Plant architecture

Quite often there are considerations here about things one can sum as plant architecture ¶.

Rose PA is obviously very important and little studied.

Recurence and length of shoots according to position are PA components as well as growing or insertion angle and branching distribution.

Any thoughts or observations ???

I think you need to further articulate what you are searching for in us since this is a new thought.

I am not exactly sure of what you are asking but I do consider “branching habit” as part of my breeding parent selection process. For example, I find Baby Love far superior in this respect than, say, Brandy or Lynn Anderson. Also, I tend to choose against something that is overly twiggy. I find this common in a lot of miniatures and seems to stop air circulation within the plant habit which does not help with disease resistance at all.

Hi Pierre,

There are many aspects like you said, but one that has been looked at to some degree is flower yield in cut flower hybrid teas by De Vries (1976) Juvenility in hybrid tea-roses Euphytica 25:321-328. There was a trend that seedlings that flowered earlier continued this trend of having a shorter time from bloom to rebloom. In the end there was little practicality for cut roses because increased flower yield was associated with shorter stems, as expected. Maybe looking over his paper considering impact on plant architecture and rebloom would be interesting. Like you said, branch angle would influence things a lot too. It might be difficult to combine all these traits into a formula by quantifying them and finding out how to weight them and combine them into desired plant architectures. There probably would be a way to look at it all and then try to boil it down and simplify it by looking at the traits with the most impact to efficiently use in a breeding program. It sounds like a good graduate school project.



Hi David,

I am rather speaking of good plant architecture for outdoor roses.

De Vries is at early selection. Cut flower roses are another problem…and are selected for vertical long stems. When cut flower roses were choosen among garden roses and viceversa it had a bad influence on HTs that are mostly rather stiff, narrow and too tall after first flush.

For most actual roses vars bad plant architecture is compensated by obliged expert pruning. Only a few shrubs do not need to be cut back at least once annualy. Such are most rugosas.

In my opinion PA is the most outstanding Knock Out feature. Here is a var that do not need more than a little shaping and downsizing.

Pierre Rutten

Hi Pierre, True, De Vries hybrid tea work offers limited information for the type of PA you (and I too) would like. It’s funny that what he selected against for stem length is probably what we will select for.

Good point about Knock Out. I think that one factor that can influence plant architecture is ploidy. Many of the great landscape roses are triploid (Knock Out, Flower Carpet Pink, Nearly Wild, Sunrise Sunset, even the floribunda Iceberg). This isn’t to say that 2x or 4x roses cannot have a relatively dense, floriferous, plant with attractive architecture, but I think it’s more common among triploids to find the architecture you are describing and what I’m trying to breed for as well. Tetraploids (referencing it to similar genetic background of course) generally have less branching, larger flowers with more substance and larger stems. At least the tetraploids I’ve generated from diploids also have a slower growth rate than the diploids they were derrived from. I believe in general triploidy is a good balance between the general benefits of each ploidy level and can be a useful tool to try to generate roses with good plant architecture to serve as landscape shrubs.

What do others think?


David, I find it interesting that you state triploids will generally have better architecture than diploids or tetra ploids. I have a few plants that I germinated from R.blanda seeds a few years ago. These plants are shorter and bushier than their mother. They also have some characturistics of R.arkansana. I suspect that they are natural hybrids, in which case then they should be triploids.

If what you are saying is correct, if I try crossing these with a tetraploid, the very characturistics (shorter and bushier) that I want to incorporate will likely disapear. Thats a little disheartening. They are nice shapely plants.

I agree with Pierre, I’m looking out my window at some HT that have been defoliated by black spot and they are just tall sticks sticking out of the ground now. It is one of my goals to create plants that are astheticly pleasing for landscape use.


Hi Paul, Your R. blanda hybrids sound great. There are multiple factors I think that contribute to plant part size and overall architecture. I think ploidy is just one factor that I think we can try to work with. I trust there are also heritable genes that can express themselves at different ploidy levels as well (maybe to different degrees at each ploidy due to dosage??) that we can work with too. Maybe your R. blanda hybrids have a nice combination of alleles from both parents contributing to the plant habit you enjoy and hopefully some offspring will get that combination and express what you like when you use them as parents no mater what ploidy level your suspected triploids provide the next generation.

Through all the chromosome counts I’ve done, I think that it is more than just a coincidence that the very popular landscape roses are primarily 3x (many of the Meidiland roses are as well). Maybe as breeders we can try to generate triploid cultivars to help meet our objectives.

Are others trying to generate 3x offspring for this purpose?


I think, with regards to architecture, I would like to see more cultivars with attractive weeping forms along the lines of R. xanthina. Some form is lost to the rigidity we require in a rose stem. I think many of us rebel against too much emphasis on big individual blooms and strong long stems. Sure, I like bushes with impact from a distance, but I want one that also invites up-close inspection. I don’t want something that looks great cut on my table to look hideous in the garden. If it’s a landscape shrub, short flower stems may even serve to greater advantage.

To PA, I might add “texture” as a feature to consider. I personally enjoy the finer textured small foliage of some species too. I seldom enjoy oversized leaves on a rose regardless of how “well-proportioned” it is to the bloom.

Twiggy, I agree, is ugly, and is one of my biggest beefs about chinas.

Not exactly the observations Pierre was soliciting – more like opinions on taste…

David. As you do I consider many aimed at landscape triploids PA to be more desirable. These roses were bred for as little maintenance as possible. Good architecture is a must then as is the strength and better health they got from near species parents. And wichuraiana may contribute.

However I see no link between diploidy and better PA among species. I would rather point exclusive selection for the (cut) flower and stem forgiving the plant for actual tetraploid cvs.

By the way did you count Baby Love and Pretty Lady chromosomes. Both behave also as triploids with limited fertility.

Philip. Foliage texture is definitely a PA component. I like xanthina ferny foliage a lot. Another different foliage texture is that of clinophylla with small leaflets and comparatively large flowers as much as I can infere from the bracteata x clinophylla seedlings I grow.

Pierre Rutten

Baby Love has highly potent pollen. It is a horrible seed parent, though. Pretty Lady is near identical in it’s reproductive qualities.

Hi Pierre,

I counted Baby Love, but I do not have ‘Pretty Lady’. BL is tetraploid. Strangely, I counted offspring of Baby Love as a female by tetraploid males that a friend raised and some are triploid. This is not that unusual in my findings. Perhaps lack of homology and pairing leave some chromosomes behind in genetically diverse tetraploids and can lead to periodic 1x and aneuploid gametes. Sometimes aneuploid zygotes can “correct” themselves and get rid of extra chromosomes or duplicate some of the missing ones. Maybe there is more difficulty for Baby Love to go through gamete formation on the female side than the male side. ‘Out of Yesteryear’ is a good example of a 3x offspring from two tetraploid parents. It’s one parent that traces back to R. bracteata, ‘Muriel’, and is 4x and the other is the 4x mini Sequoia Gold.

I agree that the generalities of plant architecture and ploidy may be hard to see in species. Some diploid species have relatively large flowers and thicker stems like R. rugosa compared to 4x R. arkansana and so on. However, when I compare my 2x and induced 4x rugosas, setigera, and polyanthas the trend is there. There are definately other factors like general genes and genetics like you pointed to in the case of breeders selecting for large, strong blooms and stems often without much concern for plant architecture. The trend of ploidy and plant architecture is relative to species and genetic background and then the variability there needs to be addressed and compared to the “optimum” plant architecture range we are trying to breed for.

Good point that many of the triploid landscape roses are pretty close to diploid species and especially very disease resistant ones like R. wichuriana.


Hi David,

I had assumed the shape and texture of the R.blanda hybrids were because of the combination of the two different genomes from each parent. The plants are intermediate in form from both parents. Even though there are twice as many chromosomes from R.arkansana you can still see the R.blanda influence in them. Like you say there could be any number of factors at work here, like dominant/recessive or additive genes. The plants produced hips this year and I was speculating that a cross with a sibling might produce a tetraploid with


The french team we know got among others a pentaploid rugosa. Very, very thick and dark green not oversized stems and leaves.

Landscape triploids ally diploid qualities that are better resistance, strength and architecture with recurence and flower qualities of the centuries long selected tetraploids.

Among species Rosa rugosa, even if quite stiff, has a very interesting PA with a very decorative foliage and naturally round shaped plant. Not twiggy at all!


When backcrosses are usually unpromising, intercrossing your seedlings can bring more variation as genomes will be broken. Quite varied and even unexpected characteristics may appear. At least it is something I observe with species hybrids sib and halfsib crosses.

Pierre Rutten

RE: the triploid landscape shrubs; this may be a atatement of ignorance here – I am relatively inexperienced – but I would assume that one reason that so many good landscape shrubs are triploid is quite simple. Triploids are often more sterile, ergo self-cleaning and quicker to rebloom.

I have noiced the progeny of many good landscapers as well, but due to their probable lack of fertility have not bothered to attempt acquiring and breeding with them. I have assumed that this same infertility was that which made them good recurrent shrub roses.

They are typically closer to their healthy species parents too, which I presume offers some advantages.

Thanks Pierre,

My biology is really rusty, but what I think what you are describing is transgressive segregation. It will be interesting see if I get the same kind results as you did.

I also was thinking along the same lines as Philip. Sterile plants don

Good points to think about Paul. Fushia Meidiland is triploid. It sounds like from Ping Lim and the Baileys breeding program that purposely generating 3x offspring is not a strong conscious goal. I would be interested in learning if it has been a conscious goal for the Meidiland roses. The Flower Carpet series are mainly 3x, but some 4x too. I suspect ‘Knock Out’, its sports, and Rainbow KO being 3x was not a conscious effort of Bill Radlers considering all the 4x parents in the pedigree (at least of the roses other than Rainbow KO, the code names of the parents are listed, but I cannot find information on them). I hope to count Double KO to see what ploidy it is since it is a full sib of KO. I sure agree with you Paul that reduced female fertility especially in many 3x roses over 4x is nice feature for rebloom. Some of my 2x poly and some 4x shrubs set so many hips they don’t rebloom well. David

About landscape roses there are tetraploids ones as well as triploids. When they began trying at groundcover breeding hybridizers rediscovered an old var that fitted perfectly: it is The Fairy. Wider than tall, full, recurent lasting flower and resistant to all deseases but the “other” spot. Even if a not very cooperating mother I have seen loaded with tags at Meilland and it is prevalent in early Kordess achievements.

Doesn’t Leen Leus write that desease resistance has to be found in species? Most landscape roses are along this line with a notable exception: Scriven’s Baby Love and Pretty Lady that have desease resistance derived from the tetraploid cinnamomae davidii elongata. Notable is that there are also tetrapoid ones promoted for long beautifull autumn hips display.

That: “It is well documented that anytime you cross a hybrid with a species that the species shows through more strongly than the hybrid.” I have not found true in diploid species crossed with tetraploid hybrid vars progenies. I did found the full range of intermediates. From close to species to close to i.e. HT. Quite close if the progeny is large enough. Somewhat along this line are Iceberg and many later introduced Kordess vars.

About Paul’s words: “species or near species does not equate to being disease resistant.” I fully agree with a strong remark.

Five years ago I had quite a number of varied crosses with foliolosa as the mother planted side by side in my rose field. About fifty meters long double row. Came first autumn they all got the heaviest rust attack I ever saw when other seedlings were no more deseased than usual. The foliolosa seedlings were red to the top and there was a red powder rain when touching them. None was less deseased than the others…

…And they never more got rust excep the usual light case here on some lower leaves!

That is why I advise not to early select for desease resistance.

Pierre Rutten

The Fairy is not quite that disease resistant. It is quite susceptible to mildew here.

I had made a previous thread about shrublets that exceeded The Fairy (Carefree Marvel and White Drift)in all aspects. Also, they set seed well enough.

Jadae I was writing about history of landscape rose breeding The Fairy introduced 1932 was mainly used for Groundcovers breeding from about 1970 to 1990 when Carefree Marvel a probable triploid no more fertile than the former was introduced 2000 and White Drift introduced 2005 in the US only.

I checked out a few of the roses by Meilland on Help Me Find Roses. I didn