Do carrots really love tomatoes? Do beans and onions hate one another? The internet (and my bookshelf) is full of anecdotal advice about which crops we should plant together, and which ones we should not.

There’s a well-known book that’s been around for ages called Carrots Love Tomatoes: Secrets of Companion Planting for Successful Gardening, by Louise Riotte. It offers page after page of “facts” about companion planting. That sounds really helpful, and I was ready to try it all, but unfortunately, when I dug in online, I discovered that there is very little science behind the advice.

Some of it is folklore. Often there is a good reason for traditions in gardening, but folklore can be wrong as well as helpful. We can’t know until we test it. Some of it is hearsay. Again, that might be right, but maybe Aunt Alma’s carrots would have done well anyway, and they just happened to be planted near her tomatoes.

According to a paper by Cornell University, most of the advice in this book is based on studies done in the 1930s by a Dr. Ehrenfried E. Pfeiffer. He took plant juices, mixed them with salts (sodium sulfate or copper chloride), and left the solutions to evaporate. He then used chromatography to look at the resulting crystals left behind. Based on whether the chromatograms were clear or cloudy, Dr. Pfeiffer declared some plants mutually beneficial and others antagonistic. It’s not at all obvious what clear vs. cloudy crystals have to do with plant compatibility. Moreover, his conclusions were never tested in the field. As the Cornell article states, “No legitimate scientist believes that this method can determine compatibility among plant species.”

The big problem is that so little legitimate research has been done on this topic. Agrochemical companies see little benefit to funding it, and it isn’t applicable to large-scale monocrop farms. (Any horticulture majors wanting a topic for their PhD thesis?)

Little by little, we’re learning what really works. In spite of its often shaky underpinnings, the idea of companion planting is far from useless. Some plants really do resent having neighbors, while others do better planted near each other. There are a number of excellent reasons why this is so.

There are several ways in which plants might benefit or harm one another. Roots give off chemicals. Some of these prevent other plants’ roots from growing nearby, which is why not much will grow under a Black Walnut tree. Some communicate warnings, perhaps of attacking insects or impending drought. (See my post from last March, “What Did You Say?”)

It makes sense to situate heavy feeders—plants which do better with extra nitrogen in the soil—next to legumes, which are able to sequester nitrogen from the air and add it to the soil. (Crop rotation already takes advantage of this, by planting beans in a spot where the previous crop might have depleted the soil.)

Beneficial insects (lacewings, wasps, etc.) are attracted to specific kinds of flowers, usually those with flat tops such as carrots and cilantro (right). Of course, you’d need to let those plants bloom, which means you won’t be eating them. Other plants attract harmful pests—to the extent that they can be used as a “trap crop”—one that distracts the insects from the more important plants. And just combining different kinds of plants can help foil pests, as a diverse planting makes their favorite meal harder to find.

Marigolds really do repel parasitic nematodes. For example the roots of African cultivars give off the chemical thiopene, which repels harmful nematodes. However, only some kinds of marigold produce this effect, and you have to grow the marigolds as a cover crop, then incorporate them into the soil to gain any benefit. Planting marigolds at the ends of your beds might make your garden look pretty, but they won’t protect your plants. (Happily, we don’t have those nasty nematodes in Colorado.)

Then there are those whose shapes fit nicely together. For example, beets and radishes have relatively small, low-growing tops that fit nicely under tomatoes, which are usually quite bare at the bottom. And we’ve already mentioned plants that repel pests, and those that fix nitrogen in the soil.

Until the research catches up, perhaps the best idea is to simply try an experiment in your own garden. If you want to be scientific about it, include a control. That is, plant two crops together, and then plant each one separately, and see if there’s a difference. Try to keep all the other factors equal—same amount of sun, same amount of water, etc. Let me know what you discover, and we’ll post it here.