| |
    
|
Furthering Landmark Discoveries at the Jane Goodall Institute
by Gregory Lauer Jane Goodall, the biologist whose highly regarded
studies of chimpanzees vastly improved our understanding of primates, is
in the American lexicon of popular and respected scientists. Her book,
In the Shadow of Man, is considered a landmark text by many. In November
of 1995, she addressed an overflowing audience of several thousand in Northrop
Auditorium and discussed her life and work in Tanzania's Gombe National
Park.
Few people, however, are aware of the Jane Goodall Institute's Center for Primate Studies located on the St. Paul campus. The Center houses more than 30 years of study notes and data in a collection of filing cabinets on the second floor of the Ecology Building. Dr. Anne Pusey, an ecology professor, is leading an intensive effort to digitize these records and utilize modern computing methods for analysis and research.
More than two decades ago, Pusey worked under the direction of Goodall as a field researcher in Gombe. Today she draws on that experience to make fundamental contributions to our knowledge of chimpanzees. "Now that the study has been going for 35 years, we're finally able to answer some long-standing questions," Pusey said.
Pusey is particularly interested in groups and their social structures. "One of the things we're looking at is female relationships among chimps because they haven't been very well studied. We're just now finding out that female chimps do indeed have dominant relationships," she said.
Jennifer Williams, an ecology graduate student researching with Pusey, explained the problem, "We're trying to determine what makes a successful female from an evolutionary perspective."
Part of the difficulty lies in the nature of the data. The Gombe study is one of the longest and most comprehensive of all such biological field studies ever undertaken. Researchers now have reams of data to analyze. Complicating matters further, many of the records are handwritten, and some of the notes are in Swahili.
Since 1960, Jane Goodall and others have collected the same type of data on the chimps. "Researchers record major events like what it's eating, where it is on the map, who it's with, various behaviors, and so on," said Pusey.
These notes are gradually migrating to the binary world of computers thanks to an array of volunteers. Unfortunately, it may take years to fully convert all of the records to an electronic format.
Currently, Pusey and Williams work in conjunction with Dr. John Carlis and others in the Department of Computer Science. Carlis, a database expert, was attracted to the project several years ago. "There are great gobs of good biology to do, and this is definitely worthwhile," he said.
Databases are powerful interpretational tools because they enable researchers to perform queries quickly and efficiently. Additionally, "Databases allow us to ask questions that were previously considered impossible," Carlis said.
For example, a biologist might ask how a community range (the area where a group of chimpanzees live) changes with respect to various environmental factors and conditions. Historically, these studies were done manually and required hours of laborious and tedious graphing. With the aid of a database, however, the most complicated part of the process is translating the question into a structured database query. Searching, extracting, and presenting relevant data is then much faster and efficient.
With databases, researchers can spend more time interpreting the data and less time sifting through it. "Some chimps are followed several times a week and may be spotted in more than 1,000 fifteen-minute intervals over the course of a year," said Scott Krieger, a senior in computer science who is analyzing the Goodall data under the direction of Dr. Carlis. Sorting through a pile of records and physically tracking chimps by hand is simply not practical.
Last year Pusey and Williams utilized the database to determine the percentage of time female chimps spend alone. Although the question seems simple at first glance, it required genuine collaboration between the biologists and computer scientists.
At first, the numbers didn't seem correct. "Females normally spend up to 80 percent of their time alone, but the results from the database were much lower," Williams said.
The problem was one of definition. In particular, what exactly did "alone" mean? From observations, female chimps spend a large amount of time with infants. Thus, the definition of "alone" was enlarged from "absolutely alone" to "absolutely alone or with immediate family." The issue then becomes one of defining "immediate family." Additionally, biologists consider other factors, such as whether a female is in estrus, before classifying a chimp as "alone."
Although translating questions into database queries is a complicated process, databases offer a great deal of flexibility. If an experimental query is unsuccessful, reformulating the question is relatively easy, and within minutes, a new data set is produced. "We can easily change our minds," said Krieger.
Creating mechanisms to foster communication is an essential part of the project. "Database modeling is about getting people to communicate precisely. We've done a lot of the spade work so that we understand each other's vocabulary, but it's not easy to be precise," Carlis said.
Another computer science professor, Dr. Phillip Barry, is examining various visualization techniques. "Right now we've done some prototype projects, taking a chunk of the data and looking at it in different ways. The end idea is to integrate it so a biologist can say, 'I want this sort of information from the database, and I want to look at it this way or with these kinds of tools.'"
Several months ago, Barry and his students created experimental visualizations of female dominance data. The primary indicator of dominance is the pant grunt, a uni-directional chimp vocalization. Although biologists do consider other behavior when evaluating chimp hierarchies, the strong correlation between pant grunts and dominance makes it a suitable basis for most investigations. A database query extracted the relevant pant grunt records.
In one visualization, nodes represented individual female chimps, and lines between nodes demonstrated how many times one female chimp dominated another. Furthermore, the display was animated with respect to time. Computer graphics allow researchers to "do some things that can't be done with traditional static displays," said Barry.
Pusey agreed, "We may be able to capture all of the data in one picture, and that will be very interesting." Chimps live in groups of 40 to 45 individuals, and they are constantly splitting up and forming subgroups. With the aid of powerful visualization packages, it may be easier to track shifting alliances and associations.
The benefits associated with interdisciplinary projects such as this are substantial. From a biological standpoint, utilizing database applications offers unprecedented opportunities to view the data in unique and novel ways.
"Without computer analysis, there's no way we could do any of this. That's pretty much the only difference between what we're doing and what others have done. People have made tallies before, but there's only so much time for that sort of thing. Once it's computerized, there's an entirely different level you can get at," Williams said.
From the viewpoint of a computer scientist, the Goodall data extends the boundaries of traditional database functionality. "The hope is that at some point the users will attempt to push the limits of the database," said Krieger. When biologists like Pusey ask questions that can't be answered with more standard techniques and methodologies, Carlis develops database extensions that deliver additional capabilities. The extensions may then be generalized, and they later become tools available for future database users in diverse fields aside from primatology. "We're not just writing chimp procedures," Carlis said.
Barry sees similar advantages in working with the Center for Primate Studies. "If you look at the state of visualization, the big questions revolve around data that doesn't have a natural analog. For example, if you're trying to visualize data from a CT scan of the human brain, the process is fairly straight-forward." However, graphically depicting relative female dominance and chimp hierarchies isn't nearly as easy. "You have a lot more challenges from the design perspective, which I appreciate," said Barry.
Pusey summarized the importance of continuing Goodall's work, "Chimps are our closest living relatives. I study them because they are a social animal, and I'm interested in the general principles that explain sociality. It's very likely that chimps can tell us something about the behavior of our common ancestors.
"For most people, it's very intriguing to look at the similarities and differences between chimpanzees and ourselves. That's certainly why chimps capture the public's imagination, and that's partly why this research is important."
