Criteria and methods for comparing the effectiveness of Rhizobium strains for pasture legumes under field conditions

Summary

Five Rhizobium strains were tested in the field with six Trifolium hosts. In order to examine criteria of evaluation of effectiveness under a range of sowing methods, all inoculant strains, plus controls, were sown with the six hosts in swards, with three of the hosts in rows and the other three as spaced plants. Swards and rows were assessed in terms of percentage plants nodulated, dry matter yield per unit area, visual rating of yield and mean dry weight of individually collected plants. Spaced plants were examined using the same criteria except that visual rating was replaced by area measurement. Harvested dry matter yield per unit area was accepted as the most reliable estimate for each of the sowing methods and served as a basis for comparison of other criteria.

Percentage nodulation was generally not related to dry matter production in rows or swards, partly because of cross-infection between plants. Although the relationship was not good with spaced plants it was concluded that the prevention of root contact provided the only opportunity for a reliable measurement of the proportion of plants nodulated.

Mean weights of individual plants removed from swards or rows did not provide a useful basis for strain comparison. The coefficients of variation of these plants commonly exceeded 100%. Although a calculated yield (mean plant weight × no. of plants) ranked strains similarly to the full harvest, the time and labour requirement was higher.

Visual rating of yield was most reliable with rows where the observer could successfully scan all treatments. It was less reliable with swards. Area measurement of spaced plants was not related to yield.

In spite of some anomalies between swards and rows, there was generally good agreement in ranking of strains in terms of harvested yield between all sowing methods. re]19730821