The Journal of Heredity. 5:415-422 (1914)
Amateur Rose Breeding1

Editor of Lyon Horticole, Lyon, France
1Translated from Lyon Horticole.

THERE are few more agreeable occupations for an amateur, and few more profitable for a professional, in horticulture, viticulture or agriculture, than that of creating new varieties of flowers, vegetables, cereals and fruits, which shall be more productive and more beautiful than those now found in our gardens, orchards, vineyards and fields.

This occupation lies within reach of all who have a little—a very little—spare time, and who are endowed with continuity of thought.

For professionals who desire to make a commercial enterprise of plant-breeding, there is one primary difficulty: they must know the varieties, the races, the sports already in the trade, of the genera which they undertake to improve.

If they ignore that detail they risk trying to break in doors that are already wide open, as the saying goes, and what they produce is likely to be an old story. It is only in the case of really sensational acquisitions that one may dispense with a knowledge of what is already in the trade. It is certain that if a horticulturist secured a blue rose, a blue dahlia, a blue carnation, he would not have to worry about the work of his predecessors or contemporaries. If the color existed in those flowers, it would be known. It has been sought long enough.

It is desirable, when undertaking to secure new varieties, that one specialize in some particular genus. That is the way followed by a great many of the most successful horticulturists, viticulturists, and agriculturists.

One might take up roses, various shrubs, gladioli, petunias, verbenas, carnations, salvias, etc., or peas, beans, beets, carrots and other vegetables. In agriculture the wheats, oats, barley, rye, the clovers and alfalfas could furnish remunerative work. This enumeration will suffice to indicate how vast is the field for exploration. As no one is expected to know everything, the breeder should direct his studies to the side which seems most profitable and, at the beginning, the easiest.


In undertaking to breed new varieties of plants, it is useful to become familiar with certain terms employed by professionals, which are constantly found in works that treat of these questions, or that will constantly be heard in conversation. Such, for example, are the words genus, species, variety, race, hybrid, cross, atavism, selection, segregation and some others. I shall summarily indicate what we mean in using these words in horticulture.

Genus.—This word has a good many different significations in French, but in botany it is applied to a group of species which differ in some of their characteristics, but which can be grouped together, for the purpose of facilitating study, because they possess certain other characteristics in common. Example: the rose is a genus which includes a great many species; the gladioli, petunias, heliotropes, clovers, alfalfas, peaches, plums, etc., are genera.

But all authors are not in accord on the definition of a genus. Formerly cheries and apricots were in the plum (Prunus) genus, and the apples in the pear (Pyrus) genus. Many similar examples might be cited. Anyhow, the genus is of secondary importance in genetics.

2The lay reader will probably be less confused if he retains the old definition, that a species is a group of organisms which interbreed freely.—The Editor.

Species.—A substantive whose definition is much disputed. It may be compared with that of a genus. If genus is a word which serves to group a number of species, species is a word which unites the varieties, races and variations of the same type.2 Today we distinguish systematic species and elementary species. The first (also known as Linnaean species or type species) are described in most of the classical Floras and works of botany. The second, which are subdivisions of the first, are also called Jourdan's species, elementary species, etc. To make a long matter short, let us say that they are races which reproduce true from seeds.

Hybrids.—Products of crossing two systematic species.

Crosses.—Products of crossing two races or varieties of the same systematic species, or of crossing descendants of hybrids.

Atavism (or Reversion).—A return of crosses or hybrids to the characters of their ancestors.

Segregation and Selection.—These are two nouns employed by genetists which generally pass for synonyms. Nevertheless they have different meanings which should be known by everyone who takes an interest in seeking new races of plants.


Segregation is a French word used as early as the sixteenth century to represent the action by which one sets apart or separates from the whole, from a mass, one or several objects which are different from it. Selection is a word first employed by English breeders to signify the choice for breeding of animals endowed with the characters which the breeder desires to fix in a distinct variety of animals. Taken in this sense, selection has the meaning of marriage, if I dare to use the word, between well-matched individuals. Selection passed from English into the French language where it popularized through the phrase selection, invented by Charles Darwin. According to the famous English scientist, natural selection refers to the predominance which nature yields to any given species or variety, because of the greater adaptation of its characteristics to those of the environment, as regards nutrition, conservation, reproduction, etc.—an adaption which allows it to persist and brings about the disappearance of less-favorably species which are unable to compete with it.

In horticultural practice, selection signifies the choice of particular subjects noticed among seedlings: extraordinarly large or small specimens, particularly early or late individuals. etc. As for trees, one can often practice selection on a single tree or shrub, selecting branches which are more fertile, precocious, late vigorous, etc., than others.

Pedigree culture or genealogical culture.—In this mode of cultivation, one must start with seeds furnished by a single individual, and must follow its descendants for a number of generations to make sure that it is fixed. If it is variable, the study of its variations should be pursued until it is decided that they are not going to furnish the desired results.

Other technical words used in plant breeding will be explained as they are employed in the course of this study. As I have promised a number of persons to discuss roses, I will now keep that promise by giving some practical directions for rose breeding to amateurs.

At the right an unopened rosebud is shown. Roses must be pollinated at this stage; if they are allowed to open before being worked, the bees are likely to interfere. The center shows the bud with sepals and petals removed, disclosing the stamens or bearers of pollen, surrounding the ovary. At the left, the ovary alone is shown, the stamens having been removed with a small pair of forceps or scissors. The bud at the left is ready to be pollinated, by the application to its sticky top of stamens cut from some other rose.
(Fig. 13)


I note in La Petite Revue a report of the exposition held at Antibes, on the French Riviera, last March, in which M. J. Lefevre makes the following comment which bears, in part at least, of the selection of branches:

"In a very different way, M. Dhumez was able to get from Nature more than she had ever wanted to give. Disregarding the field of selection of seedling variations, and confining himself to varieties which he had bought in the open market—some of them little known, such as Prince de Bulgarie, and others long since used by everyone, some of them for a long time past, such as Bastide rose—M. Dhumez was able, by a very sure selection, a judicious choice of fertilizers and, especially, of method of operation, to get such remarkable results that they would have been declared impossible: the size of the flowers, the absolute perfection of form and color of Prince de Bulgarie, Marquise de Mores, Agathe Nabonnand, Agathos and others, won as soon as they appeared not only the rapt astonishment of the public, but the highly-valued admiration of experts."

Here, then, is a grower who by a very sure selection and by other artifices of culture, was able to present well-known roses—and also the variety Prince de Bulgarie, less known—which called forth the rapt astonishment of the public and the admiration of experts.

Perhaps it will be said that the manuring and the cultural methods played a greater part than selection in M. Dhumez' success. Furthermore, one may say that his success is doubtless only momentary, incapable of reproduction by grafting, budding or marcottage. All that is possible, but the contrary is also possible. This is the reason:

Every rose grower knows that the varieties called Climbing Roses by the English are varieties whose mothers were not procumbent by nature, and that such varieties can be produced almost to order, by taking scions from branches which show a tendency to extend. Such extensions can be provoked by pruning, manuring, and especially by a shady location.


Bush roses can be produced in the same way that climbing roses are, from bushes of ordinary height. It is only necessary to cut your grafting wood in the neighborhood of the inflorescences. The procedure is well known for the Noisette Aimee Vibert. It is perhaps too well known. In attempting to procure the old, giant type I twice secured the dwarf.

I remember having seen at a rose exposition some specimens of Noisette William Allen Richardson, cultivated in pots and relatively dwarf in height, well covered with flowers; they were exhibited by M. Patichoud, a horticulturist at Lyon-Croix-Rousse. Had he practiced selection of grafting wood on the specimens of this remarkable variety obtained from Mme. Veuve Ducher? There seems to be nothing to contradict such a belief.

It seems probable, too, that many beautiful procumbent or Sarmentose varieties which have cast a bright light in the firmament of roses and then have disappeared from view, have vanished from rose gardens for no other reason than that selection of too rank-growing budwood was practiced. I could name a dozen kinds that one now finds only by accident.

An amateur, or better still a rose grower who cultivates hundreds of specimens of each variety, could certainly find either branches or whole plants presenting characters worthy of being fixed: more abundant flowers, lighter or darker shades of color, better foliage, fewer thorns, etc. How many sports are lost because not noticed?

However, they are not all lost: as witnesses, La France, which is grown in four or five different tints and one panache or striped rose (Angelique Veyssey); Malmaison, which exists in two colors; Baronne de Rothschild has three albinos; Reine Marie-Henriette origin of Madame Driout. And there are many more of the same kind, without counting a number of varieties sold as seedling variations, which are really nothing but sports.

It is especially roses of hybrid origin which present these dissociations. But there are others more ancient on which an intelligent selection would give good results, and among these are the moss roses, of which the first one known is still the best.


Rose Seedlings—Behold the infancy of the art! It is within reach of the first planter who appears. Our fathers, our uncles, have worked for us. They were expert hybridizers and cross-breeders, and when they did not hybridize, the bees hybridized for them.

Although the method is slightly played out, it still gives good results here and there. It is a lottery. Since the creation of fertile Tea hybrids, many beautiful varieties have been obtained simply by planting seeds of them. When the Hybrids Remontants flourished, it was their seeds that produced the remarkable varieties still cultivated on such a large scale in gardens and especially by commercial growers. In those days, we had to sow seeds by the thousand to produce four or five noteworthy varieties. This is the way many growers used to proceed. At the end of the season—the last of October, in the climate of Lyon—they gathered indiscriminately in collections and nurseries all the fruits that the bushes produced, without paying the slightest attention to the varieties which produced them. At least, that is the way it was done in the establishment of the late Jean Liabaud, with whom I served my apprenticeship. And that clever horticulturist had obtained in that manner beautiful varieties which are always in demand even at the present day. The rose fruits thus gathered were opened during the long winter evenings, the seeds were taken out and stratified in sand, and sowed in February in a bed not too hot, sometimes in small pots or flats, but more frequently directly in the hot-bed itself. Sometimes they were sown under a temperate glass house.

Not all the seeds which were planted germinated. Many, although perfect in appearance, contained no embryos; but enough grew each year to plant a good sized piece of ground.

If in a thousand seedlings grown in this manner, the grower found half a dozen noteworthy new varieties, he was well satisfied. It is worth noticing that the roses thus obtained from seeds did not flower at the same time. I have seen some bloom the same year they were planted, while others would only produce their first flowers after five or six years. Many had simple flowers, others half-double, some double. Some flowered scantily, others abundantly. The ones selected were grafted on eglantine (Sweetbriar), then studied, and finally put in the trade if they were worth it.

This was not the only method of procedure followed by the rose growers of Lyon. Some added seeds produced from artificial pollination between particular varieties, such as I shall discuss a little later.


To understand how ordinary seedlings of roses cultivated in gardens could produce anything profitable, you must remember that most of them are descendants of hybridization or crossbreeding. Now it is well known that plants descended immediately or distantly from hybrids or crosses do not come true from seed. They are endowed with innate variability. Their characters separate, combine to produce others, and run through a whole circle of variations which manifest themselves in diverse ways.

Simple propagation by seed of the best varieties of roses can be recommended, on one condition, which is to go at it with separate varieties and to follow the results by pedigree culture to the third generation. Some of the varieties have been so much sowed and resowed that in the last half century they have produced just about everything good there is in them.

Pedigree culture offers two advantages. First, the planter is not required to waste time sowing a second generation of seeds from plants which produced nothing good in the first generation. If he does not want to carry his experiments any farther, he will at least know that he has missed nothing good as far as he has gone. But if be has the instinct for plant breeding he may find in succeeding generations, if not remarkable novelties, at least good seed bearers or pollenizers—that is good mothers and good fathers, often simply semi-double, with which he can make productive crosses.

Good pollenizers or producers are rarely put in the trade by their originators, who keep the monopoly for themselves. Their influence is seen as a sort of trade mark in the varieties which are subsequently put into the trade.

Before taking up hybridizing and cross breeding, it is worth while noting that Nature has set a limit to the enlargement of flowers, to their duplication, florabundance, etc.

One more often attains large size in flowers by working through those of medium size, than by using large flowers to start with, unless the latter are semi double. As for duplication, if it is exaggerated, the flowers do not open well, and therefore it is to be avoided. But very double roses, when they possess good pollen, can be utilized with a chance of success to fecundate—after castration—semi-double mothers. The same is followed with other genera such as carnations, petunias, Pelargonium zonale, and begonias.

By cross pollination between hybrid remontants and the yellow rose (Rosa lutea) M. Pernet-Ducher obtained Soleil d'Or, and following it that series of varieties which approached vermilion in color.


There are plenty of cases where beautiful roses are absolutely sterile, generally because their pistils are hypertrophied or atrophied. Don't waste time in pollinating them; but if they themselves possess good pollen, it may be used on other females.

There are few genera where artificial pollination is easier to perform than with the rose. Its flowers are large, its stamens numerous, and its pistils easily seen. Furthermore, natural fecundation rarely takes place before the flowers are open.

Even if one does not know the floral organs of the rose, he can quickly learn to distinguish them. In an opened rose you see first what is called the calyx, formed of five leafy green pieces named sepals, which crown the fruit. Then the petals come, more or less numerous, and colored in different ways. In the middle of the flower are the organs of generation: stamens formed of threads carrying anthers, small yellow points containing the pollen—a sort of fecundating dust. Finally, there are the styles in the center of the flower, sometimes united, often merely crowded together. These styles are terminated by a stigma, at the top and run down to an ovule or a single seed below.

It is on this stigma that the pollen falls and fecundates the ovules. I mention these details merely for the benefit of people who have never studied the parts of plants.

To make a cross pollination, one must get a small pair of forceps, a number of small camel's-hair brushes such as are used for water colors, and as many small tin boxes as one wishes to make crosses. The reason for this is that each rose which is to be artificially pollinated may receive pollen only from the variety by which it is desired to pollinate it. If you used the same brush for a number of pollinations, you would not be sure that grains of pollen did not remain on it from one application to the next. The whole equipment costs very little.


When the roses begin to open, you must begin to collect the pollen which is to be used for fecundation. The pollen of each variety is put in a separate box, on which its name is written. The anthers (stamens) should be removed with the forceps, or if necessary with the fingers. The little box is placed indoors in a dry, shaded place, and not closed tightly, in order to prevent spoiling.

The roses to be fecundated with foreign pollen should, as soon as they have opened, be deprived of:—first all their stamens, without exception, second the petals of the center of the flower in order that these may not interfere with your own operation. When the stigmas begin to lubricate themselves, or become sticky, you bring in contact with their surfaces, by means of the camel's-hair brush, the stamens which you have previously gathered, letting them shed their pollen. After that you should cover the fecundated flower with a little bonnet of paper, in order to keep off rain. After the fruits have set, you must irrigate the plants in case of dry weather. Do not be afraid of using liquid fertilizer: one gram of phosphate of ammonia to a quart of water, and an equal quantity of saltpetre in the same amount of water.

All the fruits which were not artificially pollinated should be removed, and those which were pollinated should be allowed to ripen thoroughly before they are picked. Before the first frost in November is soon enough..

Each variety which was fecundated should be planted separately, in pots, and pedigree culture carried through two or three generations.

To find out more rapidly the probable value of certain roses as producers of novelties, it is worth while fecundating the same mother with a number of varieties of good repute as pollenizers.

It is very important to perform the operation on the first flowers which open, particularly in a cold climate where spring is late, for it is absolutely necessary that the seeds have time to ripen before the frosts of fall. Amateurs living in such climates will do well to plant the roses which they wish; hybridize in the warmest spots of the garden. In England, amateurs and nurserymen do not hesitate to cultivate their breeding stock in pots, in a well-ventilated, temperate greenhouse, in order to hasten the period of flowering, always late in that country. During the summer, the roses are planted in the open ground and, if necessary, taken back under glass in October, to finish ripening their fruits.

It seems probable that a certain number of types of roses which have given good products in the past have not yet said their last word, in view of the fact that general conditions have been modified since they were first used. If seed-producers are still the same, yet it can be said that many of the pollen carriers have been singularly modified. Since one of the parents has changed in nature, there is reason to suppose that if the old crosses were made over again, they would give different results from those originally obtained.


I am referring particularly in this passage to a well-known type of rose of the group of bracted roses (Rosa bracteata), also called Macartney roses, which produced by hybridization a rose known to the trade as Rosa alba odorata, and another known as Maria Leonida. This type has remained quite constant. But the Rose of India, one of the parents, has undergone great modification since 1850. Who can say that another hybridization would not produce a different Rosa alba odorata, which would open more easily and yet retain that precious habit of flowering in July when most roses have finished their first flowering season! I am not overlooking the fact that we have July roses on young plants grafted that same year, but on old plants they are rare.

There are other curious wild types introduced in botanical collections which seem to have been overlooked by bold and perservering hybridizers. Let me note, in passing, Rosa xanthina, which does not seem to have been "worked".

And there are still others. I will limit myself to those two. Another thing. It seems that some of the paternal and perhaps maternal ancestors of some of the finest hybrid climbing or remontants have perhaps been modified, if they have not been altogether lost. While General Jacqueminot gave rise to a numerous progeny, how many replicas of different colors—aside from sports—has Baronne Rothschild produced? And Paul Neyron? And others?

It is very desirable to know how to fight in retreating, when the offensive is no longer practicable. Generals have become distinguished by such tactics—witness Xenophon and Moreau.

What precedes is merely a morsel offered to amateurs who want to try new problems and leave beaten paths. If they attempt crosses between the old types, long crossed and recrossed, and new, untried species, they will perhaps get more failures than successes, the results of hybridization being uncertain and sometimes disconcerting. Here, however, are a few hints that may be useful:

In crossing a new type with simple flowers, which is to be used as a pollenizer, it is desirable to choose a fertile mother with double flowers, whose ancestors are also double or semi-double flowering. Duplication being highly esteemed in garden roses, the reciprocal cross may also be tried hopefully.


In making crosses with old roses of large flowers, like Paul Neyron, it seems that success might be gained by working again the variety Anna Diesbach, with hybrid remontants, or hybrid Teas, of recent Pernetiana. That is a proposition to be verified.

With Baron de Rothschild, the ancestors might found through atavism, after two or three generations of experiment. If so, these ancestors ought to be worked in their turn.

* This probably should read "once- or twice-blooming".

When one hybridizes two remontant roses, one of which although perfectly remontant has in its maternal line non-remontant roses, the products may be either single or double flowered.* I had occasion to cross the Polyantha Perle d'Or (Dubreuil) by the ordinary Bengal and obtained two huge Multifloras. The characters of the Bengal and the half-blood tea, of the Perle d'Or, had disappeared from the product. Nevertheless, every one knows that the little dwarf Polyanthas, so prolific in blooms, are the products of a cross between multifloras and tea roses, in all probability.

The Bengal type, which reproduces itself by seed, unless it has been hybridized, and which is very remontant, seems rarely to communicate its remontant character to the roses which result from its crossing. If they show, it is only after repeated crosses.

As has been said, the seeds of hybrid roses are not always fertile even though perfect in appearance, their fecundation having apparently, but not actually, succeeded. For this reason one must try various pollens, when one is certain that the mother is fecund with her own pollen, for she does not always accept the first pollen offered.

Vegetative Selection