Why a Robot?

Photo Gail Patricelli.

Yesterday we got to take off our researcher hats and put on our outreach hats (we also traded our field clothes for some clean pants). Gail and I brought “Snooki”, the latest version of our robotic female grouse, down to Lander to participate in Teen Tech Week at the public library. Word had spread among the library staff; they were quite excited to get to meet the robot in person. Some of the kids were apprehensive, but most jumped right in to give the test drive. I had loaded some videos of the fembot and the grouse on my iPad so they could see how she looked in action.

 

With the exception of the wheels, the robot makes a pretty convincing sage-grouse. This version has the mechanics tucked inside a fiberglass shell that was poured over a grouse-specific taxidermy mold. Gail artfully arranged real grouse skins over this to complete the disguise. “Snooki” can turn in place using the wheels, the body can pivot down, the neck bend up and down, and the head can swivel back and forth. You can see a video of the fembot in action at the bottom of this post.

Fembot (left) and "undressed" fembot, right.

One of the common questions we get at this sort of event is “Is the robot just for fun? What do you do with it?” Let me take a moment to answer that.

Our robotic sage-grouse

In my recent post, I talked about how we use leks as a way to study sexual selection and mate choice. Most studies of lekking animals have looked at male traits that don’t vary much over the course of the breeding season. For example, in peafowl, biologists can look at which males have the longest tails or most tail eye-spots, and see if that might relate to patterns of mate choice. You could picture this as each male holding up a sign with his “score”, and females look around until they find the male with the highest score.

Anyone who’s watched animal courtship knows that males and females aren’t always politely assessing each other from a distance. Real courtship often involves complex decisions and interactions: deciding whom to approach, how quickly, reading signals or cues and responding accordingly. For a male to succeed in the mating game, the skills required to navigate the complex world of courtship can be as important as the physical traits he carries. Some of you may even have personal experience with this situation– meeting someone who is very attractive but who comes on a little too strong, for example.

Unfortunately we don’t know much about the importance of these “social skills” in non-humans because they can be hard to measure, especially in the wild. This is where the robot comes in.  With a robotic female, we can control one side of the conversation. The fembot gives us two important tools for comparing males on the lek.

Male courting fembot.

First, it presents all males with a standardized stimulus. In live courtships, females may approach the top guys much more closely and provide signals of interest, while other males are consistently given the cold shoulder. The robot lets us measure all the males on an even playing field.

Second, it allows us to experimentally control the conditions of courtship. In a previous experiment, we could look at how male sage-grouse responded to a very basic aspect of female behavior– how close or far away she is from a male. With the more advanced version of the robot that we debuted last year, we could send either “coy” or “interested” signals to the males. I’ll describe the plan for this year’s experiments when we get a little closer to conducting them.

The older version of the robot ran on train tracks. Here we are moving her closer to two admirers.

The robot also serves an important function as a target of courtship- one of the skills we are interested in is the male’s ability to aim his best signal at the female, and in a past iteration, the robot could record what a female would hear, rather than what a biologist can record from some arbitrary position on the lek.

There are several ways that biologists can manipulate an animal’s social environment (video playbacks in the lab, audio playbacks of bird song, etc), but the robot gives us a unique way to interact with animals in wild.

Sage-Grouse leks: One of the greatest shows on Earth!

I’m kicking off my blog for the 2013 research season with a brief description of what makes sage-grouse such a great bird to study for someone interested in animal behavior and evolutionary biology.

Male sage-grouse courts a female.

One of North America’s most spectacular birds is also a species that not many people have seen. I’m referring to sage-grouse: I study the more widely distributed greater sage-grouse (Centrocercus urophasianus); I have yet to personally see the less common Gunnison’s sage-grouse (Centrocercus minimus), but that is definitely on my wish list. Given the spectacular plumage of a male sage-grouse in display, why are these birds so hard to see? A quick look at a female sage-grouse tells you the girls are built for crypsis- well adapted to blending in with their environment. For most of the year male sage-grouse also play the hiding game, so unless scared into flight they may pass unnoticed. Yet for a couple of months in winter and spring, males come into their traditional display grounds (called leks, from a Swedish word for “child’s play”) and put on one of the greatest shows on earth. These leks are often in fairly remote areas, and males typically attend them only in the early morning hours. Both of these reasons help explain why getting a look at this spectacle can be a bit of a challenge.

 

These unusual breeding clusters have captivated not only birders but evolutionary biologists as well. In about 90% of birds, both parents provide some care for the nestlings. In those cases, females are often choosing their mates at least partly on the material benefits they get from this partnership, whether it be the quality of the male’s territory or his ability to provision the female during incubation or the chicks once they’ve hatched. [Note- in many birds males and females mate outside of this pair bond, but that is a tale for another day]. Lekking species are therefore unusual among birds in that males don’t form a bond with their mate nor provide any child care. Scientists are still trying to unravel some of the puzzles that leks represent. Why do males cluster together to display, rather than searching around for females, following females around, or spacing themselves farther apart and defending larger territories like most other birds do? If males aren’t helping raise the kids, why are females so picky? What benefits do females get from choosing one male instead of another? And given that females often pick only a few among the many males on a lek, why do the “loser” males bother to stick around?

(This is a 3-hour time lapse video of the lek. The males appear as small black-and-white specks at the bottom of the frame)

Lekking animals also tend to be high on the charisma scale. Besides the spectacular sage-grouse and their cousins the prairie chickens and sharp-tailed grouse, other lek-breeders include some of the most beautiful and acrobatic birds out there, including birds of paradise, neotropical manakins, peacocks, cock-of-the-rock, some hummingbirds, and ruffs. When we see a species in which males are larger or more colorful than females, we presume these differences are related to an evolutionary process called sexual selection, where one sex- often the males- competes either directly for access to females or indirectly by producing the best advertisement among the other males. This certainly seems plausible for sage-grouse; not only are adult males almost twice the weight of females, but they have a range of specialized feathers, brightened skin patches, and other unique structures.

Male sage-grouse showing his inflated vocal sacs.

The distinctive air-filled vocal sacs are actually part of the digestive system; once inflated with air, powerful muscles just under the skin help move and shape the them during display. In spite of decades of research and routine collection by hunters, we are still finding surprising structures in these birds. Just recently we discovered that male sage-grouse have an almost songbird-like syrinx (sound-producing organ analagous to the human larynx) capable of producing two tones at once. All of this adds up to the bizarre appearance of the male that has presumably evolved through sexual selection- the males with the more elaborate versions of these unique features get to mate with more females and pass on more copies of their genes to the next generation.

 

These differences between males and females extend to courtship behaviors as well; males and only males have a characteristic “strut” display. A male’s strut serves to attract females to the lek from across the landscape, to woo females once they are on the lek, and most likely to help claim their small patch in the midst of all the other males. Each display lasts about two seconds and involves coordinated movements of the wings and body.

For sage-grouse, courtship is not just a visual spectacle. Males produce a variety of sounds during the strut display. The first two notes actually are not vocalizations at all, but instead made by rubbing stiff, pointed breast feathers against the inside of the wings. This is somewhat analogous to how crickets chirp. Making sounds with feathers may sound unusual, but it has evolved repeatedly in birds. Some common species in the Bay Area that do this include Mourning Doves (the ‘wee-wee-wee-wee’ made during take off) and Anna’s Hummingbirds (the loud chirp made at the nadir of the male’s dive display).

Spectrogram of a sage-grouse display. The two feather-produced 'swish' notes occur at about 2.6 and 3.6 seconds. The first low frequency 'coo' note is at 4 seconds, and is followed by the pop-whistle-pop at about 4.3 seconds. You can hear this in the video above.

The remaining notes are true vocalizations made by the syrinx, although unlike most birds they are made with the beak closed. These sounds start with a series of three low frequency ‘coo’ notes, and conclude with an up-down-up ‘whistle’ note sandwiched between two staccato ‘pops’. You can see these in the spectrogram- this is a visual representation of sound and you can read it much like reading music, with time progressing towards the right, pitch becoming higher towards the top of the figure, and the darkness representing something like loudness. Females may care about some very subtle differences in these sounds when they are looking for a high quality mate. Researchers have compared sound recordings of successful and unsuccessful males and found differences in the relative timing of the two ‘pop’ notes, and maybe the loudness of the whistle. What is amazing is that the differences between ‘Mr. Right’ and ‘Mr. Wrong’ are on the order of less than a tenth of a second. Females may have quite the ear when it comes to picking their mate!

 

That’s a quick introduction to sage-grouse in the spring. I feel extremely lucky to have heard and seen this show for the past several years as part of my research at the University of California, Davis. Along with Professor Gail Particelli, graduate student Anna Perry, and our intrepid field crew, we will be conducting research into several aspects of sage-grouse behavior and ecology from our field site just east of the Wind River Range in Wyoming. Over the next couple of months I’ll be discussing what it takes to set up a camp like this, how we use new technologies (including robotic birds!) to study courtship in this species, and review some of the conservation studies out of our lab. I hope you’ll join me for our 2013 field season!

Two big updates from the fall

The big updates from last quarter are probably the following:

1)   I survived teaching my first class- the 175-student Intro to Evolution course here in the Department of Evolution and Ecology at UC Davis. It was definitely a scramble to get the lectures put together each week, and I can see areas for improvement both for the course content and my teaching style, but all in all I think it went pretty well. At least no pitchforks and torches outside of my office, and reasonably good student evaluations- very few marks below average in any of the categories. Although it really is amazing how much time it takes to outline a lecture and craft a set of powerpoint slides, I enjoyed scouring the web for interesting class examples and media, and I liked the time in front of the class (at least most of the time- 7:30 AM could be rough if I didn’t sleep well or was up late putting the finishing touches on the lecture.

While I’m on this topic, a big thank you to my TAs David Luecke and Sarah Signor, as well as my graders Matt Meisner and Dena Grossenbacher. They did a fantastic job in spite of my having to learn everything from scratch my first time through as a full instructor. I’ve thanked them several times already, but can’t hurt to do it once more.

 

2)   Gail and I were awarded our second full National Science Foundation research grant on the sage-grouse system this fall! This proposal will fund another 3 years of field studies in Wyoming. The focus of this grant will be to learn more about the foraging ecology of the grouse, and how variation in energy acquisition and off-lek behavior relate to the courtship tactics available to a male while wooing females on the lek. It will extend our use of robots and sound recording by giving us a lot more information on what males are doing in the remaining 21 hours of the day when they aren’t fighting and strutting on the lek.

This project will also include both old and new collaborations. We are welcoming back John Burt (University of Washington) who was a key developer for most of our sound analysis scripts. In the past few years he has turned his attention to creating next-generation telemetry tags for use in animal behavior studies. These Encounternet tags can do amazing things like share interaction histories so when either of two animals gets near a receiving station, it will receive the data from both animals. The tags John is creating for us should have both gps logging capability as well as accelerometers to measure movements (this technology has already been used in other animal systems, such as the mountain lion study out of UC Santa Cruz). Our second collaborator is Jennifer Forbey from Boise State University. Jennifer studies the interactions between sage-grouse and their main food item- sage brush. She will be helping us study foraging success in the grouse, in particular by measuring the suite of secondary chemicals that sage plants produce.

NSF grant in, NAOC poster done

A brief mid-summer update- it has been a busy few weeks after getting back from my Wyoming/Colorado trip at the beginning of July. Aside from the usual manuscript reviews and progress on our own manuscripts, a couple of noteworthy things to check off:

Gail, Anna, and I submitted our latest NSF grant that would fund another 3 years of work on the sage-grouse. We brought on another collaborator: Jennifer Forbey is a professor at Boise State University in Idaho, and is an expert in herbivore/plant dynamics, particularly the importance of nutritional content and plant-produced toxins. With Jennifer on board, we will have a much stronger foraging component to our examination of off-lek behaviors of the grouse (in other words, what they are doing the other 20 hours of the day when they are not courting and fighting on the lek ). I feel like I usually do after writing one of these: exhausted but excited.

I also just finished making a poster for the North American Ornithological Congress meeting next week in Vancouver. I’ll be presenting some new analyses on the lateralization in behavior. Last year I presented on a similar topic at the Animal Behavior meetings, but I didn’t have much time to put together something (I was originally going to present our mechanical sound (i.e. “swish” project), but Becca was able to present this work herself, so I switched to the lateral bias analysis at the moment. This year it was still a rush, but I think we ended up with analyses that better link the existing literature on lateral biases to our own data, and hopefully have results that are pretty close to what will end up in the eventual publications.

The second Summer Session started this week in Davis, so I’m meeting with a few new prospective undergraduate researchers. We’ve made a lot of progress so far this summer- we’ve almost finished measuring mating success from our 2012 Chugwater Lek tapes and have started the Monument Lek mating success tapes. Marty, our summer program intern, has even gotten into the 2012 robot experiment tapes. It’s never good practice to get too excited over preliminary results from partially-collected data, but it does look really encouraging so far (his data look at whether males treated the “coy/disinterested” and “interested” behaviors of the fembot differently). We are also almost done with a sample of display behavior measures from the 2011 season- this will add to our understanding of males who were tested in the “environmental responsiveness” playback experiment in 2011.

Video of an experiment

We’ve started Round 2 of our experiments, measuring how responsive males are to differences in the behavior of the robot female. We want to see if we can explain some of the variation we see in how hard males are working on the lek, and one of the things we would like to know is the importance of female interest- if a female is disinterested, should a male work harder to try to win her over, or cut his losses and save his energy for the next potential mate? Does it matter what his expectation is for another female visit, and how successful he is?

Below you can see a clip of the disinterested behavior- the robot is simulating pecking at the ground. Real females frequently do this on the lek as well. This clip is from Cottontail, but we finished our second treatment on Monument Lek this morning. Chugwater will require one more experiment, and Cottontail will probably require two or three.

I love this video, not only for showing off the great job Gail has done getting the robot to look realistic, but also for the funny chasing behavior occuring in the foreground. A non-territorial male ambles from left to right, and just as he reaches the right edge of the screen, he must spot an angry territory holder. He walks quickly back to the left, and just as he leaves the screen, a fully inflated male comes sprinting across.