Logo for the Journal American Rhododendron Society

Journal American Rhododendron Society

Current Editor:
Dr. Glen Jamieson ars.editor@gmail.com


Volume 5, Number 3
July 1951

DLA Ejournal Home | QBARS Home | Table of Contents for this issue | Search JARS and other ejournals

Observations On Breeding Rhododendrons
By David G. Leach, Brookville, Pa.

        There has come into being among rhododendron breeders a philosophy of arranging crosses which, from a theoretical standpoint, seems to be wholly without support. With a deep bow to the experience of the older hybridists, I must dissent from the prevailing views on this subject.
        So far as I have been able to learn there is a strong aversion to inbreeding rhododendrons, in the fear that it will lead to degeneration. This is not necessarily true. The frequent appearance among chance seedlings of new and valuable plants in such diverse groups as roses, strawberries and raspberries is a clear indication of the possibilities which lie in inbreeding.
        It is quite probable that the breeder who sows large quantities - of self fertilized seeds from good hybrids will produce more valuable novelties than will the breeder who continually crossbreeds his plants. Surely many of the important qualities we are seeking are recessive. This being so, the chance of obtaining an expression of these recessive qualities is very much better from self-fertilized seed. There is entirely too much cross-breeding among rhododendrons, in my opinion, and the variability which results from this cross-breeding is seldom used to advantage by the hybridist because he fails to inbreed for his next generation of seedlings.
        Habitually self-fertilized plants such as beans, tobacco, barley and wheat do not exhibit degeneration after innumerable generations of inbreeding. Lack of degeneration from inbreeding simply means that the plants are "pure" in the genes they carry. The theoretical basis for degeneration from in breeding through several generations is simple enough: if a hybrid rhododendron is "impure" for a number of desirable genes, and in each case the corresponding recessive form is undesirable, for one gene three quarters of the inbred progeny will be superior and only one-quarter inferior. If the same thing is true with respect to two genes 3/4 x 3/4 will be superior and the balance inferior. If we are dealing with only five genes, 3/4 x 3/4 x 3/4 x 3/4 x 3/4 will be excellent, but this is only about one-fifth of the progeny. If there are eight genes to be considered, only one-tenth of the seedlings will be superior.
        The number of "impure" genes in a rhododendron hybrid must surely be very large indeed, so it is obvious that the breeder must grow a great many seedlings from his inbred hybrid to recover the combination of desirable qualities which he seeks. I believe that the preoccupation with degeneration has been vastly overdone but if it does occur it is easy to restore the original vigor by out crossing to another rhododendron of similar appearance but different genetic descent.
        Rhododendrons are notorious among breeders for their inclination .to- impart to their offspring qualities-. which are roughly intermediate between the two parents. Hybridizes who are inclined to think of genes in the chromosomes as equivalent to pearls on a string fail to realize that it is an outmoded concept to assign to each gene an independent influence. The genes in the chromosomes within the nucleus of a plant cell bear to each other a tremendously complicated influence on the growth and development of the mature plant. It is not strictly accurate to say that for all characters one gene is dominant over another. In such a quality as hardiness, many genes exert their influence.
        For qualities determined by only one gene the intermediate character of the hybrids between two rhododendrons which so annoys breeders may be due to the fact that a large amount of a certain gene in the nucleus of a plant cell will have a stronger influence than a smaller amount of the sane gene. In other than primary hybrids between species, the apparent intermediate character of the progeny is often due to the persistent failure of the hybridist to grow a sufficient number of seedlings to permit recovery of the desired combination of qualities.
        In breeding rhododendrons it is not difficult to get the recessive virtues isolated, and to discard the dominant faults. The problem is to get rid of recessive faults. Recessive qualities always bred "true." In a variable group of plants bred from hybrids the recessives are in the minority. The colors, shapes and other qualities found in the minority are the easiest to breed true. Here in the East where we are so much concerned with resistance to Winter cold, our primary hybrids between a hardy native species and a tender exotic species are almost always unable to endure the rigors of the climate. They are intermediate in character between the two parents. The genes involved in determining hardiness are interacting. Some are acting in a plus direction toward greater hardiness whereas others are exerting a negative influence toward tenderness and the net effect is the production of intermediate, half-hardy plants.
        It is at this point that the amateur hybridist usually goes wrong. He becomes discouraged by the indifferent quality of the primary hybrids, without realizing that the characteristics of the first generation progeny can be said to be almost without importance. It is immaterial whether the first generation seedlings are frost-resistant, or red (if that be sought) or tall or of poor plant habit.
        When two different plants are crossed with the idea of combining some good qualities of one of them with other good qualities from the second parent, the work is only half done when the hybrids flower. The next step must be to inbreed one of the hybrids, or to breed two of these hybrids together to obtain the potential variability inherent in such a cross. I believe inbreeding is to be preferred except in cases where the hybrids are incapable of fertilization by their own pollen, and in that event two of the hybrids can be crossed. Failure to take this next step of inbreeding is so widespread that it might be remarked that most of the rhododendrons in commerce today are the result of a cross that has been made with a definite object which has not been attained by the breeder.
        Having made the vital step of inbreeding the primary hybrid the next pitfall to avoid is the growing of a number of seedlings too small to permit the appearance of the desired qualities in one plant. If the primary hybrid is "impure" for only four genes (an impossibly small number), it will produce 16 different kinds of pollen and 16 different kinds of ovules so that there will result 256 different ways in which the genes might be recombined. Only one way is ideal. The chance of obtaining this ideal combination is 1 out of 256. It is necessary to grow a relatively large number of progeny from the inbred hybrid to be reasonably certain of obtaining in one plant the best qualities inherent in its two grandparents.
        The majority of breeding experiments fail because hybridists clutter their fields with an unnecessary number of primary hybrids. To save space they then grow a number of progeny from these primary hybrids which is far too small.
        Exactly the same principles apply to any combination of qualities which may be sought by hybridists in favored climates who need not be concerned about the hardiness of their productions. Whether the goal be larger flowers, better color, smaller stature or a later blooming season these characteristics can be imparted by a plant containing them to a second generation hybrid through the procedures outlined herein.
        One final question about breeding rhododendrons which is often discussed is the occasional appearance among seedlings of a plant which exhibits a quality to a much greater degree than either of its grandparents. The explanation is that the unexpected quality is determined by two different sets of genes in the two grandparents, each of which may contain factors acting as a dominant which restrains the expression of that quality. When a "pure" recessive plant appears in the F2 generation a double restraint is removed so that there is a cumulative effect for the quality in question. Thus we occasionally may get a giant rhododendron, or a super-hardy plant, or a color not to be expected from the pedigree of the seedlings.
        A related phenomenon is called transgressive variation. In the breeding of cereals the best way to obtain extremely hardy wheat or ryes is to cross different good, winter-hardy sorts. When this is done we confidently expect the second-generation plants to differ so much that we obtain extreme examples on both sides of the scale: plants that can stand no frost at all, and others that are much more frost-resistant than the parents with which we started. Inasmuch as this same phenomenon of transgressive variation has been observed in many genera and for many different qualities it is reasonable to suppose that it will occur in rhododendrons. It might be particularly valuable to extend the range of the obtusum azaleas in the East, for example.
        In a breeding project of this type, if the first hybrids from the hardiest of the standard commercial sorts are less cold-resistant than the parents, it may well be regarded as a favorable sign, strange as it may seem. That is a good indication that the two parents differ in genotype so that the hardiness factor is not due to the same combination of genes in each. In the next generation there will be a reasonable chance that we will get some plants in which both separate causes for hardiness are combined, and the result will be a super-hardy obtusum azalea which can be grown far beyond its present area of adaptability.
        I freely stress that the procedures proposed in this article have not been proved by me in many cases. I defer to the mature experience of the older breeders but, remembering the prejudices which arise in any field of amateur plant breeding, I see no reason why established genetic principles should not be tested before they are discarded as not being applicable to rhododendrons.


Volume 5, Number 3
July 1951

DLA Ejournal Home | QBARS Home | Table of Contents for this issue | Search JARS and other ejournals