Hybridization in the Genus Rhododendron
With Particular Reference to Inheritance of
White Flower Color in R. dauricum
Gustav A. L. Mehlquist
Plant Science Department, University of Connecticut
Considering that Rhododendron breeding has been going on for well over one hundred years it is perhaps somewhat disappointing to find how little exact information has been recorded relative to the inheritance of those characteristics which make a good Rhododendron. There may be several reasons for this dearth of information. In the first place, the early breeding was done in an era when new and different hybrids were much in demand and therefore very profitable so that release of information relative to productive parent plants was not considered good business. Secondly, much of the breeding in this genus always has been done by commercial nurserymen who have been primarily interested in direct, positive results and therefore did not see the need for recording failures, or the number of seedlings that were discarded for every one that was retained. Thirdly, most of the breeding in this genus has been done at the interspecific level rather than within the species. Breeding within the species would most certainly have resulted in simpler more reliable data which would have been of great value in predicting the results in further crosses but, of course, would not have led to the vast array of beautiful interspecific hybrids that adorn our gardens today.
One of the first features to be noticed in any ornamental flowering plant is usually the flowers, and therefore it is important to know something about the inheritance of the various colors of this feature. This information might be deduced from crosses made for other purposes but clear cut Mendelian ratios are not so likely to be obtained from interspecific crosses as from those made within the species.
It should not be inferred that the Mendelian ratios are important in themselves; they are simply a means to an end. From these ratios one can estimate the number of genes involved and from the number of genes one can calculate the number of plants that should be grown in order to give one a reasonably good chance of realizing the full potential of a self - or hybrid population.
The reason that good Mendelian ratios are not ordinarily obtained in interspecific crosses is that hybridization between different species is generally accompanied by various degrees of reduced fertility which most often leads to the elimination of or at least the reduction in numbers of certain categories or types, making it difficult to get a clear picture of the nature of inheritance of a particular feature.
Although little detailed genetic and cytological data from interspecific hybridization in the genus Rhododendron have been published, there is reason to believe that, were crosses to be made involving most of the taxonomic sections or series presently known, the results would be similar to those which have been obtained in most other genera which have been extensively studied. That is, the fertility of the hybrids would very likely range all the way from virtually complete to practically none.
In interspecific hybridization it is customary to recognize four categories of hybrids with respect to fertility as follows:
- Complete fertility, indicated by good pollen and eventually seed of good quality.
- Partial fertility only. Poor pollen, reduced amount of seed of lesser quality.
Fertility in back crosses only. Little or no good pollen but some seed obtained when the hybrid is pollinated by either parent species.
- Complete sterility.
Although the first category in this summary appears to be somewhat more likely in the genus Rhododendron than in many others, it would still be preferable to base ones conclusions regarding the inheritance of desirable features on data obtained from crosses within the species or from crosses between closely related species. The reason for this is that, even when an interspecific hybrid appears to be fully fertile, second generation results often indicate a lower proportion of recombinations as compared to parental or near parental types than should be expected on a random distribution basis making it difficult to ascertain the number of genes involved.
With this in mind a search was made for distinctly contrasting features in the species available and suitable for growing at Storrs, Connecticut. This locality would be generally 5a on the U.S.D.A. hardiness zone map although, as is well known, the presence of a few houses or a few coniferous evergreens strategically located might change a 5a zone into a 5b zone. At any rate, little but H-2 plants are reliably hardy in this area and some of those listed as such do not always fare too well.
It was soon realized that one of the best characteristics to work with, at least on a preliminary basis, would be white flower versus normal flower color for the species. In most genera in which inheritance of flower color has been studied at least two genetically different whites have been found, i.e. Cattleya labiata*, Dianthus caryophyllus, Lathyrus odoratus, etc. The same appears to be true in the genus Rhododendron.
* Here labiata is used as including all the taxa variously considered forms of C. labiata or independent species within the labiata group.
For reasons of easy classification of the segregates even when not in bloom, it was decided to work with that type of white which is due to complete absence of anthocyanin, the pigment ordinarily responsible for the pink, red, magenta, or purple flower colors in Rhododendrons. In such plants the leaves of the plants are usually of a lighter color than in plants destined to bear colored flowers. Thus positive identification of white versus colored can be made while the seedlings are still being grown in the greenhouse under conditions in which even the weaker seedlings are likely to survive until they are scored. This latter consideration has proved a very important feature in the study of the inheritance of recessive characteristics in plants due to the fact that seedlings characterized by certain recessive features often have a lower survival potential than seedlings characterized by the corresponding dominant features.
At the time these experiments were planned (1967) this type of flower albinism was available (to the author) only in two species; in R. kiusianum (obtusum) a member of the Azalea series and in R. dauricum of the Dauricum series.**
** The white form R. kiusianum album was a gift from Mr. Donald Richardson of Manhasset, New York and the R. dauricum album was a gift from Mr. Warren Baldsiefen of Bellvale, New York.
Several flowers on the white R. kiusianum were pollinated with pollen from a plant of the typical magenta colored form of this species. None of the resulting seedlings have flowered as yet but, since the foliage of all of them is like that of the magenta-flowered parent, it is assumed that they will all have magenta-colored flowers.
Five flowers on the white R. dauricum were pollinated with pollen from a form of R. dauricum* with typical magenta-colored
*This form was a gift from Mr. Harold Epstein of Larchmont, New York.
flowers. This form is also of a more compact habit than the white form and has a strong tendency to flower intermittently during late summer and fall. Ten flowers on the white form were self-pollinated.
The ten flowers that were self-pollinated developed into small pods which produced little seed resulting in all in thirteen seedlings all of which have the light-colored leaves of the white-flowered seed parent. Three of these seedlings flowered this year. Three years from the planting of the seed. All had pure white flowers. The five flowers which had been pollinated by the magenta-flowered form, on the other hand, developed into five well-filled seed pods. The seed was planted January 18, 1968 and resulted in 103 seedlings some of which by means of supplementary light and suitable temperatures began to flower in 11½ months time. All have bloomed by now and all have magenta-colored flowers on dark leaved plants. The first twenty plants to bloom were self-pollinated seventeen of which produced seed in varying amounts, but on the whole rather low amounts, which was planted on October 25, 1969. The seedlings were grown under supplemental light until the Spring of 1970 at which time lighting was discontinued and the plants were exposed to rather strong sunlight. It soon became apparent that the seedlings began to segregate into two categories, one with light-colored leaves like the white-flowered parent and another with dark leaves like the magenta-flowered parent. The proportions of each type are shown in the Table.
R. dauricum album x R. dauricum typicum F/1 (68014) 103 plants all with dark leaves like typical form. F/2 Progenies Colored White Totals 68014 - 3 23 9 32 - 4 89 17 106 - 5 18 6 24 - 6 58 19 77 - 9 10 2 12 - 11 19 6 25 - 13 29 6 35 - 14 9 1 10 - 15 47 8 55 - 16 10 4 14 - 17 17 5 22 - 20 8 6 14 - 21 25 5 30 - 22 27 34 - 23 1 4 5 - 24 4 4 8 - 39 95 26 121 Actual 489 135 624 Expected 468 156 624 x/a7=2 P=.0456
It will be seen from the table that out of a total of 624 plants 135 had light-colored leaves, that is somewhere been 1/5 and 1/4 of the total progeny. To date 161 plants have flowered, 136 with dark leaves and 25 with light colored leaves. All the dark-leaved plants produced magenta flowers, all the light-leaved plants produced pure white flowers. There was no exception. It thus appears safe to conclude that the lack of the usual anthocyanin pigment in both leaves and flowers in this group is due to the same gene. Furthermore, from the proportions of white to colored plants, it must be concluded that only one gene is involved. Since this gene is recessive, the designation a to identify this gene is suggested.
From the P value it may be argued that the proportion of light-leaved plants (white-flowered) is significantly smaller than should be expected on the basis of chance alone, but it should be borne in mind that recessive types generally have a lower viability potential than the normal types and that therefore, even with reasonable care, a greater number of plants will be lost from the recessive groups than from the normal groups. Another group of plants grown in connection with these experiments supports this conclusion.
In the summer of 1968 when I visited the Glenn Dale Plant Introduction Station, I was given a small handful of seedlings grown from seed from a white-flowered R. mucronulatum. It was emphasized that the plants were from open-pollinated seed and therefore might not come true. Every seedling was planted and given careful attention so as to save as many as possible.
It was soon noted that some seedlings had light-colored leaves while the majority of them had the typical dark color which could also be seen in other batches of seedlings from the R. mucronulatum group. Eighty four seedlings could be scored for this feature of which four died before the seedlings were planted in the field in the spring of 1969. This loss was entirely within the light-colored group leaving 12 light and 68 dark-leaved to be planted out. The light colored plants were marked so there could be no question later as to which was which. About one half of the dark leaved seedlings bloomed the following year. All bore colored flowers though not so deeply colored as the better forms of this species grown from other sources. This year the remainder of the dark leaved plants flowered with colored flowers as had been anticipated. Five of the 12 light-leaved plants flowered with pure white flowers as had been hoped for. At this time (1971) the light-leaved plants are only about two thirds the size of the dark leaved ones indicating that they are probably genetically destined to be smaller and slower-growing than their colored counterparts. Also it should be remembered that the loss of four plants between the time the plants were scored for leaf colors and the time of planting in the field were all from the light-colored group and none from the dark group. The significant feature of this population is not the relative proportions of light versus dark seedlings but that five of the twelve light-colored seedlings which bloomed all bore pure white flowers and that these plants probably will always be slower growing.
Crosses have now been made between white-flowered plants of R. dauricum and the white-flowered plants of R. mucronulatum. The purpose of these crosses is a twofold one, first to see whether the resulting hybrids are white thus indicating that the genes for white flowers in the two species are identical or, in the case of a colored F11 generation, are different, and secondly to ascertain whether by hybridization white-flowered hybrids with stronger constitution and viability potential can be obtained from these early-flowering and useful species.