The Co-Efficient of Humidity, 135 
Complementary Society of Pteris and Scilla in Oak Woods, 
Yorks. 
Air-d 
ry Soil. 
Water. 
Water at 1 
5"C. Humus. 
Humus. 
A.—March, 1911. 
128. Pteris rhizomes, humous soil 
at 4-5 ins. 
76-8% 
19-7% 
3-9 
129. Scilla bulbs, stiff sand, at 
8-9 ins. ... 
37-0 
8-3 
4-47 
B.—May, 1911. 
141 i. Pteris, humus at 6-8 ins. ... 
176-6 
64 7 
2-7 
ii. Ditto, sandy soil at 8 ins. 
59-2 
20-8 
2-8 
142. Scilla, stiff sand at 10-12 ins. 
29-0 
71 
4-1 
C.—July-August, 1911. 
153 i. Pteris, shale at 3-5 ins. ... 
200 
10-2 
1-97 
ii. Ditto ditto at 5-6 ins. ... 
17-3 
9‘ 1 
1-90 
174. Scilla, stiff sand at 6-8 ins. 
11-0 
5-9 
1-87 
The problem is here decidedly complex. The bracken, whose 
rhizomes lie at the junction of two very distinct soils, has a vegetative 
period extending through the season covered by the analyses. The 
bluebells, whose bulbs are deeper seated have become dormant and 
independent of water-supply before the date of the third set. But 
the selected examples, few as they are, are sufficient to indicate and 
interpret the habitats. It is true that the seasonal drying is partially 
reflected in the water content. But the coefficient demonstrates it 
much more accurately, showing a drop by August to half its value 
in the Spring. Further it eliminates entirely the diversity dis¬ 
played in the layers occupied by the bracken (141 i. and ii.) The 
zone occupied by the bluebell bulbs has a high coefficient through¬ 
out the vegetative season, whereas the bracken has to endure at 
least a partial drought before the end of the season. 1 All this is in 
strict accordance with the known preferences of these plants, but 
the water-content does not suggest that the bluebell habitat is 
physiologically the wetter one and remains so as long as necessary. 
Another pair of comparable habitats, examined on the same 
day, will suffice to prove the utility of the coefficient of humidity in 
distinguishing the ground societies within a wood. 
1 With a more extended series of analyses the method developed in Section 
VII could be applied to these woodland soils and would involve some revision 
of the values of the coefficient. 
