Years ago, I asked a friend to let me make some basic black-and-white maps for a book he was writing. The subject of the book was New Brunswick, the coastal province in Canada right across the border from Maine. The maps, naturally, included lots coastlines, islands and inlets. One of the basic design questions for the maps was whether to show the land white and the water gray, or the land gray and the water white.

People have been making monochrome maps of coastal areas for centuries, so I expected there would be some clear consensus – noted in the textbooks or in journal articles – over how to shade the land and water. I was wrong.

In the absence of a clear instruction from the cartographic literature, I decided to conduct a quick test. So I launched a small online survey asking people whether they associated light colors (or dark colors) with land or water.

The first question presented the following image and asked respondents whether it was a body of water surrounded by land or an island surrounded by water.

Then the survey presented the following image and asked the same question.

80% say this is an island surrounded by water. Again, they see white as land and dark as water.

Finally, the survey presented two versions of the same actual map, identical, except that the shading of land and water was switched, and simply asked which version they preferred:

62 percent of respondents prefer map A. That is, they want to see water rendered dark and land rendered light. Overall, the results show a consistent, but modest, preference for land to be rendered as lighter than water. These results do not constitute statistical proof; notes on a possible bias in the sample follow.

These results match well with some parts of traditional cartographic theory, but challenge others.

The theoretical arguments for different ways of portraying land and water are generally based on the old-time artists’ distinction between figure and ground. In general, in a work of art, figure is the thing the artist wants you to really notice and ground is everything else. For a quick illustration, imagine the Mona Lisa. You can picture it, right? You can recall her face, her enigmatic smile, right? Now try to recall what’s behind her in the painting. If you don’t know, that means that the face functions as figure and the scenery as ground. The other property of figure and ground is that figure seems to pop off the page (or canvas) at the viewer, but ground retreats away from the viewer.

Both of these ways of thinking about figure and ground also pertain to map-making. A long-standing theme among map makers is that light colors function as figure and dark colors as ground. Our eyes are drawn to light colors first, and light colors seem to pop off the page at us more than dark colors. So let’s think about how this applies to making maps of land and water.

If we believe that figure is the thing that pops off the page at the viewer and ground retreats, then land in a map should be treated as figure and sea as ground. Land does, after all, have a higher elevation than the water, so it is closer to the map reader than the sea. By that logic, land should always be lighter than water.

But if we believe that the most important thing on the map (and not necessarily the thing with the highest elevation) should be the figure, then that important thing should be represented in the lighter color. That, of course, only raises the question of what the most important part of a map is. For a nautical chart, for example, water is the important part and land is just the border. In contrast, for a road map, the land is important and the water is just a place where you can’t drive your car. By this logic, a map that’s telling a story about the water should make the water light and a map that’s telling a story about the land should make the land light.

Sadly, that logic doesn’t help when neither the land nor the water are clearly the most important part of the map. It also raises a possibility that people from coastal or maritime communities might have different ideas about what is important than do people from landlocked communities. At the time I did this survey, I was teaching in Kansas. Many of the survey respondents were students and faculty I knew, most of whom lived or worked in the American Midwest or Southwest. So there is a possibility that the survey results are biased toward people who grew up thinking land was essential and water was just a place for occasional recreation. That’s why the survey contained a question about the place the respondent considers his or her home. With more survey data, it might be possible to make a distinction between preferences between inland people and maritime people.