Propagating Roses
1. Know the ploidy of your roses
The great majority of species roses are diploid (14 chromosomes); the great majority of modern garden roses and European heritage roses are tetraploid (28 chromosomes). Examples of diploid species are R. rugosa, R. multiflora, R. wichuraiana, R. blanda, R. cinnamomea, and R. moschata . Notable exceptions: the tetraploids R. carolina, R. spinosissima , and the species ancestor of the gallicas and damasks; the pentaploids R. canina and its relatives. A significant minority of roses from the later nineteenth century are triploid, as hybridizers crossed tetraploid hybrid teas and hybrid perpetuals to diploid species.
2. Cross roses with the same ploidy numbers
Such crosses are not only more likely to take, but the offspring are far more likely to be fertile enough to act as parents themselves. One of the reasons that David Austin was able to create such a large group of roses in a relatively short amount of time was that he crossed tetraploid old roses to tetraploid modern hybrid teas and floribundas.
Examples: R. multiflora x Mont Blanc [diploid to diploid]; Pristine x Lilian Austin [tetraploid to tetraploid].
3. Use lower ploidy numbers as pod parents, higher as pollen
If no. 2 above doesn't fit your goals, then use tetraploid pollen on blooms of diploid plants: for example, R . wichuraiana x Peace [by convention, the pod parent is listed first]. Long experience of hybridizers over the past century suggests that you shouldn't do it the other way around (say, Peace x Frau Dagmar Hastrup). The latter is unlikely to take.
Furthermore, most diploid species (including all the more familiar ones) are "self-incompatible," meaning they prefer the pollen of other roses over their own -- they self pollinate only if other pollen isn't available. This trait makes crosses to diploids particularly easy.
Exception: see no. 4 for triploid pollen on diploid blooms.
4. Don't be afraid of producing triploids
No. 3 above will usually generate a triploid. These are generally considered sterile, which means it will be harder to follow up with a subsequent generation. However, though "triploids are sterile" is an excellent generalization and completely correct, the trick is in the definition of "sterile." The word suggests an either/or situation: either a rose can produce offspring or it can't. But the rose that is absolutely in the latter state -- or "completely sterile" -- is rare. So, "sterile" actually means "less fertile" than some standard, which could be the average for the variety, the average for the species on which the variety is based, the group, etc. By the standards of the species which are their ultimate ancestors, for example, all hybrid teas are sterile!
In triploids, fertility will always be higher in pollen than pod, simply because pollen is produced in vastly greater numbers, and it takes numbers to produce (in any useful amount) the accidents of reversion to diploid or tetraploid status that makes a triploid "fertile." In addition, some triploids seem to revert more easily, such as Iceberg, which is derived from the diploid (?) hybrid musk Robin Hood, the source of a "spontaneous" jump to tetraploid in Eva, which in turn produced the highly fertile tetraploid Pinocchio, a foundation of the floribunda class.
Examples: R. wichuraiana x ( R. wichuraiana x Peace) [pollen of a triploid placed on a diploid -- we hope for a diploid reversion]; Pinocchio x ( R. wichuraiana x Peace) [pollen of a triploid placed on a tetraploid -- we hope for a tetraploid reversion]
5. Look at parentage
It is said that to see all the possibilities inherent in a particular cross, you need to raise at least 30 plants from it (for tetraploid/tetraploid crosses -- the number is 18 or so for diploids). An easier way is to look at the parents of the roses you intend to cross. Five yellow roses in the ancestry of your pink rose suggests that yellow, pink blends, or peach are likely in the offspring.
6. Don't trust recorded pedigrees
Breeders sometimes short-circuit the information loop by identifying a parent merely as "seedling" (shame on them), but even when pedigrees are given in good faith, they are only as accurate as the breeder's records (and memory). Anybody who's been outdoors on a late spring morning juggling tags, a notebook, and other paraphernalia -- or tried to keep seedlings straight on the plant shelves in late winter -- knows how imperfect the process can be. And -- are you *sure* a small bee or wasp didn't make it under the cover you placed on the bloom after hand-pollinating it? For that matter, are you *sure* only the pollen you wanted was on that brush or that fingertip?
7. Don't assume that a cross was successful just because you get a hip
In modern roses, especially, self-pollination is a possibility unless you've been very careful to remove pistils early. And -- look at no. 6 above for small bees and unreliable fingertips.
8. Don't give up on a variety because one plant doesn't work for you
A variety is really a group, in the sense that bud mutations are known to happen and almost all are far less spectacular than the sports that have entered into commerce over the centuries. If you are using a plant of a variety that is supposed to be a good pod parent but you find yourself disappointed in its performance, get another plant from a different supplier before giving up on that variety.
9. Use species crosses rather than "pure" species
The Canadian breeder Percy Wright made this observation about fifty years ago. If you want to breed with a particular species, either cross it with another species first before going on to crosses with more complex hybrids (such as modern roses) or use a species hybrid. A direct cross to a species is far more likely to produce a plant strongly resembling the species.
Example: R. nitida x Peace is much more likely to produce a plant that strongly resembles the pod parent than is ( R. nitida x R. wichuraiana ) x Peace.
10. Choose promising seed and pollen parents
Hybrid teas, floribundas and miniatures vary greatly in the degree to which they will pollinate other varieties, as well as the degree to which they set seed and those seeds germinate. Finding the best parents can be a long process of trial and error. For that reason, there is a list of Suggested Pod and Pollen Parents for Beginning Hybridizers on this website. The list includes both classic and recent varieties. Don't assume that older varieties have been "bred out" -- the 1999 AARS winner Kaleidoscope has Rainbow's End (introduced 1984) as its pollen parent, for example. Even great classics like Peace, Charlotte Armstrong and Queen Elizabeth still have something to offer when crossed with recent introductions.
11. Have a goal
The genus Rosa is remarkably unstable -- that is to say, most of its 200 or so (?) species interbreed readily and some even vary considerably in form within the group ( R. blanda and R. multiflora are notorious for this latter trait). It's very easy to get lost and quickly overwhelmed by this abundance. In my former life as an amateur rose hybridizer, I worked toward the fabled blue rose, then turned to seeing what I could get out of one variety -- Griffith Buck's Maytime -- by putting pollen of everything available on it. In my current incarnation, I am working toward thornless hardy climbers (without any conspicuous success, to be sure). Another idea is to repeat historical crosses -- something very interesting happened once; it could happen again. Finally, of course, the most important goal is to have fun!
A selected list of rose species and their hybrids according to chromosome number.
Ploidy numbers are taken from Modern Roses 8.
Diploids (14 chromosomes)
*Also see entries under "complex species" below
Max Graf (also Lady Duncan)
Polyanthas (a few were triploid, but the majority are diploid)
R. arvensis (including the form "Ayrshire Rose")
R. banksiae
R. blanda
R. bracteata
R. chinensis (the form of the China Rose)
R. filipes (and the closely related R. brunonii)
R. foliolosa
R. hugonis (Father Hugo rose) R. laevigata (Cherokee Rose)
R. moschata (Musk Rose)
R. multiflora (this includes the Seven Sisters Rose, multiflora nana, cathayensis, and several others; early hybrids were mostly 14 or 21)
R. nitida
R. palustris
R. roxburghii (Chestnut Rose)
R. rugosa (forms; hybrids are 14 or 21)
R. setigera
R. wichuraiana (form; many of the Wichuraiana climbers or ramblers are 21)
R. woodsii
R. xanthina
Synstylae (members of a large group of Asian species: includes R. multiflora, R. filipes, and many others)
Triploids (21 chromosomes)
Hybrid Musks (some of the earliest are triploid, having come from Hybrid Perpetuals or Hybrids crossed with the R. multiflora hybrid Trier; many of the subsequent ones are tetraploid)
R. x borbonia (the original Bourbon rose; most hybrids are tetraploid)
Tetraploids (28 chromosomes)
Dupontii
Floribundas (a few early ones were triploid, but most are tetraploid)
Hybrid Teas (a few early ones were triploid, but most are tetraploid)
R. arkansana
R. carolina
R. centifolia (also its sport, the original Moss Rose)
R. damascena (the type of the Damask Rose; also Autumn Damask, Trigintipetala, and York and Lancaster)
R. davidii
R. foetida (including Austrian Copper, Harrison's Yellow, and Persian Yellow)
R. gallica (type of the Gallica roses; includes Gallica Officinalis and others)
R. kordesii
R. laxa
R. pomifera (Apple Rose; also the form Wolley Dod's Rose)
R. rubrifolia (R. glauca)
R. spinosissima (the type of the Scotch Rose; also its forms altaica, hispida, etc.)
R. suffulta
R. virginiana
Pentaploids (35 chromosomes)
*See R. canina under "complex species"
Hexaploids (42 chromosomes)
R. alba (type of the Alba roses)
R. nutkana
Complex species (forms have different ploidy numbers)
R. acicularis (14, 28, 42, 56)
R. californica (14, 28)
R. canina (Dog Rose; most are 35, but some forms are 42)
R. chinensis (China Rose; most forms are 14, but some forms and species hybrids are 21, 28)
R. cinnamomea (14, 28)
R. eglanteria (Sweet Briar; 35, 42)
R. moyesii (28, 42)
R. odorata (Tea Rose; most are 14, but some forms and hybrids are 21, 28)
The great majority of species roses are diploid (14 chromosomes); the great majority of modern garden roses and European heritage roses are tetraploid (28 chromosomes). Examples of diploid species are R. rugosa, R. multiflora, R. wichuraiana, R. blanda, R. cinnamomea, and R. moschata . Notable exceptions: the tetraploids R. carolina, R. spinosissima , and the species ancestor of the gallicas and damasks; the pentaploids R. canina and its relatives. A significant minority of roses from the later nineteenth century are triploid, as hybridizers crossed tetraploid hybrid teas and hybrid perpetuals to diploid species.
2. Cross roses with the same ploidy numbers
Such crosses are not only more likely to take, but the offspring are far more likely to be fertile enough to act as parents themselves. One of the reasons that David Austin was able to create such a large group of roses in a relatively short amount of time was that he crossed tetraploid old roses to tetraploid modern hybrid teas and floribundas.
Examples: R. multiflora x Mont Blanc [diploid to diploid]; Pristine x Lilian Austin [tetraploid to tetraploid].
3. Use lower ploidy numbers as pod parents, higher as pollen
If no. 2 above doesn't fit your goals, then use tetraploid pollen on blooms of diploid plants: for example, R . wichuraiana x Peace [by convention, the pod parent is listed first]. Long experience of hybridizers over the past century suggests that you shouldn't do it the other way around (say, Peace x Frau Dagmar Hastrup). The latter is unlikely to take.
Furthermore, most diploid species (including all the more familiar ones) are "self-incompatible," meaning they prefer the pollen of other roses over their own -- they self pollinate only if other pollen isn't available. This trait makes crosses to diploids particularly easy.
Exception: see no. 4 for triploid pollen on diploid blooms.
4. Don't be afraid of producing triploids
No. 3 above will usually generate a triploid. These are generally considered sterile, which means it will be harder to follow up with a subsequent generation. However, though "triploids are sterile" is an excellent generalization and completely correct, the trick is in the definition of "sterile." The word suggests an either/or situation: either a rose can produce offspring or it can't. But the rose that is absolutely in the latter state -- or "completely sterile" -- is rare. So, "sterile" actually means "less fertile" than some standard, which could be the average for the variety, the average for the species on which the variety is based, the group, etc. By the standards of the species which are their ultimate ancestors, for example, all hybrid teas are sterile!
In triploids, fertility will always be higher in pollen than pod, simply because pollen is produced in vastly greater numbers, and it takes numbers to produce (in any useful amount) the accidents of reversion to diploid or tetraploid status that makes a triploid "fertile." In addition, some triploids seem to revert more easily, such as Iceberg, which is derived from the diploid (?) hybrid musk Robin Hood, the source of a "spontaneous" jump to tetraploid in Eva, which in turn produced the highly fertile tetraploid Pinocchio, a foundation of the floribunda class.
Examples: R. wichuraiana x ( R. wichuraiana x Peace) [pollen of a triploid placed on a diploid -- we hope for a diploid reversion]; Pinocchio x ( R. wichuraiana x Peace) [pollen of a triploid placed on a tetraploid -- we hope for a tetraploid reversion]
5. Look at parentage
It is said that to see all the possibilities inherent in a particular cross, you need to raise at least 30 plants from it (for tetraploid/tetraploid crosses -- the number is 18 or so for diploids). An easier way is to look at the parents of the roses you intend to cross. Five yellow roses in the ancestry of your pink rose suggests that yellow, pink blends, or peach are likely in the offspring.
6. Don't trust recorded pedigrees
Breeders sometimes short-circuit the information loop by identifying a parent merely as "seedling" (shame on them), but even when pedigrees are given in good faith, they are only as accurate as the breeder's records (and memory). Anybody who's been outdoors on a late spring morning juggling tags, a notebook, and other paraphernalia -- or tried to keep seedlings straight on the plant shelves in late winter -- knows how imperfect the process can be. And -- are you *sure* a small bee or wasp didn't make it under the cover you placed on the bloom after hand-pollinating it? For that matter, are you *sure* only the pollen you wanted was on that brush or that fingertip?
7. Don't assume that a cross was successful just because you get a hip
In modern roses, especially, self-pollination is a possibility unless you've been very careful to remove pistils early. And -- look at no. 6 above for small bees and unreliable fingertips.
8. Don't give up on a variety because one plant doesn't work for you
A variety is really a group, in the sense that bud mutations are known to happen and almost all are far less spectacular than the sports that have entered into commerce over the centuries. If you are using a plant of a variety that is supposed to be a good pod parent but you find yourself disappointed in its performance, get another plant from a different supplier before giving up on that variety.
9. Use species crosses rather than "pure" species
The Canadian breeder Percy Wright made this observation about fifty years ago. If you want to breed with a particular species, either cross it with another species first before going on to crosses with more complex hybrids (such as modern roses) or use a species hybrid. A direct cross to a species is far more likely to produce a plant strongly resembling the species.
Example: R. nitida x Peace is much more likely to produce a plant that strongly resembles the pod parent than is ( R. nitida x R. wichuraiana ) x Peace.
10. Choose promising seed and pollen parents
Hybrid teas, floribundas and miniatures vary greatly in the degree to which they will pollinate other varieties, as well as the degree to which they set seed and those seeds germinate. Finding the best parents can be a long process of trial and error. For that reason, there is a list of Suggested Pod and Pollen Parents for Beginning Hybridizers on this website. The list includes both classic and recent varieties. Don't assume that older varieties have been "bred out" -- the 1999 AARS winner Kaleidoscope has Rainbow's End (introduced 1984) as its pollen parent, for example. Even great classics like Peace, Charlotte Armstrong and Queen Elizabeth still have something to offer when crossed with recent introductions.
11. Have a goal
The genus Rosa is remarkably unstable -- that is to say, most of its 200 or so (?) species interbreed readily and some even vary considerably in form within the group ( R. blanda and R. multiflora are notorious for this latter trait). It's very easy to get lost and quickly overwhelmed by this abundance. In my former life as an amateur rose hybridizer, I worked toward the fabled blue rose, then turned to seeing what I could get out of one variety -- Griffith Buck's Maytime -- by putting pollen of everything available on it. In my current incarnation, I am working toward thornless hardy climbers (without any conspicuous success, to be sure). Another idea is to repeat historical crosses -- something very interesting happened once; it could happen again. Finally, of course, the most important goal is to have fun!
A selected list of rose species and their hybrids according to chromosome number.
Ploidy numbers are taken from Modern Roses 8.
Diploids (14 chromosomes)
*Also see entries under "complex species" below
Max Graf (also Lady Duncan)
Polyanthas (a few were triploid, but the majority are diploid)
R. arvensis (including the form "Ayrshire Rose")
R. banksiae
R. blanda
R. bracteata
R. chinensis (the form of the China Rose)
R. filipes (and the closely related R. brunonii)
R. foliolosa
R. hugonis (Father Hugo rose) R. laevigata (Cherokee Rose)
R. moschata (Musk Rose)
R. multiflora (this includes the Seven Sisters Rose, multiflora nana, cathayensis, and several others; early hybrids were mostly 14 or 21)
R. nitida
R. palustris
R. roxburghii (Chestnut Rose)
R. rugosa (forms; hybrids are 14 or 21)
R. setigera
R. wichuraiana (form; many of the Wichuraiana climbers or ramblers are 21)
R. woodsii
R. xanthina
Synstylae (members of a large group of Asian species: includes R. multiflora, R. filipes, and many others)
Triploids (21 chromosomes)
Hybrid Musks (some of the earliest are triploid, having come from Hybrid Perpetuals or Hybrids crossed with the R. multiflora hybrid Trier; many of the subsequent ones are tetraploid)
R. x borbonia (the original Bourbon rose; most hybrids are tetraploid)
Tetraploids (28 chromosomes)
Dupontii
Floribundas (a few early ones were triploid, but most are tetraploid)
Hybrid Teas (a few early ones were triploid, but most are tetraploid)
R. arkansana
R. carolina
R. centifolia (also its sport, the original Moss Rose)
R. damascena (the type of the Damask Rose; also Autumn Damask, Trigintipetala, and York and Lancaster)
R. davidii
R. foetida (including Austrian Copper, Harrison's Yellow, and Persian Yellow)
R. gallica (type of the Gallica roses; includes Gallica Officinalis and others)
R. kordesii
R. laxa
R. pomifera (Apple Rose; also the form Wolley Dod's Rose)
R. rubrifolia (R. glauca)
R. spinosissima (the type of the Scotch Rose; also its forms altaica, hispida, etc.)
R. suffulta
R. virginiana
Pentaploids (35 chromosomes)
*See R. canina under "complex species"
Hexaploids (42 chromosomes)
R. alba (type of the Alba roses)
R. nutkana
Complex species (forms have different ploidy numbers)
R. acicularis (14, 28, 42, 56)
R. californica (14, 28)
R. canina (Dog Rose; most are 35, but some forms are 42)
R. chinensis (China Rose; most forms are 14, but some forms and species hybrids are 21, 28)
R. cinnamomea (14, 28)
R. eglanteria (Sweet Briar; 35, 42)
R. moyesii (28, 42)
R. odorata (Tea Rose; most are 14, but some forms and hybrids are 21, 28)
