57 



Owing to the abnormally wet weather experienced darinj^ the 

 summer of 1920, the sowing of this experiment was postponed 

 until the latter half of August : unfortunately the season proved 

 to be too late for sowing white clover, meadow foxtail and rough 

 stalked meadow gi'ass, and as the germination was too poor the 

 seedlings of these species were not counted. 



The numbers of surface seedlings of the other three species 

 were counted four weeks after sowing, while the produce from 

 the duplicate beds was cut and weighed about eight months 

 after sowing (8th April, 1921). At the same date the number 

 of tillers per plant was counted on 50 typical plants from each 

 bed. 



The number of seeds employed in these three experiments 

 was : — 



Pot Cultures 20,800 seeds. 



Box „ 27,000 



Bed „ 54,000 „ 



Total ... 107,800 „ 



Soil. — The soil used in the pot and box experiments was taken 

 from the plot on which the beds were sown. It was a light loam, 

 which is the typical soil of the district. 



Seeds. — The seeds used in the three experiments were care- 

 fully selected from the same bulks. All light and imperfect seeds 

 were discarded, while the clover seeds were rubbed between sand- 

 paper so as to reduce the number of hard seeds present; the 

 rubbing had the effect of increasing the percentage germination 

 of red clover from 81 to 94. The selected seeds were tested for 

 germination before commencing the investigations. 



Red Clover.— A glance at the Table will show that the number 

 of surface seedlings at depths of | in. and over was very consider- 

 ably lower in the pot cultures than at corresponding depths in 

 the box and bed cultures. The poor results given by the pot 

 cultures can be explained by the fact that the seeds were sown 

 much too thickly in the pots — about 10 times as thickly as 

 normal lield seeding. As a result, the layers of soil overlying the 

 seeds were completely raised by the pressure exerted by the very 

 broad cotyledons of the seedlings, so that practically the only 

 seedlings that reached the surface were those which were for- 

 tunate enough to work their way up tlu'ough the cracks in the 

 soil. 



A comparison of the results givon by the diflforent depths in 

 the box and bed experiments proves very conclusively that red 

 clover seeds should never bo left uncovered. Covering the seeds 



