Abstract Woody pioneers invade disturbed tropical forests but can be patchily distributed. We investigated causes of this patchiness in selectively logged forests near the Danum Valley Field Centre (DVFC) in Sabah, Malaysia. In a recently logged forest, we compared seed sources and seedling establishment between two logging systems (conventional practices [CL] and reduced impact logging [RIL]) and between two soil disturbance classes (topsoil or subsoil exposed). Tree seeds were less abundant than shrub and treelet seeds, and seed banks contributed many more viable seeds than did seed rain. Topsoil removal reduced seed and seedling counts and growth rates of planted seedlings. More pioneer seedlings established on RIL than on CL plots, but survival of planted seedlings was lower under RIL, perhaps due to denser canopy cover. Broadcasting seeds increased seedling recruitment for five animal-dispersed pioneer trees but not for two wind-dispersed trees. Longer term survival and growth of pioneer seedlings were compared between logging gaps (canopy but not soil disturbed) and skid tracks (canopy and soil disturbed) at a second site. Relative numbers in gaps vs. skid tracks varied by species, but across species, seedlings grew 40% more rapidly in gaps than in skid tracks. Survivorship differed among species but not between habitats. The three most common species experienced >95% mortality within 42 months. At a third site, we compared effects of mammalian herbivores on pioneer seedling performances in exclosures and open (control) plots centered at the edge of recently abandoned skid tracks. Fewer pioneer seedlings established in skid tracks than in less-disturbed edges, but abundances did not differ between exclosure and control plots. Over 18 months, pioneers survived better and grew on average 44% taller with mammalian herbivores excluded. Pioneer colonization can be seed limited, but high seedling mortality, exacerbated by herbivore damage, can limit recruitment even when seeds are available.