30 The Art of Landscape Gardening 



bank at c ; and this difficulty would be increased by 

 raising the ground to the dotted line d, to set the 

 large house on the same level with the smaller one. 

 It therefore follows that if the house must stand on 

 a natural hillock, the building should not be larger than 

 its situation will admit ; and where such hillocks do not 

 exist in places proper for a house in every other respect, 

 it is sometimes possible for art to supply what nature 

 seems to have denied. But it is not possible in all 

 cases ; a circumstance which proves the absurdity of 

 those architects who design and plan a house, without 

 any previous knowledge of the situation or shape of 

 the ground on which it is to be built. Such errors 

 I have had too frequent occasion to observe. 



"When the shape is naturally either concave or per- 

 fectly flat, the house would not be habitable unless the 

 ground sloped sufficiently to throw the water from it. 

 This is often effected, in a slight degree, merely by the 

 earth that is dug from the cellars and foundations ; but 

 if, instead of sinking the cellars, they were to be built 

 upon the level of the ground, they may afterwards be so 

 covered with earth as to give all the appearance of a 

 natural knoll, the ground falling from the house to any 

 distance where it may best unite with the natural shape, 

 as shewn at e, f, and g: or, as it frequently happens 

 that there may be small hillocks, h and i, near the 

 house, one of them may be removed to effect this pur- 

 pose. This expedient can also be used in an inclined 

 plane, falling towards the house, where the inclination 

 is not very great, as shewn at l ; but it may be ob- 

 served of the inclined plane that the size of the house 

 must be governed in some measure by the fall of the 

 ground; since it is evident that although a house of 

 a hundred feet deep might stand at k, yet it would 



