Maataloustieteellinen Aikakauskirja Vol. 60: 235—254, 1988 Breeding of frosthardy rhododendrons M. UOSUKAINEN" and P.M.A. TIGERSTEDT 2 * " Agricultural Research Centre Healthy Plant Center SF-41340 Laukaa, Finland 21 Department of Plant Breeding, University of Helsinki SF-00710 Helsinki, Finland Abstract. In 1973 a Rhododendron breeding program was started at the Department of Plant Breeding, University of Helsinki. The program was based on the extremely hardyrhododen- dron material that had been naturally selected at Arboretum Mustila. The aim was to create new frosthardy cultivars that could tolerate minimum temperaturesbelow —35°Cin mid-winter. Rhododendron brachycarpum subsp. ligersledtiiNitz. was used as the maternal parent in the majority of crosses. Also R. smirnowii and R. calawbiense were used as sources for winter hardiness. Totally 148 different combinations were done between species, species and hybrids and between hybrids. The total number of matings was 496. Over 20 000 seedlings were obtained and a total of 13 752 plants were planted in field trials on eight different test sites between 1975—1979. After two extremely cold winters, 1984/85 and 86/87, about 60 % of the plants died or were severely damaged. R. brachycarpum subsp. ligersledtii crosses with either R. smirnowii or R. calawbiense gave the best genetic material for better climatic adaptation. The progenies of these crosses survived without any severe damage when minimum temperature was as low as 37°C. Tissue culture methods were developed and used in vegetative propagation of selected ortets. By the end of 1987 sixclones have been released as new cultivars for commercial propa- gation. Index words: Rhododendron breeding, R. brachycarpum subsp. ligersledtii, R. smirnowii. R. calawbiense. selection for cold tolerance Introduction Rhododendrons belong to the Ericaceae- (heath) family that consists of about 70 genera and 1900 species (Lawrence 1951). They are widely distributed on acid soils throughout temperate regions of northern and southern hemispheres and to a lesser extent in the subarctic. A number of ericaceous species cover the coniferous forest floor of northern 235 JOURNAL OF AGRICULTURAL SCIENCE IN FINLAND https://www.c-info.fi/en/info/?token=3bil68MM1B6gHBaU.2wAFWh7BscD5GaRVNK4LTg.RCUR4t-tjAeJXtUr02-GXLQ-t8IrJ2q8Zf_dxsh9KrDX7XtoEUNlUjg7Iqi5mv57_6aN5hSVCS0HDUxCrXZ_z0hDz5FnxtnCdWv4c7in3U_viSDOI3RRjd3y3jmgT2ONT33tr6lnUVPcCZJwdNg-HPJDIN9V0Uply2h5PAjtYUQ3tn5ukPDiMtyLXgy04HsQug Europe including Finland. However the genus Rhododendron is restricted to a single species, R. lapponicum L., naturally distributed in the subarctic under open tundra-like conditions. The genus Rhododendron comprises about 900 species in the cool and temperate regions of the northern hemisphere with a concentra- tion of species, about 700, in the mountains of east and southeast Asia. The genus Rhododendron was first recog- nised by Linnaeus in Species Plantarum in 1753. Recently there has been a total rever- sion of the genus into eight subgenera. Some of them are further split into sections and even subsections (Cullen and Chamberlain 1980). In fact, a revision of azalea types is currently undertaken and may result in the division of these rhododendrons into seven sections. Rhododendrons have been used as or- namentals in the gardens of central Europe ever since plant explorations began their in- troduction from China and other parts of the world early in the 19th century. A number of famous plant hunting expeditions (Forrest, Rock, Kingdon-Ward and others) in the be- ginning of the 20th century introduced a large number of new rhododendron species into the gardens of Europe, particularly from China and other parts of southeast Asia (Kruss- mann 1968). This also promoted plant breed- ing efforts to improve the ornamental value and hardiness in new hybrids. Species within sections of rhododendrons are easy to cross and new character combinations of growth habits and flower colours can be produced almost ad libitum. Intersectional crosses are however often difficult to make and resulting hybrids are often totally sterile. Sterility may depend on strictly morphological malforma- tions, but also ploidy levels may vary from diploid to decaploid in different species thus causing hybrids to have genomic imbalance. For instance crosses between diploids and tetraploids give sterile triploid hybrids. Ploidy level manipulations by colchicine treatments give promising results to produce new fertile alloploid rhododendrons. Thus more than a hundred years of selection and hybridisation has produced over 4 800 named »cultivars» (Salley and Greer 1986) and many more man made products have been left un- registered. Most of this plant breeding work has been done in the United Kingdom, Unit- ed States and in various countries of central Europe. Thus the released cultivars are gener- ally not hardy in northern Europe. At present some 50 000 rhododendron plants are yearly imported to Finland (Anon. 1986). The imported plants are mainly winter green Rhododendron catawbiense Michaux and deciduous R. japonicum (A. Gray) Sur- ing. The first mentioned is the most commonly cultivated rhododendron species in Finland. Only some plants, grown in southern and cen- tral Finland, have shown to be of adequate hardiness (Kallio 1966). This presumably de- pends on genetic variation in hardiness be- tween different individuals of the true species in combination with particularly favourable local growing conditions. The average winter minimum temperature that R. catawbiense can stand is —32°C (Cox 1979). The mini- mum temperatures in southern and central Finland have been varying between —3O,6°C at Hanko to —3B,5°C at Jyväskylä (Il- matieteen laitos 1986). Also following culti- vars are imported and grown regularly: ’Caractacus’, ’Cunningham’s White’, ’Dr. H.C. Dresselhuys’, ’F.D. Godman’, ’Nova Zembla’ and ’Scarlet Wonder’. The frost tolerance of their buds varies between —23°C and —32°C (Salley and Greer 1986). It can be concluded that imported species and cultivars are generally not adequately adapted to the Finnish climate. Plants suffer particularly from cold spells below —3O°C that damage flower buds or sometimes vegeta- tive buds or whole shoots. The incentive to launch a major rhododendron breeding pro- gram in Finland is based on three major facts. Firstly, it has been well proven, that import- ed material is not adequately adapted to the climate. Secondly, it has been shown in ar- boretum and garden trials that there are spe- cies of rhododendron that can stand even the coldest of winters in Finland. Thirdly, erica- 236 ceous species are generally well suited to grow on the acidic forest soils and particularly on the organic peat soils in Finland. The breeding program was started in 1973. It was based on the extremely hardy rhododendron material that during 40 years had been grown and naturally selected at Ar- boretum Mustila, south east Finland, 60 deg. 44 min. N. lat. and 26 deg. 29 min. E. long (Tigerstedt 1986). This collection was estab- lished by Dr. C.G. Tigerstedt who had a clear notion of the importance of seed origins and variability in hardiness between plants of a population. Thus introductions were, when- ever possible, made in the form of a number of individuals representing the species. In the years 1930—1950 Mr. B. Schalin, city gard- ner in Helsinki, established a number of rhododendron plantations in the Helsinki city parks and on his private property near Hel- sinki (Schalin 1953). At the Brödtorp plant nurseries, Mr. B. Knape started nursery production of rhododendrons and azaleas in 1953 (Knape 1984). Thus in 1966 Kallio reported that 40 different species and cultivars had been grown in Finland. These plantations have shown that there are a number of rhododendron species that adapt fairly well to the climate in the south-central part of the country. The aim of the program was to produce superior species hybrids and particularly to find desirable recombinants in populations of crosses between species and hybrids. Frosthardiness has invariably been the num- ber one trait selected for, but in addition a lot of effort has been put on selection for differ- ent growth habits of the bush, including foliar characteristics. A whole spectrum of flower colours from pure white to deep red has also been a collective target including the details of flower morphology and inner colour and feature of the corolla. Long term testing at Arboretum Mustila has shown that R. brachycarpum subsp. tigersted- lii Nitz. is an exceptionally hardy species (Fig. 1). This species was named after Dr. C.G. Tigerstedt, who introduced it to the world in 1935. That year he received a seed source originating from the Kongo-San mountains in eastern-central Korea. Plants withstood —43,5°C at the arboretum in February 1940 (Tigerstedt 1986). This species was the only tall, evergreen species that survived without any visible damage and flowered normally af- ter the extreme cold winters 1984/85 and 1986/87 (Uosukainen 1988). At Anjalan- koski, the official weather station of the Finnish Meteorological Institute, about 20 km from the arboretum, —39.3°C was recorded close to the ground in January 1985 and —39,B°C at the same spot in January 1987 (Ilmatieteen laitos 1985a, 1987). According to Cox (1979) it is probably the hardiest rhododendron in the world. The exceptionally good adaptation of the »Tigerstedt» rhododendron is well document- ed in its ability to produce natural regenera- tion at Arboretum Mustila. Some of the natural seedlings have grown to large flower- ing shrubs. In fact, R. brachycarpum has been the only introduced rhododendron species able to reproduce generatively. At Arboretum Mustila there are several observations made on coniferous trees, indicating that the ulti- mate proof of good adaptation can be con- cluded from how well the generative cycle, in- cluding meiotic division and formation of pollen, functions in the new environment. The ability to regenerate naturally also indicates that the introduced species has found an eco- logical niche in its new environment. Some other species of rhododendron are worth special mentioning for being hardy at Arboretum Mustila. The true form of R. brachycarpum, R. metternichii, and R. smir- nowii are completely hardy under normal winter conditions. However, severe damage is caused by temperature around —35°C. Ac- cording to Cox (1979) they should stand about —26°C. Also so called Seidel-hybrids have been hardy at Mustila. Rudolf Seidel used R. smirnowii, R. catawbiense and R. ar- boreum in his hybrids produced in Germany at the beginning of this century (Krussmann 1968). Particularly R. arboreum contributed 237 deep red flower colours to his hybrids but it was also a source of frost sensitivity. These species and hybrids are not general- ly available in commercial plant nurseries be- cause of their unsatisfactory ornamental properties or difficulties in their vegetative propagation. Their flower colours vary from pink to almost pure white and the inflores- cence is often small or loose. Also theiroften tall and loose growth habit causes problems already during the nursery cultivation. In ad- dition theplants are often slow flowering, re- quiring several vegetative years before flower- ing commences. The comparison of climatic constraints in various places where rhododendrons are grown show that in Finland, in northern Scan- dinavia and in East Baltic, rhododendrons meet the marginal limits of their survival (Cox 1979, Galle 1985). In south-central Finland weather conditions are extremely vari- able from year to year due to fluctuations be- tween a maritimeand continental climatic pro- file. Night frosts in early June and again in mid August, heat and drought spells in the summer and cold spells in the winter demand a very high adaptive tolerance of successful plants. This is particularly the case when Fig. I. Rhododendron brachycarpum subsp. tigersted- tii is a tall white flowered bush. It has been used as the maternal parent in the majority ofcrosses in the breeding program. Fig. 2. Meristem tipculture methods were developed and used for clonal propagation of the selected or- tets. Fig. 3. The crosses between R. b. subsp. tigerstedtii and R. smirnowii resulted in very winter hardy off- springs. The Fl-hybrids were medium tall bush- es with a fine growth habit, attractive foliageand pinkish white flowers. Fig. 4. Fl-hybrids from the cross between R. b. subsp. tigerstedtii and R. yakushimanum are dense, low bushes. They have pubescent leaves and young shoots and white or pinkish flowers. 238 growing conifers or evergreenrhododendrons. They may suffer heavily from desiccation in March—April when the solar radiation is al- ready strong, evaporation from leaf surfaces is considerable, but water uptake is virtually nil due to frozen ground. The taxonomic and systematic classification used in this presentation needs special com- menting. Latin names and names of the cul- tivars have been taken from treatises by Stevenson 1930, Nitzelius 1970, Cox 1979, Encke et al. 1980 Salley and Greer 1986 and Krussmann 1968. There has been a great confusion in the classification of the Japanese Rhododendron species in Subsection Pontica. The confusion concerns R. degronianum, R. metternichii and R. yakushimanum. Accord- ing to recent papers, these species are all either subspecies or varieties of R. degronianum (Chamberlain and Doleshy 1987). General- ly the older classification has been used in this paper, because of the difficulties in determin- ing the correct names of the used parental in- dividuals. Thus the paternal individuals used in the crosses carry the names that were in use in the rhododendron plantations from where the pollen was collected. Fig. 5. The matings between a species and a hybrid, such as between R. b. subsp. tigerstedtii and R. /or- resta var. repens-hybrid, gave very variable off- springs. Fig. 7. The first named cultivar ’Elviira’ originates from the cross R. b. subsp. tigerstedtii x R.forrestii var. re/ww-hybrid. It is a low, dense bush with red buds and flowers. Fig. 6. Open pollinated seed was a very good source for variation and selection. The mortality among young plants and the number of weak individu- als were very high particularly in the offsprings of so called Seidel-hybrids. Fig. 8. In the extremely cold winters, 1984/85 and 1986/87, differences in winterhardiness between hybrid progenies became clearly visible. In the picture taken from the field trial at Mikkeli, the frostbitten line in the middle resulted from the cross R. b. subsp. tigerstedtii x R. arboreum. 239 Parental breeding material All crosses were made on mother plants selected among the hardiest species, hybrids and cultivars at Arboretum Mustila. Totally 53 mothers, belonging to 8 species and 27 hybrids were used. The hardiest species and hybrids at Arbore- tum Mustila were initially used as pollinators. Also less hardy species at Mustila, but with more colourful flowering and more decorative foliage and growth habit were used as polli- nators. Pollen was also procured from Bengt Schalin’s garden. In 1975 new pollen sources was brought in from three rhododendron col- lections in south Sweden (Alnarp, Christinelund and Sofiero) and from the Roy- al Botanical Garden of Copenhagen in Den- mark. In 1976 pollen was procured from the rhododendron collections of the Proefstation voor de Boomkwekerij at Boskoop in Hol- land. Totally 114 individuals belonging to 23 species and 48 hybrids were used as fathers in the crosses. It is generally very easy to handle rhododen- dron pollen due to the fact that it can be preserved viable even at room temperature from several months up to a whole year. Thus species flowering at very different times can readily be hybridised. The total list of paren- tals used in the crosses is given in appendix 1. Methods Time table The duration of different stages of the breeding program are illustrated in Table 1. The program was started in 1972 with prelimi- nary crosses. On the bases of this orientation the first large scale hybridisation effort was made in 1973. The whole crossing plan was completed within six years. Progenies were raised from 1974 to 1980 simultaneously de- veloping an optimal nursery technique. Sub- sequent field trials were established 1975— 1980. First selections in the trials were made in 1978 and this work phase is planned to be completed by 1989. A research program for optimizing micropropagation of rhododen- drons was carried out in 1980—1984and sub- sequently applied to cloning of selected hybrids. Thus the cloning of selected ortets was started in 1982 and this work phase is es- timated to take nine years. The planting of clonal field trials was started in 1983 and the work is planned to be finished by 1990. The first cultivar was named in 1986and the clonal selection for new cultivars is estimated to last from 1985 to 1995. Thus the breeding pro- gram proper is estimated to take about 24 years. Pollination Most of the hybrids were produced by con- trolled hand pollination. Flower buds were emasculated just before opening and isolated in terylene bags, specially designed for tree breeding work (Duraweld Ltd., U.K.). Polli- nation was done after 7—14 days when the stigma had opened and when a drop of sticky nectar revealed its readiness to accept pollen. In some cases seed after open pollination was collected as it was found that at the arbore- tum, pollination between species flowering simultaneously often takes place. Totally 496 cross batches were obtained in the breeding program (Appendix 2). Hybrid progenies Seeds were sown and progenies were raised using methods described by Uosukainen (1976). The hybrid populations were planted at seven locations in southern and central Fin- land (Appendix 3). No plantations were done north of 65 deg. N. lat., which was considered to be the northern limit of garden rhododen- drons. In the west east direction, the cli- mate becomes gradually more continental as the influence of the western Golf-stream gradually decreases and the influence of the vast land masses to the east of Finland in- creases. These facts were taken into account in selection of experimental sites for the winter 240 Table I. Timetable of the breeding program from the first crosses to the release of new cultivars. Duration of stage Year Stage of program 0 123456789 1972 Preliminary work 0 1973 Crosses 1 1974 Raising of progeny 2 1975 Planting of progeny tests 3 1976 1977 1978 Selection of progeny 4 1979 1980 Research program in micropropagation . 5 1981 1982 Cloning of ortets 6 1983 Planting of clone tests 7 1984 1985 Clone selection 8 1986 New cultivars 9 1987 1988 1989 1990 1991 1992 1993 1994 1995 Total tenure of breeding program: 24 years. hardiness trials. Also the hybrid populations planted were classified in hardiness groups avoiding the planting of sensitive groups in the most severe test locations. The experimental sites also represented varying edaphic conditions. Soil types varied from glacial moraine and podsolic woodland soils to drained sphagnum swamp. All sites had a coniferous crown canopy of Scots Pine or Norway Spruce to prevent direct solarradi- ation. Selection criteria The most important selection criterion was adaptation of hybrid plants to climatic con- straints. This involves several single traits, such as low temperature tolerance, growth ini- tiation and cessation periods and drought resistance. The duration of field testing was determined to 10—15 years depending on the occurrence of critical »bottleneck» years (Ti- gerstedt 1970). In addition following charac- ters were considered in selection: Flowering characters: colours, flower and inflorescens shape, durability of flower clusters and their weather resistance, flower- ing period, first flowering, flowering propen- sity. Foliage characters: colour, position, shape, morphology, ageing and their resistance to nutrient disorders. Growth habit: creeping, shrublike, tree- like, compact or loose. Phenology: vegetative and generative cy- cles. General plant health. Ageing of plants. Cloning of selected material In 1981 a research project was started in order to match each selected hybrid to an op- timal meristem tip culture method (Uosukai- 241 nen and Niskanen 1985). This was necessary as it was found, that genotypes reacted differ- ently to hormones and nutritional components in the culture medium. The selection and vegetative propagation (Fig. 2) of the hybrids is illustrated in figure 9. The vegetative propa- gation terminology used (ortet ramet clone) is adopted from tree breeding (Zobel and Talbert 1984). From each ortet 10—20 ramets were planted for clone testing in Hel- sinki, at Överby, at Kaarina, at Mäntsälä, at Arboretum Mustila, at Jyväskylä and at Haapamäki (Appendix 3.). These clonal tri- als were started in 1983. Results The total number of matings in the breed- ing program was 496. There were 442 cross- pollinated batches that consisted of 148 differ- ent combinations between species, species and hybrids and between hybrids (Table 2 and Ap- pendix 2). In addition there were 30 selfpolli- nated batches, 14controls (no pollination) and 10 openpollinated seed batches. (Table 3). More than 20 000 seedlings were obtained from the seed batches and about 25 % of them died during the first year. There were also plenty of week individuals that did not sur- vive transplanting outdoors. Between 1975—1979 a total of 13 752 plants were planted in field trials on eight different test sites. By 1982, 37 % of these plants had died (Table 4) due to different reasons; un- favourable planting sites, replantings, theft and failure of care during the establishment of field trials. During the first 5—9 years af- ter planting, poor winterhardiness was seldom the actual reason for mortality. Selection in the trials was started about three years after planting, when most plants had reached the height of 75 cm and thus their shoots were no longer sheltered by snow cover in mid-winter. The frosthardiness of the breeding materi- al was tested for the first time during the winter 1984/1985 and again two years later during in 1986/1987. These two winters were in many test sites the coldest in this century Table 2. Number of different combinations and num- ber of crosses (batches in parentheses), which were made in the breeding program. Paternal Maternal Species Hybrid Total Species 41 51 92 (174) (98) (272) (62 intra specific) (112 inter specific) Hybrid 27 29 56 (108) (62) (170) Total 68 80 148 (282) 060) (442) Table 3. Number of self pollinated, control and open pol- linated seed batches. Type of pollination Self Controls Open Total �� *«* Species Hybrids 23 323 7 6 87 22 Total 30 14 10 54 * Self pollination = the emasculated and isolated flow- ers with pollen from the same individual. ** Controls = emasculation, isolation or no isolation, no pollination. *** Open pollination = without emasculation and con- trolled handpollination. Table 4. Number of planted hybrids in the field trials and the percentage of survival after establish- ment (1982) and after the first and second »bottleneck» winters (1985 and 1987). Number Survival % of plants °! J P'am .ed 1982 1985 1987Hybrids Test sites Mustila 713 61 46 42 1 600 68 31 26 2 917 74 57 51 Helsinki I Helsinki 2 Vähämäki Piikkiö 751 94 47 40 3 129 69 59 54 211 92 84 80 2 475 72 57 41 1 995 13 5 1,5 Lappeenranta Mikkeli Oulu 13 752Total 63,5 46,2 39,3Mean % Helsinki I = Kaivopuisto. Helsinki 2 = Haaga. 242 5 (Table 5). During these winters 30 % of the breeding material was either severely damaged or killed (Table 4). In winter 1986/87 par- ticularly, there was severe root damage due to thin snow coverage during the lowest tem- perature period. In January 1987 the snow cover was only 10—20 cm in southern and central Finland (Ilmatieteen laitos 1987). There were large sib-group differences be- tween the offsprings from different mother species and cultivars. Winter survival was in- variably superior in the offsprings of R. brachycarpum subsp. tigerstedtii. Particular- ly hardy were its progenies after crossing with R. smirnowii. These hybrids survived 100 % on all test sites and over 80 % had no visible damages (Fig. 3). Even the flower buds toler- ated —36°C without damage. Also »Tiger- stedt» progenies after crosses with R. cataw- biense and its variety album ’Glass’ gave very hardy FI-hybrids. In Flelsinki, all individuals of these crosses survived, 22 % without damage, 56 % slightly damaged and 22 % with severe damages on their leaves, buds and branches. However, the reciprocal crosses in- dicateda strong maternal effect for hardiness inherited from the »Tigerstedt» rhododendron mother. The matings between »Tigerstedt» and a number of other species gave somewhat less hardy offsprings than the particular ones men- tioned above. Crosses with R. metternichii and R. yakushimanum gave offsprings (Fig. 4) that survived —37°C but mostly with slight or even severe damages. All crosses with R. williamsianum, R. wardii or R. orbiculare as fathers gave offsprings with poor winter hardiness. Plants were often killed within two or three years after planting in test fields. The matings between two hybrids or be- tween a species and a hybrid gave extremely variable offsprings (Fig. 5). There was large variation both in morphology, flowering characteristics and in frost tolerance. Mortal- ity was generally high in recombinant plant populations but very interesting material could be selected from such groups for further clone testing. Table 5. The minimum temperatures at seven weather stations near the test sites (ILMATIETEEN LAITOS 1985 (a,b) and 1987). Weather Lowest Minimum Minimum stations recorded temp. temp. min.temp. °C °C °C 1984/85 1986/87 Mustila —37,3 —33,2 —37,3 Piikkiö —36,7 —36,7 —34,5 Helsinki 1 —34,3 —26,5 —34,3 Helsinki 2 (—52) —35,9 —31,8 —35,9 Lappeenranta —38,3 —31,0 —36,8 Mikkeli —42,4 —34,7 —36,8 Oulu (—53) —40,1 —36,5 —37,5 Mustila: Weather station of Anjalankoski. Helsinki 1: Weather station of Kaisaniemi. Helsinki 2; Weather station of Helsinki— Vantaa airport. Open pollinated seed, collected at the Ar- boretum Mustila rhododendron valley, was a very good source for variation and selection. The mortality was very high particularly in the offsprings of R. smirnowii and the so called Seidel-hybrids (Fig. 6). In many cases the off- springs died already as seedlings due to fun- gal diseases, such as Botrytis cinerea. Some of these recombinant populations died total- ly in the field within the first five years. Among the surviving recombinants, beautiful flower colours and shapes were often com- bined with satisfactory winter hardiness. By the end of 1987 about 80 individuals from hybrid populations had been selected to be used as ortets for micropropagation. The ramets were planted in clonal trials for fur- ther selection. By the end of 1987 a total of 2 410 ramets from 35 ortets have been plant- ed in the field for repeated clone testing. Six hybrid clones have been released as named cul- tivars to commercial production. There are two red flowered hybrids, ’Elviira’ (Fig. 7) and ’Hellikki’, two hybrids with pink flow- ers, ’Flaaga’ and ’Universitas Helsingiensis 350’ and two white hybrids, ’St. Michel’ H42 and ’P.M.A. Tigerstedt’. Discussion Evergreen rhododendrons are typically 243 244 plant species that prefer maritime climates. In Europe, rhododendron breeding and use is concentrated to Ireland and the U.K., the Benelux and coastal northwest Germany. In North America the Pacific Northwest is the optimal area with some additional areas on the east coast. In Asia, Japan is the center of breeding and use although they have concen- trated their efforts on the deciduous azaleas. Extremely low winter temperatures, below —35°C, under conditions of little snow cover- age is particularly detrimental to flower buds of rhododendrons. Critical low temperatures for flower bud damage may vary within spe- cies depending on the seed origin. Just as in conifers, a difference of one degree latitude or 100 m altitude may be critical for frost hardiness. Thus it is obviously somewhat un- critical to classify species of rhododendrons as hardy or not hardy in a certain location without simultaneously noting the seed origin, including the specific site from where seed was initially procured. Particular attention must here be given to the seed origins of Rhododendron brachycar- pum. There seem to be controversial opinions about the taxonomic classification of this spe- cies and some scientists have questioned the justification to separate it into subspecies (Chamberlain and Doleshy 1987). From a strictly population genetic aspect, it may well be that this species varies gradually (clinal variation) within its natural distribution in Japan and Korea. Thus the subspecies tiger- stedtii may well represent the extreme end of a dine where hardiness and even morpholog- ical traits deviate considerably from the typi- cal R. brachycarpum. The fact that there is considerable genetic variation in hardiness of different seed sources of a species is worth special consideration from the plant breeder. In fact, there may even be variation in hardiness between in- dividuals belonging to the same seed source. Thus the breeder must a priori select the bas- ic material for hybridization from seed sources originating from climatically marginal low- temperature areas of the species. Secondly, the breeder may considerably improve hardiness breeding by using parental material that has already been naturally selected for winter hardiness in a new critical environment. This was the case in the present rhododendron breeding program, where the parental materi- al, particularly on the maternal side, has been naturally selected in the arboretum for peri- ods up to 40 years. Particularly hybrid popu- lations, where both parents had been through selection at Arboretum Mustila exhibited out- standing climatic adaptation. Much hardiness was usually lost by using imported pollen from ornamentally interesting plants that had grown in better climatic conditions in southern Scandinavia or central Europe. Critical low temperatures also damage the vegetative bud, although its hardiness is usual- ly greater than that of the flower bud. Ulti- mately low temperatures may damage or kill annual shoots thus totally destroying rhododendron plants. This often occurs only after plants have grown above the snow cover, to a height of 50—100 cm in south central Finland (Fig. 8). In addition to the direct effects of low tem- perature on bud or shoot damage, evergreen rhododendrons may, as mentioned above, in- directly suffer from low temperatures due to frozen ground and subsequent desiccation. This »sunburn-effect» usually occurs in late winter or early spring when the ground is still frozen, but solarradiation is already intensive, particularly due to the strong reflection from the white snow around the plants. This par- ticular condition is typical of the Finnish cli- mate. In fact the Finnish language has a spe- cial word (ahava) for this kind of »sunburn- effect». It is particularly fatal to evergreen rhododendrons that grow on south slopes un- der heavy solar radiation but it may even damage indigenous evergreen conifers. From the description above we conclude, that rhododendron field trials following hybrid breeding must continue at least until the test plants reach a height well over the snow cover. This mayrequire a test period of at least 10years for high-bush types, while this question may be uninteresting for procumbent growth types. However, observations at Arboretum Mustila have indicated, that each 15-year peri- od during this century has experienced at least one critical year for determining woody plant Fig. 9. Selection of ortets in hybrid progenies, micropropagation of ramets, planting of clone tests and the release of new cultivars. 245 hardiness. In the rhododendron program we have been lucky to have two such years with one year interval, but generally a 10—15 year test period is mandatory so that progeny test- ing may with reasonable precision determine hardiness of new cultivars. Breeding for extreme hardiness is inevita- bly a very long-term enterprise. It actually starts with the geographical selection of mar- ginal seed sources of wild species. The seed sources are then grown as plant populations in the experimental area (Arboretum) where- by natural conditions further select hardy in- dividuals within thepopulation. This phase of »passive breeding» should take the population through at least one critical »bottleneck» year at a stage when the plants are already far above snow coverage. The rhododendrons used in the breeding program at Arboretum Mustila were introduced in the 1930 s and have gone through 3—4 very critical winter periods where natural selection has been ruthless. Hybrids produced in the breeding program have now been in field trials for B—l 38—13 years and breeder selection is still going on. Select- ed and cloned material has been out in repli- cated field trials (clone tests) for about 5 years. These tests are planned to continue until 1995 (Table 1). Considerable time-saving can be achieved in the breeding program by effective use of meristematic cloning. Field trials of new hybrids must also be given proper time (10—15 years) to obtain reliable hardiness selection and also to select for ornamental values. However, meristematic cloning of pos- sible cultivar candidates (ortets) can be start- ed during the field trial period so that materi- al is available for replicated clonal testing once the selection decision has been made. Also op- timal cloning methods can be worked out in- dividually for each cultivar candidate so that they are available when cultivar decision is made. Deliberate breeding for climatic adaptation, such as hardiness, is a long process, often covering half a century to be on a firm genet- ic resource basis. In therhododendron breed- ing program described here, the crosses be- tween R. brachycarpum subsp. tigerstedtii and either R. smirnowii or R. catawbiense gave the best genetic material for the future efforts to breed for better climatic adaptation. Acknowledgements. We wish particularly to express our gratitude to Mrs. Sylvi Lehtonen for her valuable as- sistance, professional skill, enthusiasm and personal de- votion to this program. We are grateful to Arboretum Mustila for allowing us to use the livingcollection of rhododendrons for the breed- ing program. We wish also to express our gratitude to the Academy of Finland and to the horticultural research foundation of Nikolai and Ljudmila Borisoff for financing this research program. We are also grateful to the City Park Department of Helsinki, the City Park Department of Mikkeli, the Department of Horticulture at the Agricultural Research Centre, the Research Farm Kotkaniemi of Kemira Oy, the University of Oulu and to Mr. Niilo Karhu. They have all given us land for the field trials of the hybrid proge- nies. Without their help in planting and careing of the test fields the establishment of the field trials in their present extent would not have been possible. References Anon. 1986. Maatilahallituksen tuontitilastot. (Import statistics. Govt. Agric. Bureau). Helsinki. Chamberlain, D.F. & Doleshy, F. 1987. Japanese mem- bers of Rhododendron subsection Pontica: Distribu- tion and classification. J. Jap. Bot. 8: 225—243. Cox, P.A. 1979. The larger species of rhododendron. 352 p. London. Cullen, J. & Chamberlain, D.F. 1980. A preliminary synopsis of the genus Rhododendron. Notes R. Bot. Card, Edinburg 36: 105—126. Encke, F., Buchheim, G. & Seybold, S. 1980. Zander Ilandworterbuch der Pflanzennamen. 844 p. Stuttgart. Galle, F. 1985. Azaleas. 438 p. Oregon. Ilmatieteen laitos 1985a. (Finnish Meteorological Insti- tute). Kuukausikatsaus Suomen ilmastoon. Tammikuu 1985. 79: 1. 1985b. (Finnish Meteorological Institute). Kuukausi- katsaus Suomen ilmastoon. Helmikuu 1985. 79: 1. 1986. (Finnish Meteorological Institute). Kuukausi- katsaus Suonien ilmastoon. Yhteenveto 1986. 80; I, 246 1987. (Finnish Meteorological Institute). Kuukausi- katsaus Suomen ilmastoon. Tammikuu 1987. 81: 1. Kallio. T.K. 1966. Koristepuiden ja -pensaiden levin- neisyydestä jamenestymisestä Suomessa. Ann. Agric. Fenn. 5, Suppl. I. 107 p. Helsinki. (In Finnish with German summary). Knape, B. 1984. Sädd- och odlingsförsök med Rhododendron i Brödtorp 1953—1983 (Sowing and cul- tivation experiments on rhododendrons at Brödtorp, SW Finland, in 1953—1983). Sorbifolia 15: 27—33. KrUssmann, G. 1968. Rhododendren, andere immer- grtine Laubgehölze und Koniferen. 190 p, Hamburg, Berlin. 1986. Manual of cultivated broad-leaved trees & shrubs. Vol 111, PRU-Z. 510 p. London. Lawrence, G.H.M. 1951. Taxonomy of vascular plants. 823 p. New York. Nitzelius, T.G. 1970. Rhododendron brachycarpum D. Don ex G. Don ssp. tigerstedtii, eine neue Unterart. Dent. Baumschule 22: 207—212. Salley, H.E. & Greer, H.E. 1986. Rhododendron hybrids. A guide to their origins. 391 p. London. Schalin, B. 1953, Koristepensaista kauneimmat. 141 p. Porvoo, Helsinki. Stevenson, J.B. 1930. The species of Rhododendron. Rhodod. Soc. 855 p. Edinburgh. Tigerstedt, P.M.A. 1970. Dendrological Experiments at Arboretum Mustila. Lustgärden 1970: 141—174. (In Swedish with English summary). 1986. Arboretum Mustila. Pubi. Finnish Dendrol. Soc. Voi. 2. 28 p. Helsinki. Uosukainen, M. 1976. Alppiruusun (Rhododendron sp.) siemenlisäys ja taimikasvatus. Dendrologian Seuran Tiedotuksia 7; 44—50. & Niskanen, A.-M. 1985. Meristem tip culture of woody perennials. Hereditas, Suppl. Vol. 3: 153—154. 1988. Utvintringskador pä Rhododendron och andra Ericaceae växter vintern 1984/85. Manuscr. 6 p. (Avail- able at Agric. Res. Centre, Laukaa, Finland). Zobel, B.J. & Talbert, J.T. 1984. Applied forest tree improvement. 505 p. New York, Chichester, Bisbane, Toronto, Singapore. 247 SELOSTUS Talvenkestävien alppiruusujen jalostus M. Uosukainen 1 * ja P.M.A. Tigerstedt2* 1 * Maatalouden tutkimuskeskus Tervetaimiasema 41340 Laukaa 2 * Kasvinjalostustieteen laitos Helsingin yliopisto 00710 Helsinki Alppiruusut (Rhododendron sp.) kuuluvat Ericaceae- eli kanervakasvien heimoon ja sukuun kuuluu lähes 900, etupäässä aasialaista lajia. Alppiruusujen jalostuskäyn- nistyi Keski-Euroopassa jaPohjois-Amerikassa 1800-lu- vun loppupuolella. Kohta sata vuotta kestäneen jalos- tustyön tuloksena on yli 4 800 nimettyä lajiketta rekisteröity virallisesti. Suuresta lukumäärästä huolimatta ei näiden lajikkeiden joukosta löydy Pohjois-Euroopassa hyvin menestyviä lajikkeita. Heikosta talvenkestävyydestä huolimatta tuodaan Suomeen vuosittain noin 50 000 alp- piruusun tainta Keski-Euroopan taimitarhoista. Helsingin yliopiston kasvinjalostustieteen laitoksella käynnistettiin vuonna 1973 alppiruusujen (Rhododendron sp.) jalostusohjelma, jokaperustui Mustilan arboretumin kestävään alppiruusuaineistoon. Erityisen kestäväksi alp- piruusuksi osoittautui korealainen kookas, valkeakukkai- nen laji R. brachycarpum subsp. tigerstedtii Nitz. eli mustilanalppiruusu. Erinomaisen talvenkestävyytensä vuoksi tämä laji valittiin jalostusohjelman keskeiseksi lajiksi. Pyrkimyksenä oli yhdistää mustilanalppiruusun talvenkestävyys ja muiden sitä arempien lajien ja lajik- keiden koristeellinen kasvutapa ja kaunis kukinta. Mui- ta keskeisiä lajeja ohjelmassa olivat R. calawbiense, R. mellernichii, R. smirnowii ja R. yakushimanum. Risteytysohjelmassa käytettiin äiteinä kaikkiaan 53 Mustilan arboretumissa kasvavaa alppiruusuyksilöä. Ne kuuluivat 8 lajiin ja 27 niistä oli hybridejä. Pölyttäjinä käytettiin kaikkiaan 114 yksilöä, joista osa oli Mustilan arboretumista tai muista suomalaisista alppiruusukokoel- mistä ja osa siitepölystä haettiin Ruotsista, Tanskasta ja Hollannista. Pölyttäjien joukossa oli 23 lajia ja 48 hybridiä. Käytetyt hybridit on lueteltu liitteessä I. Risteytysohjelmassa tehtiin kaikkiaan 148 erilaista lajien, lajien jahybridien jahybridien välistä risteytystä. Niiden lisäksi tehtiin 30 itsepölytyserää sekä 14 kontrol- lierää, joissa emaskuloitua kukintoa ei pölytetty. Vapaapölytteisiäeriä oli 10. Kaikkiaan risteytysohjelmassa oli 496 risteytyserää. Risteytysten tuloksena saatiin yli 20 000 siementainta, joista ensimmäisen elinvuoden ai- kana kuoli noin 25 %. Jälkeläiskokeet perustettiin seitsemälle paikkakunnal- le: Elimäelle, Helsinkiin,Piikkiöön, Kotkaniemen Vähä- mäkeen, Mikkeliin, Lappeenrantaansekä Ouluun. Vuo- sina 1975—79 jälkeläiskokeisiin istutettiin kaikkiaan noin 14 000 tainta. Vuoteen 1982 mennessä oli istutetusta ai- neistosta kuollut 37 % eri syistä johtuen.Huono talven- kestävyys oli ensimmäisinä vuosina vain harvoin syynä taimien menehtymiseen; useimmiten syynä oli epäsopiva istutuspaikka tai taimien väärä hoito istutusten jälkeen. Hyvin usein taimia myös varastettiin koepaikoilta. Aineiston valinta käynnistyi, kun taimet olivat noin 75 cm:n korkuisia, eli niiden latvat olivat lumipeitteen ylä- puolella keskitalvellakin. Tärkein valintaperuste oli ilmas- tollinen sopeutuminen, kuten pakkasenkestävyys, hallan- kestävyys jakasvukaudenaikaisen kuivuuden kestävyys. Muita valintaperusteita olivat kukintaominaisuudet, leh- distön koristeellisuus, pensaan kasvutapa, kasvullisten ja suvullisten jaksojen rytmittyminen, kasvien terveys ja kas- vien ikääntyminen. Risteytysaineiston talvenkestävyyttä testattiin kylminä talvina 1984/85 ja 1986/87. Nämä kaksi talvea olivat useimmilla koepaikkakunnilla vuosisadan kylmimmät. Erityisesti talvella 1986/87esiintyi poikkeuksellisen run- saasti juuristovaurioita. Mainittujen kahden kylmän tal- ven seurauksena noin 30 % risteytysaineistosta kuoli tai vaurioitui niin pahoin, että ne poistettiin koekentiltä. Risteytysjälkeläistöjen kesken oli huomattavia eroja tal- venkestävyydessä. Mustilanalppiruusun (R. b. subsp. ligersledlii) jälkeläiset olivat talvenkestävyydeltään sel- västi parhaimmat. Kun se risteytettiin nukka-alppiruusun (R. smirnow'n) kanssa saatiin erityisen kestäviä jälkeläis- töjä. Samoin, kun pölyttäjänäkäytettiin puistoalppiruu- sua (R. calawbiense), saatiin hyvin kestäviä, joskaan ei kovin koristeellisia jälkeläisiä. Resiprookkiset risteytyk- set osoittivat mustilanalppiruusun voimakkaan äidinvai- kutuksen talvenkestävyydenperiytymisessä. Sen sijaan sen kookas kasvutapa ei periytynyt voimakkaasti jälkeläis- töön. Risteytyksistä kääpiölajikkeiden kanssa saatiin hy- vin hillittykasvuisia, usein kasvutavaltaan kompakteja hybridijälkeläisiä. Risteytysaineistosta valittiin kaikkiaan noin 80 paras- ta hybridi yksilöä tarkempia lajikekokeita varten. Niiden lisäämiseksi kehitettiin meristeemilisäysmenetelmä, jon- ka avulla kloonattiin lajikekokeita varten tarvittavat tai- 248 met. Lajikekokeita alettiin perustaa vuonna 1983 seit- semälle paikkakunnalle Etelä- ja Keski-Suomeen. Vuo- den 1987 loppuun mennessä kuusi hybridi-kloonia oli laskettu kaupalliseen lisäykseen. Lajikkeet ’Elviira’ ja ’Hellikki’ olivat punakukkaisia, ’Haaga’ ja ’Universitas Helsingiensis 350’ olivat vaaleanpunakukkaisia sekä ’St. Michel’ H42 ja ’P. M. A. Tigerstedt’ olivat valkeakuk- kaisia. Jalostusohjelman tuloksena on Suomeen saatu alp- piruusulajikkeisto, jonka alhaisten lämpötilojen kestävyys on ainakin —35°C ja osa lajikkeista kestää jopa -40°C:een pakkasia ilman mainittavia vaurioita. Täten alppiruusujen viljelyaluetta voidaan Euroopassa ja Amerikassa laajentaa sekä pohjoiseen että ilmastoiiaan mantereisille alueille. Monivuotisten puuvartisten koristekasvien jalostus vaatii runsaasti aikaa. Alppiruusun jalostusohjelma vie arviolta noin 20 vuotta. Talvenkestävyyden testauksessa kenttäolosuhteissa on otettava huomioon se, että Suomes- sa esiintyy keskimäärin 10—15 vuoden välein ns. »pul- lonkaulavuosi», eli normaalia kylmempi vuosi. Täten jälkeläiskokeitasuunniteltaessa on niiden kesto arvioitava siten, että ainakin yksi »pullonkaulavuosi» osuu kokeen ajaksi. 249 Appendix 1. Parental species, subspecies and hybrids that have been used in the breeding program. The maternals are all from Arboretum Mustila. The paternals originate from seven different Rhododendron collections. Names are according to collection labelling. I. Maternal plants at Arboretum Mustila: Number of indi- viduals Names Species and subspecies: Hybrids; R. ’Cunningham’s White’ (Cunningham, 1850) I R. ’Dr. H.C. Dresselhuys’ (den Ouden, 1920) 1 R. ’English roseum’ (A. Waterer) 1 R. ’Scarlet Wonder’ (Hobbie/Le Feber 1960) 1 R. Brachycarpum-hybrid (bright red n:o 81) 1 R. Calawbiense ’Album’ 1 R. Repens-hybrid (N:o 92, Mustila) 1 R. x fraserii (W. Wats.) 1 R. Orbiculare-hybrid (N:o 8, Mustila) 1 R. Oreodoxa-hybrid (N:o 66, Mustila) 1 R. Seidel’s hybrid (N:o 57, Mustila) I R. Smirnowii Seidel-hybrids (Mustila) 11 R. Smirnowii-hybrid (N:o 76, Mustila) 1 R. Williamsianum-hybrid (N:o 6, Mustila) 1 R. hybrid (bright red, N:o 16, Mustila) 1 R. (hybrid 73—109—29) (Elimäki) 2 R. brachycarpum D. Don ex G. Don 2 R. brachycarpum subsp. tigerstedtii Nitz. 10 R. catawbiense Michaux 1 R. dichroanthum subsp. scyphocalyx (Balf.f. & Forrest) Cowan 1 R. japonicum Suringar 1 R. luteum Sweet 1 R. schlippenbachii Maxim. 2 R. smirnowii Trautv. 8 Total 26 Hybrids: R. ’Dr. H.C. Dresselhuys’, (den Ouden, 1920) 1 R. ’English RoseunT, (A. Waterer) 1 R. ’Roseum Elegans’, (A. Waterer, before 1851) 1 R. Catawbiense-hybrid (dark red, n;o 42 at Mustila) 1 R. x fraseri W. Wats. (= R. ’Fraseri’) 1 R. Oreodoxa-hybrid, (n:o 66 at Mustila) 1 R. Smirnowii Seidel-hybrids (Seedlings from 1930’5) 17 R. Smirnowii-hybrid, (n:o 76 at Mustila) 1 R. Williamsianum-hybrid, (n:o 6 at Mustila) 1 R. hybrid (bright red, n;o 16 at Mustila) 1 R. hybrid (dark red, n:o 20 at Mustila) 1 Total 27 Number of mother plants: 53 Total 27 2.2. Bengt Schalin, Jorvas, Finland 1973 and 1975 Species: R. albrechlii Maxim. I R. forrestii var. repens, Reute, (Balf.f. & Forrest) Cowan & Davidian R. schlippenbachii Maxim. I 1 1R. smirnowii Trautv. Total 4 Hybrids: R. ’Elisabeth Hobbie’ (Hobbie, 1945) R. ’Gertrud Schäle’ (Hobbie, 1951) R. ’May Day’ (A.M. Williams, 1932) R. Calophytum-hybrid R. Williamsianum-hybrid Total 5 2.3. The Swedish Agricultural University, Alnarp, Sweden 1975: Species: R. yunnanenseFranch. 1 2. Paternal species and hybrids: Names 2.1. Arboretum Mustila; Hybrids; R. ’America’ (M. Koster & Sons, 1920) R. ’Bow Bells (Rothschild, 1934) R. ’Dir. E. Hjelm’, (D.A. Koster) R. Fortunei-hybrid Total 1 Number of indi- viduals 1 I 1 1 Species: Total 4 R. adenophorum Balf. f. & W. W. Sm. 1 2.4. Christinelund, Sweden 1975: R. arboreum Sm. R. brachycarpum D. Don ex. G. Don 2 R. brachycarpum subsp. tigerstedlii Nitz. 16 1 IR. calawbiense Michaux 1 R. hybrid (very dark, unknown) R. calawbiense Michaux var. album Glass 1 Total 2R. dichroanthum subsp. scyphocalyx (Balf.f. & Forrest) Cowan 1 2.5. Sofiero,. Helsingborg, Sweden 1975: Species: R. luteum Sweet 1 R. metternichii S. & Z. 2 R. schlippenbachii Maxim. 2 R. orbiculare DC., 5224 Kew R. yunnanense Franch. 1 1R. smirnowii Trautv. 9 Total 36 Total 2250 Hybrids: 2.7. Proefstation voor de Boomkwekerij, Boskoop, Holland 1976:R. ’Britannia’ (C.B. van Nes, 1921) R. ’Elisabeth Hobbie’ (Hobbie, 1945) R. ’Margarete Waterer’ Species: R. decorum Franch. 1 R. ’May Day’ (A.M. Williams, 1932) R. ’Old Port’ (A. Waterer, 1865) R. ’Pink Pearl’ (J. Waterer, before 1897) R. Williamsianum-hybrid, Böhlje 1957 R. fbrreslii var. repens (Balf.f. & Forrest) Cowan & Davidian 1 R. impeditum Balf. f. & W. W. Sm. 1 R. lutescens Franch. 1 Total 7 R. metternichii S. & Z. 2 R. orbiculare DC. I R. williamsianum Rehd. 1 R. yakushimanum Nakai 2 R. yunnanense Franch. 22.6. The Royal Botanical Garden, Copenhagen, Denmark 1975: Species: Total 12 Hybrids: R. campylocarpum Hook. f. R. decorum Franch. 1 R. ’Dr. Ross’ (Boskoop) 1 I R. ’Thunderstorm’ 1 R. degronianum Carr. 4341/52 R. degronianum Carr. 1 R. ’Vulcan’ (Waterer, Sons & Crisp, 1938) 1 1 R. orbiculare x R. dichroanthum R. metternichii S. & Z 1 (Boskoop) 1 R. metternichii S. & Z. hondoense R. wardii W. W. Sm. 1 R. Thuinhybrid scyphocalyx (Boskoop) 1 2 R. Wardii-hybrid astrocalyx (Boskoop) 1 Total 8 Total 6 Number paternal plants: Species 64 Hybrids 50 Total 114 251 252 Appendix 2. Summary of controlled crosses between 1973—1978. 253 Appendix 3. Map of hybrid progeny tests and clone tests in Finland. 254