Despite our understanding of chemical defenses and their consequences for plant performance and herbivores, we know little about whether defensive chemicals in plant tissues, such as alkaloids, and their spatial variation within a population play unappreciated and critical roles in plant‐herbivore interactions. Neighboring plants can decrease or increase attractiveness of a plant to herbivores, an example of a neighborhood effect. Chemical defensive traits may contribute to neighborhood effects in plant‐herbivore interactions. We examined the effects of nicotine in leaves (a non‐emitted defense chemical) on plant‐herbivore interactions in a spatial context, using two varieties of Nicotiana tabacum with different nicotine levels. A common garden experiment demonstrated that visits by grasshoppers decreased with increasing density of neighboring plants with a greater nicotine level. In contrast, visits of leaf caterpillars were not affected by neighbors, irrespective of nicotine levels. Thus, our results clearly highlighted that the neighborhood effect caused by the nicotine in leaves depended on the insect identity, and it was mediated by plant‐herbivore interactions, rather than plant‐plant interactions. This study demonstrates that understanding of effects of plant defensive traits on plant‐herbivore interactions requires careful consideration of the spatial distribution of plant defenses, and provides support for the importance of spatial context to accurately capture the ecological and evolutionary consequences of plant‐herbivore interactions.

There is increasing evidence that herbivore–plant interactions on a focal plant species are influenced by interspecific neighbourhood effects via neighbouring plants (i.e. an associational effect). However, intraspecific neighborhood effects imposed by plant traits have been less appreciated. Specifically, the significance of intraspecific neighbourhood effects in population‐level consequences of plants has been totally overlooked. Using two varieties of Nicotiana tabacum (high‐ and low‐nicotine), we evaluated the neighbourhood effects based on patch‐level interactions in a split‐plot 3 × 3 factorial experiment that manipulated number of plants (4, 9 and 16 plants) and culture type (monoculture plots with high‐ and low‐nicotine plants, and polyculture plot) in an experimental garden. We found that herbivore visits on plants varied depending on the number of plants per patch and culture type. Presence of more high‐nicotine plants decreased herbivore visits in the four plant plots, and presence of high‐nicotine plants in the nine plant plots decreased herbivore visits on both high‐ and low‐nicotine plants. In contrast, in the 16 plant plots, herbivore visits on high‐nicotine plants in polyculture plots were lower than others, including those on high‐nicotine plants in monoculture plots. Our findings clearly demonstrated that the intraspecific neighbourhood effect could occur depending on the aggregation of highly defended plants (i.e. high density and/or plant‐spacing). This study suggests that multiple mechanisms for the neighbourhood effect simultaneously worked, depending on the patch size and composition of defensive traits of individual plants, and that intraspecific neighbourhood effects may influence population‐level consequences for plant–herbivore interaction. A free Plain Language Summary can be found within the Supporting Information of this article.