THe Carson DIoxIDE OF THE Sort AIR 345 
crops of oats, had a considerably higher content of carbon dioxide, even 
after the removal of the crop and especially for about two months after 
harvest, than did cans 5, 6, 7, and 8, which remained uncropped for the 
two seasons. 
After the oat crop from cans 1, 2, 3, and 4 was harvested, on fey 1, 
1918, cans 2 and 3, and the ancropped cans 6 and 8, were iced to tnillot. 
Cans 1, 2, 3, and 4 are here designated as the high-carbon-dioxide series, 
while cans 5, 6, 7, and 8 are called the low-carbon-dioxide series. Thus, 
in the high-carbon-dioxide series, cans 1 and 4 were bare and cans 2 and 3 
were cropped to millet; in the low-carben-dioxide series, cans 5 and 7 
were bare and cans 6 and 8 were cropped. All these cans were sampled 
in the usual way for carbon dioxide, and the data obtained are given 
in table 7 (appendix, page 360). The samples were taken twice a week 
at first, and later they were taken daily. The moisture in the soil was 
maintained at or near 30 per cent (oven-dry basis). 
Within a month of seeding, the crop was thinned to thirty plants to 
a can; so that at the time when the effect of the plants on the carbon 
Bioxide became noticeable (a month after seeding), the number of peas 
was the same > for all cans. 
Resulis 
Tn table 7 it is shown that the differences between the percentages of car- 
bon dioxide in the cropped soil and those in the uncropped soil in the high- 
carbon-dioxide series, were approximately the same as the corresponding 
differences in the low-carbon-dioxide series. .In table 8 (appendix, page 361) 
it is seen that the majority of the differences in carbon-dioxide production 
by the crop in the two series (as determined by the difference between the 
amount of carbon dioxide produced by the cropped soil and that produced 
by the uncropped soil) was well within the limits of the experimental error. 
It seems, therefore, that the crops produced carbon dioxide quite independ- 
ently, and that this production was not affected by the amount of carbon 
dioxide in the soil, at least not within the limits set by this experiment. 
How closely the difference between the curves for the cropped soils cor- 
‘responded with those for the bare soils is shown in figure 58 (page 343). 
The relationship between the temperature of the soil at the time of 
sampling, and the carbon dioxide in the bare soil and also that due to the 
crop on the low-carbon-dioxide series (determined by the difference as 
