The host selection behaviour of the walnut twig beetle, Pityophthorus juglandis, was assessed by monitoring the landing rates of the beetles with sticky sheet traps on the host and nonhost hardwood branches. 2 Sticky sheet traps were deployed for 8 weeks from 6 June to 2 August. 2017 in the Putah Creek Riparian Reserve, Davis, CA. Branches from host northern California black walnut, Juglans hindsii, were paired with branches from six nonhost hardwood species. 3 The landing rate of P. juglandis (412 beetles trapped/8 weeks; 389 on host branches, 23 on nonhost branches) was significantly greater on the host branches for all nonhost hardwoods except Populus fremontii. Proportional comparisons of beetle presence also revealed a significant preference for the host branches compared with all but two nonhost species, Acer negundo and P. fremontii. 4 Capturing P. juglandis without the use of an aggregation pheromone was a rare event, underscoring the difficulty of studying the initial phases of host selection behaviour in bark beetles. Unbaited funnel traps adjacent to selected host trees in the experiment only captured five individuals over a 19-week period. None were captured in traps adjacent to nonhost trees. 5 This study provided evidence that P. juglandis discriminates between host and nonhost branches while in-flight. This directed flight behaviour is likely informed by the recognition of both host and nonhost volatile cues. 6 This study established an ecological context for the development of a semiochemical-based repellent system for protecting walnut trees from future attacks from this invasive bark beetle.