design it

"... research is never completed ... Around the corner lurks another possibility of interview, another book to read, a courthouse to explore, a document to verify." ~ Catherine Drinker Bowen, U.S. biographer

After my literature search, several of my questions were still unanswered.

What is the mussel distribution at Nahant?

What percentage of those are juveniles?

Is there a different distribution for juveniles and adults?

What I did learn is that mussels spawn (have babies) in the summer months, and that by the time late fall/early winter comes around, most juveniles have already settled into a permanent mussel bed and are too big to move around. Fall semester is not a good time to do a distribution study of adult and juvenile mussels. But, I could still look at the distribution and size of mussels, and even survivorship, since winter brings a lot of harsh storms to Nahant, and there’s always predation going on. Ok, so now I have an idea that I’d like to investigate. How do I go about doing that?

Once a research question or objective is decided on, the real work begins. There are three major steps to conducting an experiment: Planning and Design, Fieldwork, and Analysis. This is the real meat of science, the whole process of trial and error, calculated guesses, failed results, confusing data, going back to the drawing board and starting all over again until meaningful results come through. This is the saga of my experiment.

Designing an experiment is one of the most creatively demanding tasks out there. There are many different types of experiments, and each requires that you take into consideration lots of different parameters. I was interested in studying the distribution of mussels, which meant I needed to do a population survey. A population survey is just a fancy way of saying "counting how many individuals there are in this area." Ecologists have come up with many ways to figure out how many individuals make up a population. I decided to use one of the most basic approaches, quadrat samples. A quadrat is a marked off area of known size that can be used to select a sub-set of an entire population. Quadrats can be large (several square miles) or small (a few square millimeters). It depends on the area you're sampling. In our class, we use 0.5m² (half square meter) quadrats and 0.25m² (quarter square meter) quadrats. A quadrat sampling system works by taking many small, random samples of a population in a larger area. If you know how much area your quadrat covers, you can figure out things like organism density, (for example, how many individuals there are per square meter). If you know the size of your quadrat and the size of your larger sampling area, you can figure out rather accurately the total number of individuals in your sampling area.

However, I not only wanted to look at density, but I was interested in looking at the size of mussels and survivorship as well. This posed a couple of interesting problems. Normally to measure density and size, I would lay down a quadrat, physically collect all the mussels I found in my quadrat, count them, measure them with calipers, and then release them. However, to look at survivorship, I would want to leave the mussel beds intact and be able to find them again so I could compare the number of mussels across time. When I presented this problem to one of my professors, Dr. Robertson, she suggested I try photoquadrats. I was referred to a paper by Jon Witman (1987) which details his construction and use of an underwater camera designed to take high quality pictures of quadrats. This was great! I could record mussel beds without disturbing them, and I would have photo-evidence to compare them later on. But how would I measure the size of mussels? Camilo Khatchikian, our TA, introduced me to some open-source software (TPS-DIG) that would allow me to not only count the mussels in my photos, but also measure them. I was all set, and I set out to construct my own method for taking terrestrial photoquadrats.

When it comes to design, I typically follow the KISS method: Keep It Simple Stupid. My photoquadrats were nothing more than a photo (using a digital camera) of a quadrat (picture at right: reflection of me taking a photo of a quadrat). With Camilo's help, I experimented in the lab with our two quadrat sizes, and decided to use the smaller one, 0.25m², because I could focus the camera on it better, and ended up with clearer pictures. For reference points in the picture, I super-glued a 10cm ruler to one edge of the quadrat, and marked off half centimeters with a permanent marker on another edge.

My next challenge came in figuring out how I was going to re-locate my quadrats. It was going to be critical that I be able to take pictures of the exact same mussel beds three weeks apart in order to look at survivorship. Camilo and superglue came to the rescue again. Camilo suggested I use labeling tape to number and mark my quadrats. I could punch out the quadrat number, and superglue the tape to the rocky substrate in a specific orientation to my quadrat (under the upper right hand corner). This way I would even be able to re-orient my quadrat into the exact same position three weeks apart.

The last element of my experimental design was to help me investigate the question, were there more mussels in the upper intertidal, or the lower intertidal? This meant I had to figure out a way to determine the "upper" intertidal and "lower" intertidal. To do that I decided to measure tidal height, or the vertical change in height moving up the shoreline. Fortunately, we already had a method worked out to do this. Standing at the low water mark with a laser sight and a vertical pole that had every half meter marked off, I could start at the substrate and move the laser sight up until it hit the height of my quadrat, and then take that reading from the pole.

I was all set. The plan was simple: I was going to go into the field, pick a random vertical transect (a line to place my quadrats on), then pick random sampling locations a meter to two meters apart starting at the top of the ascophyllum zone and working toward the water line. At each sampling location, I would separate the ascophyllum down to the substrate and place my quadrat, regardless of how many mussels there were. I'd take a few pictures of each quadrat, to make sure I got at least one good one, and then I'd mark the upper right hand corner of the quadrat with a piece of labeling tape glued to the rock. When I got to the bottom of my transect I would use the laser sight to determine the height of all of my quadrats. Finally, I would repeat this process for as many transects as I had time for (the more transects I could get the more data I would have to look at). All that was left was to put the plan into action ... but you know what they say about the best-laid plans.

The Cycle of Science

home ask questions read design it do it look at it ask more questions about me