1. The work presented in this paper covers a period of 13 years and gives the results secured, by means of cylinder experiments, in a comparative test of nitrogen in nitrate of soda, stable manure, and leguminous green manure crops, on eight types of soil, seven of which are Coastal Plain soils.

2. The work was carried out at the Experiment Station in galvanised iron cylinders similar to those used in the availability work which was started in 1898. 320 such cylinders were used, making it possible to run a four-year rotation on the eight types of soil with five different treatments for each type.

3. The five treatments are as follows: Series I—no fertiliser. Series II—minerals only (phosphoric acid and potash). Series III—minerals and leguminous green manure crops to supply nitrogen for the four main crops. Series IV—minerals and stable manure.

All cylinders receive liberal applications of ground limestone every five years.

4. The rotation consists of rye, corn, potatoes and oats, all grown every year on the eight soils.

Crops are harvested at or near maturity, dry weights recorded, and samples prepared and nitrogen determinations made in duplicate.

5. Under all of the treatments the largest crops were obtained during the first two or three years of the period. From the point of maximum yield which in most cases occurred within the first three years, there has been a gradual decline in yields, both of total dry matter and total nitrogen.

This decline, however, is not without some exceptions.

Without exception the average yields for the 13 years are less than the average for the first seven years.

6. Of the eight types of soil used Quinton sandy loam has consistently given the largest crops and the largest nitrogen return, followed closely by Penn loam. Norfolk sand has likewise consistently given the lowest returns.

Windsor sand has shown a remarkable response to the green manure treatment.

7. Of the five different treatments the green manure series has given the highest average returns on all the soils, although in a few cases scattered through the 13 years (and for the majority of the soil types during the first two years) the nitrate of soda series gave the largest yield. In a very few cases the yield on the stable manure series has exceeded that of the green manure series.

The average yield of the nitrate series stands between the green manure and stable manure series.

8. It has thus been shown that for a period of 13 years nitrogen supplied by leguminous green manure crops, grown between the main crops of the rotation, has been more effective in crop production than 15 tons of manure every two years, while at the same time the nitrogen content of the soil of the green manure series was maintained, for a period of five years at least, on a level with that of the stable manure series.

9. It has also been shown that the green manure series yielded larger crops than the nitrate of soda series (160 lb. of nitrate per acre annually), while at the same time the nitrogen content of the soils of the former was maintained at a higher level than that of the latter.

10. It is pointed out that it may be possible to maintain crop yields at a rather high level, even when the total nitrogen content of the soil is not kept at quite so high a level as was found in the original soil. Under such conditions, however, a constant turnover of readily available nitrogen is necessary.