PS 31-123 - Does soil pooling often alter the results of plant-microbe interaction experiments?

Background/Question/Methods

The study of plant-microbe interactions has become a foundational focus of plant ecology. A persistent issue in this area is whether soil pooling is an appropriate technique to quantify strength of plant-microbe interactions. Potential problems of sample pooling include altering the magnitude of plant-microbe interactions and an artificial reduction in the variance across replicates. To address this issue, we exposed nine herbaceous species to field-collected soil inocula that were either pooled or kept separate in a greenhouse study. The impacts of soil microbes were determined by comparison to autoclaved controls. We then compared the results of pooled and unpooled treatments to determine whether 1) the strength of the plant-microbe interactions changed, 2) the relative impacts of soil microbes were consistent in both methods, and 3) the variances across individual plants were altered.

Results/Conclusions

Sample pooling altered the strength of plant-microbe interactions in four of nine herbaceous species. Pooling reduced microbial inhibition in two species, reversed the direction from negative to positive in another, and generated a positive interaction in the fourth. The relative strengths of plant-microbe interactions were strongly altered, leading to no correlation between results from pooled and unpooled soils. However, the average magnitude and direction of plant-microbe interactions was almost identical between the two approaches (ln (response ratio) pooled: -0.183; unpooled: -0.182). Variances often differed between pooled and unpooled live soils. Opposite to the expected pattern, variances were typically higher in pooled soils. This effect nearly disappeared in autoclaved soils, suggesting soil microbes were the driving factor. Overall, our results suggest that, while unpredictable, sample pooling of inocula in experiments alters the strength, direction, and variance of impacts often enough to limit their utility in assessing the strength and direction of plant-microbe interactions.