In the first part of this series, I discussed examples of asymmetry—both directional asymmetry (right-versus left-handedness) and anti-symmetry (differences between sides, but in a random direction)—and raised the problem of how directional asymmetry, like the enlarged left tusk of the male narwhal or the higher left ear of the barn owl, could evolve. In other words, how would a gene know whether it was on the left side of an organism or the right?
In yesterday’s installment, I discussed two recent pieces of research, based on the directional movement of cilia and the asymmetrical operation of proteins, that could produce a directional gradient in a bilaterally symmetrical organism and thus lead to the evolution of handedness in a trait.
The potential difficulty of a gene somehow “knowing” it was on the right versus left side of an organism that’s bilaterally symmetrical led me to this question:
One question that occupied me when I was younger was this: if you take an organism that is, by and large, bilaterally symmetrical, like Drosophila (though there is a bit of handedness in a couple of its traits), could you impose artificial selection on it to produce handedness? That is, could you select for a line of flies whose right eyes were bigger than their left, or who had more bristles on their left side than on their right (and vice versa in both cases)? How hard would that be? Given the absence of marked bilateral asymmetries in species like Drosophila that could act as developmental cues for the successful selection of directional asymmetry, you might think it would be hard—even though virtually every other trait in Drosophila can be successfully changed by artificial selection. Tomorrow we’ll learn the answer to my question.
Today I’ll give the answer, which is that yes, it’s very hard to select for directional asymmetry in organisms. Drosophila, for example, are pretty bilaterally symmetrical, though there are some slight differences in morphology on the right versus left sides that are directional. I’d predict, then, that it would be hard to select for a line of flies that would be directionally different: say, one that had the right eye always bigger than the left, or had more bristles on the left side than on the right. Or, at least harder to do that than to select for other traits that are obviously variable in populations, like a simple increase in the number of bristles, or the ability to move more toward the light than the dark.
In fact, of all the artificial selection experiments I know about in Drosophila, the only ones that have ever failed are those selecting for directional asymmetry. You can increase antisymmetry by selection fairly easily: that is, you can make flies more asymmetrical for traits like eye size and bristle number, but not directionally so.
The paper by Ashley Carter et al. given at the bottom gives a history of selection experiments in Drosophila for directional asymmetry, and then adds a new experiment (this was published in 2009). Here are the experiments preceding that of Carter et al.; they are a dismal history of failures:
- In 1960, Maynard Smith and Sondhi published an experiment in which they tried to select for directional difference in the number of ocellar bristles (there are usually three, one anterior and two posterior, both on top of the head). They showed no significant increase in the directionality.
- In 1965, Beardmore selected for directional asymmetry in the number of sternopleural bristles, which are located on the sides of the fly (below). The two lines that were selected in opposite directions (more bristles on right vs. more bristles on left) diverged very slightly but significantly over 50 generations (a long time for selection). The selection response, however, was tiny compared to other experiments that selected simply for an increase or decrease in total bristle number regardless of the side.

- In 1973, Purnell and Thompson selected for directional asymmetry of wing folding. A given fly always folds its left wing over its right, or vice versa. But different flies have different folding directions, so this is a case of antisymmetry. These workers selected for a line of flies in which left folded over right consistently, and vice versa. While the authors claimed a modest directional response, Carter et al. say that there was in fact no difference achieved in either line.
- In 1987, I published a paper in which I placed the eyeless mutant of Drosophila into a genetically variable line. This mutant makes the eyes very small, and often asymmetrical. I then selected for lines having the left eye bigger than the right and vice versa. As a check on the general presence of genetic variation, I also selected for reduced eyes on both sides of the head. The last experiment succeeded greatly, showing there was genetic variation for eye size, but both experiments selecting for directionality failed: there was no sign, after 30 generations, that I had produced a line with eyes consistently bigger on one versus the other side of the head. Here’s some of the variation in the expression of the eyeless mutation:

- Finally, in 1990 Tuinstra et al. selected for a directional difference between left and right bristle numbers on the scutellum in a line containing a mutation that destabilizes bristle number. (There are usually four scutellar bristles, as shown below). After 12 generations they saw no response to selection for directionality, but the line was also depauperate in general genetic variation.

In the experiment of Carter et al., the authors selected for directionality in the distance between the posterior crossvein of each wing from the tip, trying to create lines in which the distance was larger on the left than on the right—and vice versa. Here’s a diagram of the trait they selected for: the average of the distances from the intersection of the posterior crossvein with the longitudinal veins on either side to the tip of the wing; in other words, the average of two distances shown in the dark black lines in the bottom figure.

Carter et al. practiced 15 generations of selection for right wing distances bigger than left, and in the opposite direction as well. There was no significant response. Again, selection for directional asymmetry (which could have reflected either a difference in either the vein configuration or the general size of the wing) failed.
So out of these six experiments for directional asymmetry in Drosophila, only one succeeded, and that was a very modest success.
Why the failure? The authors suggest three possibilities:
1). There is no left-right axis of asymmetry that could allow genes to cue on whether they’re on the left or right; therefore there could be no variant genes that could produce directional asymmetry. The authors reject this hypothesis because flies do show some directional asymmetry in their guts, genitals, and a very small amount in wing size.
2). The amount of genetic variation that exists for directional asymmetry—that is, genes that can recognize what side of the body they’re on given that some slight directional asymmetry already exists—is small. Given this lack of available genetic variation, selection would often be unsuccessful. This would be my favored hypothesis given the pervasive bilateral symmetry in Drosophila. There just aren’t many asymmetries for genes to cue in on.
3). It’s easier to evolve directional asymmetry if the trait is initially antisymmetrical, with one side different from the other, though in a random direction. This would already show that the trait is sensitive to developmental differences on the sides (though not directionally)—and perhaps that sensitivity could be leveraged into a directional response. This seems unlikely to me (and the authors) because at least two experiments (mine and Tuinstra’s) created big antisymmetry by destabilizing a trait via mutation. Despite that increase in antisymmetry, selection for directional asymmetry was unsuccessful in both cases.
This lack of success is in stark contrast to the pervasive success of other selection experiments in flies. I know of no failures of selection for other traits that don’t involve directional asymmetry, though of course some unsuccessful selection experiments might not have been published.
To conclude, then, we do see directional asymmetries in some organisms, so selection has, in some species, picked out genes that distinguish right from left. But in organisms like Drosophila that are bilaterally symmetrical on the whole, it’s hard to produce flies that are right-handed or left-handed for some traits. Carter et al., and I, suggest that this is because there are very few genetic variants that can take advantage of the very slight directional asymmetries that pre-exist in Drosophila. This, in turn, would suggest that such selection might be more successful in species having more pronounced directional asymmetries to begin with—organisms like humans. But of course we can’t select on our own species, and even in mammals the long generation time (and lack of interest in directional asymmetry!) makes such experiments impractical. Nevertheless, I still find directional asymmetry fascinating, simply because I want to know how a gene can tell whether it’s on the right versus left side of the body.
______________
Carter, A. J. R., E. Osborne, and D. Houle. 2009. Heritability of directional asymmetry in Drosophila melanogaster . International Journal of Evolutionary Biology 2009:7.











