BOOK III. GLAZING OF HOT-HOUSE ROOFS. 819 



copper. The use of sheet-copper, however, may now be considered as exploded in hot- 

 house building, wrought-iron being a much more economical, wholesome, durable, and 

 equally elegant substitute for timber. In general, it may be observed, that where sashes 

 and rafters are used in the formation of moveable roofs, a mixture of timber and metal is 

 better than timber alone, the former in extremes of temperature being liable to expansion 

 and contraction. Thus sashes with iron bars, and the outer frame or the styles and rails 

 of timber, move readily in the grooves of cast-iron rafters, because when the metal expands 

 with great heat, the timber in a slight degree contracts. The reverse is also the case, and 

 cast-iron sashes slide readily in timber rafters. In both cases small rollers should be in- 

 serted, either into the sides of the sash or the fillet or groove of the rafters, or both. Cast- 

 iron rafters need not, for general purposes, be more than half an inch 264 

 thick, and six or eight inches deep ; where the house is wide, they 

 require to be supported by slender pillars. Wrought-iron rafters 

 may be rolled out of broad bar-iron (jig. 264. ), so as to present as 

 light and elegant an appearance as our moulded wrought-iron sash- 

 bar. (Jig. 252.) 



1621. Arrangements for covering the roofi of hot-houses by boards, canvass, or mats, to be 

 lifted or rolled up or down, might be easily contrived and advantageously used ; but ex- 

 cepting in pits and low hot-houses, they are not thought worth attending to, it being con- 

 sidered better to gain the admittance of all the light possible, than lose it for the sake of 

 a little economy in fuel. 



1622. The pillars or props which are placed on the parapet, to support the rafters, whether 

 of timber or iron, are generally formed of the same thickness as the rafters, because 

 similar sashes are placed between them. 



1623. Interior uprights to support wide roofs are almost always of iron, either wrought- 

 metal or small cast-iron columns, sometimes forming intersecting arches, or treillaged 

 capitals, or connecting imposts for training creepers. 



1624. The wall-plate, or coping of the parapet, is sometimes a plate of timber, some- 

 times of stone, and occasionally of cast-iron. Wherever upright glass is not employed, 

 it must of necessity form also the guttering for the water of the roof, and at the same 

 time for the water which condenses on the glazed inside of the house. 



1 625. Objections to metallic roofs. In general it may be observed, that till lately gar- 

 deners had a prejudice against metallic roofs. Of authors, who have avowed this, Aber- 

 orombie, Mean, and Nicol, may be mentioned ; others have adopted a cautious neutrality, 

 as M'Phail, Forsyth, Aiton, &c. Philosophical and amateur gardeners have generally 

 approved of their introduction ; among which may be named Knight, Sir George Mac- 

 kenzie, Loddige, and others. We shall here, as briefly as possible, enumerate the ob- 

 jections to metallic roofs, which are expense, rust, breakage of glass, abduction of heat, 

 and attraction of electricity. 



Expense. Metallic houses are, in general, rather more expensive than wooden ones ; but they admit 

 more light and are more durable and elegant. 



Rust. That all ordinary metals are liable to rust is undeniable. This objection cannot be got rid of. 

 The reply is, balance against it the advantages of light and durability ; and take into- consideration that 

 careful painting will in a great degree prevent it. Knight observes, if one third of the sum requisite to 

 keep a wooden roof properly painted be expended upon an iron roof, no injury will ever be sustained from 

 the liability of that to suffer from rust. (Hart. Trans, v. 231.) 



Breakage of glass. This is altogether denied, as respects cast or wrought iron at least, and if appli- 

 cable at all, can only be so to copper or compound metallic roofs, where weakness produces a bending of 

 the sash ; or where corrosion or unequal expansion of improper mixtures of metals as iron cased with 

 copper, occasions a twisting of the bar. Cast-iron or solid wrought-iron frames, have never been known 

 to occasion the breakage of more glass than wood. The grand cause of the breakage of glass, arises in 

 almost every case from glazing with broad laps. The expansibility of copper is greater than that of brass, 

 and that of brass greater than the expansion of iron in the proportion of 95, 89. 60. (Young's Lect.} Con- 

 sequently copper is above one third part more likely to break glass than iron ; but when it is considered, 

 that a rod of copper expands only one hundred thousandth part of its length, with every degree of heat, 

 and that iron only expands the one hundred and sixty-six thousand six hundred and sixty-sixth part, 

 the practical effects of our climate on these metals can never amount to a sum equal to the breakage of 

 glass. 



Abduction of heat. The power of metals to conduct heat is an objection, which, like those of rusting 

 and additional expense, cannot be denied. The reply is, the smaller the bars, the less their power of con- 

 ducting ; and a thick coat of paint, and the covering of half the bar by the putty requisite to retain the 

 glass, also lessens this power ; it is added, heat may be supplied by art, but solar light, the grand advan- 

 tage gained by metallic bars, cannot, by any human means, be supplied otherwise than by the transparency 



Attraction of electricity. To this objection it is replied, that if metallic hot-houses attract electricity, 

 they also conduct it to the ground, so that it cannot do any harm. Also that no instance can be produced 

 of iron hot-houses having been injured by the effects of this fluid. 



SUBSECT. 4. Glazing of Hot-house Roofs. 



1626. Glazing was formerly performed with the very worst description of glass, called 

 green glass; and accordingly, Adanson, in 1710, recommends the adoption of Bohe- 

 mian glass, then the best in Europe, but now equalled by our best crown or patent crown 

 tables. If, as Bouguer has shown, one fortieth part of the light which falls perpendicu- 

 larly on the purest crystal is reflected off, or does not pass through it, it may safely be 



