of spokes, or roots, over the old ones each year, and as they grow Culture of 

 the old set of roots, which has done its work, decays away. Eremuri 

 The nucleus has, after one or two years, a tendency, especially 

 in E. Himalaicus, to split itself into two crowns, each of 

 which when strong enough, sends up a flower stem and be- 

 comes a separate plant, and each of these puts forth its 

 own system of roots which tangle and intertwine with each 

 other. These crowns subdivide again year by year and 

 form a dense mass of detached but interlaced plants which 

 are too weak to produce flower stems. On digging up and 

 investigating one of these clumps which for four years had 

 produced no blooms, I found that it consisted of no less than 

 fifteen distinct crowns, each with its own set of roots, and it 

 was a labour of patience and some difficulty to disentangle these 

 one from the other with as little injury as possible to the brittle 

 roots. My practice now is to take up in August or September 

 every plant that has during the previous springjproduced two 

 flower spikes, separate them, and replant. Although it is not 

 possible, even with great care, to avoid breaking the roots to 

 some extent in the process of lifting, I do not find that the 

 plants die or suffer appreciably : sometimes a year elapses before 

 they flower again, but I think that is more probably due to 

 their not having reached flowering size. 



I do not protect them from the cold of winter and have 

 not observed that they have suffered in consequence. The 

 period of risk comes with spring, for they start into growth very 

 early in the year, often in February. A most fat and interesting 

 head pushes itself up through the ground and thickens and 

 strengthens from day to day : it almost seems to grow as one 

 watches it. The broad and handsome but tender leaves speedily 



141 



