Root development is crucial for plant growth and therefore a key factor in plant perfor-mance and food production. Arabidopsis thaliana is the most commonly used system to study root system architecture (RSA). Growing plants on agar-based media has always been routine practice, but this approach poorly reflects the natural situation, which fact in recent years has led to a dramatic shift toward studying RSA in soil. Here, we directly compare RSA responses to agar-based medium (plates) and potting soil (rhizotrons) for a set of redundant loss-of-function plethora (plt) CRISPR mutants with variable degrees of secondary root defects. We demonstrate that plt3plt7 and plt3plt5plt7 plants, which produce only a handful of emerged secondary roots, can be distinguished from other genotypes based on both RSA shape and individual traits on plates and rhizotrons. However, in rhizotrons the secondary root density and the total contribution of the side root system to the RSA is increased in these two mutants, effectively rendering their phenotypes less distinct compared to WT. On the other hand, plt3, plt3plt5, and plt5plt7 mutants showed an opposite effect by having reduced secondary root density in rhizotrons. This leads us to believe that plate versus rhizotron responses are genotype dependent, and these differential responses were also observed in unrelated mutants short-root and scarecrow. Our study demonstrates that the type of growth system affects the RSA differently across genotypes, hence the optimal choice of growth conditions to analyze RSA phenotype is not predetermined.