6 Burled Weed Seeds 



the wind. As far as possible the ground in the immediate neighbourhood 

 of the glasshouse was kept free from weeds and nothing was allowed to 

 flower. After the experiment had been going on for some long time a fair 

 number of Senecio vulgaris appeared in boxes in which they had not been 

 evident in the early days of the test. Such seedhngs were looked on with 

 suspicion as being wind carried in all probability, and no account has 

 been taken of them in the final results. After about 18 months a number 

 of Sonchus oleraceus seedlings began to appear, and these also were 

 considered as derived from wind carried seeds, as the species is so abun- 

 dant in the immediate neighbourhood of the Laboratory. The conditions 

 of the experiment, the t3'pes of the various seeds, and the distribution of 

 the seedlings in the boxes render it highly improbable that any of the 

 other seeds were of external origin, and it may be taken that the seedlings 

 were derived from seed already in the soil when the samples were taken 

 from the field. 



Several samples were taken in each field selected for experiment and 

 the fields were so chosen as to include land of as varied history as possible. 



A. Old Pasture (never under arable as far as is known). 



(1) Harpenden Common. Sampled June 16th, 191.6, 3 holes. 



(2) Park Grass. Sampled April 11th, 1916, 4 holes. 



B. Pasture, origiualh/ arable. 



(1) Meadow at Laboratory house. Sampled April 28th, 1916, 4 holes. 



(2) Barn Field Grass. Sampled Oct. 19th, 1915, 4 holes. 



(3) Geescroft Field. Sampled about April 20th, 1916, 4 holes. 



(4) New Zealand Field («). Sampled Aug. 24th, 1915, 4 holes. 

 New Zealand Field (b). Sampled Sept. 24th, 1915, 4 holes. 



C. Arable Land. 



(1) Long Hoos. Sampled April 25th, 1916, 2 holes. 



(2) Barn Field (8-0). Sampled April 6th, 1916. 2 holes. 



(3) Agdell (5). Sampled April Sth, 1916, 2 holes. 



Li the tables relating to each field the number of seeds which ger- 

 minated at each inch are added together for all the holes sampled, 

 except in the case of the totals, which are given for each individual hole 

 for each species. 



A. Old Pasture. 



(1) Harpenden Common. 



So far as is known this is genuine old pasture and has never been under 

 tillage. The soil is of a fairly light nature and is deep, as the lowest inches 



