You are given data that consist of two measurements made at each of several different depths. When depth (or altitude) is a variable, most scientists plot their data so that depth (or altitude) is vertical, and so that depth increases downward (or altitude increases upward). In this way, the viewer sees the data in a way that mimics the real-world position of the samples: the deeper ones are lower.
When scientists want to compare two measurements, as we do here, they commonly plot the two measurements on the same diagram but with different symbols. You can do that here, with an open symbol for the clear bottle and a dark symbol for the dark bottle (pretty clever, eh?).
You aren't ask to plot the data in Part 2, but note how the data are plotted. The variables are years and numbers of fish caught. When one of the variables is time, it's conventional to make time the horizontal axis, with earlier to the left and later to the right. Think how silly the plots would look if time were on the vertical axis.
You are given data that seemingly consist of three measurements made at each of several locations, each of which is designated with a letter. However, there are essentially just two measurements, because % suspension feeders and % deposit feeders total to nearly 100 in each case. Furthermore, the letters designating the locations presumably have no ecological significance, so we needn't treat them as data.
With that said, we're now down to essentially two variables: % silt & clay and % suspension feeders and deposit feeders. With no spatial variable (as in Part 1) and no temporal variable (as in Part 2), the default convention in the sciences is to plot the independent or controlling variable on the horizontal axis. That's what should be done here.
Putting the independent or controlling variable on the horizontal axis requires a judgment on our part: - which is the controlling variable? For example, if you were given data on class attendance by students and those students' grades, you would assume (we hope) that attendance would be the variable that at least partly determines grades, and so you would make attendance the horizontal axis.
In our case, with data on % silt & clay and % suspension feeders and deposit feeders, we have to ask "Would the kind of seafloor determine the kind of organisms, or would the kind of organisms determine the kind of seafloor?". If you've been to lecture lately, you'll know the most likely general answer.