Arguably, research in speciation has focused more on how the process begins, that is, how the first barriers evolve, than on the later stages of the speciation process (i.e., the evolution of [nearly] complete reproductive isolation). A recent special issue in Philosophical Transactions of the Royal Society B seeks to partially rectify this issue. We analyze this issue in the context of host shifts in Timema stick insects. In particular, in an article in this special issue we show that host shifts to closely related plant species are common, but cause only weak reproductive isolation. In contrast, much rarer host shifts from flowering plants to gymnosperms generate strong reproductive isolation. Thus, such rare and difficult events might be important for generating fully isolated species.