284 SCIENCE AND FRUIT GROWING 



(five to six inches of rain) coincide with any increase in the 

 deleterious effect of the grass (XIII, 53). 



However conclusive the water determinations in the Harpenden 

 soil were against a deficiency of water being the explanation of 

 the grass-effect, they were still more so in the case of the Ridg- 

 mont soil, for there it was found that the grassed soil was actually 

 wetter than the tilled soil, at any rate as regards some of the 

 chief plots where the action of grass had been studied. These 

 were four in number, consisting of plots which had been under 

 grass since 1895 or 1899; each plot was a strip of ground 213 

 feet long by n feet wide, and there was one tilled plot in the 

 middle of the grassed plots, and others on either side of them. 

 The plots were all sampled to a depth of nine inches, on three 

 dates during August, 1907, and on one in September, 1910, 

 when it was found that, on three out of four of these occasions, the 

 grassed ground was the wetter of the two, the maximum difference 

 being 2*5 per cent, of water, with an average of 0*7 per cent, for 

 the four dates. Similar, though even larger (2*9 per cent.) 

 differences were found at two other stations in the farm, and, 

 though such differences did not appear to hold good throughout 

 the whole of the ground for the ground varies considerably in 

 different parts their existence at the very stations where the 

 effect of grass on trees is most noticeable, is conclusive against 

 this effect being attributable to lack of moisture (XIII, 59). 



Direct experiments on the effect of increasing the water-supply 

 in the soil were carried out, giving trees under grass weekly 

 allowances of water through pipes placed in the ground under 

 their roots, the supply being adjusted so as to represent about 

 one -and- a- half times the previous week's rainfall. The experi- 

 ment was continued for eight years in one case, and for two years 

 in another. In both cases (XIII, 63) the extra water had 

 increased the vigour of the trees, but had failed to raise this 

 vigour to that of trees grown in tilled ground without additional 



water; the results were 



Grass. 

 Tilled 



ground. No extra Extra 



water. water. 



Vigour of tree . . 100 5*6 1^6 



ioo 13 44 



Field experiments such as these are not always satisfactory, 

 owing to many complications which may arise when attempts 

 are made to alter the conditions in the small patch of ground 

 occupied by a tree, so as to make them different from those in 



