Tritrophic experiments with plants, herbivorous insects and birds – Part 8: Aviary behavioural experiment

After planning the tritrophic experiment with plants, herbivorous insects and birds, and preparing the aviary, it’s time to conduct the actual experiment. I will use in this blog post as example my own studies (Mäntylä et al. 2004, 2008, 2016) which are quite simple. Inside an aviary there is always one bird and two branches (one systemically induced by herbivores and one intact control). All the photos and video in this blog post are from my experiment with great tits, blue tits, pine sawfly larvae and Scots pines. I have in my experiments used birds captured from the nature, so they were naïve to the study set-up. The capture was done with necessary licences and the birds were released unharmed close to the capture place right after the behavioural test. The average time in captivity for the birds was around one hour. If you are using birds that are kept in captivity, some of the things I mention below could be a bit different.

Objective observation: I mention this already here in the beginning of this blog post: it is very good if the observational data is collected objectively. This means that person watching the bird behaviour from the videos does not know which branch is herbivore-induced and which is control. And also that it is the same person watching all the videos and collecting the data. I have always wanted to do that myself. In my experiments with birches I had always an assistant that cut the branches from the herbivore-tree and from the control tree, and colour-coded those before bringing them to the study booth. Therefore, I only knew at the time that there was an orange branch and a blue branch. After watching the videos of the experiment I was given the information of those colours in each bird test. In my experiments with pines, I didn’t have an assistant. It felt useless to colour-code the branches when I had prepared those myself from the very beginning. But my co-author Sven Kleier came up with a perfect solution. He said that he can use logo removal for the videos. I had always a paper on the floor of the booth telling the number of the bird, which branch was which and the date. He said that it would be easy to remove the text from the videos because it is all the time in the same place. And if the bird flew over or jumped on the paper, it wouldn’t be removed, only slightly blurred. This worked very well. Sven re-named the videos after the logo removal and gave the list with real bird/video identities after I had watched the videos.

Duration: In my very first experiment with willow warbles, mountain birches and Arge sawflies, I didn’t decide any test time per bird beforehand because no one knew how they would behave. But I quickly noticed that 15 minutes was enough for most of the individuals. So, therefore in my following experiments the test time has been 10 or 15 minutes per bird. When there is nothing to eat for the birds in the test booth, they quite quickly lose their interest to search for anything between the leaves/needles.

Ringing of the birds: To get reliable data, you can use each bird individual only once in each experiment. And therefore you need to mark the birds. Ringing them is the best way. For that you need a licence (and the rings) from the local bird ringing authority. Another thing that I quickly learned when doing the tests with birds was that you cannot ring the birds before the behavioural experiment. Especially great tits and blue tits are too interested in the shiny metal ring in their feet, and don’t do anything that they are supposed to do. In any case I had to capture the bird again after the experiment from the booth, so I ringed them after it. At the same time I measured them and took a photo if needed.

Calming down: When the bird is released inside the test booth it is of course first a bit stressed about being in captivity. Some individuals fly around the booth and some just sit frozen on the floor or some other place. But quite quickly they notice that there is nothing too threatening and they calm down and start to get interested in the experimental set-up (the two branches). This calming down is very easy to notice. The bird quickly ruffles all its feathers (see the video clip below). The bird individuals are personalities with different levels of exploratory behaviour. In the experiments we wanted to know the choice of the birds, so we wanted them to be calm when doing this and not just randomly landing on either branch. Some birds calmed down in 30 seconds and some needed a bit more time, but almost all of them calmed down in less than 5 minutes. If the bird did not calm down during the pre-designated test time (10 or 15 minutes), it was released and left away from the data. Maybe 5 % of all the individuals were like these.

0:02 Bird accidentally lands on a branch, 0:13 Bird ruffles its feathers = calms down, 0:18 The first choice of the bird

Data from the video: As I told, in the experiments we were interested which branch the bird chose. We called this the first choice. So, after the bird had calmed down, the first branch it jumped on or flew to was its first choice. I have below figures from my experiments showing the first choice of the birds (systemically herbivore-induced or control branch). I won’t lie, it is very slow (and sometimes boring) to watch the bird videos and collect the data. I’d say on average you need 30 minutes for a 15-minute video. There are nowadays also computer programs to help you with data collection from the videos. In my experiments the problem was that I wasn’t sure if the program would notice if the bird was on the branch or under it. The bird needed to sit on the branch (and not on top of the pot holding the branch) to count it as a visit. Therefore, I did the data collection the old-fashioned way. In addition to the first choice, I also recorded where the bird calmed down, how many visits the bird made to both branches and how much time it spent on those branches. There are also lots of other things you can record, depending on the study set-up. I wanted to keep my study set-up simple, so the bird had three possible places it could be: herbivore-induced branch, control branch or elsewhere. Here is an example of an Excel sheet (example_bird_data) showing information of one bird individual. The first sheet (‘times’) shows what the data looks after I had watched the video and the second (‘final’) shows it in the end with calculations and information of the bird and branch identities. The third sheet is a legend trying to explain what the different columns, rows and cells include. Please, ask me if you want more information. 🙂


© Elina Mäntylä (, 19 January 2017


Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s