American Rose Annual (1956) 41:123-125
New Approach to the Breeding of Hardy Roses
Dr. F. L. Skinner, Dropmore, Manitoba

"Rosa virginiana was one of the American rose species that I used many years ago with rather indifferent results; however in the spring of 1950 I had in bloom, in pots, plants of R. damascena Celsiana, R. d. rubrotincta and a double white form of R. alba as well as a plant of R. virginiana alba. The flowers of the latter were fertilized with the pollen of the three old roses and from the seed secured about twenty seedlings germinated in 1952. Two of these seedlings that flowered for the first time this year had double flowers, one had white flowers resembling R. alba while the other had clear pink flowers like Celsiana in form and only slightly smaller; the foliage of all these R. virginiana alba hybrids is clean looking and nice. Pollen from these hybrids was used this summer on some of the old roses with apparently satisfactory results."

When fusion hybrids of human and mouse cells are maintained as tissue cultures, the human chromosomes are soon lost. However, if a strain of mouse cells with a genetic defect is used, the human chromosome carrying the missing "gene" will be retained.

Modern advances in DNA analysis have revealed that hybrids do not always receive full sets of chromosomes or genes from both parents. In some cases, only a small group of genes from the pollen parent are retained in most of the offspring (e.g. hybrids of diploid Helianthus species). In other cases, neither parent contributes a full set of genes to the hybrids (e.g. Aegilops species hybrids).

I have noticed several cases where hybrids selectively retain paternal "genes" that correct a genetic fault in the seed parent, much as occurs in human-mouse cybrids.

1) Millardet (1894) crossed the White Four-Season strawberry
  (a white-fruited cultivated form of Fragaria vesca) with pollen of the "Chili velu" (a cultivated form of F. chiloensis). One of the four seedlings resembled the seed parent exclusively, except that it bore red fruit and was pollen sterile.

2) East (1934) crossed a white-fruited Fragaria vesca x F. virginiana. One seedling duplicated the seed parent almost entirely, and was diploid like that species. It bore red fruit (like Millardet's), and appeared to be pollen sterile because the anthers failed to dehisce. However, when the anthers were forced open, the pollen proved to be fertile. East studied the odd hybrid for three years before noting that central leaflet was dentate only halfway to the petiolule — like the octoploid pollen parent, but unlike any known form of F. vesca.

3) East (1930) used a double-hybrid F. vesca (heterozygous for both pink/white flowers and red/white fruit). There was a distinct deficiency in white-fruited seedlings from the crosses.

4) Van Tubergen (1906) Crossed Lilium martagon var. album x L. Hansoni. Several hundred of the hybrids looked like ordinary purple Martagons. None bore white flowers. Other plants were more obviously hybrid, though in varying degrees.

5) To this brief list we now may add Skinner's work with Rosa virginiana alba. Apparently the normal form of R. virginiana was less successful as a parent than it's sport. It may be that the genetic weakness (i.e., albinism) of the variety allowed the hybrid cells to retain more of the paternal inheritance.Rosa alba, by the way, is a naturally white-flowered form rather than a "mutation".
     It may not always be necessary to use a "genetically defective" seed parent. There is some evidence that subjecting a seed parent to adverse conditions (transplanting shortly before flowering, withholding water, exposing to brighter light or higher temperatures than the plant prefers) can increase the proportion of paternal inheritance. Michurin used such techniques. More recently, Christie Hensler has raised some remarkable Iris hybrids in similar fashion.
     Partial hybridization is wide-spread. Wherever it occurs it may be possible to influence the inheritance of the progeny, to some degree, by choosing seed parents that have more reason to retain some paternal influence, or by forcing the seed parent to endure conditions the pollen parent favors.

6) Christie Hensler also informs us that Norris finally succeeded in crossing Iris dichotoma with Belamcanda chinensis (now called Iris domestica, of all things) by using a crossbred Belamcanda (normal orange crossed with mutant yellow) as pollen parent.

7) Burbank (1914): " usually seems to make no practical difference whether you take pollen from flower A to fertilize flower B, or pollen from flower B to fertilize flower A.
     This observation, which was first made by the early hybridizers of plants more than a century ago,—notably by Kölreuter and by Von Gartner,— fully confirmed by my observations on many hundreds of species. Nevertheless it occasionally happens that the plant experimenter gains some advantage by using one cross rather than the other. In the present case it seemed that by using the Lawton as the pollenizing flower, and growing berries on the brownish white species, a race was produced with a more pronounced tendency to vary."

8) Halstead (1903): The author assumed (incorrectly) that the pollen parent determines the color of the kernels. In crossing two sweet corns, Black Mexican and Country Gentleman, he saved only the white kernels on Black Mexican and only the black kernels on Country Gentleman. As it happens, Black Mexican carries a transposon. When the transposon is silenced (the normal condition in the strain) the kernels are black but the silks, stamens and glumes are light green. When the transposon is not silenced (a common occurrence when the strain is crossed with another), the distribution reverts to its (presumed) normal condition: white kernels and pigmented silks, stamens and glumes. In this case, the progeny from BM x CG (white kernels only) were much more variable than those from the reciprocal cross (black kernels only). It would be interesting to repeat the cross and to raise the black kernels separately from the white. It may be that the "genomic shock" that provokes the un-silencing of the transposon is also responsible for the increased variability.

9) Report of the Third International Conference 1906 on Genetics p. 87.
Specimens Exhibited by Dr. John H. Wilson, St Andrews University, Scotland.
4. Digitalis lutea x D. purpurea, D. lutea x D. purpurea alba, and D. purpurea alba x D. lutea. These specimens showed that the hybrids having the white foxglove as a parent were more vigorous than those having the purple foxglove, and further that the reciprocal crosses between the white foxglove and D. lutea differed in respect of form and colour of the flower.