An informal chat with a
Biologist [B.CI.0] who works with Butterflies.

1:30-4pm, October 26, 2004

This is not a “formal” contextual inquiry, as this biologist is quite involved with our BioACT project. Nevertheless, the observations I make are useful to help frame what some biologists do in a day’s work.

EcoEvo Lunch

I visited this lunch in Herrin Hall to recruit more biology researchers for contextual inquiries. When I dropped by, a lady was giving a lunchtime talk on Butterfly Dispersion. Through some quick “googling” I decided that she probably was Prof. Carol Boggs (http://www.stanford.edu/group/CCB/Staff/carol.htm). The idea was that she was comparing her results with results from a much earlier study (late 1970’s) by Paul Ehrlich & Cheri Holdren on dispersion of butterflies. She showed satellite maps of the area (near the four corners states) where she was studying, including overlaid imagery of wind direction and magnitude.

Her talk included tables of dispersal rates, which she claimed showed no significant difference (for a reason I could not tell). There was also discussion of pre-diapause larval web mortality, and also mention of stochastic lambda (of some data I didn’t get).

Her first conclusions were that Ehrlich and Holdren might have been asking the wrong questions.

Summary: Biologists present research results as all scientists do, with arguments, and the tables, graphs, statistics, maps, and pictures that support those arguments.

Then, we walked over to B.CI.0’s office where we proceeded to chat about her work and my intent and interests.

Rate Limiting Factors

One of the first topics we discussed was that there are many rate-limiting factors in her work. Some are things that can be helped with automation—for example, transcribing data from her notebook into an Excel spreadsheet or Access database. Some rate-limiters cannot be helped, such as the current weather.

Data that was never used

One of the biggest factors that limits her ability to work effectively is the transformation of data from one form to another. This can be done for various reasons. For one, it may be easier to take notes in the field with a paper notebook, but it’s easier to work with the data in Excel. Then, she might have to import Excel or Access data to Systat or Splus, her statistical analysis software.

A concern she identifies in biology is that getting more data is getting easier and easier, but searching through it is difficult. She cited one instance where a colleague tried videotaping a field excursion. Afterward, they tried working through it, but gave up because there was just “too much information.” But she does firmly believe that the more data she can get, the better. If she had a tool to organize all her data, it would help her do her work better and faster.

Summary: In the ideal case, we would record everything, and have amazing tools to help us filter out all the garbage, and give us what is most important to our work.

The Lab Notebook

She uses her lab notebook to keep very careful notes on procedures, track experiments, and paste results (graphs and charts). She will use the graphs and charts to help her solve problems, and pose questions. The way she arranges and lays out her graphs is important.

The lab notebook serves as her temporal record of her work. Much like a designer’s idea/sketch log, it is a robust, tangible record of the steps she has taken in analyzing the results of her experiments.

Most of the content is actually generated on a computer (procedures written in MSWord, graphs generated by Systat or Splus, tables printed from Excel). Why? First, computer-generated procedures are easy to duplicate. If she wants to run the same experiment, barring a change here or there, a tweak of a single step, she can go to MSWord, change the procedure, and print it out again. Second, computers can generate visualizations that she cannot do by hand. Systat and Splus are software that allow her to explore her data. She needs tangible copies of her data visualizations because it is much easier to manipulate and compare side-by-side on an 11x17” paper notebook than it is to compare on a computer screen. What if she had dual Apple cinema displays? Who knows?

From Mackay’s Biology Notebook paper, we see how Mackay and her team discovered that biologists value their paper notebooks very highly. One great reason for it is because it serves as a temporal record for patent reasons. Institut Pasteur of France has had problems in the past with arguments over patents, so they now keep very careful notebooks to serve as evidence for patent cases.

Summary: Lab Notebooks are important for many reasons. Computerized solutions can never fully replace paper notebooks, but can definitely help.

Calendar on the Wall

A second artifact that she showed me was her project calendar that she used for a project she completed in the summer. It was a three month calendar drawn in marker on 11x17” glossy paper. There were actually two copies. One was created to plan the project, and the other was to record the actual project.

The calendar was used as a collaborative record of tasks done in a project, in this case by two people – the biologist and her undergraduate research assistant. It served as a visual, temporal record of the project. If they wanted to see if a task had been completed, they can look back. If they complete a task, they can cross it out with a large X. The paper calendar allows quick scanning, to get an overview of the project at hand. It also allows portability, because they can bring it into the field if necessary.

One issue I should have probed more deeply was if it is necessary to reference back-and-forth between the calendar and her lab notebook. The link in this case would be the date, of course, as she can quickly flip to the correct place in her lab notebook.

Summary: Temporal records of work appear in both the laboratory notebook and the calendar-on-the-wall. The temporal layout helps to remember what has been done, and what needs to be done.

Coin Envelopes, and Collected Specimens

One project that she is working on involves her going out into the field and collecting leaf samples into little coin-envelopes. The envelopes are marked with unique identifiers to keep track of them, to maintain a link between the physical sample and any data she collects from it.

The samples are then sent on to a technician to analyze the concentration of Nitrogen (N₂) samples. The data will be sent back to her, and she can do her analysis.

Figure 1. Sample Coin Envelopes

Path of Samples/Data

Biologist [sends samples to] à Technician [sends data and samples back to] à Biologist

Again, one question we might ask here are which are the rate limiting steps to her work? Where are the weak junctions? Bottlenecks in data processing? What’s the critical path?

Non-Optimal Situations

A unique aspect of field biological work is the presence of rough conditions. Sometimes, when you are out taking data and measurements, it will be windy, dusty, or rainy. The biologist showed me her notebook made of “Rite in the Rain” paper (http://www.riteintherain.com/). The paper is tear-proof and water-proof, and is used in extreme conditions. You can use pencil to write on this paper. You can use special water-proof pens.

Many of her artifacts are crinkly, with coffee stains and signs of age and use. Some of it serve as evidence of work habits… some of it are evidence of her standard work conditions. Notebooks that are taken out into the field inevitably get put on the ground, accidentally dropped in the mud/water, and always accumulate dust.

These non-optimal conditions do not lend well toward technology use. She recounts PDAs she has used in the field, that have become damaged because of the dusty conditions, which led to severely scratched screens. Inevitably, electronic equipment will be dropped on the ground, or in the water. There are some solutions, of course, such as waterproof casings for PDAs.

Figure 2. Waterproof casings for PDAs.

Why paper? She prefers paper over most forms of note taking because it provides a robust way to record pretty freeform annotation. She has thought of using video or audio, but decided against it because there is just “too much junk to filter.” It is difficult to work through a video/audio account of a field trip and find what were the good and useful annotations, and what were the other 95% of junk. In addition, voice is difficult to capture in the wind without a good microphone. And finally, more electronic equipment means heavier loads when hiking around trails.

Summary: Computing solutions must be able to survive non-optimal situations. However, the added value of a computing solution might be so great that the biologist might be amenable to some added hassle.

The Butterfly collection

She then brought me to see the butterfly collection she works with. They are organized by species, stored in wooden drawers that can be pulled completely out of the enclosing shelf. She will find herself taking out a specimen to collect samples from, to do DNA analysis or spectral analysis of the sample.

Information related to the specimen is stored directly under the specimen (attached to the pin which also holds the butterfly to the foam bottom). Usually all that is attached is a number and name identifying the species. Sometimes, a vial of alcohol containing parts that must be preserved will be glued onto the pin. But that is about it.

Our biologist feels that she would like to be able to attach all the data that is related to the species to the pin. Currently, she has to go back to her database and enter the unique identifying number, and look at the results. It is difficult to ask questions about how species relate because she has to go back and forth between her computer. She would like a means to do queries at the site of the collection (for example, to see which species were found near the Rocky Mountain preserve).

An anecdote she related to me was an experiment to time the movements of butterflies. Someone had made a computer system for her to do so in the field (w/ PDAs)… in the end, all she used was the “stopwatch” feature, to tick the moments that the butterfly flew in a certain pattern. She said that technology is usually flexible enough that you can use it for something other than its primary, or intended purpose.

Summary: It seems the first step to take would be to provide hyperlinks between physical specimens and their digitally stored data.

Technology in the Field

She described to me the CyberTracker system, used by scientists to track animals in the field in Africa. The scientists gave a bunch of PDAs to (often-times illiterate villagers) to help them track animals. Thus the system was very graphics/icons based. Here is a link: http://news.bbc.co.uk/1/shared/spl/hi/picture_gallery/04/africa_south_african_innovators/html/6.stm

Journals of Natural History

“The Naturalist’s Field Journal”

The biologist then described to me the need to keep detailed field logs/notes. It is standard practice for natural historians to keep logs of their travels. These logs include the date/time, location, description of the land, species they see/hear, etc. These artifacts are especially important when they submit specimens to museums. Museums request very complete field notes, to help document the situation surrounding the capture/collection of the specimen.

She gave me this binder of notes photocopied out of Steven Herman’s book describing how to keep a field journal based on a system established by Joseph Grinnell. Here’s a website that describes field journals (http://www.pacifier.com/~mpatters/details/details.html). My notes on the book by Herman can be found in the other document linked to the biology project page. However, I do note here that sketches are probably made obsolete by digital cameras nowadays.

Summary: Good field journals are ideal, but few biologists go to the great care needed to keep them.

Equipment

She referred to some piping and tubing she had to bring out to the field once. Regularly, she’ll have anywhere from a backpack/notebook/pencil to that plus laptop (to upload measurements), dichotomous key, and equipment to set up experiments.

Data Flow

Paper à Excel à MS Access à Splus / Systat à Results in PPT

Earlier records are used as backups, to verify results or data in later stages of the pipeline. If results don’t look right in Splus, the data points in Access or Excel, or even Paper might be checked to make sure.

Note to Self

Here are some EcoEvo faculty that I should contact… perhaps I can get them and/or their grad students to participate in the contextual inquiries. http://biology.stanford.edu/faculty_eco.html