30 THE BOOK OF THE ROSE chap. 



ous in luw-lying places and that a covering overhead is a 

 better 2:>rotection than one at the side, because evapora- 

 tion upwards towards the sky produces cold. My 

 neighbour, a quarter of a mile off on a little hill, has 

 always from three to five degrees less of frost than I 

 have ; and even if it were not so I believe that the same 

 amount of fi'ost would be more destructive to vegetation 

 to me in a river valley than it would be to him on the 

 upland. Valleys or low-lying ground, especially if near 

 water, should therefore be avoided, and the uplands in 

 all cases be prefen^ed. 



Mere height above the sea level would not, in most 

 cases, be a matter of much moment ; though on the one 

 hand the top of a mountain would not of course be a 

 desirable spot, and on the other a very flat plain with 

 little height above the sea would probably be subject to 

 severe frost : thus the flats of Cambridgeshire, which 

 have such a slight fall to the sea, are well known as 

 registering very low degrees of temperature. Rather 

 high ground, not necessarily the top of a hill, with valleys 

 in the neighbourhood for the cold air to fall to, would 

 probably be a good situation as to comparative immunity 

 from frost. 



In dealing with situation, I do not allude to different 

 parts of England, though it is well known that Marechal 

 Niel cannot be successfully grown out of doors in York- 

 shire, while in some parts of Devonshire and Cornwall 

 even sub-tropical plants will live through the winter. I 

 am not supposing that the Rosarian can choose his 

 county for Rose-growing : if he could, for combination of 

 climate and soil Herefordshire would probably make a 

 strong bid for the pride of place, at all events for H.P.s. 

 Teas, which depend less upon actual fatness of earth 

 and more upon dryness in atmosphere and soil, may 



