THE EFFECT OF GRASS ON TREES 271 



to show that the presence of this surface crop reduced the vigour 

 of the tobacco plants from 100 to 



71 in Ridgmont soil, 

 52 in Long Ashton soil, 

 63 in Stackyard soil, 

 55 in sand. 



These values (XIV, Appendix, p. 141) certainly show no reduc- 

 tion in the action of grass in the case of the Long Ashton soil as 

 compared with the Ridgmont soil, nor even in the case of the 

 sandy soil and of pure sand. Numerous other experiments were 

 also made wherein the comparative effect of grass on fruit trees 

 was determined when these were grown in the Ridgmont soil or 

 in sand, and eleven such series gave 51 and 48 as averages 

 of the vigour of the trees in the two cases practically identical 

 (XIII, 105). It may be concluded, therefore, that the action 

 of grass is almost independent of the actual nature of the soil. 

 But differences in the consolidation of a soil in situ, differences 

 in the depth of soil available for growth, differences in drainage, 

 and differences in the nature of the subsoil are not reproduced 

 in pot experiments, and it is to these, therefore, that we must 

 look for an explanation of the different effect of grass on trees 

 in different localities (see p. 307). 



The action of grass on trees is evident from an examination 

 of any part of the tree ; the weight of the crop is affected, so is 

 the leaf-size, the total weight of. foliage, the length and vigour of 

 the new wood formed, the increase in weight of the stems, and 

 the development of the roots. The roots of grass-grown trees 

 have been repeatedly examined to see if any definite affection or 

 modification in structure could be detected, but without success ; 

 the action of the grass appears merely to consist of a reduction 

 in the number of new rootlets formed by the tree, and still 

 more a reduction in the rate of development of these. Thus, 

 in a number of cases, where both established and recently trans- 

 planted trees were examined, grass was found to have reduced 

 the number of rootlets forming in the season from 100 to 71 (a 

 reduction of 29 per cent.), whilst the weight of them was reduced 

 from 100 to 36 ; so that these 71 new rootlets had a compara- 

 tive weight of only 36, and their rate of development must have 

 been reduced by nearly 50 per cent. (XIII, 16). 



The effect on all the features of grass-grown trees which have 

 so far been mentioned can be expressed numerically, but there 



