(A) Experimental setup: a single leaf was enclosed in a plastic clipcage with (clipcage+ T. notatus; solid line) or without 20 T. notatus (clipcage, dashed line). Additionally, unenclosed control leaves (control, dotted line) were also collected. (B–H) CK values in leaves at different time-points after the start of herbivore exposure: (B) sum of cis-zeatin (cZ), trans-zeatin (tZ), N6-isopentenyladenine (IP) and their ribosides (cZR, tZR, IPR). (C) cZ, (D) tZ, (E) IP, (F) cZR, (G) tZR and H) IPR. Statistically significant differences were identified by two-way ANOVAs (C, E) or by ANCOVA with mirid as factor and time as continuous explanatory variable (B, D, F, G, H). (B) sum of CKs: time F1,51 = 0.041, p=0.841; mirid F1,51 = 1.934, p=0.170; time*mirid F1,51 = 5.270, p=0.026. (C) cZ: time F7,39 = 5.855, p<0.001; mirid F1,39 = 31.43, p<0.001; time*mirid F7,39 = 1.596, p=0.166. (D) tZ: log transformed time F1,51 = 3.024, p=0.088; mirid F1,51 = 0.986, p=0.325; time*mirid F1,51 = 2.572, p=0.115. (E) IP: time F7,39 = 11.538, p<0.001; mirid F1,39 = 7.134, p=0.011; time*mirid F7,39 = 2.279, p=0.048. (F) cZR: log transformed time F1,51 = 1.083, p=0.303; mirid F1,51 = 5.066, p=0.03; time*mirid F1,51 = 0.256, p=0.615. (G) square root transformed tZR: log transformed time F1,51 = 9.910, p=0.003; mirid F1,51 = 1.164, p=0.286; time*mirid F1,51 = 0.688, p=0.411 hr IPR: log transformed time F1,51 = 11.909, p=0.001; mirid F1,51 = 1.484, p=0.229; time*mirid F1,51 = 1.813, p=0.184. Error bars depict standard error (N ≥ 3). FM: fresh mass. For raw data see Raw_data_FIGURE_3_S3 (Dryad: Brütting et al., 2018).