804 REPORT—1904. 
on the intensity of the struggle, while if the struggle be very intense those 
biaiomorphoses which were from the beginning useful, although but accidentally 
so, will survive. These are the adaptations proper. 
SuB-SECTION—AGRICULTURE. 
The following Papers were read :— 
1. Analysis of the Soil by means of the Plant. 
By A. D. Hatt, ILA. 
In view of the many difficulties attaching to the interpretation of soil 
analyses as a guide to the manurial requirements of the soil, attempts have 
been made from time to time to use the living plant as an analytical agent. 
It is well known that while the ash of a given plant possesses a characteristic 
composition, variations in the proportion of constituents like the potash or 
phosphoric acid will take place to a certain extent in response to the manuring. 
Investigations on the utility of the analysis of plant ashes to ascertain the needs 
of the soil for specific mineral manures have been undertaken by Heinrich, 
Helmkampf, and others, and particularly by Atterberg, who used oats as his test 
lant. 
To try the agreement between this method and analysis of the soil, further 
experiments were begun in 1902 with oats grown in pots containing six soils of 
very different types. Although in certain striking cases both methods agreed in 
their results, there was no strict measure of consistency between the two sets of 
figures, while the variations between the material grown in duplicate pots of the 
same soil were often greater than that between different soils. 
For further information the data accumulated in the Rothamsted experiments 
were consulted, and analyses of wheat, barley, mangels, and potatoes from certain 
of the plots were compared with the analyses of the soil of the same plots. In 
dealing with cereals it is necessary to examine the whole plant, the composition 
of the grain fluctuates but little with the manuring ; any deficiency of a particular 
constituent will result in less grain being formed, while any excess will be left 
behind in the straw. From these Rothamsted results it seemed that though the 
composition of the plant did reflect that of the soil, yet the range of variation 
shown by the plant was less than that indicated by soil analysis. The ash of 
the root crops showed, however, a wider range of variation, and, in view of the 
greater sensitiveness of root crops to the lack of mineral plant foods as compared 
with the comparative indifference of the cereals, they seemed likely to prove better 
test plants to indicate the need or otherwise of specific mineral manures. 
Samples of potatoes, mangels, and of swedes were obtained during 1903 from 
experimental plots in various parts of the country, where the field trials indicated 
a reaction to phosphoric acid or potash manuring; analyses of the ash were made 
and compared with the analyses of the soil. The results indicate that the 
analysis of the ash of the swede plant would often provide a better indication of 
the phosphoric acid requirements of the soil than does the analysis of the soil 
itself, and that similarly the mangel plant will serve to test the state of the soil 
as to potash. A greater number of data as to the limits of normal variation in 
the composition of the ash are, however, wanted before the method can be 
employed for practically testing the soil. 
2. The Probable Error of Agricultural Field Experiments. 
By A. D. Haun, M.A. 
Field experiments upon crops are subject to various sources of error: (1) errors 
incident to the operations of measuring the plot, applying the manure, &c., 
gathering and weighing the produce ; (2) errors due to permanent inequalities of 
the soil brought about by differences in texture and composition, variations in the 
subsoil, drainage, &c., which errors again may vary with the season; (3) errors 
