4 On the Rain and Drainage - Waters at Rothamsted. 
the drainage-waters being contaminated bj the waters from 
other plots, and it was necessary, after baling the water out of 
each pit, to allow the pipe to run for some time before a trust- 
worthy sample of water could be collected. 
These arrangements are now much improved. The surface- 
drain running across the bottom of the field has been abolished ; 
the alternate furrows, forming the boundaries of each plot, have 
been prolonged on to the headland, and connected by a 4-inch 
pipe with the main drain, thus removing the surface-water 
between each plot without passing it over the drain-pipes 
of other plots. The main drain has also been considerably 
enlarged, and relaid 1 foot below its former level. In con- 
sequence of this improvement the outfall-pits now never fill 
with water, and no difficulty is experienced in collecting at 
all times the drainage-water flowing from the pipes. These 
improvements were completed in February 1879. 
It has been necessary to go into this detail because the value 
of the analyses of the drainage-waters entirely depends on these 
waters truly representing the drainage of particular plots of soil. 
It will be gathered from what has just been stated, that the 
earlier collections of waters were somewhat more liable to oc- 
casional contamination, both from surface-water, and from 
mixture with the waters of adjoining plots, than the collections 
made since February 1879. 
Since the opening of the drains, a careful record has been 
kept of the dates on which each drain ran, and of the size of 
the flow when observed. In the driest season experienced 
(October 1873 to September 1874) the largest number of run- 
nings recorded for any pipe was three ; while in the wettest 
season (1878-9), the largest number of runnings from one pipe 
v\ras forty-six. In Table XXXVI. is given the total number 
of daily runnings of each drain-pipe during fifteen years. 
In summer time, while the field is covered by the crop, 
drainage rarely takes place. A commencement of running is 
usually made in October, and the drainage reaches its greatest 
vigour in December and January. 
It will be noticed that the drains from the various plots do 
not run with equal frequency. Plots with a similar winter 
drainage show a dissimilar summer drainage if one bears a 
heavier crop ; the relative amount of drainage from two plots is 
even in some cases reversed at different seasons of the year, the 
plot having the greater winter drainage giving a less summer 
drainage by reason of its larger produce. Such facts will be seen 
by comparing the records of drainage for Plot 3&4 with those 
for Plot 7; those for Plot 5 with Plot G; and Plot 11 with 
Plot 13. The average crops grown on all these plots will be 
