Fig. 3

Results of analysis of thermal response patterns in the two species. The color of each cell (blue = excess, orange = deficit) represents the deviation of the observed from the expected number of transcripts based on hypothetical equivalence of the marginal frequencies. The units are number of transcripts. For each temperature extreme, the species expected to be less well adapated to that extreme is displayed on the x-axis for the two response categories corresponding to upregulation (Bimodal and Low for the low temperatures, or Bimodal and High for high temperatures). The distribution of response categories for those transcripts in the better-adapted species is arrayed along the y-axis. The dashed boxes indicate the matched responses (e.g. High - High). a Low temperature extreme: there is an excess of shared Low and Bimodal expression types and a bias away from all other categories in A. picea, consistentwith the enhanced response hypothesis (Fig. 1). b High temperature extreme: in addition to an excess of matched categories, there is an excess of High and Bimodal transcripts in A. picea that are not upregulated in A. carolinensis (Intermediate and Not Responsive), partially consistent with the tolerance hypothesis. The complete set of matched observations is shown in Additional file 1: Figure S1. Expression types are defined in Table 1