Models of the Fisher‐Lande process (FLP) have been used successfully to explore many aspects of evolution by sexual selection. Despite this success, quantitative tests of these models using data from sexual radiations are rare. Consequently, we do not know whether realistic versions of the FLP can account for the extent and the rate of evolution of sexually selected traits. To answer this question, we generalize the basic FLP model of sexual coevolution and compare predictions of that basic model with patterns observed in an iconic sexual radiation, birds of paradise. Our model tracks the coevolution of male and female traits (two in each sex) while relaxing some restrictive assumptions. Using computer simulations, we evaluate the behavior of the model and confirm that it is an Ornstein‐Uhlenbeck (OU) process. We also assess the ability of the FLP to account for the quantitative aspects of ornament evolution in the genus Paradisaea using published measurements of display traits and a phylogeny of the genus. Finally, we use the program OUwie to compare model fits to generic OU and Brownian motion processes and to estimate FLP parameters. We show that to explain the sexual radiation of the genus Paradisaea one must either invoke extremely weak stabilizing selection on female mating preferences or allow the preference optimum to undergo Brownian motion at a modest rate.