240 Correspondence. 
is obtained by the identical arithmetical process that yields 96.96 parts of 
Organic and Volatile matter per cent. of moisture-free peat and all the 
other percentage values in the second and fourth columns of his table 
of the analysis of peat. As they have all been deduced by the same process 
from the amounts in terms of the wet peat, their relative values remain 
unchanged. But there is a great gain in adopting the dry-peat values. 
‘To correct any false impressions about the composition of peat,’ 
Mr. Johnson states :—‘ Peat usually contains 80-90 per cent. of moisture 
and when air-dried, 15-25 per cent.’ Whilst literally true the statement 
is misleading, for the standard of comparison is changed midway and the 
numbers are not comparable. Who will realise that the second quality 
(Say 15 per cent.) is little more than a thirtieth of the first (80 per cent.) ? 
But so itis. If a wet peat loses 80 per cent. of water on air-drying, there 
will be a further loss of 3 per cent. when it is oven-dried; and these 
numbers become 400 per cent. and 15 per cent in terms of the air-dry 
peat. Either one scale or the other must be adhered to throughout. I 
will try to illustrate briefly the principle underlying ecological methods 
of soil analysis. 
Of course, there are other factors to be considered, but what I am 
concerned with is to find an accurate quantitative expression for the 
relative wetness or humidity of a plant habitat. To take a specific case. 
Heather growing on the moors usually has its roots partly ina layer of 
moderately pure peat and partly in a nearly pure coarse sand, and these 
layers differ enormously in their water-holding capacity, and therefore 
in their water-content, at all times of the year. I had on that account to 
reject the water content as an unsuitable index of the soil humidity and to 
look for a more reliable expression that would be generally applicable to 
natural soils. Numerous analyses have shown that, excluding clay soils, 
the water retained in the soil is held almost entirely by the humus; so 
much so that as soon as the humus in any soil is reduced to unity the corres- 
ponding value of the water becomes a true measure of index of the wetness 
of the soil. This ratio I have called the co-efficient of soil humidity 
and subject to a correction in covtalt cases the co-efficient of soil humidity 
water-cont : 
vater-content here the water-content is the water lost when the 
humus-content 
soil is allowed to dry in an ordinary room at about 15°C. 
In the abstract published in The Naturalist in 1911 this was called the 
water co-efficient and its values were shown to be distinctive of various 
types of moorland. <A couple of analyses will show how it is of service 
in another field. Both samples were taken in the same wood on the 
same day for the purpose of distinguishing the habitats of (1) hair-grass 
(Deschampsia flexuosa) growing over grit boulders and (2) the great 
brome or fescue (Bromus giganteus) typical of much damper places and 
actually growing on a little alluvial flat by a stream side. In the latter 
case the sample was in two portions, one from the ball of roots, the other 
just below. To remove difficulties that have arisen I will set out the an- 
alysis in full. 
equals 
(A) In terms of Wet Soil. 
Soil No. 172 1731 173 
Desc. flexuosa. Bromus giganteus. 
W ATER— per cent. per cent. per cent. 
(a) Loss at 15° 19.91 55-55 22.55 
(6) Further loss 
at 100°C. 4.74 2.82 1.63 
Humus=Loss on 
combustion 20.25 11.28 4.13 
Mineral Residue 55-10 30.35 71.68 
100.00 100.00 99.99 
Naturalist, 
