Figure 2.

Relationships between the timing of spring temperatures and timing of breeding in wheatears. (a) Temporal trends (years 1993–2012) for thermal progression of spring (open symbols show date TS200b3, that is, the date when thermal sum based on T _base = 3°C reached 200, dashed line shows linear regression: date TS200b3 ~ year) and median lay date (filled symbols, solid line shows weighted linear regression: median lay date ~ year, weight = 1/SE _{lay date}). The difference between median lay date and thermal progression of spring describes the degree of phenological matching. (b) Annual estimates of the degree of matching between wheatear timing of breeding and thermal progression of spring resulting from the difference in point estimates for data shown in a. Absolute larger values correspond to a greater difference in timing, where positive values correspond to a relative later timing and negative values to a relative earlier timing of breeding with respect to the estimate for thermal progression of spring. All dates are dates since May 1. (c) There was no direct link between wheatear demographic rates and the annual degree of matching as exemplified for number of local recruits. Analysis details for this and all other demographic rates are shown in Table 4. Median TS, the yearly median thermal sums at hatch date, was estimated from individual thermal sums at hatch date and represents the annual degree in population‐level phenological matching. Dots show mean values, and error bars show SE of the raw data. Lines show the predicted relationship generated using bootstrapping implemented in the R package “ez” (see Supporting Information; solid: median, dashed: 95% CI)