For the main experiment, Simms concentrated on the yellow bush lupine (Lupinus arboreus), and chose to inoculate each plant with two of three rhizobial strains selected from the 10 strains that Povich tested. The three were chosen to include a poor nitrogen sharer, a mediocre sharer and good sharer, so as to give each plant a choice between two distinctly different strains.
Povich, with assistance from UC Berkeley undergraduates Yla Tausczik and Mona Urbina, then planted lupine seeds collected from six different stands of yellow lupine on the reserve site. As the seeds sprouted, they inoculated the potting soil with various pairs of the three strains.
Using a technique called the polymerase chain reaction to identify the bacteria in individual nodules, the researchers found that many plants nodulated both strains that were offered. On average, however, within individual plants, the poorly performing bacteria inhabited small nodules, while the more effective inhabited larger nodules.
"There's not a big difference between strains," Simms said, "yet the plant can distinguish them."
In a subsequent experiment, they confirmed that the larger the root nodule, the more nitrogen-fixing bacteria it contained. Only rarely did they find more than one strain of microbe in a nodule, though a single plant might nodulate more than one strain in separate nodules.
"These results support the hypothesis that legumes can favor more cooperative rhizobia by manipulating bacterial fitness in the nodule," the authors wrote.
Source:University of California - Berkeley