American Rose Annual pp. 33-36 (1924)
Breeding Hardy Roses for Northeastern America*
*Paper No. 128, Department of Plant Breeding, Cornell University, Ithaca, N. Y.
By A. C. FRASER. Ithaca, N. Y.
Assistant Professor, Department of Plant Breeding

EDITOR'S NOTE.—Through the kindness of Prof. E. A. White, formerly the efficient secretary of the American Rose Society, contact was made with the interesting advance work here set forth. Particular attention is called to the footnote on page 86.

THERE seems to be considerable justification for Mr. McFarland's statement that "Any rose is winter-hardy anywhere if sufficiently protected." ("The Rose in America," p. 185.) When numerous varieties of Hybrid Teas and even Teas can be brought through the winter in the cold, wind-swept areas bordering on the Great Lakes, in parts of Canada, and at Sitka, Alaska, it would seem that almost anyone who really wants fine roses could have them. There are many people, however, who would like a few rose bushes "about the place" and who are either unwilling or unable to devote much time to them. They may now grow the old reliable roses that do well in spite of neglect, chiefly the Rugosas and their hybrids, forms of R. cinnamomea and R. gallica, some Hybrid Perpetuals and the various Briers. The development of hardy door-yard roses of high quality will increase the number of varieties which can be grown in this way.

A few years ago the writer began the breeding of roses with a twofold purpose: first, to produce roses of high quality which would do well in the northeastern part of this country, and, secondly, to study the inheritance of plant and flower characteristics in this genus. In connection with the first of these aims, three distinct problems present themselves. The first involves the development of hardy bush roses with desirable bud shape, flower color, and foliage characters. Then there is the need for a hardy yellow climber which will hold a bright yellow color from bud to full-blown flower, and particularly in strong sunlight. Then the Pernetianas, which give us such attractive yellow and salmon blooms, must be made more dependable.

The wild species of roses offer the greatest promise as sources of hardiness, and it would seem that the new hardy roses must come from crosses of these with the more desirable garden types. Here at Ithaca, the procedure has been to cross the best of the garden roses with such of the wild species as are known to be hardy. The first generation hybrids are grown (first crop of seedlings), and, where possible, their blooms are sell-pollinated and a second generation is raised from the same cross. According to plant-breeding theory, and this has been demonstrated many times experimentally, there are better chances of getting desirable recombinations of the characteristics of the parent plants in the second generation than in the first. It is quite possible in this way to obtain choice seedlings from mediocre, or even poor, first-generation hybrids.

It is not always possible to use this method, however, as hybrid plants are not uncommonly sterile. For example, there is an attractive rose known as Rosa rugosa repens alba, apparently a hybrid of Rugosa with a climber. One plant of this at the local test-garden of the American Rose Society has resisted all efforts to make it set seed. No hips have ever been observed on it, and it has failed to set fruit when carefully hand-pollinated. Repeated crossings with twenty-five different varieties of roses have failed to give any seed.

In other cases, the hybrids which are not self-fertile can often be out-crossed successfully to other plants. This method has been used a great deal by rose-breeders. Hybrids which have not been altogether satisfactory have been "stepped up" by crossing again with desirable types. The method is especially useful when the new type of rose used in the cross is itself fairly satisfactory. There is this objection to it, however, that it may result in the covering up of certain desirable traits that might otherwise have appeared in the second generation of the original cross. Where the breeder desires to get everything possible from the first cross, and particularly where he wishes to get a shuffling and recombining of rather diverse parental traits, he will do well to grow the second generation.

The rose-breeding work at this station has been under way for but a short time, and, consequently, no second-generation plants have been raised. The practicability of the method, as judged by results, is yet to be demonstrated. In many of the crosses which have been made here, Rosa rugosa is one of the parents. This species has appealed to many rose breeders. Budd used it some years ago in breeding hardy roses for Iowa. It has been used in crosses at the Alaska Experiment Stations, and Director Georgeson says that it is perfectly hardy in all parts of Alaska (Alaska Agricultural Experiment Station Report, 1922). Of a large number of our crosses of R. rugosa with other types, about six series have bloomed. These are all rather disappointing as they stand, although it is hoped that some of their seedlings will be more satisfactory. In several of these crosses, the hybrids are scarcely more hardy than the Hybrid Tea parents. This is particularly true in one large family of plants of R. rugosa X Golden Spray (HT.). For the past two winters these plants have lost, to our surprise, from one-half to two-thirds of the season's growth by winter-killing.

Other crosses involve R. gallica, R. rubiginosa, R. Wichuraiana, the native roses R. nitida, R. blanda, and R. carolina, and such of the European and Asiatic species as have been brought into flower. Work with roses progresses slowly, especially at the start, and so far nothing of special value has been developed here. Many of the crosses seem to have great possibilities, however. The conditions under which the hybrid plants are grown generally furnish an adequate test of hardiness.

In connection with some of this breeding work, it was desired to make crosses of early-blooming wild species and other types flowering later in the season. As the periods of blooming in many cases failed to overlap, an attempt was made to store some of the pollen. Tests were made with pollen of a number of species, not merely those which stopped blooming early in the season. Buds were collected when just about ready to open, and were allowed to shed their pollen on clean paper. This pollen was then poured into gelatine capsules of the No. 2 size, and one or two capsules of each kind of pollen were sealed up in a glass vial containing calcium chloride as a drying agent. Temperature seems not to be an important factor in the keeping of rose pollen, while dryness is important. When this pollen was used later on in the season, a number of hips set. It is too early as yet to conclude that all of the seeds were hybrid. They may have resulted from development without fertilization (apogamy), a phenomenon that is not uncommon in the rose, as shown by the work of Dingler of Germany and Almquist of Sweden. Some of the seeds of the cross R. carolina X R. rugosa have given very thrifty plants which are clearly of hybrid nature. In this case, the R. rugosa pollen was thirty-eight days old when used.

Roses do not furnish the most satisfactory material for a study of heredity, yet it seems very much worth while to learn as much as possible regarding the mode of inheritance of such things as flower color, doubleness, winter-hardiness, bud shape, thorniness, foliage characters, and the like. Studies of this kind will serve as a basis for more exact rose-breeding work.

*Only three days before the receipt of Prof. Fraser's paper a letter was received from an Ontario rose worker, Mr. Frank Flett, mentioning the work of Dr. Crocker in an important way. Mr. Flett writes thus of Dr. Crocker: "In this letter he states he has germinated a number of Dr. Van Fleet's hybrids. Where seeds had dormant embryos, he found they went through the most rapid change in preparing for germination if the germinators were hold at about 41° Fahr., and that practically every seed germinated in from 90 to 140 days. As the seeds germinated they were transferred to flats, for the germinated seed, can be transferred immediately to 65° Fahr. without returning to dormancy." Here is an important discovery, concerning which it is regretted that the closing hour of this Annual does not permit the seeking of direct information from Dr. Crocker.—EDITOR.

There are many difficulties to be encountered in making a genetical study of roses, chief among them being the fact that most cultivated varieties are of hybrid origin to begin with, and, therefore, one does not start with material that will breed constant. Presumably most of the wild species come true to seed—at least all of those tested here have done so. The crossing of such types with the garden varieties will bring one constant element into the cross and should facilitate the study of the breeding constitution of the hybrid types. Other difficulties lie in the slowness of growth and reproduction, in self- and cross-sterility of varieties, in the relatively small numbers of plants to be obtained from any cross, and in the difficulties of propagation. The slow germination of the seed has been a great handicap to all rose-breeders. Some recent work by Dr. Wm. Crocker,* of the Boyce-Thompson Institute for Plant Research, promises to eliminate much of this difficulty.

In spite of these handicaps some progress is being made in a study of inheritance in Rosa. One or two families in particular are furnishing some interesting data. The results are incomplete as yet and do not warrant the drawing of conclusions.