I do a lot of selection in my plant breeding for traits that I don't even know that I am selecting for. Here's a couple of examples:
Last summer, a collaborator gave me some maxima squash that she had grown. Thinking that they might get added to my landrace. Two of the three varieties were so hard that I couldn't even cut them open! One had a woody skin. The other was just hard fleshed in general. I hadn't realized that over the years I have been selecting for soft fruits that are easy to handle in the kitchen.
Today, the germination test for Solanum habrochaites finished. It germinated at 100%, and was the quickest to germinate of any of my tomato varieties... That was startling to me, because I have a note in my seed catalog that germination on S habrochaites is erratic, and that germination can be expected in flushes every few weeks. Ha! I suppose that during the 4 generations that it has grown on my farm, that I inadvertently selected for quick germination.
What sorts of inadvertent selection have you noticed with your seed saving?
There are a couple of major genes in cucurbits associated with the woody skin. These can literally have a very hard wood layer that actually blunts your knife if you try to pound it into the squash. They are best opened by dropping them on the driveway from the right heights so that they break cleanly in two. Most of the varieties we grow don't have the woody skin, because grocery store shoppers don't know how to open them. Those hard shelled varieties can be very nice to have if you have a problem with rodents eating your squash as they finish maturing in the field.
Automatic selection in squash seems to be for bigger seed cavities, thinner flesh, and smaller seeds but more of them. I select every fruit to counter these in every generation. I think this is the reason why so many commercial varieties start out good and then deteriorate year by year after released.
If you save seed from volunteers you can select for seed dormancy mechanisms. At least, where the seed represents seed you planted in some earlier year that did not come up that year.
Usually, where we save seed only from the patch we planted that year, we automatically select against seed dormancy mechanisms.
Usually, wild species are outcrossers and often have incompatibility mechanisms. But the domesticated derivatives have often lost incompatibility mechanisms. So we seem to often automatically select for inbreeding tendency when we domesticate plants. My guess is that what happens is that inbreeders are more likely to present us with new phenotypes based upon homozygous recessive mutations, and we tend to save and preserve and play with anything different. Such new recessive mutations have a much harder time presenting us with the new homozygous phenotype when in obligate outbreeding material.
I think people often times accidentally select for slow germination and wimpy growth by letting transplants overgrow so much that those that grew fastest become severely pot bound and are so badly damaged in in transplanting that the plants are rogued in the field.
I think you can also select for slow germination and wimpy growth by planting frost-tender plants too early. Then you get a hard frost, and all those that came up first are killed. And your entire plot is derived from plants that had not yet emerged when the frost hit.
I routinely get automatic selection against bush and semi-bush squash. I sow about 3X to 10X as much seed as needed so as to select for the fastest growing plants. This pretty effectively eliminates all homozygous and heterozygoust bush plants, since they don't develop root systems as big as full vine plants, and the plants just don't grow as fast. If I actually want any bush plants from segregating material, I have to take special precautions not to eliminate them automatically.
