How do political lotteries affect outcomes? We study the monthly lotteries used to exhaustively assign legislators to smaller preparatory committees in 19th century Belgium. We focus on the period of democratization following universal, male and plural suffrage, which had allowed a substantial number of new, third-party Socialists to enter. We ask whether random group composition, or more extensive exposure to certain types of members, affected voting over the universe of more than five-hundred roll-call votes between 1897 and 1902. We consider for example the effect of debating more Socialists, more politically experienced incumbents, or more representatives of majority Flemish- vs. French-speaking districts. We find small but significant exposure effects on rebelling against the party majority, in unexpected directions. We further illustrate the relevance of the political lottery for some measures which passed only after attracting cross-partisan support, including on the PR electoral system for national elections, a first in 1899.