1:30 to 3:30pm, November 5, 2004
On Friday, 5 November 2004, I visited the workplace of Biologist B.CI.2. He is a 3rd year PhD student in Biology, studying Landscape Ecology. In particular, he studies bees and pollination in Central and South America. His most recent work was in Costa Rica, where he researched how different land use and land types affect the amount of pollination that bees perform (i.e. spatial studies of pollination). He has spent two seasons in the field.
He hasn’t published papers on bees yet, but has published papers on other topics in the past. Currently, he is working on a chapter for a book on bees. He has also published some material in spanish; he’s working on translating it and getting it published here. He’s working a grant proposal to get some more funding for his research.
He used his personal field notebook for “mental notes.” He did not record lots of data in the field, but would note things like from why he chose a certain site, to how many hours his hired help worked….quite varied in use. He didn’t really look at it after he got back to his lab station.
The other piece of “Rite-in-the-Rain” paper that he used was his daily field log. This was important paper that recorded information regarding each site. It would be split up per site, per person, and describe the conditions/variables, such as what type of bait (N day old fish – N was important to normalize data). This paper was the most important sheet, as it was the only record of weather data and times, locations, etc.
Equipment, Techniques, Day-to-Day Schedule in the Field
They’d wake up at 5:30am, get out by 6:30, work 6-7 hours in the field, and get back to the station. Then, they’d spend 8 or 9 hours mounting, identifying, labelling, and entering data.
In the field (in Costa Rica), he would drive to 6 sites a day. The equipment consisted of bee traps (about 1.5 feet tall) and bee nets (think: butterfly nets). They’d also have field notebooks and a sheet to record their site-to-site information (such as who caught how many bees; what types of bait; time of day; weather; etc). Weather would be logged with handheld weather stations, and then written down. The weather stations didn’t have enough memory to hold all the data.
They would drive to the sites in the morning, set out traps, then come back ~3 hours later, before the rains came at or after noontime. They recorded how long the traps were out, to normalize results of sampling.
They sampled using both traps and timed netting techniques. As a sidenote, timed sampling is a very standard technique in field biology. He used a timed sampling technique for catching bees with nets. He would allot 15 minutes and everyone would catch as many as possible with their net. These numbers would be normalized per person’s average, of course, as some biologists were faster than others.
After they caught a bee, they would put it in a killing jar (~10cm high, 2 cm wide, stuffed with some cotton moistened with ethyl acetate—nail polish remover). These small jars were labeled the night before, with site, date, and who would catch the bee. They’d come back with about 10-12 jars for the day, about 90 bees in total.
At night, they would go through and pin, mount, and label all the bee specimens. They fill out data sheets as they identify the specimens. They use dichotomous keys to do identification… and many times they only get to a genus or subgenus. They used Michener’s key, but had backup keys in case they needed a second opinion. For example, venation might be a character that was ambiguous… but the second key might skip venation altogether. Sometimes, they would ask another of their team to do the identification, to confirm their own beliefs. Bees are not studied as well as many other insects, so many species are unknown. They would finish the labelling and identification that day, so as to never fall behnd.
Bees whose species he was not entirely sure of was marked in the Excel spreadsheet with an asterisk (*). Experience with bees is very important to successful identification. B.CI.2 had to get trained from another expert (grad student also) before he he went on his first expedition in 2003. In 2004 he was more comfortable with identification.
Now, his whole collection is stored in boxes like at the California Academy. Each bee has pinned below it 3-4 labels, printed with a printer from an Excel spreadsheet entry. He has about 2500 specimens, 1600 of which are labelled completely in his Excel spreadsheet.
He is quite an expert with MS Excel, and does most of his analysis in Excel. He uses it to pose questions, and find answers and correlations in his data. He doesn’t think he has a “set way” to process the data… he will organize it based on any questions he has. For example, if he were to wonder, “How does the area of forest fragments affect pollination?” then he would have to rearrange his spreadsheets to allow him to see the answers. I told him about Pat Hanrahan’s Tableau software.
The most powerful feature of Excel in his opinion, is the ability to arrange data using pivot tables. This is the same kind of interaction that Tableau offers for large DBs. Basically, you can drag and drop data labels to different parts of your spreadsheet, and Excel will recalculate and aggregate sums as needed. Read more about pivot tables here: http://www.cpearson.com/excel/pivots.htm
His current work has 1600 entries, and can be managed well with Excel’s pivot tables. He does not use a database. He sometimes uses Jump (a Mac software for stat analysis).
His files are pretty organized, and he backs up his data. He will have multiple versions of spreadsheet files. (INQUIRE FURTHER)
We spent some time talking about research tools. He uses Endnote, which he says is useful, but is still limiting. (INQUIRE FURTHER) He mentioned that he wished Steve Jobs would come out with some killer research app that was exclusive to Apple. He said researchers would switch to Mac if this happened. I agreed, and said they could call the software, iGraduate.
He said he’d like some software that could keep track of your research process, and help you remember why you made some choices, why you chose some references over others, etc.
He plans to do more work with the specimens, possibly identifying them further, by asking specialists. Right now, many of his specimens are only identified to subgenus…. He will also do pollen analysis, to count the pollen, and identify them. As his research question is to find out how pollination is affected by different variables, he will have to go through the laborious task of counting and identifying pollen grains.
To extract pollen grains, he plans to sonicate them, putting them in test tube or centrifuge tubes, and pulsating the bath with sound waves, to knock out the pollen without destroying the bees. He will then count the grains with visual inspection of images. He can also use a stereoscope if necessary. He will use dichotomous keys of pollen to help classify them…. He would be extremely grateful if someone would make a pollen counting and classification system for him. J I told him this was entirely possible (to some degree), and referred him to Andrew Ng. For pollen, he’d also like to sort them on size, color, other characters… in addition to counts. He also mentioned paleo-ecology, of studying the past… (SEE AUDIO)
We also spoke briefly about his experience at the New York Botanical Gardens. He also mentioned the Cybertracker project, which B.CI.0 also mentioned, where they gave PDAs with pictorial software to illiterate Africans villagers to identify animal species.
He has not done much collaboration in terms of data sharing, etc. He worked in Costa Rica with 2 assistants, and 1 hired helper, who did not participate in the identification of bees.
The Audio Transcript (and my Interpretations)