Red-Winged Blackbird Analysis


Data from 52 RWBL that we caught in 2016 & 2017 in Ithaca between 18 April and 22 July. We tried to take a 7 point stress series with samples at 1, 3, 7, 15, 30, 45, 60 minutes. The exact time points vary a little bit though and we use the exact measures in all the plots and analyses. Some samples are also missing from the end or middle of some time series.

Note that I have removed one bird from early season that has an astronomically high cort value even starting at baseline and through the full series. It doesn’t really make any qualitative difference, but including that bird just makes it hard to plot because it is so much higher than any others.

Descriptive Plots

First, I’m just plotting the full stress series on the raw and log scale. There are a lot of possible choices here for how best to display this, so I’m dropping in a few different options at this point. The y axis scale is very different for the three breeding stages, which makes it hard to compare the shapes. There are a few options with changing the scales or log transforming or plotting in different facets that make it easier to compare.


Figure x. Time course of corticosterone measurements in each life history stage by individual (gray lines and points) and by group (red line and confidence interval). Note that y axes differ between the plots.


Figure x. Same as above but on log scale.


Figure x. Lines on the same panels. Left is normal scale right is log scale.

Response Shape by Life History

A lot of classic studies ask about differences in absolute cort values at different life history stages (e.g., find some Wingfield studies). We can see here that there are big differences between early breeding vs. late breeding vs. molt. For the most part, early breeding is higher than the other two across the board. We can also ask though whether the shape of the stress response differs or not between stages. For this I’m plotting the responses as a percentage of maximum for each individual over our seven point time series. This lets you see very clearly how the shape of the time course differs between stages.

I could do some actual statistics on this, but basically here you can see that:
- Early breeders start with a baseline value that is already at a higher percentage of their maximum value.
- Early & Late breeders achieve a higher percentage of their maximum value faster (i.e., they have a ‘faster’ cort response).
- Molting birds achieve maximum value later (i.e., slower response).
- As a result of those last two points, molting birds actually sustain their maximal rate of increase longer than the other stages.


Figure x. Time course of corticosterone as a percentage of maximum corticosterone achieved. Panel A shows loess regressions for each individual sampled. Panel B shows the combined loess regressions with 95% confidence interval for each life history stage.