How does district magnitude affect electoral outcomes? This article addresses this question by exploiting a combination of two natural experiments in Argentina between 1985 and 2015. Argentine provinces elect half of their congressional delegation every two years, and thus districts with an odd number of representatives have varying magnitudes in different election years. Furthermore, whether a province elects more representatives in midterm or concurrent years was decided by lottery in 1983. I find that district magnitude (a) increases electoral support for small parties, (b) increases the (effective) number of parties that gain seats and (c) reduces electoral disproportionality. The last two results are driven by the mechanical rather than the psychological effect of electoral rules.