Provisioning versus Nonprovisioning Conditions
Overall, fGCM levels were lower in association with provisioning experiments (mean ± SD of untransformed data: 627 ± 1101 ng/g) than during periods without provisioning (1173 ± 1478 ng/g). Patterns were similar for frequent callers (provisioning: 753 ± 1409 ng/g; no provisioning: 1408 ± 1792 ng/g), infrequent callers (provisioning: 508 ± 561 ng/g; no provisioning: 829 ± 1067 ng/g) and noncallers (provisioning: 551 ± 909 ng/g; no provisioning: 1187 ± 1385 ng/g). The interaction between caller type and presence/absence of provisioning thus showed no significant association with fGCM levels in the simplified model (mixed-effects linear regression: z = 0.89, N = 29 individuals, P = 0.374), whereas the effect of presence or absence of provisioning was highly significant (z = −9.67, N = 29 individuals, P < 0.001) but in the opposite direction than predicted ( Table 1, Fig. 1; see Table A1 in the Appendix for the results of the full model).
Provisioning versus Nonprovisioning Conditions
Overall, fGCM levels were lower in association with provisioning experiments (mean ± SD of untransformed data: 627 ± 1101 ng/g) than during periods without provisioning (1173 ± 1478 ng/g). Patterns were similar for frequent callers (provisioning: 753 ± 1409 ng/g; no provisioning: 1408 ± 1792 ng/g), infrequent callers (provisioning: 508 ± 561 ng/g; no provisioning: 829 ± 1067 ng/g) and noncallers (provisioning: 551 ± 909 ng/g; no provisioning: 1187 ± 1385 ng/g). The interaction between caller type and presence/absence of provisioning thus showed no significant association with fGCM levels in the simplified model (mixed-effects linear regression: z = 0.89, N = 29 individuals, P = 0.374), whereas the effect of presence or absence of provisioning was highly significant (z = −9.67, N = 29 individuals, P < 0.001) but in the opposite direction than predicted ( Table 1, Fig. 1; see Table A1 in the Appendix for the results of the full model).
การแปล กรุณารอสักครู่..
