Interaction project end-of-season summary

Thank you to Eliot Miller for this end-of-season summary about a new project investigating bird interactions at your feeders. And thank you to all FeederWatchers who sent in data! Read on to see how Eliot has put them to amazing use.

This FeederWatch season, we ran a pilot project aimed at documenting interactions at feeders. The project started late in the FeederWatch season; however, in that short time, we received a tremendous show of interest and data. Thanks to everyone who participated! The project was led by me, Eliot Miller, and I’d like to share our preliminary summaries and answer your most frequently asked questions.

Preliminary results

More than 200 FeederWatchers (209 to be exact!) submitted an impressive 1994 observations of interactions at feeders. This map illustrates where interactions were observed:

If you have interactions that you haven’t added yet, it’s not too late. We will be writing up the results of this first season in the coming weeks. If you add observations soon, they could still be a part of the forthcoming scientific paper. Not sure how to get started? Read this introductory page.

Of the nearly 2000 observations, only 2.5 percent were flagged as unusual and were excluded from this preliminary analysis. If you do not see your observation, I will be following up with you via email. Of the remaining observations, 1874 of these were records of one bird displacing another, 37 were of one bird catching and eating another, and 23 were of one bird mobbing another. There are, of course, many other interesting behaviors we could quantify at feeders, but of the three included in this pilot season, clearly displacement was the most common.

In total, 96 species were observed in these interactions. Species varied in how frequently they appeared in the submitted data. This variation reflects both the commonness of some species and how frequently individuals of some species interact with others. For instance, even though Carolina Wrens and doves are commonly seen at feeders, they are much less well represented in the interaction data when compared with other species. These birds tend to arrive at your feeders, eat quietly, and don’t cause much of a commotion. American Goldfinches, House Finches, House Sparrows, Northern Cardinals, and Downy Woodpeckers, on the other hand, appeared very frequently in the database. Perhaps this is because these species are much more likely to interact with others at the feeders–we will find out as we continue to analyze the data.

Preliminary analysis of aggressive behaviors

Birds that were particularly aggressive to members of their own species included Anna’s Hummingbirds, Common Redpolls, and American Goldfinches. Birds that appeared commonly in participant submissions but were infrequently aggressive to members of their own species included Blue Jays, Red-bellied Woodpeckers, and Common Grackles. We have not yet accounted for how common these birds are at feeders in general. Doing so will give us a much better sense of how aggressive different species are towards other birds, both those of the same species and those of other species.

Displacement observations

The following graph illustrates displacement observations (click to enlarge):

This graph may seem overwhelming, but it isn’t as complex as it seems. In this graph, arrows run from the source to the target of aggression. The thickness and color (from blue through pale yellow to red) indicate the frequency with which that interaction was seen. Thick red lines represent interactions that were very commonly observed, for instance of Hairy Woodpecker displacing Downy Woodpecker. Finches and sparrows can be found towards the left of the graph, woodpeckers and doves towards the right, small birds towards the bottom, and large birds towards the top. In some cases, it might look like the arrow runs both ways–that’s because it does! A good example is between Northern Cardinal and House Finch. Most of the time Northern Cardinals displaced House Finches (big red arrow), but a few times House Finches displaced Northern Cardinals (skinny blue arrow, mostly hidden under the red arrow).

Looking at the graph, a few trends emerge. Many of the arrows run downwards, indicating that larger birds tend to displace smaller birds from feeders. But this pattern doesn’t always hold true. Bushtits displaced chickadees, Downy Woodpeckers displaced Mourning Doves, and many more counter examples can be found. There’s more to it than just size. What drives the direction of these interactions? Answering that question will be a major objective in analyzing the data submitted by FeederWatchers!

Predation observations

Below is a diagram of observed predation (click to enlarge):

Predators are colored in red, prey in yellow. Looking at this diagram, it’s clear that Sharp-shinned Hawks tend to focus on smaller prey and Cooper’s Hawks tend to focus on larger prey. Interestingly, although much of Sharp-shinned Hawk prey are forest birds, much of the Cooper’s Hawk prey species are considered invasive or expanding beyond their traditional ranges. Can this observation explain recent research showing Cooper’s Hawk counts increasing in urban areas?

Mobbing observations

Finally, here is a diagram of mobbing observations (click to enlarge):

Note that many of the flagged observations were for things like Red-winged Blackbirds and European Starlings mobbing other species. Although these birds travel in “mobs” (flocks), mobbing behavior usually refers to instances where one or many birds harass a potential predator (for instance, a group of crows chasing a hawk). This Wikipedia page provides added detail, but it considers mobbing to only be an anti-predator behavior. While that is by far the most common way you’ll see this behavior, we define mobbing more broadly: a bird or multiple birds harassing a bird, usually a predator, that the harassing birds wouldn’t otherwise be able to remove (displace). So, it may well be that mobbing adequately describes the flagged observations of blackbirds and starlings “mobbing”, but for now, those observation have been excluded until the behavior is verified. Of the remaining mobbing observations, it’s probably no surprise that corvids (jays, crows, ravens, and magpies) make many appearances here. Ultimately, we may be able to see if these mobbing behaviors vary over time and space.

Thank you FeederWatchers!
The level of interest this project has been extremely encouraging. The initial blog post received 119 comments. More than 200 of you have participated, and all in just the last few months of the FeederWatch season. I have traded personal emails with many of you, and I have learned more about these interactive behaviors than I ever could have learned by myself, watching my own feeders. We now have data from all over North America, which is really exciting. From here, the next step will be digging deeper into the data, trying to understand what drives these patterns. After we account for what birds were at the feeders when you saw these interactions, we should have a much better sense of how often the different species engage in these sorts of interactions. I will be presenting these initial findings at the upcoming North American Ornithological Conference, then summarizing our findings in a forthcoming scientific paper. And that’s what we’ve accomplished just from the first season! The project will almost surely continue next season. As we build the database, we will eventually be able to look at how these interactions vary across time and space. Are there cases where a species is dominant in one area, but subordinate in another? Are there times of year when a species is more aggressive than others? I hope to be writing a blog post with answers to these questions by next year. Thank you all so much for your participation!

Frequently asked questions

Q: How do I participate?
A: First, you need to be an active participant in FeederWatch. Then read how to get started. If you are interested in contributing data for the 2016-17 FeederWatch season, sign up now. The upcoming FeederWatch season starts November 12, 2016, and we start shipping instructional kits in September.

Q: How do I find my checklist ID, and is this necessary?
A: Yes, we link the interactions you submit to the actual FeederWatch checklist via that ID. You can see examples of how to get the ID in this YouTube video.

Q: Are you only interested in interactions between different species of birds (interspecific), or are you also interested in interactions between the same species of birds (intraspecific)?
A: Both! Initially the project was intended to focus on interspecific interactions. However, before I had a chance to clarify the instructions, so many intraspecific observations were submitted that I decided to leave the option open to submit intraspecific observations. It turns out, as shown above, these observations might provide a good indicator of a species’ overall level of aggression.

Q: I saw a behavior (for instance, failed predation, chasing, etc) that does not seem to fit into one of the three categories you are collecting data for. What should I do?
A: Because this was the pilot year and we wanted to restrict our data collection to a few easily defined behaviors, our list of behaviors is admittedly quite limited. We hope to add additional behaviors next year, but for now your observations have to fit into one of the three behavioral categories: displacement, mobbing, or predation. If the behavior you observed was not one of these things, please keep your notes, and if we add that behavior as a category in the future, submit your observation then!

Q: Will the project continue next year and/or be expanded to eBird?
A: Many folks enjoyed participating in the project, and many others wanted to continue submitting data after the FeederWatch season ended. In terms of FeederWatch, the project will almost certainly continue next year in some fashion. Either much like it is or perhaps in a slightly more integrated way. There is some chance it will be expanded to eBird, but that remains a distant possibility at the moment.

Q: Can I collect data after the FeederWatch season?
A: Although there is no formal way to submit observations you collect outside of your FeederWatch counts at the moment, you can still continue to collect them as part of eBird counts! You can email your observations to me along with your eBird checklist number. Or you can keep them in your files, and someday, I hope, you will be able to add them to an eBird checklist.

Q: I enter all my information, and everything seems right, but when I enter the observation it doesn’t show up in the table below. What’s wrong?
A: This was a significant problem this season. Unfortunately, we do not know the cause. It appeared to be the system getting overloaded, and it tended to happen over the weekend when people were submitting lots of data. Next year, we will strive to have a more integrated system. Please be patient as we work out the kinks. In the meantime, to assure yourself the data did indeed get entered, reload the input site and paste in your checklist ID. Your observations will probably show up in the table of submitted interactions at the bottom of the page. If they don’t, hang onto them and send me an email. It’s important to us that your data are properly recorded.

Follow the FeederWatch interactions research on my personal webpage and on Twitter Follow @EliotITMiller