iv.] NATURE'S EXPERIMENTS 37 



described structures are due to the direct action 

 of air, water, soil, etc., my experiment with 

 the water crowfoot showed that both types of 

 leaf appeared by heredity when the seed was 

 sown in earth and air, and not in water. No 

 further proof is required. 



Proserpinaca palustris is like Cyrilla (fig. 9). 

 They both grow in water, but the stems reach 

 into the air, so that they habitually produce 

 both kinds of leaves. One leaf of Cyrilla is 

 represented with a laciniate margin. This is due 

 to its growing partly in air and partly in water. 

 It is a common feature in the floating leaves 

 of Ranunculus heterophyllus. A leaf may be 

 found standing obliquely, so that half the blade 

 has grown in air, the other half being sub- 

 merged. The result is that the former half of 

 the blade is completed, but the latter dissected. 

 Another feature is noticeable in these two 

 plants (Cyrilla and Proserpinaca). Always form- 

 ing both kinds of leaves, they are only formed 

 in air or water respectively, and neither can 

 be formed under reversed conditions. But in 

 Ranunculus the dissected type has become so 

 rigidly fixed that, as I have shown, it is now 

 formed in air. So that while only the capacity 

 of forming them is heredity in the former, the 

 actual structure itself is hereditary in the water 

 crowfoots. 



