probably tend to economy of heat. Our fire 

 was lighted or revived, as the case might be, once a day, exactly 

 as in the management of a vinery at work, and during the winter 

 only three times a week. The fuel was coal-slack and cinders. 

 Shelter from prevailing winds should be provided for, and tall 

 plants moored by cords to tent-pegs in the ground. In the 

 matter of plants recommendable for an experiment, I should 

 omit Canna and HedycMuyn, the marbled-leaved Begonias, and 

 all winter flowering things, unless possessing great beauty of 

 foliage. Only the very best Solanuras to be used, and these 

 treated as described above. I have recorded the most successful 

 subjects in ray own experiment, and strongly recommend their 



trial. 



As fresh and likely subjects for trial, I may suggest tender 

 annuals, such as the beautiful painted Amaranths, tricolor and 

 hicolovy with the common Cockscomb and the rarer sorts of 



■ 



Impatiens ; IpomcBa ; Nil and Hederacm, with their handsome 

 derivatives now to be had in variety at the shops, to say nothing 

 of the hopes of seeing the exquisite QuamocUt in its true cha- 

 racter. Every Convolvulaceau should be tried. Of the annuals, 

 some to be sown where they are to remain, and others raised first 

 under glass. The perennial sorts to be first grown, at least for one 

 season, in large pots, under glass also. Iporncea JicifoUa, omitted 

 in my report, succeeded perfectly thus, and the same treatment 

 would apply to Phaseolus Caracalla^ a plant well w^orthy of trial. 

 I can recommend no other Phaseolus or Dolichos, unless perhaps 

 D. ligiiosus. Bongainvillcea will suggest itself, with Passiflora 

 and Tacsonia, Aralia papyrifera^ Mmidevillea, Schuhertia^ with 

 Lapageria, and other Alstrcemerian forms, all likely; Stepha- 

 notiSy Alamanda^ Dipladenia, EcJiites, more uncertain, but how 

 amply they would repay if successful. Campylolotrys, Clero* 

 dendroHy Citrus (plunged), Curcimia, Dioscorea discolor. Gar- 

 deniaj Gloriosa, Papyrus, Pergidaria, all nearly certain. A 

 trial of Cyanopliyllum will be indispensable, with summer flower- 

 ing Melastotnads and the new variegated Cyperus, 



I may here call attention to the evident suitability of the treat- 

 ment to Cucurhitacem of all kinds as a cheap and effectual way of 

 producing a collection for exhibition this year. By setting apart 

 a bed for the express purpose, and restricting each plant to one 

 runner, a great number of sorts might be grown. The plants to 

 be grown in rows on each side of the bed, and the vines trained 



