The increasing availability of empirical adaptive landscapes in microbes pro- vides an opportunity, for the rst time, to study evolutionary theory on real- istic adaptive landscapes. We demonstrate this new approach by simulating adaptive evolution of a haploid, asexual population on the A. nigeri and the E. coli TEM empirical landscapes. Our results demonstrate some interesting evolutionary phenomena. First, adaptation starts with a big leap forward, but additional tness increases become smaller and further apart. Second, in the smooth E. coli TEM landscape, every genotype along the evolutionary trajec- tory reached high frequency before the next genotype emerged and increased in frequency. In contrast, in the rugged A. niger landscape, only four genotypes reach a high frequency, although adaptation proceeds through seven di erent genotypes. These di erences have profound e ects on our ability to predict the trajectory and pace of adaptation. We consider this new approach critical for our understanding of many biological phenomena, from emergence of drug resistance to formation of microbial communities.