488 
Journal of Agricultural Research 
Vol. XXIV, No. 6 
(8) The solubility of this compound in carbonic acid. 
(9) The action of the organic compounds on neat cement. 
(10) The increase in quantity of acids as decomposition proceeds. 
PART IV. FIELD WORK 
Before drawing a definite conclusion from the results of these labora¬ 
tory investigations it was decided to make most careful field observa¬ 
tions to ascertain if the laboratory findings were borne out by actual 
conditions. These observations have covered two years and during that 
time the subject has attracted considerable public attention. Reports 
have come in, notably from the muck lands of the South, which con¬ 
siderably broaden the scope of the investigations. 
representative locauties 
The field work was planned so that it should cover the greatest pos¬ 
sible range of soil conditions. No attempt was made to get mere num¬ 
bers of observations. On the other hand, an attempt was made to 
investigate the peats overlying dissimilar geological formations, first in 
Wisconsin and later in Minnesota. 
EIMESTONE SOIE 
In Wisconsin, eastern Dane County was taken as typical of the glaciated 
limestone country. This includes University Marsh, where the peat was 
underlain by marl. This has been described earlier in this paper. Dur¬ 
ing 1920 a considerable amount of the tile in the eastern part of the 
marsh was relaid at a greater depth. Among these were alternate 
lines of Group VII. They included the lines opened up in 1919 from 
which the samples were taken. It was found that about half the tile 
were either collapsed or were not fit to be put back in the ground. This 
is well illustrated in Plate 3, A. This rapid final collapse, after disin¬ 
tegration was once well under way, later proved to be characteristic. 
A large marsh of the alkaline type south of Madison, Wis., was also 
investigated. It lay in the bottom of a stream valley and in its natural 
condition would be subject to more or less periodic overflow and seepage 
and wash would be heavy. Water would either be standing or running 
in the tile the greater part of the time. A portion of this marsh was 
drained in 1914. The tile were examined in the spring of 1920. A 
pile of unused tile showed that the quality was not of 5 ie highest, as 
judged by present standards, but was a very good average at the time 
it was made. On opening up the tile lines an interesting condition was 
disclosed. At the outlet of the lines fresh drainage water did not enter 
the tiles in large quantities but found its outlet into the main ditch. 
The inside of the tile was covered with sediment and the concrete thereby 
protected. Under these conditions there was little disintegration. 
Farther up the lines, where water either stood or ran a larger part of the 
season, the cement of the tile had almost completely disappeared. On 
account of their fragility, there was considerable diflSculty in taking 
out samples. The tile apparently retained their perfect shape and were 
functioning quite properly as long as they were not disturbed. The 
act of digging, however, caused their complete collapse. At the upper 
