American Rose Annual 51: 150, 155-160 (1966)
The Old Problem of Species in Rosa With Special Reference to North America
Dr. Eileen W. Erlanson Macfarlane
Columbus, Ohio

*Cancer Research Associate (Cytologist), Pathology Department,
The Ohio State University Medical Center.

IN THE excellent 1965 edition of Modern Roses 6, Dr. G. D. Rowley (Il) with the help of Dr. Peter S. Green, has provided an alphabetical list of the rose species of the world which is very comprehensive. It includes a great number of names printed in italics because they refer to plants that are now recognized to be identical with previously described species. They are therefore no longer used by botanists and are called synonyms (= same name).

Dr. Rowley has also included the synonyms for each of the now recognized species after the official Latin name. Because of the variability of rose species many minor forms have been given specific status and are described in the literature, some from dried herbarium specimens only; this was done especially in the period from 1880 to 1920, and caused great confusion. The task of sorting out these so-called species is a formidable one. Dr. Rowley followed his own decisions and those of other botanists and made a great contribution by sinking so many names in synonymy.

In Europe much of the chaos was reduced by the work of the great Belgian taxonomist Dr. George A. Boulenger, who, after retirement from the British Museum, worked at the Jardin Botanique Royale at Brussels, on the large Rosa herbarium of his countryman Francois Crepin, as well as extensively in the field. His conclusions, with hundreds of synonyms, were published in Les Roses d'Europe (1). Soon after I started to study the American rose species I visited European herbaria and had the pleasure of spending some time with this distinguished rhodologist. He was among the first to stress floral characters and phenology (flowering periods) for distinguishing rose species, noting that they should be considered before the minor things like prickles, hairs, glands and foliage.

When I blithely undertook to revise the North American roses at the urgent request of Charles C. Deam (then the State Forester and Botanist of Indiana), the old professor who was Chairman of the Botany Department of the University of Michigan, protested that it was an impossible task and feared that I would give up in despair. Dr. Boulenger told me that Francois Crepin studied the rose species of the world for thirty years "and then went mad." Certainly he was never able to publish any revision of the Genus Rosa.

1) Dr. J. Horace McFarland requested an outline of my conclusions for the 1932 American Rose Annual. Instead of completing a revision, I went into other fields of cytology and genetics, first of all among the people of India. In the mid-century Dr. Walter Lewis (7) took up this study, using methods similar to mine in order to revise the American species. He tackled them one by one and completed monographs on R. acicularis, R. foliolosa and R. setigera before he went abroad. He published a monograph on the subgenus Hesperhodos in 1965 (R. minutifolia and R. stellata, rare species of the southwest) which I have not included in this article.
    I was very fortunate in the care that my rose plants received, and still receive, at University of Michigan (8) where I had at one time up to 10,000 plants of known origin. Dr. Lewis took plants of many of the inter-specific hybrids which I had produced (9) to Blandy Farm, Boyce, Virginia, but the help available there was inadequate. Wild roses easily become choked by volunteer tree seedlings, they spread by suckers and cultures become badly mixed. My collection of about a thousand plants for progeny study at the experimental garden of the California Institute of Technology, was disbanded after the mature plants had been studied and herbarium specimens made.

2) The Caninae or dog roses have a unique and peculiar system of breeding which shows them to be hybrids. This was revealed by microscopic study of their cytology and chromosomes. Although they are polyploids (4n or 5n mostly), the pollen always carries only n or 7 chromosomes and can therefore be used to pollinate diploids successfully. The balance of the chromosomes are in the ovule (egg), and for this reason the dog roses are called unbalanced polyploids. A successful cross was made by me (4, 8, and 9) with pollen from hexaploid (6n) R. nutkana on a tetraploid (4n) dog rose R. mollis. Both parents contributed 21 chromosomes and the one large F1 hybrid plant raised was hexaploid (6n). Hybridizers please take special note.

Dr. Boulenger and I both agreed that there are actually a relatively few species of Rosa, but that several of them vary enormously and interbreed in the field. Professor H. H. Bartlett, as chairman of my doctorate committee, was hopeful that the new techniques of experimental taxonomy and cytogenetics would solve the problem. We started immediately to transplant wild roses from the field to the Botanical Gardens of the University of Michigan, of which he was the director. We wrote to botanists and foresters over the country requesting plants. I also travelled to collect seeds and plants and a large living collection was built up which eventually covered an acre or more of ground. After nearly ten years devoted mostly to this study, and after many field trips across the American continent, including visits to herbaria, I published my conclusions (5) as "experimental data for a revision of the North American Roses" and supplied an analytical key for the identification of 27 species in the section Cinnamomae of Rosa, in which I included the Carolinae species (footnote 1).

Chromosome Numbers

After Täckholm's report in 1922 we had the help of chromosome studies (Cytology) in Rosa. The characteristic number of chromosomes in each cell of a species is a multiple of 7 called "n" but varies from 2 x n (14) in diploids to 8 x n (56) in octoploid species. In Modern Roses 6 Dr. Rowley has included in his list the vernacular names, the country of origin and the chromosome number for each species. Now that rose breeders are aware of the importance of the chromosome number of the parent plants for successful hybridization this record of the normal chromosome number of each species will be most helpful. Unusual and abnormal chromosome numbers which have been reported in the literature, are due to hybridization in nature and to cytological accidents (2, 4). These are special cases and can occur in any biological species, even Homo sapiens (man) himself (footnote 2).

What is a Species

The modern concept of a species in plants and animals goes back two hundred years to the Swedish biologist Carolus Linnaeus. The best definitions of a Linnean species are perhaps those of: (1) Vavilov (1931) "A separate morpho-physiological system connected in its genesis with a definite environment and area"; (2) Babcock (1930) "A natural group possessing relative stability, combined with a definite tendency to vary." Everyone knows that a species is supposed to breed true from seed, generation after generation.

Among American rose species there are two types: (A) Easily delimited Linnean type species, which are quickly recognized. In North America these have a circumscribed area of distribution: e.g., R. setigera, R. foliolosa, R. nitida, also introduced species such as R. laevigata and R. bracteata; (B) Highly variable, called polymorphic species, collective species or species complexes, which range over thousands of miles. They exhibit parallel variations, that is: tall and dwarf; armed and unarmed; pubescent and naked foliage; no glands or glandular leaves; globose or pear-shaped hips; characters that can be found in all of them, which indicates a high degree of genetic similarity among them. Many of the minor variants have been given specific names. They hybridize in the field, a fact that was not verified when I started my study (2)

This free crossing in areas where the ranges and flowering times overlap, naturally causes another maze of variant types, as well as pollen and seed sterility.

Those of you who are interested in the local native roses can soon recognize the few main species of the region. Then you will notice unusual individuals or clumps (clones), which are interesting for special features. They may be natural hybrids or "forms" or sports (mutants) of a species. It has been suggested they could be identified by a series of letters or numbers to distinguish their minor characters, a sort of formula (5). They can be given varietal names if distinct enough.

From the beginning I noticed that the phenology of wild roses was helpful in distinguishing some species. That means the date of the opening of first flowers and the time taken for the hips to ripen. There are some features of the flower itself which are specific (10) and help in identification such as:

  1. Number of stamens and their position after flowering.
  2. Length of stamen filaments (whether all about equal or varying).
  3. Texture of ripe hip and time from flowering to ripening of fruit.
  4. The nutlets: number, size, color, whether hairy or not.
  5. Pollen grain size (microscopic).

Boulenger also stressed the importance of the width of the disc at the base of the stamens and the width of the orifice through which the styles and stigmas emerge. These characters of the flower have been largely ignored by rhodologists, chiefly because they are difficult or impossible to study on dried herbarium specimens. But modern classification of higher plants is based fundamentally on the reproductive organs. Dr. Lewis (7) emphasized the importance of studies in the field and experimental gardens. He undertook a painstaking statistical study on dried herbarium material of eight quantitative characters and also recorded 16 minor qualitative characters. The only floral character among the five that I have listed above, that he studied, was pollen grain size.

Pollen Size

Like chromosome number, pollen grain size is a microscopic characteristic and it varies with the former, increasing with chromosome number. It is fairly easy to determine and I found that it could be done on dried material (3). It should be possible for any rose breeder to do this with an ordinary student microscope. Before a crossing is made the pollen should always be examined microscopically (6) in order to be sure that the grains are not shrunken, empty and sterile, which is the case in many individual plants of Rosa setigera (The Prairie Rose), as well as in hybrid plants both wild and cultivated.

The North American Species in 1964

3) In my opinion R. rudiuscula is a natural hybrid between R. arkansana and R. carolina. It is a definite taxon (entity) and can be designated R. X rudiuscula. It was re-created by me at Ann Arbor from the parent species. I also recognized four other species of hybrid ancestry in this geographic area (R. X naiadum, X R. obovata, X R. michiganensis, and X R. schuettiana). Also four "ecotype species" of the ten main eastern species. These are distinct forms of the species each associated with a particular habitat, as follows: of R. arkansana: R. alcea (a dwarf of the northern great plains, probably called R. suffulta by Dr. Rowley); also R. housei, an eastern swamp form: of R. blanda: R. williamsii on limestone, Bic., Quebec, only; of R. woodsii: R. macounii (Greene) of dry sub-alpine plains.

After reading about this difficult problem it will probably be a surprise to hear that east of the Rocky Mountains, Dr. Lewis and I both agree that there are only eleven main native species of Rosa; or ten and one natural hybrid (footnote 3) The North American rose species east of the Rockies are:

Diploid (2n = 14) -6
R. blanda, R. foliolosa, R. nitida, R. palustris, R. woodsii (Section — Cinnamomeae) and R. setigera (Section — synstylae)

Tetraploid (4n = 28) -4
R. arkansana, R. carolina, R. rudiuscula and R. virginiana

Hexaploid (6n = 42) -1
R. acicularis

Dr. Lewis has not published his conclusions for the western United States.

The following are my own.

North American Species of Cinnamomeae west of the Great Plains

The following six species were included in my final list and key.

Diploid (2n = 14) -2
R. gymnocarpa, R. pisocarpa

Tetraploid (2n = 28) -3
R. californica, R. durandii, (3)R. spithamea

4) R. acicularis is also in this region and hybridizes with R. nutkana in the Rockies to give hybrid species R. X engelmannii. Other hybrid species will doubtless be recognized when this area is studied more intensively. I was only able to identify three ecotype species in the west: (1) of R. gymnocarpaR. calvaria of the southern Sierra Nevada; (2) of R. durandii Crepin — R. yainacensis, a dwarf form of dry habitats in California (synonym R. pinetorum); (3) of R. nutkana — R. manca, a dwarf form of the southern Rocky Mountain region.
     In my conclusion I included R. rugosa (2n) as a native of North America. Seeds were brought by Dr. Alfred Warthin from a wild bush in Alaska. In addition there are three native species of the interesting Minutifoliae Section: R. minutifolia, R. mirifica and R. stellata. The type localities of all of these I was able to visit. Dr. Boulenger removed this section from the genus Rosa altogether.

Hexaploid — 1 (footnote 4)
R. nutkana

All are polymorphic species.

With these relatively few species in the section Cinnamomeae it is surprising that a revision of the North American species is such a formidable task, but it most certainly is. In this section Dr. Rowley listed 31 species (including R. rugosa as an escape, and R. rudiuscula) for North America and omitted R. durandii Crepin (even from synonymy); whereas I recognized only 16 main species including R. durandii, plus six species of hybrid origin for a total of 22 species. Of these six species of hybrid origin Dr. Rowley does not list three at all, even as synonyms and gives one of them full specific rank (R. rudiuscula). Of the other two, he calls one (R. engelmannii) a variety of R. acicularis, one of the parent species and another (R. naiadum) a synonym of one of its parent species. This gives an idea of the difficulties that Dr. Lewis, or someone else, must resolve before the long awaited revision can be made. As one who worked on this problem for ten years and then "chickened out," I wish him, or her, all success.

  1. Boulenger, G. A., 1924-32. Les roses d'Europe. Brussels.
  2. Erlanson, Eileen W., 1929, Cytological conditions and evidences for hybridity in North American wild roses, Bot. Gazette 87: 443-506.
  3. ———————— 1931 A group of tetraploid roses in central Oregon. Bot. Gazette 91: 55-64.
  4. ———————— 1933. Chromosome pairing, structural hybridity and fragments in Rosa. Bot. Gazette 94: 551-566.
  5. ———————— 1934. Experimental data for a revision of the North American wild roses. Bot. Gazette 96: 197-259.
  6. ———————— 1934. Pollen analysis for rose breeders. American Rose Annual 1934, 19: 63-68.
  7. Lewis, W. H., 1957. Revision of the genus Rosa in eastern North America: A review. American Rose Annual 1957, 42: 116-126.
  8. Macfarlane, Eileen W. E., 1957. American rose species and their hybrids at the Botanical Gardens of the University of Michigan. American Rose Annual 1957, 42: 90-98.
  9. ———————— 1961. Report on hybrids of American wild rose species at the Botanical Gardens at Ann Arbor, Michigan. American Rose Annual, 1961, 46: 107-16.
  10. ———————— 1963. A self-pollination mechanism and other items in rose species. American Rose Annual, 1963, 48: 188-93.
  11. Rowley, G. D., 1965. Rosa in Modern Roses 6.