American Rose Species and their Hybrids at the Botanical Gardens of the University of Michigan
Dr. Eileen W. Erlanson Macfarlane Dr. Macfarlane is an associate research professor with the Institutun, Divi Thornae. Foundation, a scientific research center in Cincinnati, Ohio.
The first comprehensive experimental study of the native North American rose species was undertaken by me at the request of Mr. Charles C. Deam, Indiana state forester for many years. Mr. Deam wanted to know the rose species in Indiana and he found that the experts did not agree on their identifica. tion. Enthusiastic support and valuable guidance were given to the idea by Professor Harley H. Bartlett, director of the Botanical Gardens of the University of Michigan at Ann Arbor. Building the Living Collection and Methods of Study Rose plants were assembled at Ann Arbor from wild localities all over North America with the help of 150 botanists, foresters, students and friends to whom requests were made. Collections of seeds were also obtained from individual wild plants and from rose species in the botanical gardens of Europe.
Each collection of one or more plants or seeds was given a serial accession number and was catalogued upon arrival; for the garden is actually a living museum. Each accession was also given a card in the file under Rosa, bearing the accession number, where it was collected, by whom, on what date, also the species identification if known. In the 10 years from 1922 to 1932, 1400 accessions were made by the staff of the garden under the personal supervision of Dr. Frieda Cobb Blanchard, the assistant director.
Each accession was also given a permanent zinc label and the collection was then planted outside, or in pots in the greenhouse or cold frame (tender plants and seeds) under the care of the resident horticultural superintendent, first Mr. A. P. Wezel, then Mr. Walter Kleinschmidt and their corps of gardeners.
At first the plants were grown in narrow beds in a lawn but it was later found to be more convenient to grow them in rows in a cultivated field as is usual in nurseries.
Detailed studies of species characteristics were made on the growing plants of: stems, foliage, flowers, fruit and fertility; the reproductive organs were processed for microscopic study of the heredity-bearing chromosomes. Each fall I made extensive field trips, usually alone, to observe the plants in their natural habitats and to collect roots and seeds from across the United States. Herbarium collections of wild roses were examined at American universities and also in London and Brussels in order to study the dried specimens of famous botanists and rhodologists including Linneus, Banks, Crépin, Greene, Jepson, Rehder, Rydberg and Watson. When seeds were taken from wild bushes, dried specimens of the parent plant were made for the herbarium at the University of Michigan Museums. Eventually, dried specimens were prepared from the seedling offspring at the botanical gardens for comparison.
Some of the many ways in which the living specimens were investigated in the hope of helping Mr. Deam, and others with the same problem in wild rose identification, are briefly described below.
1. Pollen fertility was carefully investigated and was found to be variable, sometimes even among the offspring of a single wild bush of such common species as Rosa blanda, R. californica and R. Woodsii.(7) Rosa setigera, the Prairie Rose which is often used by hybridizers, frequently showed only aborted pollen - especially those bushes which set a lot of good fruits (hips). This surprising discovery was published Originally in the American Rose Annual in 1934.8 Another extensive study of rose pollen fertility was made by Dr. Walter S. Flory, Jr., in Texas in 1950,11 but little is known of ovule (egg) fertility.
2. Apomixis, or production of seeds without pollination was tested for each year on both American species and Euro-' pean Dog Roses. My results with the latter were always negative. They were confirmed by Dr. H. D. Wulff in the American Rose Annual of 1955.16 Only once was apornictic fruit obtained, -from a R. blanda with 53% sterile pollen.7
3. Microscopic studies of chromosome numbers and behavior were extensive and have been published in technical journals. (3,7 )
4. Phenology, or the seasonal appearance of flowers, was carefully recorded and was found to be helpful in distinguishing many species.(4,7)
5. Vitamin B and C content of hips was investigated in some preliminary work by Dr. Felix Gustafson of the University of Michigan in 1950 but not published.
6. Hybridization was started as soon as a few different species became well-established in the gardens. Seeds from 128 successful crossings were set up for germination between 1926 and 1931. The methods used were almost identical to those described by Roy Shepherd in the 1956 annual.15 The purposes of this work were mainly to:
(a) discover whether some wild "species" were actually spontaneous hybrids between other wild species. This was verified.(6,7)
(b) to throw light on the evolution of the geographical races and of our species, especially of the chromosome numbers which vary from 14 to 56 in cell nuclei (5,10.)
A diagram of several of my interspecific hybrids to show the relative fertility of their pollen was published in the 1937 annual.9. Thirty-seven of these interspecific hybrids involving 17 species are still alive and have been listed alphabetically by seed parent species in Table I in the appendix to this article, together with the maternal chromosome number and the species of the pollen parents.
Female Fertility and Germination
Recently a notebook was found with unpublished data on the number of flowers pollinated, hips and plump seeds (achenes) which set and percentage of germination for several of my inter-specific crossings. At the kind invitation of Dr. William Crocker, director of the Boyce Thompson Institute for plant Research, many of the batches of hybrid seeds were sent to him for germination. They were kept in peat moss in the refrigerator for two or three years and the tiny seedlings were returned to Ann Arbor as they appeared. Dr. Crocker described his rose seed germination methods in the 1926 annual. This is another example of the wonderful help and cooperation I received from many botanists.
Articles in the 1956 annual show that germination is a problem for rose breeders. In my own experience, in 1929 good seeds (achenes) which sank in water were obtained from 54 interspecific crossings although only 25 were finally accessioned because garden numbers were not assigned until after germination. In 1930 only 15 lots of seedlings were accessioned from 30 crossings. In 1931, 13 lots of hybrid seedlings were accessioned from 16 crossings that I made in California.
Data on hip and seed setting and on germination for 29 interspecific hybridizations are given in Table II in the appendix They show a great and apparently unpredictable variation from five to 85 percent germination.
California Wild Rose Garden
Dr. Thomas Hunt Morgan, Nobel Prize winner and head of the Department of Biology at the California, Institute of Technology, became interested in the rose species investigations through Professor Bartlett, and generously made space and facilities available for me, at Pasadena. I made an extended rose collecting trip in the western U. S. in 1928-29, and seedlings of species that are not hardy in Michigan were sent to California to be raised in the experimental gardens under the Supervision of Dr. E. G. Anderson. These cultures were studied in detail by the methods described above in 1930-31 when was a visiting National Research Fellow at Cal. Tech.
This species collection was later discarded, but seeds from some of the crossings were germinated at the Boyce Thompson Institute and the seedlings of a few are still alive in Ann Arbor.
The wild rose garden at Michigan was seldom visited from 1934-41 or from 1943 until after the war in 1946. It had been maintained in good order and over 1000 plants were living. With the help of old maps of plantings and the permanent labels, I was able to pull out invading suckers and check identification of cultures. The collection has since been visited each season and a few specimens and data collected; labels renewed; invaders, duplicates and sickly plants discarded.
Renewed Interest in American Rose Species
A few years ago Mr. Walter Lewis of British Columbia took up the study of the rose species of the northern Rockies and wrote his master's thesis on his observations. Mr. Lewis came east to study with Dr. Walter S. Flory, Jr. at the Blandy Experimental Farm of the University of Virginia, an authority on rose species."
In 1954 some plants of species and interspecific hybrids from the collection at Ann Arbor were donated by me to the Park of Roses in Columbus; a second set of plants was sent to Blandy Farm. The following summers other large batches to plants at Ann Arbor were prepared for transportation, and in 1955, Dr. Flory and Mr. Lewis visited the botanical gardens and personally took the plants to Blandy.
After the visit Dr. Flory wrote a very gracious tribute saying, "The botanical garden is to be congratulated on maintaining the plants, and the plant labels in such excellent condition. This is certainly one of the best collections of rose species and species hybrids in the world. I am sure this collection is not only a decided asset to the university botanists, but since it is so carefully labeled - is also a source of instruction, interest and pleasure to the many rosarians in Michigan, as well as the public in general." These sentiments were endorsed by Mr. Lewis, who said that the plants have provided "invaluable material for any comprehensive study of the genus Rosa,
The gift of many plants to the Blandy Experimental Farm has been a very great help in my study of the North American roses. The collections' continued availability for research on the genus Rosa cannot be over-emphasized when one realizes how much data were obtained from a few plants and consequently how much remains to be uncovered."
Professor Bartlett retired as director of the botanical gardens in 1955 and his successor, Professor A. G. Norman, has assembled all the rose species in one area. This summer I made a final check of identifications, labels and records. Cards for all dead and discarded collections are filed separately and a map has been made by Dr. P. A. Hyypio of the living cultures. There are now 538 plants under 205 accession numbers. These include co-types of seven of my new varieties and species,2 and 120 hybrid plants from 35 of my interspecific crossings (see Table 2, appendix). Conclusion A scientific revision of our species was not made and it is indeed gratifying to know that Mr. Walter Lewis has undertaken this enormous task. In brief my conclusions were that there are probably only a few true species of Rosa (20) in North America and that they show great variation in minor characteristics as explained in the earlier annuals.(6,11,12) These investigations of species formation, chromosomes and sterility provided a valuable background for subsequent research on human races in India, and for research in cancer and toxicology at the Institutum, Divi Thomae.(13) It was an arduous, fascinating task and now it is a joy to know, not only that the investigation is being continued, but that the representative collection at Ann Arbor will be care fully maintained and that it will survive. In what other genus than Rosa would experimental plants persist for 25 and 30 years and give so much information? TABLE 1. Inventory of 37 interspecific hybrids growing at University of Michigan Botanical Gardens, 1956. Number of crosses in which each species was involved as seed and/or pollen parent. Basic chromosome number, N = 7. Specific chromosome No. = 2N, 4N, etc. No. of Hybrids as
**Rowley, 1955 American Rose Magazine, reported an octoploid (8 N) seedling grown from seed (open pollinated) of this hybrid sent to John Innes Hort. Inst., 1946.
Plump achenes all float in R. palustris and its hybrids; in other species those with good seeds sink in water.