96 THE LIFE OF THE PLANT 



substances that exist in a plant ; to the other the same 

 ash has been added and also nitrogen in the form of a 

 nitrate, namely saltpetre. Sunflower seeds exactly- 

 similar in weight were planted in both pots, two in 

 each. They have come up, but a distinct difference is 

 evident at the end of the experiment : the first flower- 

 pot contains two miserable, unhealthy plants, scarcely 

 rising above the soil; the second contains two healthy 

 specimens with flowers and seeds, leaves and stem being 

 as well developed as those of sunflowers grown simul- 

 taneously in the best garden soils.^ Yet the only 

 tliff^erence between the two experiments consists in the 

 fact that to the second flower-pot some saltpetre, i.e. 

 nitrogen, was added. Similar results might have been 

 reached had nitrogen in the form of an ammonium salt, 

 instead of saltpetre, been used. The inference is that 

 plants need nitrogen. 



Here is another experiment. We take several jars 

 containing nutrient solutions (fig. 25) ; some of the jars 

 contain all the necessary salts, others the same salts 

 minus that of potassium. We place exactly similar 

 buck-wheat seeds in every jar. After a certain time 

 we notice that the former jars contain healthy plants 

 which flower and produce ripe seeds, while in the rest, 

 instead of developing well, the plants are weak or 

 have perished. We may repeat these experiments 

 many times over, and always get the same results. 

 The inference is that the plant needs potassium, that 

 it cannot exist without potassium. 



You see on this table the results of similar experi- 

 ments also made with buck-wheat seeds (fig. 27). 

 The first, third, and fifth row received a complete 

 nutrient solution, the second received no nitrogen, 



1 Fig. 26 represents on the left-hand side plants grown with saltpetre 

 (for the sake of comparison a leaf of a garden specimen is given), and on 

 the right plants grown without saltpetre. This is the classical experi- 

 ment of Boussingault. 



