Mammalian browsers can cause ecologically and economically significant damage to important tree species, particularly at the vulnerable seedling stage. Some tree species or particular parts of trees are less preferred as a result of chemical defences - harnessing such defences could reduce the appeal of more preferred trees and thus their loss to browsing. We tested the potential of birch bark extract, a forestry by-product, to protect seedlings of Scots pine Pinus sylvestris from two globally significant browsers: moose Alces alces and red deer Cervus elaphus. We quantified the responses by captive animals to extract application as a function of both the distance between seedlings and whether the extract was applied to all or alternate seedlings. Both moose and red deer consumed pine seedlings less frequently when seedlings were treated with birch bark extract, but their responses differed when only alternate seedlings were treated. At inter-seedling distances equivalent to those used in forestry, red deer browsed untreated seedlings more frequently when alternate seedlings were treated than when none were treated (associational susceptibility), whereas moose browsed untreated seedlings less frequently with treated than untreated neighbours (associational refuge). These neighbourhood effects were not evident at inter-seedling distances three times larger or smaller than those used in forestry. There was also no significant difference in the frequency of browsing on treated seedlings in alternate compared to completely treated arrangements at any inter-seedling distance. Red deer removed significantly less foliage from treated than untreated seedlings once browsed. Browsing by moose followed a similar but non-significant trend. The relative number of untreated to treated seedlings browsed was 2:1 for red deer and 4:1 for moose. Red deer used a finer scale of selection than moose for choosing between seedlings, consistent with their smaller body mass. However, the coarser-scale of selection used by moose resulted in consumption of more untreated relative to treated seedlings. Our study highlights the potential of existing plant defensive chemistry, harvested from forestry by-products, to reduce herbivore damage to tree seedlings. We also demonstrate the need to consider both plant associational effects and the foraging strategies of the dominant herbivores in applying repellents.