Source DocumentPrevious PageTable Of Contents

Plant improvement in horticulture: the case for fruit breeding

Michael G. Mullins

Department of Agronomy and Horticultural Science, University of Sydney, N.S.W. 2006.

Horticulture is a word with many meanings. It refers to a specialized form of agriculture concerned with the commercial production of fruits, vegetables, flowers and nursery crops and it refers also to the cultivation of ornamental plants for the enrichment or repair of the environment. The two branches of horticulture, production horticulture and environmental horticulture, make use of a very wide range of plant species. In this paper I will concentrate on the genetic improvement of woody perennial fruit crops, but same of my observations will apply also to woody ornamentals.

The breeding of fruit crops is one of the most important but most neglected areas of horticultural science. Progress in fruit breeding has been far from spectacular. In the apple, all but a fraction of the world's production is based on cultivars which arose before 1910 (Knight and Alston 1969). Few important genes have been identified in citrus (Cameron and Soost 1969; Soost and Cameron 1975), the grape cultivar Chasselas was known at the tine of Joan of Arc, Cabernet Sauvignon is said to be Roman (Penning-Rowsell 1971), and newer crops such as avocado and macadamia are founded on the vegetative progeny of a few chance seedlings.

Nevertheless, there have been substantial improvements in both scions and rootstocks in most of the major crops, but selection within and among existing genotypes has been the main basis of improvement rather than the creation of new genotypes by hybridization. Clonal selection and phytosanitary procedures have had a major impact on crop productivity.

In the present economic and social climate, however, the prospects for fruit breeding are very favourable and the aim of this paper is to create an awareness of, and perhaps enthusiasm for, the idea of genetic improvement in horticulture. I hope to show that the technical problems which confront the fruit breeder, and which hitherto have been such a strong disincentive to research in genetics and breeding of woody perennials, are by no means insurmountable.

Technical problems in fruit breeding

Most woody perennial fruit plants are outbreeders and are highly heterozygous and they do not breed true form seed. Accordingly, the subtle combinations of genes which constitute a good cultivar must be perpetuated by vegetative propagation. Most of the traits which make up a good fruit are polygenic in their inheritance and few are simple Mendelian characters. Seedlings of woody perennial plants have a protracted juvenile phase during which they do not normally form flowers. With conventional techniques the minimum generation times are from 4-15 years depending on the species. To this must be added the tine required for orchard or vineyard evaluation of advanced selections, perhaps another 5 years. Other biological constraints in fruit breeding include self-incompatibility, polyploidy and parthenocarpy.

Having, at last, raised a progeny the fruit breeder is faced with the problem of selection criteria. Fruit crops tend to be bred for quality rather than yield and there are few simple objective tests of fruit quality. Finally, there are conservative forces which tend to resist the introduction of new cultivars irrespective of excellence. Certain cultivars have become firmly entrenched in the market-place (e.g. Granny Smith) and consumers have little interest in unfamiliar fruits. wholesalers tend to dislike new cultivars because marketing is made more complicated. In certain crops, notably grapes, the introduction of new genotypes is opposed by schemes such as "Appellation d'Origine Controlee" which are designed to stabilize quality and reinforce the image of the existing products.

Manipulation and the cosmetic principle

A basic difference between broad-acre agriculture and horticulture is that innovation in the production of field crops is primarily in the genotype but innovation in horticulture has been almost exclusively at the level of husbandry. In cereals, for example, new cultivars are introduced to accommodate changes in economic or biological factors of production. Disease resistant wheats with improved yields or quality are continually being bred and the lead-time for the production of a new cultivar is about 10 years. For reasons that I have explained, fruit cultivars change slowly, or not at all, and the response to biological or economic constraints in fruit production is to manipulate the existing genotypes by applying higher inputs of husbandry. Included are changes in standard husbandry (rootstocks, pruning, training), in chemical-based husbandry (fertilizers, pesticides, growth regulators), in mechanization (mechanical harvesting and pruning), and in post-harvest husbandry (controlled atmospheres, maturation regulators, pesticides).

In fruit production we tend to superimpose 20th century technology onto ancient genotypes so as to bring them into line with contemporary requirements. This is what is meant by the "cosmetic principle". The only significant element of genetic innovation in fruit production in this century has been the exploitation of spontaneous and induced somatic mutants in a number of fruit crops.

The need for manipulation of growth and development in the production of fruit crops has been the rationale for much investment in horticultural research in the sphere of plant physiology and biochemistry. The argument employed has been that improved understanding of the mechanisms of plant growth and development will enable us to manipulate more effectively plant growth and development in the field. It is becoming clear, however, that this argument is losing its potency and that cosmetic treatment of archaic genotypes by high inputs of technology is an outmoded approach.

Pressures for change

The costs of modern intensive fruit production are, of course, very high and include a substantial energy component. For the future there is a strong economic incentive to reduce inputs and to simplify the husbandry of fruit crops. There is also a strong social pressure for simplication of the methods of production of fresh foods. Consumer attitudes to chemical plant protection have hardened in recent years. The fact that grapes may have to be sprayed 20 times in a season for control of downy mildew and bunch rot, or that control of apple powdery mildew requires a dozen or more sprays, is regarded with increasing concern for the risks of toxic residues and damage to the environment. An expression of this concern is seen in legislation such as the new Pesticides Act (1978) of New South Wales. The message is clear; there is urgent need for new cultivars of fruit crops in which, as far as possible, genetic attributes can take the place of sprays, and which can be grown with minimum inputs. Disease resistance, pest avoidance and tolerance of environmental stresses such as drought and salinity are all characters which are amenable to selection. To these can be added attributes such as nutritional value, ease of storage, seasonality of production, size, form and bearing habit.

Few would dissent from these propositions but, knowing the technical problems of tree breeding, are they merely forlorn hopes? In providing an answer to this question I will, first, describe what is being done in the breeding of fruit crops, how it is being done, and what needs to be done in terns of research.

Conventional methods in fruit breeding

The traditional and most widely-employed approach to breeding clonally-propagated crops is to select within large Fls produced by Dossing existing cultivars. The best of the segregates are multiplied vegetatively, and tested in replicated trials (Simmonds 1979).

The method has had some notable successes in each of the major fruit crops (Janick and Moore 1975), but it has some strong disadvantages. The turnover is slow and the genetic base is narrow. Little genetic information is forthcoming because generations seldom go beyond Fl. Genetic gain in a breeding population is not exploited in following generations and the method is founded on the personal skill of the individual.

In France, much study has been made by biometrical geneticists and fruit breeders of "strategies d' amelioration" in fruit crops. Recurrent selection procedures have been proposed for grapes in which information can be accumulated on the inheritance of the characters concerned and on the breeding value of parents. Essentially, these schemes separate the creation of cultivars, which can occur at any time during successive cycles, from the long-term genetic improvement of the breeding population. Schemes such as recurrent selection are not so firmly founded on the skill of an individual because, theoretically, the successive cycles can proceed with different caretakers. This has distinct advantages in the case of plants with long generation times. It remains to be seen, however, whether recurrent selection (Bouquet 1977), "panmixie controlee" (Rives 1977), or other slow-turnover long-term breeding plans with woody perennials (Wagner, 1975), will survive the administrative fatigue that is a cannon hazard of fruit breeding programmes.

In apple breeding, considerable progress has been made in producing disease-resistant cultivars by the insertion of a number of single gene characters from wild Malus into polygenic backgrounds, followed by backcrossing to scion cultivars to upgrade fruit size and quality (Hough et al. 1953). There has also been progress in the acquisition of genetic information. More than 40 genes with dominant alleles have now been identified in the apple, chiefly by the group at East Mailing Research Station in England (Alston 1971).

Many apple cultivars have arisen as spontaneous mutants of existing genotypes, e.g. the numerous colour variants of the original Delicious apple which arose in 1881. Artificially-induced mutants of apples and other fruits are also of considerable potential importance. Work in Canada (Lapins 1965), England (Campbell and Lacey 1973), Italy (Donnini and Rosselli, 1977) and France (Decourtye, 1971) has shown that irradiation is highly effective in producing useful variants of existing cultivars. The so-called "compact mutants" are an example.

Novel methods in fruit breeding

Reviews on the potential contributions of aseptic culture methods to plant improvement in agriculture and horticulture appear in the literature with irritating regularity. Most authors are academic botanists or biochemists with little knowledge of crop plants or crop production systems and many of the claims of these tissue culture devotees need be taken with a pinch of salt.

It must be stressed that most of the manipulations which have been accomplished in vitro with protoplasts, pollen, somatic cells or other explants have been with plants such as tobacco and wild carrot. These experimental systems have been specially selected for their high degree of regenerative competence in vitro and for their suitability as tools for basic research on the nature of embryogenesis and organogeresis.

It is a big step from herbaceous annual test plants to the breeding of commercially-important trees. A characteristic of most fruit plants is that they are very difficult to regenerate in vitro. So far, high frequency somatic embryogenesis and organogenesis has been achieved only in grape (Mullins and Srinivasan 1976; Rajasekaran and Mullins 1979; Barlass and Skene 1978) and citrus (Button and Kcchba, 1977). The ability to produce very large numbers of individuals from somatic tissues is a prerequisite for the application of genetic engineering to crop improvement and it is also the basis of clonal propagation - a subject of special significance in horticulture.

Aseptic culture techniques may be of value for rapid multiplication of scarce clonal material, or for production of pathogen-free plants, but enthusiasm for the technique should not obscure the fact that there have been substantial advances in recent years in the conventional methods of "cloning" - the induction of adventitious roots in cuttings of fruit plants (Howard 1971).

The creation of genetic variability by induced mutations in vitro, protoplast fusion, organelle transfer or DNA-uptake, is potentially of great interest in crop improvement. In fruit crops, however, lack of genetic variability is seldom a serious concern. The genus Vitis, for example, contains more than 60 interfertile species which are distributed from cool-temperate to tropical environments. Similarly, in Citrus, Malus, Pyrus, Prunus and most other genera of fruit plants, there are great reserves of unexplored genetic variation (Janick and Moore 1975).

The essential problem in the improvement of fruit crops is not the creation of new variation but the development of new strategies for the efficient exploitation of existing genetic variation. The haploid method of breeding is of special interest in this regard (Winton and Stettler 1974). Techniques for producing large numbers of haploid and homozygous diploid plants from major fruit species are not yet available. Some progress has been made in growing plantlets in vitro from anthers of hybrid grapevines (Rajasekaran and Mullins 1979) but the origin of the plantlets has yet to be confirmed.

It is clear that aseptic culture techniques are an interesting addition to the fruit breeders' armoury but, if their potential is to be realized, there must be an increased investment in research on regenerative phenomena in vitro in horticulturally-significant species. Hitherto, there has been excessive extrapolation from simple, contrived, experimental systems with herbaceous plants to complex horticultural situations with woody perennial species.

Limiting factors

As a generalization there are no intractable genetic problems which are limiting factors in fruit improvement. There is much genetic variation within genera of fruiting plants that is potentially useful to the breeder and there are already several genetically-sound breeding procedures by which this variation could be exploited. The major problems in fruit breeding are long generation times and lack of pre-selection criteria. These problems are in the sphere of plant physiology rather than in genetics and an important direction for horticultural science in the next decade is the application of plant physiological research to breeding rather than to the manipulation of existing genotypes.

Juvenility and precocious flowering

The juvenile and adult phases of the life cycles of woody perennial plants, and the mechanisms which control the expression of juvenile and adult characters, have been studied in detail in only a few species (Zimmerman 1976). The juvenile and adult forms of fruit species are generally less distinctive than those of Hedera helix, a favoured subject for physiological studies of phase change, and control of phase in temperate, sub-tropical and tropical fruits has received relatively little attention.

In the field, the dwarfing rootstock M27 induces precocious flowering in scions of apple seedlings (Tydeman and Alston 1965) but it is in controlled environments that greatest progress has been made in curtailing the juvenile period. Apples have been made to flower in 16 months from seed (Aldwinckle 1975) and by a combination of genetic (Visser 1970) and physiological manipulations there is a distinct possibility that flowering in apple seedlings can be induced within 12 months of germination.

Morphologically, the juvenile phase of the grapevine is short-lived (Mullins et al. 1979) but even with highly intensive methods of cultivation the first bunches on a grape seedling may take nine months to make their appearance (Huglin and Julliard 1964). Recently, flower initials have been induced in grape seedlings within four weeks of germination (Srinivasan and Mullins 1979) by conversion of the first-formed tendril into an inflorescence by cytokinin treatment. These results, together with developments in the culture in vitro of the vine and its relatives (Mullins and Srinivasan 1976; Rajasekaran and Mullins 1979), hold the promise of at least 1.5 generations per year.

These new techniques for induction of precocious flowering in grape seedlings arose from an unrelated programme of research on the physiology of flowering in cultivars of the grapevine. Similar spin-off could well result from intensive research on flowering in other important fruit crops. Control of flowering in citrus, for example, is virtually an untouched field and the problem of juvenility in citrus breeding is especially severe (Soost and Cameron 1975).

It is curious that detailed information on the control of flowering is available for numerous annual and biennial plants, including species of little or no economic importance, but that such information is scarce or unavailable for many of our major fruit crops.

There are several reasons for this paradox. The first is that woody perennials are inconvenient as experimental subjects. They are large and are difficult to manage in glasshouses or growth roars. With normal techniques the turnover of experiments is slow and this has severely limited the acquisition of knowledge on the physiology of flowering.

Second, the control of flowering was first studied in certain herbaceous plants which are highly responsive to changes in light or temperature. Subsequently, much research on the mechanisms by which apices are transformed from the vegetative to the reproductive mode of development has been with plants which are specially selected for sensitivity to photoperiod or to vernalization.

A third factor is the tacit assumption that the information which emerges from research on these "plants of convenience" has widespread if not universal applicability. Textbooks in plant physiology, for example, seldom refer to the fact that flowering in many species, including economically-important fruit plants, is regulated by mechanisms which are either unknown or which seem to lack a close relationship with the familiar light- or temperature-controlled systems.

Pre-selection criteria

Few research institutions have sufficient land and money to grow large populations of fruit tree seedlings to maturity, especially if plants are grown at orchard spacings. In breeding tree fruits it is essential to pre-select promising material at the seedling stage so as to reduce the progeny size. Equally, pre-selection is needed for early elimination of inferior genotypes so that resources can be used to best effect for the evaluation of fruiting behaviour. Simply-inherited characters which are readily-identifiable must be selected as early as possible in the life of the seedling so that later selection can be concentrated upon polygenically-inherited characters where identification is normally more difficult.

Several pre-selection indices, characters of seedlings which are correlated with characters or performance of adult trees, are available for apples (Watkins 1974) and grapes (Wagner and Bronner 1974). There are many tests for resistance to pests and diseases which may be applied to seedlings (Watkins and Werts 1971) and there are methods for predicting the vigour of potential rootstocks from the anatomy of the roots and leaves of apple seedlings (Beakbane and Thompson 1939; Beakbane 1965).

Tests are available which can predict certain aspects of fruit quality at the time of first fruiting, for example, the bromcresol-green-test for presence of the malic acid gene in apples (Nybom 1959; Brown 1975), but tree performance in general has to be tested under orchard conditions. The development of new pre-selection tests and the induction of precocious flowering in seedlings are complementary problems in fruit breeding, and both are amenable to investigation.


A recurring theme in this paper has been the need to regard our fruit species as subjects worthy in their own right of intensive scientific investigation. In the last 25 years the plant sciences have advanced very rapidly, especially in the spheres of physiology and biochemistry. Great emphasis has been given to the study of processes in growth and development and reductionism has been the prevailing philosophy. The intensive study of individual crop species has been largely restricted to a few highly specialized research institutions, for example, East Malling Research Station in the case of the apple.

The time is now ripe for a renewed attack on the physiological problems which retard the progress of fruit improvement. However, this work must go hand-in-hand with exploration, in a genetic sense, of the variation which exists in the wild relatives of fruit plants, and with studies on the mode of inheritance of economically-significant characters.

Hitherto, the idea of fruit breeding has invoked knee-jerk objections in many administrators and policy-makers on the grounds that it is too difficult, too expensive and too long term. These objections are rapidly losing their substance. The modern approach to the improvement of woody perennial plants, involving the speeding of generations, pre-selection and the application of genetics, has the potential to solve many of the problems which up to now have seemed insoluble. The prospects for fruit improvement are highly favourable.


1. ALDWINCKLE, H.S. (1975). HortScience 10:124.

2. ALSTON, F.H. (1971). Proc. Angers Fruit Breed Symp. INRA. Angers (1970), 231.

3. BEAKBANE, A.B. (1965). Rep. E. Mailing Res. Stn. for 1964. 102.

4. BEAKBANE, A.B. and THOMPSON, E.C. (1939). J. Pamol. 17:141.

5. BARLASS, M. and SKENE, K.G.M. (1978). Vitis 17:335. -

6. BOUQUET, A. (1977). Ann. Amelior. Plantes 2/:/5.

7. BROWN, A.G. (1975). In: "Advances in Fruit-Breeding" (Ed. J. Janick and N. Moore). Purdue University Press, West Lafayette.

8. BUTTON, J. and KOCHBA, J. (1977). In "Applied and Fundamental Aspects of Plant Cell, Tissue and Organ Culture" (Ed. J. Reinert and Y.P.S. Bajaj). Springer-Verlag. Berlin.

9. CAMERON, J.W. and SOOST, R.K. (1969). In: "Outlines of perennial crop breeding in the tropics" (Ed. F.P. Ferwerda and F. Wit). Misc. Papers Landbouwhogeschool Wageningen 4:128.

10. CAMPBELL, A.I. and LACEY, C.N.D. (W73). J. hort. Sci. 48:397.

11. DECOURTYE, L. (1971). Proc. Angers Fruit Breed Symp.

12. INRA. Angers (1970). DONNINI, B. and ROSELLI, G. (1977). Riv. Ortoflorofrutt. It. 61:175.

13. HOUGH, I.F., SHAY, J.R. and DAYTON, D.F. (1953). Proc. Amer. Soc. hort. Sci.

14. 62:341.

15. HOWARD, B.H. (1971). Sci. Hort. 23:116.

16. HUGLIN, P. and JULLIARD, B. (1964). Ann. Amelior. Plantes 14:229.

17. JANICK, J. and MOORE, J.N. (1975). "Advances in Fruit Breeding". (Purdue University Press, W. Lafayette, Indiana).

18. KNIGHT, R.L. and ALSTON, F.H. (1969). Rep. E. Melling Res. Stn. for 1968 (1969), 125.

19. LAPINS, K. (1965). Can. J. Plant Sci. 45:117.

20. MULLINS., M.G. and SRINIVASAN, C. (1976). J. exp. Bot. 27:1022.

21. MULLINS, M.G., NAIR, Y. and SAMPET, P. (1979). Ann. Bot.- (In the press).

22. NYBOM, N. (1959). Hereditas 45:332.

23. PENNING-ROWSELL, E. (1971). 'The Wines of Bordeaux" (International Wine and Food Publishing Co., London).

24. RAJASEKARAN, K. and MULLINS, M.G. (1979). J. exp. Bot. 30:399.

25. RIVES, M. (1977). Ann. Amelior. Plantes 27:61.

26. SIMMONDS, N.W. (1979). "Principles of Crop Improvement" (Longman Group, London).

27. SOOST, R.K. and CAMERON, J.W. (1975). In: "Advances in Fruit Breeding" (Ed. J. Janick and N. Moore). Purdue University Press, West Lafayette.

28. SRINIVASAN, C. and MULLINS, M.G. (1979). Planta 145:187.

29. TYDEMAN, M.H. and ALSTON, F.H. (1965). Rep. E. MalIing Res. Stn. for 1964 (1965), 97.

30. VISSER, T. (1970). Euphytica 29:193.

31. WAGNER, R. (1975). Ann. Amelia. Plantes 25:151.

32. WAGNER, R. and BRONNER, A. (1974). Ann. Amelior. Plantes 24:145.

33. WATKINS, R. (1974). Indian J. Genet. 34:1248.

34. WATKINS, R. and WERTS, J.M. (1971). A. Appl. Biol. 67:153.

35. WINTON, L.L. and STETTLER, R.F. (1974). In: "Haploids in Higher Plants: Advances and Potential" (Ed. K. Kasha). University of Guelph.

36. ZIMMERMAN, M.H. (Ed.) (1976). Symposium on juvenility in woody perennials. Acta Hort. 56:317 pp.

Previous PageTop Of Page