Antibiotic resistance genes are frequently encoded and disseminated by mobile genetic elements such as conjugative plasmids. Despite their clinical importance, the evolutionary forces that maintain these plasmids - and how conjugation influences their success - remain poorly understood. I used experimental evolution to test when conjugation benefits the stability of antibiotic resistance plasmids. By varying the availability of potential recipient cells and the strength of positive selection for plasmid-encoded traits, I tested 12 combinations of experimental conditions. The stability of a conjugative plasmid and a non-conjugative variant in E. coli was compared during an 18-day serial passage experiment. As evolutionary theory would predict, conjugation’s evolutionary benefits were most apparent when recipient availability was high and positive selection for plasmid carriage was weak. But interestingly, the effect of conjugation was time-dependent: conjugation was initially detrimental to plasmid fitness, yet became associated with increased plasmid prevalence in the long term. I will discuss competing evolutionary mechanisms that could explain this result, and present further evolution experiments and whole-genome sequencing data to distinguish between them. This work contributes to efforts to better understand the role of plasmids in microbial genomes, and the importance of horizontal gene transfer in the maintenance of antibiotic resistance.