Relation to Cultivation. 
407 
beds, each of 12 square metres area ; the one with 360 very 
vigorous plants, the other with 410 weakly plants, from seeds 
of the same parent. The weakly plants were small, and 
therefore required less space. When the plants shot up in 
the following year, the one bed gave 8 per cent., the other 
only 3 per cent, of twisted stems. 
C. Influence of the Soil. 
The richer the soil, and the more vigorous the plants, the 
better is the prospect for a high proportion of twisted stems. 
In the year 1891, in addition to the culture on good loose 
soil (see p. 401), I had a control-experiment on poor hard 
ground. The good soil was manured with 2 kilog. of 
guano and dried cow-dung per square metre ; the poor soil 
received only one-eighth of a kilog. of guano per square 
metre ; the treatment in other respects was identical. The 
result was — 
34% twisted stems 
I 4/o )> )> 
On the good manured soil 
On the poor soil 
In very poor soil the proportion of twisted stems may sink 
to nothing. I made an experiment of this kind in 1894 with 
a bed which consisted, to a depth of half a metre, of nothing 
but sand. For seed I used two samples, gathered in 1891 
and 1893, which gave, in control-experiments, 10 per cent, 
and 25-30 per cent, of individuals with spiral phyllotaxis ; 
half a bed (4 square metres) was sown with each of the 
samples. The seeds were covered with garden-soil in order 
to ensure germination. The two half-beds gave respectively 
94 and 124 plants, that is 24 and 31 per square metre, which, 
as they were for the most part small, only touched each other 
here and there. The plants were examined in May, 1895, 
and it was found that, without exception, the phyllotaxis was 
decussate. 
This is a convenient opportunity to direct attention to 
a circumstance which has hitherto been insufficiently recog- 
E e 2 
