LETTER Differential resistance to freezing and spatial distribution in a chemically polymorphic plant Thymus vulgaris Justin Amiot, Yann Salmon, Christian Collin and John D. Thompson* UMR 5175 Centre d’Ecologie Fonctionnelle et Evolutive, CNRS, 1919 Route de Mende, 34293 Montpellier cedex 5, France *Correspondence: E-mail: john.thompson@cefe.cnrs.fr Abstract Secondary compounds play multiple ecological roles. In this study, we present novel experimental evidence of differential tolerance to freezing temperatures among chemotypes of a chemically polymorphic plant, Thymus vulgaris. Non-phenolic chemotypes showed a significantly better survival and re-growth after early-winter freezing ()10° in early December) than phenolic chemotypes. Comparison of temperature data (1971–2002) at a phenolic and non-phenolic site showed that whereas early-winter freezing occurred in 6 years in the non-phenolic site they never occurred at the phenolic site. Observations of trichome morphology (where the essential oil is stocked) with and without intense freezing indicate that non-phenolic chemotypes may escape any negative effects of freezing by releasing their essential oil into the atmosphere during severe freezing. The correlation between tolerance of freezing and local temperature regimes strongly suggests that differential freezing resistance is a key ingredient of the distribution of thyme chemotypes in space. Keywords Chemical polymorphism, freezing resistance, Lamiaceae, secondary compounds. Ecology Letters (2005) 8: 370–377 INTRODUCTION Plant secondary compounds have multiple biotic and abiotic functions. Well known to act as a chemical defence against herbivores and pathogens (Levin 1976; Bryant et al. 1991), secondary compounds may modulate interactions with competing plants (Rice 1979; Ehlers & Thompson 2004) and pollinators (Beker et al. 1989; Ayasse et al. 2000) and provide UV protection (Close & McArthur 2002) or carbon/nutrient balance adjustment in response to sunlight (Shure & Wilson 1993). The existence of genetically based polymorphisms for secondary compound production has been reported in various species, several of which show spatial variation in the occurrence of different chemical forms, e.g. Pinus halepensis (Schiller & Grunwald 1987), Trifolium repens (Daday 1954a,b), Mentha citrata (Murray & Lincoln 1970), Origanum vulgare (Vokou et al. 1993), and a range of Thymus spp. (Stahl-Biskup 2002; Thompson 2002). Such spatial variation provides an appropriate situation in which to evaluate the different biotic and abiotic functions of secondary compound production. However, untangling the ecological role of different factors has proven complex. In several species which show spatial variation in relation to temperature and other abiotic environmental features, chemical forms also show differential resistance to herbivory, e.g. Trifolium repens (Jones 1973; Dirzo & Harper 1982a,b; Hughes 1991), Lotus corniculatus (Ellis et al. 1977; Compton et al. 1983), and Thymus vulgaris (Linhart & Thompson 1995, 1999). It thus remains unknown whether abiotic factors directly cause spatial segregation of chemical morphs or whether their impact is via an effect on herbivore abundance. Thymus vulgaris L. (Lamiaceae), a small aromatic woody shrub in the western Mediterranean, has natural popula- tions which contain one or more of six different chemical morphs, or chemotypes (Granger & Passet 1973). Each chemotype is named after the main component of its essential oil, which is either a phenolic monoterpene, thymol (T) or carvacrol (C), or a non-phenolic mono- terpene, linalool (L), thuyanol (U), a-terpineol (A) or geraniol (G). All six are genetically controlled by an epistatic series of five loci and are part of the same biosynthetic pathway (Passet 1971; Vernet et al. 1986). In Ecology Letters, (2005) 8: 370–377 doi: 10.1111/j.1461-0248.2005.00728.x Ó2005 Blackwell Publishing Ltd/CNRS