112 SOIL CONDITIONS AND PLANT GROWTH 



If either the rye, buckwheat or spinach excreted any toxin the amount 

 accumulating during the growth of six successive crops was insufficient 

 to cause any appreciable depression in yield in the next crop ; the ex- 

 ceptional result given by buckwheat in sand could not be confirmed. 



These and similar experiments show that no lasting toxic effect is 

 produced by any of the crops studied, and they rule out any toxin 

 hypothesis as an explanation of the advantages of rotations ^ where there 

 is always a lengthy interval between the crops. They do not, however, 

 show that there is no transient effect, and they are thus quite consistent 

 with some remarkable observations by Pickering on the effect of grass 

 on apple-trees (227). It was found at Woburn — and the observation 

 has since been confirmed elsewhere — that the effect of sowing grass 

 round apple-trees is to arrest all healthy growth and absolutely stunt 

 the tree. The leaves become unhealthy and light coloured, the bark also 

 becomes light coloured, while the fruit loses its green matter and becomes 

 waxy yellow, or brilliant red. Where the grassing was done gradually 

 the trees accommodated themselves somewhat to the altering condi- 

 tions, but never grew so well as when grass was absent. 



This effect might have been due to various causes : changes in 

 aeration, temperature, water supply, food supply, or physical condition 

 of the soil, but careful experiments failed to show that any of these 

 factors came into play. Covering the soil with cement excluded ^ir at 

 least as thoroughly as grass, and yet did not produce the grass effect, 

 nor was it suppressed by wet seasons, liberal watering, or a supply 

 (in pot experiments) of sufficient water or nutrient solution to keep the 

 soils of grassed and ungrassed trees equally moist, or equally well sup- 

 plied with food. On the other hand, the grass effect was produced 

 when perforated trays of sand containing growing grass were placed on 

 the surface of the soil in which trees were growing, so that the washings 

 from the grass went straight down to the tree roots. There seemed no 

 possibility of the grass roots in the trays abstracting anything from the 

 soil, and the only explanation appears to be that a toxin is excreted by 

 the grass. Such a toxin, however, must be very readily decomposed, 

 because no toxic properties could be discovered by laboratory tests 

 either in soil that had been removed from the grass roots or in the 

 washings from the above-mentioned trays. These results show that 

 we must be prepared to consider possible toxic effects of one plant 

 on ano\iier growing alongside of it, and they raise the question whether 

 such effects may not play a part in determining plant associations. 



' Some curious problems are thus left unsolved, some of which are discussed more fiilly 

 by the author in Science Progress, July, 1911, p. 135. 



