266 TPIE BIOLOGY OF FLOWERING PLANTS 



passing on to the next generation. Infection of the leaves 

 takes place through the hydathodes on the leaf margin and 

 is confined to the epitheme, a hypertrophy of which produces 

 the knots on the leaf margin. After the entrance of the 

 bacteria the hydathode is closed by ingrowth of cells, but 

 this may occur in the absence of infection. 



Two bacteria were isolated from the seeds, and one of 

 them, named Bacillus folticola, is supposed to be the 

 symbiont, though this is not certain. It may assimilate 

 very small quantities of free nitrogen. Plants free from 

 bacteria were obtained from seed sterilised by heat. They 

 show curious malformation — a tuberous swelling of the 

 buds, and a failure to develop leaves. As plants occasionally 

 found naturally with these characters are free from bacteria, 

 it is likely that the teratological forms are due to the absence 

 of the symbiont ; or rather that normal development can 

 take place only when the bacillus is present. Normal 

 plants do not grow well in soils free from nitrates. The 

 symbiosis thus seems to be necessary for normal develop- 

 ment, but the way in which the bacterium acts is quite 

 obscure. Its importance does not seem to lie in 

 supplementing the nitrogen ration of the plant. 



General Remarks. — Not many years ago the case of the 

 Leguminosae was regarded as almost unique, but we now 

 know that bacterial symbiosis with higher green plants, 

 though not common, is widespread. In all cases except 

 one, the advantage derived by the higher plant lies in 

 increased nitrogen supply. One cannot but believe that 

 many more cases will be discovered. The interesting leaf 

 symbiosis of Ardisia and Pavetta has further widened our 

 view by showing that symbiosis is not confined to the root 

 system. In general the symbiosis is not obligate ; the higher 

 plant gets on very well without the bacterium if it is suitably 

 nourished otherwise. Ardisia seems an exception to this 

 rule too. Ardisia and Pavetta present a further remarkable 

 development in that the symbiosis is congenital. We may 

 relate this with the fact that chance infection of the shoot 

 system in the air is, of course, much less likely to occur than 



