Classic biological control can be a powerful option for those tasked with managing biological invasions; however, some biocontrol releases lead to non-target attack – feeding, damage or development on species other than the target species. The cinnabar moth, Tyria jacobaeae L. (Lepidoptera: Erebidae) was introduced to Western Oregon as a biocontrol agent, and established on the native herb Senecio triangularis (commonly: arrowleaf groundsel) in a case of sustained non-target attack. In this thesis, I examined whether seed loss caused by cinnabar larvae feeding on flowers and seeds has the potential to impact S. triangularis reproductive rates. I evaluated this potential impact as ecological risk posed to S. triangularis, using the framework risk = sensitivity x exposure. Varying seed rain density and vegetation removal treatments in a field experiment showed that seedling recruitment rates were seed-limited (sensitive to seed availability): plots with increased seed rain had increased seedling recruitment for two seed-addition levels compared to control plots. This finding was robust across plot conditions, though plots with clipped vegetation had reduced seedling recruitment response. Seedling recruitment was particularly sensitive to seed quantity at low seed rain densities; increasing seed rain from none added to approx. 360 seeds per 0.25 m2 increased probability of a plot recruiting from 10% to 85%, and increased mean seedling counts from 0.15 to 3.01 seedlings. Observational data from five sites in the Oregon Cascades and Coast Range showed that phenological matching was low across sites in two regions, quantified as a pairwise overlap index for virulent larval stages and vulnerable flowering stages. Despite low percent overlap, plants experienced a wide range of floral damage by two metrics: proportion of capitula damaged, and proportion of capitula lost (fallen from pedicel due to larval feeding, fate unknown). Mean proportion of capitula damaged was around 40% for the two sites with highest percent overlap (approx. 3-3.5%); while mean proportion of capitula lost per site was as high as 60%, but had no apparent relationship to percent overlap. Combining this experimental and observational evidence, I find that seedling recruitment for S. triangularis is sensitive to seed loss, particularly at low levels of natural seed rain. Exposure to seed loss is relatively low due to a phenological mismatch between the peak feeding stages of cinnabar larvae and vulnerable flowering stages of S. triangularis, with many capitula escaping herbivory. Together, this suggests low to moderate risk to S. triangularis from floral and seed damage by cinnabar larvae. Further investigation could determine natural seed rain densities for known populations, and investigate abiotic drivers of phenological mismatch for these two species in order inform how risk to S. triangularis is expected to change given climate trajectories for the region.