Choosing where to live is an important decision for animals and many species that prefer to stay within a confined area, so their habitat is limited to a specific range. This means that adaptation to that particular environment confers higher fitness than other non-adapted habitats. On the face of it, the theory appears sound, but is there any evidence that this is the case?
Linda Partridge (1974, 1975) examined this theory by testing two species of songbirds: the coal tit and the blue tit. The coal tit prefers pine woods, while the blue tit prefers oak woodlands. It was hypothesized that each species would be more efficient gathering food in their preferred habitat, as measured by the amount of time the birds took in gathering food using various habitat-specific techniques.
Partridge tested these ideas by hand-rearing young birds of each species. They were then tested on a number of food gathering tasks. In the wild, each species uses different techniques and strategies that appear well suited for the type of environment they inhabit. The blue tits, for example, live in dense oak woodlands and are good at obtaining food by hacking and pulling covered food, hanging upside down from branches, etc. Coal tits do not use these techniques because they live in more opened spaces. When tested, the blue tits outperformed the coal tits in tasks that involved these skills, indicating that birds were not generalists capable of foraging in any habitat but had adapted to their specific environment. There are theoretical and experimental limitations to this study, but in general, it showed that adaptation to a specific habitat made a species more efficient at food gathering and thus more fit (if we take food gathering as a measure of fitness) (Alcock, 1993).
As I was reading this article, I wondered if I could use AI software (Grok Imagine) to create a relatively accurate illustration of the two species doing their thing in their natural habitats. I would not have attempted this a few months ago for several reasons:
Species identification. From past experience, trying to generate an accurate animal species was a hit and miss affair, particularly if the animal was not well known.
Environmental context. I wanted to show each bird in their preferred habitat. Would the software understand this from a general description?
Split screen. This is a relatively new function in videos generated with Imagine. The command allows you to split the image into several segments that can be independently animated to show split-screen sequences.
Incorrect 'tit' output. Given the risky name of this bird, I was concerned (though slightly curious) about the software output.
First, I tested the software with a general prompt:
A dual split screen with the first one showing a blue tit bird foraging for food in its preferred environment and the second screen showing a coal tit bird also foraging for food in its preferred environment.
Base image for first animation
Notice the vagueness of the prompt. I was not aiming for intricate artistic flair in this illustration but scientific accuracy with minimal input. The software generated multiple versions of the birds side by side, and though I cannot vouch for the biological exactitude of the output, they looked good to me. I chose the one that more closely fit what I was aiming for and animated it without any further prompting. The chirping sounds were a bonus, but I cannot confirm whether these are actual sounds from those species and not just some random bird sounds.
The one problem that I noticed in this particular image and video was that the coal tit bird is the one foraging up on the tree and is not really hanging upside-down. Given that blue tits were the ones who excelled in the experiment, then that is the species that should be hanging around on the tree branch. I edited the prompt and used more specific language:
A dual split screen with the first one showing a blue tit bird foraging while hanging upside-down on a branch and the second screen showing a coal tit bird also foraging for food on the grass
Base image for second animation
In this version, the species remained in the same positions, but their environment and foraging behaviour was altered. The agile blue tit now hung from the branches like a pro.
Broadly, I made the following observations.
Species identification. The software seems to have generated the correct species for the blue and the coal tits. It does require careful verification.
Habit/foraging behaviour- It appears accurate, at least good enough for demonstration purposes. It would be interesting to find out if the software understands that different birds exhibit different foraging behaviours and techniques.
Split screen. This function worked flawlessly and could be made even more complex with varying dimensions and effects.
Incorrect 'tit' output. I was relieved (phew!) that the software understood I was referring to a bird species in my prompt and not something more spicy. :P
Sounds. I had honestly forgotten about sound, so I was pleasantly surprised by the chirping, though I cannot confirm its biological accuracy.
The ability to generate images based on artistic styles is one thing that is amazing in and of itself, but adding factual scientific information to images and videos is mind boggling. There are many educational topics that are not so easily understood and graphical visualizations help clarify these complex ideas. Traditionally, developing such videos required a lot technical and info-design knowhow, time, and resources. A tool like Grok Imagine streamlines the process and puts vast amount of design power in a user's hands.
Alcock, J. 1993. Animal Behaviour: An evolutionary approach.
Partridge, L. 1974. Habitat selection in titmice. Nature 247:573-574.
Partridge, L. 1976. Field and laboratory observations on the foraging and feeding techniques of blue tits (Parus caeruleus) and coal tits (Parus ater) in relation to their habitss. Animal Behaviour 24:534-544).
| X | InLeo | NFT Showroom |
▶️ 3Speak
