Fig. 5

Effects of analysis method on reproducibility of w m (α,β,t) projections (raw data). Continuous lines display w m (t) averages of leaf dry weight in shoots. Dashed lines in panel a show w m (α,β,t) projections produced by log-linear transformation. Dashed lines in panel b display those projected via nonlinear regression. Nonlinear regression estimates support greater reproducibility of observed w m (t) values through w m (α,β,t) proxies (Table 2)