Another thing to consider is whether we may reasonably assume that two diploid species could give rise to a new tetraploid species under natural conditions. Natural hybrids are known in Rosa, of course, but these seem to be rare and short-lived. At least, we do not see new tetraploid species spring up wherever the ranges of two diploid species overlap.
But something different may occur during times of changing climate. It is known, for example, that a mature specimen of a species may survive in a region where its seedlings have little or no chance of surviving to maturity. And when the climate is changing, a lone specimen may hang on as another species moves into the region.
I recently found a paper by Nakamura et al. [Molecular based evidence for a lack of gene-flow between Rosa x hybrida and wild Rosa species in Japan, Plant Biotechnology 28, 245–250 (2011)].
In the present study we analyzed the seedlings of seeds from three wild native > Rosa > species (> R. multiflora > Thunb., > R. luciae > Rochebr. et Franch. ex Crép. and > R. rugosa > Thunb.) selected from several locations across Japan where the wild rose was growing in close proximity to cultivated rose plants (> Rosa > x > hybrida> ). To determine whether gene flow from cultivated rose had occurred, young leaves of 1,296 seedlings from the wild > Rosa > plants were analyzed by PCR for the presence of the KSN locus. This locus originated from a sport of > R. chinensis > Jacq. var. > spontanea > (Rehd. et Wils.) Yu et Ku and is involved in the recurrent flowering phenotype observed for cultivated rose hybrids, but is absent in Japanese species roses. The KSN locus was absent in all seedlings sampled, indicating no gene flow to wild > Rosa > species from the cultivated rose had occurred, and providing evidence that the probability of gene flow from cultivated to wild > Rosa > species in Japan is low or non-existent.
https://www.jstage.jst.go.jp/article/plantbiotechnology/28/2/28_10.1217a/_pdf
That seems definite enough, but I was reminded of something else.
The Garden p. 136 (Feb. 16, 1878)
Rosa polyantha.—Mr. George Paul having told us of the beauty of this Rose as a vigorous climber in Mr. J. Sisley’s garden at Lyons, we, in reply to an inquiry, received the following note from that veteran amateur, who has enriched our gardens with so many good varieties of plants:—
“As to the principal object of your letter, I can say but very little. I have possessed > Rosa polyantha > since it has been introduced from Japan. It is a very hardy shrub, having stood with me the severe winter of 1870, when we had, in December, 24° Centigrade below freezing point. The flowers are white, small, and single, but they are borne in clusters, like the Lilac. It seeds freely. It needs no pruning; quite the contrary. I have obtained from its seeds several varieties with double flowers, which are about the form and size of those of > R. multiflora> ; they are generally white, but some are tainted yellowish or pink. I must suppose that the mother plant has been fertilised, through insect agency, with the pollen of other Roses, because my son, who is an engineer in Japan, wrote to me that in the neighbourhood where he resides, where the > Rosa polyantha > grows wild on the hills, it always comes true from seed, but that may happen because there are no other Roses in the neighbourhood.”
Clearly, the Japanese form of R. multiflora sometimes received pollen from garden roses when a lone specimen was growing in a French garden, but it is much more likely to be pollinated by other specimens of its own species while growing in the wilds of its native Japan.
The same is likely to have occurred during ancient times of climate change, whether of warming or cooling. An old specimen clinging to life could receive or share pollen with an invading species. The hybrid could exceed either parent — perhaps gaining tolerance of local soil conditions from the old one, and tolerance of the new climate from the other. Occasional tetraploid offspring from the diploid hybrid could then continue on where neither of its parents were fully adapted.
Hurst (1932) touched on this subject:
Mechanism of Creative Evolution (1932) pp. 112-113
C. C. Hurst
Several external factors may be the cause or occasion of this duplication of chromosomes. A severe frost, for example, will temporarily suspend divisions in the pollen grains and cause some germ-cells to form (provided the frost has not been too severe), bearing twice the normal number. These on fertilising normal egg-cells will produce triploids or, if the egg-cells have been similarly affected, tetraploids. In gardens where many plants are out of their natural environment, having come from countries with more regular climates, such occurrences are not infrequent. In their own country, once the winter is over, they produce their flowers with no set-backs, but in England, where we get a warm spell in the spring long enough to bring out the flowers, often followed by a severe frost for several nights just as the flowers are forming, many aberrations and abnormalities arise. Similarly in a wild state, an unusual season may upset many of the normal mechanisms of the plant. De Mol has discovered that many of the different chromosome types in bulbous plants have been due to the custom of drying off the bulbs after the flowering season and, in the case of bulbs for commerce, the subsequent forcing to produce early-flowering for culture in pots. These bulbs, being subjected to various degrees of temperature at the time of the formation of their germ-cells (these being formed deep down in the bud during the previous summer or autumn before flowering), produce various irregularities of division. Many of these irregularities fail to carry on but the few that struggle through will give rise to new races, sometimes of great beauty, for our gardens.
)