PROTECTION OF PLANT-HOUSES DURING COLD NIGHTS. 311 



ciently, is that of enclosing a complete stratum of air between 

 it and the glass, this body of air being entirely shut off from the 

 surrounding outer atmosphere, as far as may be practicable to 

 do so; and as air is a bad conductor of heat, the warmth of the 

 interior is prevented from passing to the exterior atmosphere, by 

 means of direct radiation from the glass ; or, in other words, the 

 exterior atmosphere, being prevented from coming in contact 

 with the glass, cannot absorb from the interior any sensible por- 

 tion of its heat. To secure this advantage, it will be evident 

 that the covering must be kept some distance from the glass, 

 and should be on every side where the structure is formed of 

 glass ; the coverings, in fact, should form a complete case to all 

 the glazed portion of the structure.* 



So far, so good. As a matter of protection, and nothing else, 

 this is all very well. The advantages of such a covering will 

 be obvious to every one ; and, as a matter of protection alone, it 

 deserves every word that can be said in its favor. Whether it 



* In the different experiments, it appears that the cooling effect of 

 wind at different velocities on a thin glass surface, is very nearly as the 

 square root of the velocity. In these experiments, the velocity of the 

 air Avas measured by the revolutions of the vanes of a fan. The tem- 

 perature of the air was 68^, the time required to cool the thermom^eter 

 20^ was noted for every different velocity, and the maximum tempera- 

 ture of the thermometer in each experiment was 120^. In still air, it 

 equired 5' 45'' to cool the thermometer this extent, and Table VIII. in 

 the Appendix shows the time of cooling by air in motion. 



In consequence of the large quantity of glass used in the construction 

 of horticultural buildings, the cooling effect of wdnd is of considerable 

 importance. We find, however, that, with an increased velocity, the 

 cooling effect is considerably less in proportion, on glass, than on metal, 

 and it will be very much less on window-glass than even what is stated 

 in the table. As glass is an extremely bad conductor of heat, the 

 increased thickness which vrindow-glass possesses over that which 

 forms the bulb of a thermometer, will make a material difference in 

 the quantity of heat lost by the abduction of the air, there will be, as in 

 this case, a greater difference between the temperature of the external 

 and the int^ernal surface. The cooling effect of wind, therefore, is not 

 near so considerable as is generally supposed ; and the effect of wind in 

 hot-houses is very much increased by open laps and accidental fissures 

 in the glazing of the sashes. 

 27 



