Alternating Annual and Biennial Habit. 301 



PLANTS PER PERCENTAGE OF PLANTS 

 RKn PT ANTS 



SQUARE METER WITH STEMS 



No. 1 1350 100 23 % 



" 2 and 3 630 + 650 50 43 " 



" 4 380 30 58 " 



That is to say, the closer the plants are together, and 

 the less room each one has, as a result of this, for the 

 unfolding of its leaves, the smaller is the number of 

 annual plants. 



In the following year I repeated this experiment, but 

 this time with the seeds of annual plants. The result 

 was, however, the same. There were 1188 plants on one 

 bed of 13 square meters, that is, about 90 per square 

 meter; of these 20% were annuals. On the other bed 

 of the same size there wxre 348 plants (or 27 per square 

 meter) and 54% developed stems. 



I repeated the same experiment once more, in 1890, 

 with the seeds of an annual plant of 1889. On tlie one 

 bed there were 40 plants per square meter, of which 

 17% were annual. On the control bed there were only 

 ten plants in the same area, and of these 72% produced 

 stems in the first summer; the extent of the bed in both 

 cases was 5 square meters. 



In 1891 I investigated the influence of the distance 

 between the plants in an experiment with Oenothera 

 laez'ifolia, raised from the seeds of an annual race which 

 had been selected for three generations.^ The two beds 

 were of the same size, had the same aspect and the same 

 soil, and both received a similar and liberal dressing of 

 guano. They were sown in the middle of May on the 

 same day, but at the end of July they contained 195 

 and 638 plants respectively (per each 6.S square meters). 

 As a result of this, the bed in wliich the plants were far 

 See the pedigree in Vol. I, p. 224. 



