I have read (quickly) the paper. I see nothing there that addresses the idea that the stripes might be simply an identifier the flies use. Rather, the paper seems focused on the effect the stripes have on the flies rather than the idea I suggested. For example: "If, as our results indicate, stripes that span the range of interspecific stripe width variation repel flies to the same extent, adaptive explanations for stripe width variation would have to pertain to some other selective pressure."

The use of "would have to" completely ignores the potential of stripes to be an identifying feature. If my hypothesis were correct, there would not have to be any other selective pressure; since all stripes identify "zebra," any stripes would work.

For example, to test my hypothesis, it might be sufficient to check whether, in those zebras where some of the hide is tan rather than striped, the tan portion of the hide is attacked by the flies disproportionately to the striped portion of the hide.

In sum: the paper reflects on the mechanism of the stripes in repelling flies, and seems to disregard the possibility that the stripes simply allow the flies to identify "not a good target."