286 SCIENCE AND FRUIT GROWING 



ments were precisely similar, but the trays contained earth only. 

 Yet, with this complete separation of the grass from the tree, 

 the effect on the latter was almost as great as in other cases. 

 A summary of the results obtained in some twenty series of 

 experiments gave the following values for the vigour of the 

 trees under the different conditions of experimentation (XIII, 

 93, and XVII) 



Grass. 



No grass. No gauze Trays and 



or trays. Gauze. gauze. 



ioo 59 61 66 



There is, it is true, a small progressive decrease in the effect 

 of the grass, as the separation of it from the tree-roots becomes 

 more complete ; but this is so small that it is probably not due to 

 the more perfect separation at all, for it is not greater than would 

 be accounted for by the diminution of vigour of growth of the 

 grass itself owing to the restriction of its root-development. 

 Whether this is so or not, it is abundantly clear that the tree 

 suffers owing to what passes down to it from the grass, and not 

 from the grass abstracting from the soil about the tree-roots 

 anything which those roots require. 



Whatever the deleterious substance formed by the grass 

 may be, it appears to be formed, and to act, equally under 

 varying conditions. Many experiments were made in which 

 the trees or the grass, or both of them, were grown in sand, 

 instead of earth, or in sand with different admixtures of earth; 

 but the results were substantially the same in all cases, though 

 grass grown in sand appeared generally to be rather more de- 

 leterious than grass grown in earth, probably owing to the greater 

 facility with which any deleterious substance formed in it would 

 be washed down ; but strict comparisons were difficult, because 

 it is impossible always to get an equally good growth of grass 

 in sand and in earth. Also it was found that but little difference 

 was made whether the grass was grown continuously in the 

 trays, or whether relays of trays with fresh grass seeds germinat- 

 ing in them were used (XIII, 94). 



These pot experiments were generally allowed to run either 

 for two or for three years; after three years the trees suffered 

 from the limitation of earth space. In many cases the effect 

 of the grass during the first season was but small, for the 

 growth during this first season, before the tree has become 

 well established, is largely dependent on the stock of material 

 stored in the tree itself. To allow the grass to exert its full 



