PLANTS, WITH GLASS ROOFS. 193 



Angle of incidence 85, 80, 70, 60, 50, 40, 30, 20, 10, 1. 



Per centage of rays reflected. 50, 41,22, 11, 5, 3, 2, 2, 2, 2. 

 Now if we suppose a roof in one plane with the sun shining on it at six 

 o'clock in the morning, and at six o'clock in the afternoon, at an angle of 85, 

 which would be the case in March and September, fully one half the rays 

 which fell on the roof would be reflected ; while, in the case of a ridge and 

 furrow roof, if he shone on half the roof, that is on one half of each of the 

 ridges, at any angle with a perpendicular not exceeding 30, at the same 

 periods, only 2 per cent, of the rays would be reflected. Suppose, then, the 

 area of the entire roof taken as one plane to be 100 square yards, and, to 

 facilitate calculation, that only 100 rays fell on each yard, then the total 

 number which would enter through the roof in one plane would be 50,000, 

 while those which would enter through the ridge and furrow roof would be 

 99,000, or very nearly double the number. If we compare a roof in one 

 plane with the framework in wood, with a similar one with the framework 

 of iron, and take the space rendered opaque by the wood at 21 per cent., 

 and by the iron at 7 per cent., then the greater number of rays admitted at 

 all times by the iron roof over the wooden one will be as three to one. 



487. Iron roofs have been objected to from their somewhat greater 

 original expense, from their supposed liability to break glass by contraction 

 and expansion, and from the iron being liable to conduct away heat in winter, 

 and to become hot to such a degree as to be injurious to the plants in sum- 

 mer. With respect to expense, that is, we believe, now considered the chief 

 objection ; but though it may be greater at first, yet it is amply compensated 

 for by the greater durability of iron houses, when properly constructed, and 

 when the iron is never allowed to become rusty for want of paint. As a proof 

 of the durability of iron houses, we may refer to the iron Camellia house at 

 Messrs. Loddiges', erected in 1818, and the iron houses in the Horticultural 

 Society's garden, which were erected, we believe, in 1823. The breakage of 

 glass supposed to result from the contraction or expansion of the metal was 

 at one time considered a very weighty objection ; but the severe winter of 

 1837-8 did not occasion so much broken glass in iron as it did in wooden 

 houses. A bar of malleable iron 819 inches in length, at a temperature of 

 32, only increases in length one inch, when heated to 21 2 s ; but this differ- 

 ence of 180 of temperature is more than plant-houses are liable to ; indeed 

 50 or 60 are as much as is necessary to be taken into account. If we sup- 

 pose the iron-\vork is fitted at a period of the season when the temperature 

 is 55, then 50 lower would be within 5 of zero, and 50 higher would be 

 105 ; extremes which the iron roof of a hothouse will seldom exceed. Now, 

 according to the above data, a bar ten feet in length would extend or contract, 

 by the addition or reduction of 50 of heat, l-25th of an inch as nearly as 

 possible. An iron sash-bar, half-an inch thick between the two edges of 

 the glass, would not expand in thickness, from 50 of heat, much more than 

 one six-thousandth part of an inch. It may easily be conceived, there- 

 fore, that the lateral expansion of sash bars, which are in general not quite 

 half an inch in thickness, by any heat which they can receive on the roof of 

 a hothouse, will never have any effect on the glass between them. To guard 

 against all risk of breakage from this cause, however, it is only necessary 

 not to fit in the panes too tightly. Indeed, the objection may now be con- 

 sidered as given up by all experienced hothouse -builders. The liability of 



