Last year we started an experimental project using our chickens to help us plant crop seeds. The idea came from the observation that sometimes seeds from spilled chicken feed would germinate in the pasture a week or two after the chickens moved on to a new spot. So we’ve been broadcasting seeds as we move the chickens to fresh grass, allowing them to trample the seeds into the soil surface and then to fertilize the planting by dropping manure all over the ground. Some fraction of the seeds inevitably get eaten by the birds, but when the grass is thick enough, plenty of the seeds manage to escape their beady-eyed attention.

It is interesting that even after trying to add new varieties to the mix (peas and sunflowers), I ended up circling back to the three seeds that I ended last year with: oats, daikon radishes, and purple top turnips. I really thought that peas would be a good addition because peas and oats have long been a favorite companion planting combination. But it seems that peas get off to a good start but they are inhibited later in life by competition from the perennial plants in the pasture. So after a promising start, they all stalled out, becoming pale, and spindly. And the sunflowers? Well, they never germinated at all. I have seen a few sunflowers from the chicken feed escape from the chickens in the past, but this year either we had bad seeds or the chickens were really good at finding all the seeds and eating them.

I wonder at the inexact alignment between “learning” and “knowledge” in farming. After two years of trial and error, I know a few things about planting seeds with chickens. But really, I only know what happened in one pasture in certain specific weather conditions with specific batches of seeds and within a particular poultry management system. This year we had a dry spring followed by an overwhelmingly wet summer. If next year were to experience a reversal with a wet spring and a dry summer, would my seed selection still be optimal? Perhaps not. I am nowhere close to having generalized knowledge based on a large set of data. I’m just one little point of experimentation here, out in the fringe of farming. It seems that there are endless opportunities for learning, but knowledge is always conditional, grading away quickly from its solid core to its gauzy outer edges.