Table 1 Univariate generalized linear mixed models fit by maximum likelihood using a negative binomial distribution to account for overdispersion of elemental counts.

Element	Variable	Estimate	SE	Z-value	P-value
Al	Status	−0.074	0.107	−0.69	0.491
Ba	Status	0.096	0.123	0.781	0.435
Br	Status	−0.149	0.149	−1	0.318
Ca	Status	−0.758	0.222	−3.42	<0.001
Cl	Status	−0.605	0.34	−1.778	0.0753
Cr	Status	−0.108	0.073	−1.49	0.137
Cu	Status	0.014	0.114	0.12	0.905
Fe	Status	−0.108	0.512	−0.211	0.833
Hf	Status	−1.407	1.843	−0.764	0.445
K	Status	−0.323	0.4	−0.807	0.419
Mg	Status	−0.766	0.279	−2.743	0.006
Mn	Status	−0.171	0.246	−0.696	0.486
Nd	Status	0.012	0.035	0.35	0.726
Ni	Status	−0.138	0.044	−3.11	0.002
Pb	Status	0.645	0.666	0.969	0.332
Pd	Status	0.29	0.362	0.801	0.423
S	Status	−0.32	0.14	−2.28	0.023
Si	Status	0.165	0.264	0.626	0.531
Sr	Status	−0.276	0.354	−0.778	0.437
Ti	Status	0.028	0.316	0.087	0.931
V	Status	−0.304	0.242	−1.257	0.209
Y	Status	−0.241	0.206	−1.172	0.241
Zn	Status	−1.225	0.27	−4.543	<0.001
Zr	Status	0.231	0.586	0.394	0.693

Each element was modelled separately to test for differences in wild samples from captive samples (Status), where samples were nested within each quill. Coefficient estimates and their standard error are presented along with Z and P values. Those tests significant at the 0.05 level are italicised.

Quick links

Search