Table 1.
Precision and accuracy of enzyme-linked immunosorbent assay (ELISA) and radioimmunoassay (RIA) measurements of condor plasma, urate extract, and feather extract.
All samples were run in duplicate and the number of samples analyzed in is in parentheses.
Fig 1.
Plasma corticosterone (CORT) (A and D), urate glucocorticoid metabolites (GCM) (B and E), and feather CORT (C and F) levels measured by radioimmunoassay (RIA) are more reproducible than enzyme-linked immunosorbent assay (ELISA) across a range of sample dilutions. Symbols (open and filled squares, circles, triangles, and diamonds) represent a sample from an individual condor that was analyzed over a range of sample dilutions. CORT and GCM levels in diluted samples are expressed as a percent difference from expected values (horizontal dashed line at 0% diff.), based on levels measured in the most dilute sample (i.e., the lowest amount of sample in milligrams per 100 μL assay solution), and assumes sample matrix interferences are minimized in this most dilute sample. The vertical hash-marked region in each panel reflects the range of sample dilutions (x-axis) used for all samples in this study. The horizontal grey-shaded region reflects the CORT or GCM measurement uncertainty (± 2 RSD, based on intra-assay precision) for each assay and sample type; symbols within this region do not measurably differ from expected GC levels.
Fig 2.
Measurements by radioimmunoassay (RIA) vs. enzyme-linked immunosorbent assay (ELISA) are different but significantly correlated for plasma corticosterone (CORT) and urate glucocorticoid metabolites (GCM).
Each data point represents a condor sample measured by both RIA and ELISA. The dashed line indicates idealized agreement (y = x) between the RIA and ELISA values. In all sample types ELISA measurements trended lower than RIA measurements. (A) Plasma CORT concentrations by ELISA and RIA. (B) Urate GCM concentrations measured by RIA and ELISA, levels measured by ELISA are systematically lower by ~50%-600% compared to RIA. (C) Feather CORT concentrations measured by RIA and ELISA appear to agree only for lower CORT concentration samples (<12 ng/g).
Fig 3.
Condor urate glucocorticoid metabolite (GCM) concentrations significantly increased within 2 hours of a handling stressor.
GCM concentrations (ng/g dry wt.) in condor urates collected sequentially following a physical handling and venipuncture event. Panels A-D and J are zoo-captive condors, while panels E-I and K are wild condors. X-axis shows the elapsed time since handling start; note y-axis scale difference between condor panels A-L versus J and K. See S7 Table for additional information on individual urate samples.
Fig 4.
Feather section corticosterone (CORT) concentrations vary over the period of feather growth in free-flying California condors.
CORT concentrations measured in sections of flight feather (2 cm lengths along rachis axis) collected from five individual wild condors. (A) Contiguous 2 cm sections from condor 631’s #6 primary feather show changes in CORT concentrations over the time of feather growth. (B-D) Non-contiguous #3 primary feather sections from condors 192, 312, and 401 show changes in feather CORT concentration over time of feather growth and variation between birds. (E) Retrix (tail) feather CORT levels from condor 336, who died of lead poisoning while receiving clinical treatment (feather growth day 0). The estimated duration (days) of feather section growth is represented for each section by the width of each section line, determined using a primary feather growth rate of 4.4 ± 0.28 mm/day in California condors [28] (See S2 Table for details). Total CORT per feather section (pg) is normalized to feather section length (mm along rachis axis) to represent integrated plasma CORT levels over time of feather section growth [33,46]. Grey shaded area indicates timing of the estimated 18 day period within which the condor was trapped, held captive in a flight pen, and handled.
Fig 5.
Relative influence of predictor parameters on (A) plasma corticosterone (CORT) values, and (B) urate glucocorticoid metabolite (GCM) values in wild California condors. Model averaged estimate of beta coefficients for all top model parameters with error bars depicting 90% confidence intervals. Parameters with confidence intervals including zero do not have sufficient support for predicting the response variable [54].
Table 2.
Ranking of candidate multiple linear regression models describing variation in plasma corticosterone (CORT) concentrations and glucocorticoid metabolite (GCM) concentration of first collected urates in California condors.
Within sample type, the subset of models accounting for 90% of AIC weight and the null model (intercept) are shown.