THE DEPTH IN THE SOIL AT WHICH*PLANTS OCCUft. 495 



this connection. He says : — " The seed under natural conditions 

 germinates near the surface of the ground; but the fully 

 matured corm is rarely found at a less depth than 3 in., and 

 often occurs 4 or 5 in. deep. The small one-year-old corm is 

 always found near the surface, and it is evident that in the 

 annual process of reproduction it possesses some power of 

 descending deeper into the ground. This would, however, seem 

 inconsistent w T ith the mode of reproduction, in which the seed 

 corm is found on the top of that which it replaces." He 

 then describes an experiment in which Crocus seeds were sown 

 (in August 1878) half an inch below the surface of the soil in a 

 flower-pot 5 in. deep. On investigating matters in September 

 1884 the corms were found at the bottom of the pot, and above 

 each a string of empty corm-tunics representing the downward 

 progress of the corms. He concludes : — " The process of the 

 descent of the corm from near the surface to the necessary 

 depth is difficult of explanation, and it must be viewed as one of 

 the many self-protective phenomena in plant-life, the modus 

 operandi of which we do not understand." 



It is my impression that these 4 'ephemeral " or contractile 

 roots are of much commoner occurrence on Crocus corms than 

 Maw suspected. This opinion is based on the fact that I have 

 never for many years had any difficulty in procuring ample 

 material of these roots for class demonstration, whilst on at least 

 one occasion the difficulty was to obtain developing corms from 

 which they were absent. Nevertheless, the conditions of their 

 variable and perplexing occurrence have yet to be determined. I 

 am inclined to think them especially frequent (in C. aureus and 

 minimus) in the case (1) of w r eakly non-flowering corms, and (2) 

 of corms which have from any cause reached a level too near 

 the surface. This view is advanced with all reserve, as the 

 examples that have come under my own observation are too few 

 to justify any general conclusions as yet. 



Other examples of the type in which only a limited number 

 of the roots formed is contractile include Tigridia pavonia, 

 Gladiolus communis, and several other Iridacese. From amongst 

 the Dicotyledons, Oxalis eleyans behaves in conformity with 

 the above. 



(3) A third type, common amongst dicotyledonous plants, 

 may be described. In this the primary tap-root persists, and it 



