And writing of regulatory mechanism, I got to thinking about Batesian mimicry in butterflies. One excellent example is Papilio dardanus, an African species that can mimic at least 11 distinct model species. Within a given population, however, it only manages (or needs) three forms.
The various forms of mimic (female only) involve multiple pigments in various patterns. And yet, the mimic-pattern is inherited as a “unit character”, or nearly. As I recall, about 8 or 9% of females are imperfect in their mimicry. This is not an altogether bad thing, because this break-down allows the raw materials for building a new mimicry, when required.
This sort of modular control is found in other areas. And by a rather nifty coincidence I have been observing a break-down in regulatory control in a tall bearded iris.
The typical form of this variety is a very ordinary two-tone, with light blue-violet standards with darker and bearded falls.
On April 24, while walking my dog in the park, I saw a freak flower with six horizontal, bearded falls. I have seen pictures of this sort of flower (e.g. ‘Clementina’), but had never seen one in person. The next day I saw two more of these odd flowers, but on a different plant of the same variety. Clearly these were not the result of a mutation.
Looking more closely, and taking more pictures, I saw that some of the later blooms were imperfectly modified. Some of the organs that should have been standards were only partially bearded, and the tepal was streaked with darker pigment over the lighter.
Photographing one yesterday I saw a stamen sticking out where it should not be. Looking closer, I pulled up a style-arm and found three more stamens where there should have been only one.
http://bulbnrose.x10.mx/Heredity/FreakyFlowers/Freaky_flowers.html
It appears that the initial function of this regulatory network was to distinguish the inner tepals from the outer. This provided a basis for further differentiation: form and disposition (standard or fall), beard, colors, etc.
The relevance, here, is that some aspects of species distinction are in large clumps that tend to hand together. Hurst discussed these groups of traits at great length, but his critics thought he was being silly for not spending more time considering such pressing matters as the quantity and distribution of pubescence on the underside of the leaves. /s
I am suggesting that these large sets of traits are analogous to the mimicry patterns of the African butterfly, and to the regulatory control of standard/fall differentiation in the tall bearded iris. The contrary position would be to assume that all Rosa species are merely collections of coincidentally associated unit characters.