492 
Journal of Agricultural Research 
Vol. XXIV, No. 6 
I and 2 on the map, i to the westward and 2 to the eastward of the main) 
have a few inches of marl under them below the peat. 
There is very heavy seepage from the sand hills down through the sand 
underlying the peat. Natural escape of the seepage water at the main 
channel of Coon Creek is blocked by the comparatively impermeable 
marl. The main outlet drain of the tract cuts across the marsh and sup¬ 
plies the necessary outlet. The laterals at the time they were laid were 
from 3X to 4 feet deep, and are now 3 to 3K feet below the surface of the 
peat. There is from i to feet of peat between the bottom of laterals 
^ I and 2 and the sand. The main outlet, however, since it was first dug 
has always been about a foot below the outlets of the laterals. There 
has been a tendency, therefore, in times of little flow for the seepage 
water to escape directly from the sand into the main outlet without pass¬ 
ing upward into the laterals. This condition would be accentuated both 
at the lower end of the laterals near the main and at the upper end of the 
laterals where their grade carries them above the almost flat ground 
water table. Lateral i carries heavy seepage from a bay indenting the 
high land, and the tendency for the ends of the lateral to get above the 
ground water level would be partially neutralized. In lateral 2, however, 
the tendency would be most marked at all low water stages. 
The tile system was laid out and installed in 1918 under the direction 
of Mr. H. B. Roe, drainage engineer of the experiment station. Alternate 
lines were of clay and concrete. When the concrete tile arrived they were 
freshly cured and did not meet the strength requirements. As they also 
appeared to be quite uniformly porous the entire lot was condemned by 
Mr. Roe. Another test of the tile was made 13 days after the first, which 
indicated a much greater strength. Under the circumstances it was 
decided by Doctor Alway that it would be better to use the tile than 
lose a season. The tile were therefore installed. 
TABre XVI .—Tests of Coon Creek tile Oct. 2, 1918, one week after delivery (5 and 6 inch 
tile tested) 
Test No. 
Wet 
weight. 
3-point 
breaking 
load. 
Calculated 
breaking 
load. 
I. 
Lbs. 
9 K 
10 
loX 
10 
ioj 4 
Lbs. 
375 
375 
413 
325 
425 
Lbs. 
k62 
' 563 
620 
488 
638 
2. 
A . 
c;. 
Average. 
10. 025 
382. 6 
574 
TABtK XVII .—Tests of Coon Creek tile Oct. 15, 1918, three weeks after delivery (s-inch 
tile tested) 
Test No. 
Dry 
weight. 
3-point 
breaking 
load. 
Calculated 
breaking 
load. 
1 
! 
Lbs. 
SH 
SH 
Lbs . 
538 
675 
Lbs . 
807 
2 .' 
Average.:.' 
8. 56 
606.5 
910 
